Home

TMQ Vol I.iv - tim

1. ney DI DI XOR A XOR A JP 3015H OUT 84H A JP 4000H JP 3015H NOP I may look into altering SET2RAM to add this patch on the fly however I don t have it at this time RSHARD s ARCHIVE RESTORE Fm LDOS Support jjkd Roy the new MISOSYS hard drive support package does the HD backup utility for large files preserve the LRL on a restored file HARDCOPY BAS does not and even though this can be fixed via COPY C N tis inconvenient Fm MISOSYS Yes it does Part of the file s header record is the directory info After the restored file is written to the targeted disk and closed the directory is updated to match the Ist 20 bytes of the original DIR record Thus the LRL is preserved even if the restoral is to an old file of the same name with a different LRL Trouble with DEVICE Fm William Chao Got a copy of LDOS 5 3 few weeks ago and now I have it running on my Mod 4 with a 15Meg hard drive I noticed LDOS Information THE MISOSYS QUARTERLY SPRING 1987 23 Volume I iv something peculiar sp whenever I issue a command such as DEVICE or DIR When given DEVICE it will show up all the logical drives I have 8 but it can t finish in 1 screen so it pauses for the second screen on the second screen I would get the column header for the DIR command and then device would finish On DIR sometimes the column header would appear twice if pauses between screens and logical drives Any
2. Fm Paul Rehberg I enjoyed lying by the pool yesterday sunny a hot 91 degrees in Houston reading my Winter edition of TMC Shall I stock a supply of cord wood for the fireplace in anticipation of the Spring issue lt gri Letters to the Editor Volume I iv I would gladly pay 59 95 again for a new Pro Wam that included a business calculator i e one that knows how to perform subtraction like us 9 to 5ers use all day long Any chance of one in the update And what ever happened to 3 ring holes in the TMQ Fm MISOSYS We gave up on 3 hole punching after about 50 copies of issue I i were ruined by the hole punching operation We do not expect to take another look at hole punching However we do expect to continue the perfect binding introduced in issue I iii really makes a class appearance I assume that a you mean subtraction as in number number result Have you not used AFPCALC That is algebraic Do you want something else Kind of hard to squeeze any floating point float routines If you want an then that can not on the more functionality in a calculator application as the take up quite a bit of space algebraic INTEGER calculator be arranged however one is drawing board for release 2 0 Fm Adam Rubin Yes the perfect binding on TMQ does look much nicer I also liked the Separate sheet for postage and addresses Any plans to get an ISSN Oh and I liked the content too Incid
3. INSTALLable Software Fm Byron P Peebles Am I imagining things or are the VAST MAJORITY read all commercial of microcomputer 8 and 16 bit a little on the defensive side when sugges tions are made on improving their products I get the feeling I m hearing a whining sound while I read the replies I know budgets are tight for the small business person I know theft of software is know there must be fifty problems I m not aware of for the commercial developer I also know there exists a person called the end user client This mythical person is NEVER WRONG is TREATED WITH RESPECT and has the WORLD S MOST RESPECTED OPINION I also know that this is the person that keeps everyone in business rampant I And once in a blue moon this end user client has experiences that could provide great insight to the high tech developer were that developer willing to listen Letters to the Editor THE MISOSYS QUARTERLY SPRING 1987 44 s Volume I 1v Sound of soapbox being shoved back under the desk The above message relates to the coming death of the 8 bit microcomputer Fm MISOSYS Byron I think when a developer spends a lot of time revamping a package and the first thing folks ask when it gets released is why don t you add this and add that and when s the next release it gets sort of frustrating Very few folks are considerate enough to wait some degree of time before they start asking for the moon
4. The first is easily handled in one of two ways The data may be moved to a disk readable by the new machine or it may be transferred over an RS232 link The disk method is limited by the fact that a special software will MS DOS Information THE MISOSYS QUARTERLY SPRING 1987 set of questions for me probably isn t practical EA E Volume I iv need to be purchased and b you transfer a diskfull at a time can only If your database is on a hard disk you will have to break it into chunks move each chunk over and then re assemble the chunks on the hard disk at the other end The RS232 method can generally be accomplished using public domain software at both ends so no special software need be purchased Also the entire file can be transferred from hard disk to hard disk no segmentation required The disadvantage is that you have to have both machines in one place at one time Modeming and you can t trade in the old machine on the new machine Massaging the data to fit into the new program is another matter entirely You need to decide what database you are going to use on the destination machine first then we can examine the options in more detail Editor kudos Fm Marc Nowell to H Brothers You speak of a new configurable editor I m using BRIEF now and am blown away by it I set mine to a WEIRD mix of commands from the Norton Editor WordStar SideKick and about 30 homebrew What do
5. IF CHAR amp H8 THEN IF COUNT gt O THEN DEC COUNT POKE POSIT ION COUNT 32 GOSUB OUTBYTE GOTO GETKEY ae A RE e art aP Se TES ELSE OVO T Khi Ton GOTO GETKEY RA Rd i P RJA ENDIF P Q BOX 297 E ENDIF PADSTOW N S W 2 2241 pg OT POLIO KE areng Fm MISOSYS There were 5 lines in the program which were causing EnhComp to crash Each was of the form IF expression THEN The syntax of IF is either IF exp or IF exp THEN action Of course EnhComp should not crash Patches BC53 FIX and BC63 FIX to BC CMD which change compilation from a crash to a displayed Syntax error message are in The Patch Corner After applying the patch you will correctly get a syntax error Fm Bob Connors I assume here I go again assuming that the culprit is the first THEN of each IF statement shown So instead of IF exp THEN IF exp I should have IF exp IF exp THEN ELSE is that correct Fm MISOSYS That s sort of correct If you look at page 4 50 of the manual you will see that the IF statement has two forms You were trying to use a combination of both of them IF cond progcode ELSE progcode ENDIF is one of the forms IF cond THEN action ELSE action is another You can t have IF cond THEN IF since the second IF starts a new statement That doesn t match either of the acceptable forms The compiler didn t check MISOSYS Products Tidbits PEE i f PA E 3 f Vo
6. 0 PRINT Retries sent RI PRINT Correct answers CA PRINT Wrong answers WA PRINT Passes SPA PRINT Percent correct with retries gt CA Q 100 PRINT Percent correct without retries n CA Q RT A 100 a 1000 PRINT PRINT Press lt gt to return to saat 1010 A INKEY IF A G THEN 530 ELSE 1010 ee GU ee ee mp p ee Astme tangu Gang ee ee ee ee ee ee ee Alan a e le ee ee Gini ee ee ee ee ee ee ee ee ee ee tep tep ee ee ee ee ee tep ee ee ee ee mn Applications for the User pl Applications for the User Volume I iv Data Indexing using VICTORIA AUSTRALIA BASIC by Paul Wade The not quite so recently arrived Misosys Quarterly was again up to the professional standard that I have come to expect of yourself please keep up the good work Accompaning this letter is a program written in BASIC using The Basic Answer TBA Files are as follows 1 LIBENTRY TBA TBA source of the cross reference data entry programs 2 LIBREAD TBA TBA source of the cross reference data retrival program 3 LIBENTRY BAS compiled and packed basic program 4 LIBREAD BAS compiled and packed basic program 5 LDOSINFO DAT title file names for cross reference programs to access 6 LDOSINFO TXT data file of cross referencing to the title file 7 INSTRUCT TXT instructions on using the cross reference programs in ASCII but will run as a basic program This program was originally written by myself
7. In fact that is exactly MS DOS Information Volume I iv the point I m make and understand trying to Is it true that a COM file will be placed into absolute memory by the loader as a function of what has gone on before In other words at different times a COM file might find itself with different Code Segment CS i e be located in some DIFFERENT PLACE in ABSOLUTE memory But here is the important thing to me NOTHING IN THE PROGRAM has been changed corrected modified etc every byte of working code is identical to the previous time it was used EVEN IF AT A DIFFERENT ABSOLUTE ADDRESS If that concept is 1 then I would say that a COM program is RELOCATABLE On tother hand an EXE file would be code that is relocated Do I have the right concept I hope so I ve got to get on with bigger things Fm LDOS Support jjkd Yes and no lt lt Grin gt gt Under that definition of relocatable then yes COM are and the EXE are not Only the CS register need change Assuming that s settled let s look at this from a slightly different point of view The 280 provides two basic kinds of jump instructions absolute and relative Absolute jumps go to a particular spot in the 64K addressable by the Z80 The relative jumps on the other hand go to a particular spot relative to the current instruction pointer Thus no matter what the absolute IP may be code written only using JR may be moved aro
8. Sy E A H S 1 ag fA fer i i kad gt 70 ay 29 4 fe i f A N S Y 4 tS IO onma m t ae Our Compuserve Forum THE MISOSYS QUARTERLY SPRING 1987 104 We Mt tly The LDOS 5 3 upgrade kit is now available to take your Model IIl or 4 in 3 mode to the year 2000 LDOS 5 3 provides complete media compatibility with LS DOS 6 3 the newest Model 4 DOS released by Logical Systems Inc With LDOS 5 3 you can add 12 years to the life of your software Just look at these improve ments over version 5 1 4 Only 34 95 Date support through December 31 1999 time stamping for files Enhancements to LDOS now free up 14 additional file slots for data disks On line HELP facility for DOS and BASIC 117 screens of help DOS Enhancements LIBRARY Enhancements New FORMS lets you change printer filter parameters New SETCOM lets you change RS 232 parameters Improvements to UST add paged displays full screen hex mode and fiexible tab expansion MEMORY displays directory of terminate and stay resident modules SYSTEM lets you direct the SYSGEN to any drive adds a flexible drive swap subcommand SMOOTH for faster disk throughput DIRectory display enhanced with time stamps file EOF and more e We ve also improved AUTO COPY CREATE DEBUG DEVICE DO FREE KILL and ROUTE and added CLS and TOF commands UTILITY Enhancements e We ve added TED a full screen text
9. THE MISOSYS QUARTERLY in this issue e UNDATE reverses DATECONV by L M Garcia Barro e 80x86 assembly language by Phil Oliver Converting LDOS filters to LS DOS e Previewing output from SCRIPSIT e Announcing PRO WAM Release 2 e Learn MORSE with CODE BAS Volume I Issue iv Spring 1987 10 tn S rumma ee eee ee a rt a a a a ama re M Bo ee ge ee Se ae a we eg i FA oh F f x co ocom Se THE MISOSYS QUARTERLY Volume I Issue iv Table of Contents The B lurb a 6 M e e e s e e a Announcing New Products RATFOR 86 and RATFOR M4 8 8 8 o PRO WAM Release 2 e e gt o 8 o e Letters to the Editor g ssa a 8 neh DR ee ek ee v LDOS Informat ion e a al e e e e LS DOS Information s s 6 6 6 n s o The LSI Colum a gt e LSI Patches to LS DOS 6 3 e o MS DOS Information gt e e o 2 9 e Applications for the User UNDATE by Luis M Garcia Barro STRIP BAS by James K Beard Ph D MORSE code from a Model 4 ess s Data Indexing using BASIC by Paul Wade Previewing output with SCRIPSIT by Roy Soltoff The Programmer s Corner Using I O Redirection with MC by M Johnston Writing filters for LS DOS by Roy Soltoff 8086 Assembly Language by Phil Oliver MISOSYS Products Tidbits s s
10. UNDRLN DC 80 ETX since the statement does not adhere to the correct syntax the DC pseudo op does not permit repetition You would need to recode the statement as a pair of statements MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 94 Volume I iv UNDRLN DC DB ETX 80 I don t feel it is necessary to work up a patch for this It will be corrected when MRAS gets reassembled to another release Fm John M Pantone I have been using your MRAS assembler now for several months and I am extremely pleased MRAS is without a doubt one of the finest assemblers micro or mainframe based I have ever used I have however run into a rather obscure bug which I would like to bring to your attention It relates to the DW pseudo op specifically when multiple arguments are provided and where the first arguiment is the address of a label If the first of many arguments to DW is a constant or if DW is arguments only the bug doesn t appear I have enclosed four figures which illustrate the problem I have tried to isolate the bug to as small an example as possible while still illustrating the Situation used with single In the process of preparing this letter I ran into 2 other less disturbing bugs 1 The CI flag of MRAS appears to do nothing the REL file produced not a CIM file as documented is identical to one produced without the CI flag 2 The I y flag of MLINK
11. argl alog s sqrt s 2 1 Compute acosh s dval exp argl n 1 Compute cosh argl n 1 dval 5 dval l dval _qfroms dval Convert back to single precision return end acos x function required for Model 4 The above example shows how RATFOR extends the power of FORTRAN while at the same time making programming simpler Another example shows how the use of macros set up with the define statement can extend the fundamental keyword set in the language The example selected is INPUT which performs a function similar to the BASIC language INPUT keyword A prompt is defined and a variable is read in from the keyboard and an error in input results in a repeat of the prompt until the keyboard input is successful Note that local labels are not repeated this process is similar to that used in better macro assemblers which use non repeated local symbols in macros Following the example are samples of its usage INPUT macro tested on MS DOS microcomputer using Microsoft FORTRAN version 4 0 Syntax for INPUT macro INPUT lt prompt gt lt variable gt lt format label gt define INLAB 1 define INPUT define INLAB PERL ME repeat ierr l print 1 read ifelse 3 3 err INLAB 2 ierr 2 INLAB continue until ierr 2 define READLUN define PRINTLUN Sample usages INPUT MDL for range computation 0 to end mdldb Floating point variable INPUT Input file name fname 100 ASCII string
12. emphasis by the operating system I therefore have the idea that an EXE file can and will be loaded into ANY place in memory by DOS HOW does it decide WHERE to load and that DOS addresses so it will play that the correct view fixes the necessary where it 15 Is 2 If I use MASM and LINK on my XT CLone to produce an EXE file a simple file with only DOS commands and proper interrupts NO direct hardware access shouldn t it be executable on a Tandy 1000 HD as is If not WHY not 3 I wrote a very short 8088 ASM program which was eventually to be merged into a BASIC program Following procedures outlined in Assembly Language Primer for the IBM XT amp PC I used MASM no problem followed by the use of LINK with its HIGH switch e g LINK file HIGH The accompanying text Until now LINK created EXE files AT THE LOWEST AVAILABLE MEMORY ADDRESS ed This will now be occupied by BASIC HIGH option will cause an EXE file to be installed in High memory directly below the transient part of DOS My VERY poorly written DOS manual confirms this Fortunately Use of the HIGH switch causes MS LINK to place the run file as high as possible in memory Otherwise MS LINK places the run file as low as possible Now how does THAT fit in with the idea that an EXE file is NOT tied to any part of memory What does that mean installed in LOW memory These are all real English words but
13. 56 Applications for the User Volume I iv 5217 521A 521B 521D 521F 5222 5225 5228 5229 522C 522F 5232 5235 5238 523B 523C 523E 5241 5244 5246 5248 524B 524E 5251 5252 5253 5255 5258 525B 525D 525F 5261 5263 5264 5267 5269 526C 526D 526F 5272 5274 5275 5277 527A 527C 527F 5282 5283 5286 5289 528C 528D 528E 5290 5291 5294 5297 5298 5299 529A 529B C23E53 7E D630 FE08 D23E53 320D53 321F53 4F CD0942 C23E53 DA2E53 CD8F47 FD E09 325252 57 1E00 CD1B53 3ACD55 CB5F 2009 214554 CD6744 C33853 1600 1803 CD1B53 3A1455 D602 FE21 3802 3E20 3C 329852 1E02 CD1B53 E5 CB66 CA8652 DDE 5 El 3E 96 DD7712 3E42 DD7713 3A7054 3C 327054 CDE 752 C23253 El 7D C620 6F C26C52 CD0953 3E00 1c BB D26952 00035 00036 00037 00038 00039 00040 THE MISOSYS QUARTERLY 00041 00042 00043 00044 00045 00046 00047 00048 00049 00050 00051 00052 00053 00054 00055 00056 00057 00058 00059 00060 00061 00062 00063 00064 00065 00066 00067 00068 00069 00070 00071 00072 00073 00074 00075 00076 00077 00078 00079 00080 00081 00082 00083 00084 00085 00086 00087 00088 00089 00090 00091 00092 00093 00094 00095 NEWTYPE DIRCYL FILES GET AGAIN NEXT DIRSIZE Applications for the User JP LD SUB CP JP LD LD LD CALL JP JP CALL LD LD LD LD CALL LD BIT JR LD CALL JP
14. 86 worlds comes from the segmentation but so far as programs are concerned this is largely invisible being handled by the loader unless you have to be concerned with multiple program segments Fm Nate Salsbury to H Brothers I have successfully merged my EXE file into my BASIC program The major problem pointed out by you and others was that HIGH on a 640K machine is different from HIGH at 256K Seems to me that I should have known THAT lt grin gt Perhaps it is the BIG NUMBERS that I am not used to yet T once reported to you that I was having a Turned out to worked just problem with ML in a string be a clerical error on my part fine when I did it the right way Thanks for participating in The electronic education of Nate Hyman Kaplan Salsbury BIW with all your Sysopping writing product reviewing are you ALSO STILL teaching Fm Nate Salsbury to LDOS Support Joe Sorry to break the thread but I fell quite sure that your earlier reply would have scrolled off by now First is a semantic problem re relocatable COM files Acknowledging that COM files must start at an offset of 100H in this code segment it IS true that CS can be located anywhere in memory subject to the 16 byte steps Doesn t this mean in practice that I can have my COM program loaded into and run almost ANY place in memory at 16 byte start intervals by suitably diddling CS If so to MS DOS Informat
15. EDO 11 A minor modification of this will allow you to do either output with no CR output with both CR and LF or output to both devices When you ve digested this we ll go on to serial input The problem with input is that routines in BASIC are hosed up choices the byte I O There are two 1 Link the keyboard to the RS232 You can then scan for RS232 input via INKEYS The problem with this is that you can t distinguish between keyboard input and RS232 input but is usually adequate for bulletin board programs and the like LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 to disable 2 amp 3 on 39 Volume I 1v 2 Use at least a little assembler for input With the tools available in the new 6 3 0 release you can use the SVC interface USR call without really writing any assembler at all LS DOS 6 3 questions Tim Clute 1 Why wasn t QFB cmd added instead of the new DISKCOPY CMD 2 What is the reason for the distribution disk only having 2 floppies enabled Love 6 3 hope that some of the suggestions make it into 6 3 1 or 6 4 lt grim lt hint gt Fm MISOSYS Only 2 floppies are enabled on the 6 3 distribution disk because only 2 were enabled on 6 2 Most Model 4 folks have only two floppy drives If drives 2 and 3 were enabled then there would be some delay during global search for the system to ascertain that you had no drive 2 nor 3 Thus it makes sense release That speeds up global
16. FETET ENCES e aia 74 95 4 S amp H PRO DUCE A2 pass labeling Z80 disassembler from disk or memory with screening input for data areas Generates ASM E E PEEN EEAO EA EEI E SEEE 34 95 2 S amp H PRO MRAS Powerful relocating macro assembler devel opment system REL module compatible with Micrasoft Includes full screen text editor REL librarian VM linker with Overlay capability nenevaasovae 89 95 4 S amp H UNREL T80 Converts MRAS or M 80 REL object files to ASM Use on your own REL modules FORLIB GRLIB BASCOM BASRUN ete 49 95 2 S amp H PRO EnhComp An enhanced BASIC compiler with a built in assembler for Z80 in line code mixed with BASIC LOGO iike turtle graphics strings to 32767 chars muiti line functions keyed tagged SORT REPEAT UNTIL structured IF ENDIF labeled statements double precision functions AAN E I NEE E E 124 95 4 S amp H LS TBA A structured BASIC translator Labeled statements Conditional translation pseudo global and local variables 14 char var names ccssecesceaveeeaes 39 95 4 S amp H PRO MC A full K amp R C compiler with nearly 200 functions Structs unions bitfields enum dp floats and functions Wildcards I O redirection args overlay support Requires PRO MRAS or M 80 o oo eee scccccecscceeesceeeceseees 124 95 4 S amp H RATFOR M4 A professional implementation of RATional FORtran
17. I realize that I ve caused a lot of confusion by not saying what I really meant When I said variables I didn t mean all variables I really meant variables declared at the outer lexical level global Sorry it was late Fm Byron P Peebles I found the whole thing enlightening I am relearning C after several years away for it It sent me back to the books since I could vaguely recall some guaranteed zeros nulls and some guaranteed trouble One thing this all did for me was force me to locate the books No small task around this office Brothers MSC guarantees that it will unitialized external and static Fm H zero all variables Fm Les Mikesell to LDOS Support That quote can be taken as an absolute guarantee that globals and static values equivalent to the permanent variables mentioned in the AT amp T book are always supposed to be initialized to zero K amp R s comments about style may be safely ignored Lint doesn t even object to this assumption A sample program test for uninitialized global include lt stdio h gt int x ee O T G e E ad A ii r P Ndal O To if y 0 puts Y 0 Kl return 0 bel a a Tae Pap The lint output o OR ga OORE function returns value which is always ignored puts Note that there is no complaint about x being used before set A test run on the 3B2 showed both x and y as zero though the y value was of course accidental MISOSYS P
18. OUTCF 5323 Fl 00081 POP AF 5324 C30000 00082 OUTP2 JP 5327 AF 00083 OUTCF XOR A 5328 CD0000 00084 OUTPL CALL 532B D5 00085 KBD1 PUSH DE 532C CD2B00 00086 GALL KBD 532F Dl 00087 POP DE 5330 FEOD 00088 CP ODH 5332 20F7 00089 JR NZ KBD1 5334 Fl 00090 POP AF 5335 AF 00091 XOR A 5336 C9 00092 RET 00093 0029 00094 LENGTH EQU PAWSTRT 5200 00095 END 5200H Let s take a look at some of this code in more detail The first part of the program from line 1 through line 35 should be pretty obvious The three tests at lines 26 32 are integrity checks of the device being filtered Note that I am not going to discuss the ramifications of filtering in general that you can get from THE PROGRAMMER S GUIDE The code from line 36 through 40 POP IX LD L 1X 1 LD H 1X 2 LD OUTP 2 1 HL LD OUTP1 1 HL pops the address of the DCB originally passed in register DE by the FILTER command into register IY Then the current driver address which could be a filter entry point is picked up from the DCB and put into register HL This vector is then poked into the pagepaws filter code so that the filter can chain to the current device routine The code from line 41 through 56 relocates the pagepaws filter code to high memory Note that the actual filter code itself is in lines 71 through 92 The rest of the program consists of messages and error handling routines The front end of the filter PAWSTRT JR START OLDHI DW DB
19. PRO WAM now resident to bring up a calendar a card file Letters to the Editor Volume I iv and a bringup file Steve Milliken s excellent SHELL runs fine with a short patch supplied by Steve and I have what has to be a VERY good WP Setup Might add I ve had the Model 4 since sometime in 82 It is the old non gate array model that I ve added a beefed up power supply and two teac DS DD drives to TANDY aka Radio Shack True I have not always been treated like a returning prodigal son each time I walk into the store and I have done most of the hardware hacking and repair myself Often tuning to maximum smoke I have met the usual battery salesman in many RSCC s but lets face it folks they make darn good machinery and without them we wouldn t have a thing to grouse about would we THE PRICE OF DISKS Have found that usually cheaper is better Think more depends on the condition and quality of your drives than on how much you spent on the media Have had 3 00 a shot disks with pretty jackets fail to format on a PC clone with a wobbly drive It is not difficult to maintain and adjust your own floppy drives and you otta learn Will never be able to get really precise like a mechanic with expensive test racks but you can keep it in the ball park I have bought nothing but the 29 specials for some time and have had no trouble I really think that the Model 4 and LSDOS could format a flat rock if you get the wholes punched right
20. available will determine how hard it is how many instructions it takes to accomplish a particular data movement And the amount of memory addressable by the processor determines the possible size of programs and data which may effectively be used with it I A Comparison of Instruction Sets First a look at the 2Z80 following registers A B C IY and SP Editor note registers I and R are The Z80 has the Di Es 8 L IX two additional special purpose registers which are not usually dealt with by the applications programmer The single letter registers are 8 bits wide IX IY and SP are 16 bits Additionally certain combinations BC DE and HL of 8 bit registers may be treated as one 16 bit register The two letters of the 16 bit register name corresponds to the simple concatenation of the two 8 bit registers involved HL for example is registers H and L made up of The addressing modes of the 2Z80 are not symmetrical that is certain operations may use only certain registers and not others One example common to most processors is the The Programmer s Corner THE MISOSYS QUARTERLY SPRING 1987 280 and 8086 Additionally 5391 s Volume I iv exclusive use of SP for stack oriented operations such as PUSH POP and CALL Also most arithmetic operations are oriented for use with the accum lator or A register or the HL register pair You cannot for example directly
21. define PRINTLUN 3 program dcplot Plot Dolph Chebychev window parameters dimension h 128 print 1000 1000 format No of points in window output LUN s read 1005 n lun lunl 1005 format 3i10 if lun gt 4 call open lun plot dat 256 write lun 5000 5000 format Windowing Tradeoff _ Two Way Sidelobes _ Lobe Broadening if lun gt 4 call open lunl plotl dat 256 write lunl 5010 5010 format Windowing Tradeoff Two Way Sidelobes _ N Lobe Crossover dB do isdb 20 100 sdb isdb call dolchb n sdb h sbf dbd write lun 1010 sdb sbf 1010 format 2g15 6 write lunl 1010 sdb dbd end subroutine dolchb n sdb h sbf dbd Dolph Chebychev window weights Inputs n Number of weights sdb Sidelobe level gt 0 for weights h i lt 0 for sbf and dbd only Out puts h Weights not computed if sdb lt 0 sbf Sidelobe broadening factor 3 dB width 1 n dbd Bin crossover with FFT or beam crossover with delta sine l n Frequency response is T n 1 q cos PI f s where T n 1 is Chebychev polynomial of order n 1 f is fraction of sample rate F q is constant gt 1 found from q cosh acosh s n 1 s is sidelobe height as amplitude ratio dimension h n x 256 Announcing New Products 7 THE MISOSYS QUARTERLY SPRING 1987 Volume I iv asdb abs sdb q qfroms asdb n Compute q q3db qfroms asdb 3 0103 n Compute broadening factor sbf 2 n acos q3db q PI dbd 20 alogl0 dcf n q
22. for a friend who is a social worker at a nearby hospital He required some method of being able to compile data on successful methods of treatment The example in his case was that the title was the name of the patient and the data for the txt file was the treatments provided e g drugs ect accomodation help welfare payments etc with all the data entered by specifing treatments recived he could find the patient As this is an ongoing method updating had to be quick and easy The example that I have included with the disk contains the references to all your previous NOTES the LDOS publications and now TMQ I realise that BASIC is not the best language for such a program but looking around I could not see any other program which would allow such a file structure since the size of the search is only limited by the disk space versafile is about the next closest program that I know about perhaps nobody has ever done it before The biggest problem I have with it is the search time perhaps EnhComp would speed it up for cooks state the ingredients and the program will sort out the possible recipies imagine Gargleblasters You have my permission to include it in contributions public domain but rights reserved if your laws are as I understand them Because of the extent of the contribution it will appear only as part of DISK NOTES 8 Ed Applications for the User THE MISOSYS QUARTERLY SPRING 1987
23. should have memdisk I thought I had already added that to the README TXT file LDOS Information Volume I iv I 11 double check it was on April 28th Also the BOOT command will be inoperative unless I patch SET2RAM to alter the first 5 or so bytes of the ROM image Problem is that BOOT does a RST 0 however in the RAM image mode that doesn t bring in the BOOT code since the ROM startup code has been de activated It s on my low priority list to add such a patch With SET2RAM installed you have to push the ORANGE button to invoke a re BOOT Fm Gary Phillips Just wondering Roy The Model 4 hardware interface package arrived and like all your products works beautifully Now since I m using a 4P can I install SET2RAM then dump the ROM image to a file and avoid having to reinstall every time I boot up in model 3 mode Once SET2RAM is applied to the 4P what I would like to do is dump the memory image from 0000H through 37FFH to a file call that file MODELA III and have the BANK routines load into the system along with the ROM image rather than going through all those motions each time I boot Does SET2RAM have to initialize itself with other parts of the DOS this is probably a dumb question how could it not or will the functions work if they become a permanent part of the image I am planning to use the software on my BBS which runs on a 4P I want to use the MEMDISK as a system disk for faster overlay op
24. 00 00 00 ASCII to binary conversion DO3 E9 C3 C8 68 F03 E9 83 18 E6 D08 F4 83 57 C3 BA 63 F08 F4 00 00 00 00 00 check step value for initialization DOO D6 06 F00 D6 6E RSH61 FIX 05 27 87 Patch to RSHARD6 DCT Apply via PATCH RSHARD6 DCT RSH61 test direct invocation DOO 1A C3 09 39 F00 lA CA 81 35 D09 39 CB 5F CA 81 35 C3 11 30 F09 39 00 00 00 00 00 00 00 00 enable JCL D00 9C C3 11 39 F00 9C B7 20 27 D09 41 B7 C2 BA 30 C3 96 30 F09 41 00 00 00 00 00 00 00 ASCII to binary conversion D04 0C C3 18 39 F04 0C 83 18 E6 D09 48 83 57 C3 DD 33 F09 48 00 00 00 00 00 check step value for initialization DOO B1 06 F00 B1 6E RSH62 FIX Patch to RSHARD6 DCT 06 11 87 Apply via PATCH RSHARD6 DCT RSH62 Corrects installation into high memory when low is not available DO2 CD CD 1D 39 F02 CD 22 2B 35 DO9 4D 21 00 00 45 3E 64 EF 22 2B 35 C9 F09 4D 00 00 00 00 00 00 00 00 00 00 00 SDAT53 FIX 04 14 87 Patch to EnhComp SUPPORT DAT Corrects USING on certain single double fractional values For LDOS 5 3 PATCH SUPPORT DAT D60 14 21 F60 14 C2 For LDOS 5 1 PATCH SUPPORT DAT D60 14 21 SDAT62 FIX 04 14 87 Patch to PRO EnhComp SUPPORT DAT Corrects USING on certain single double fractional values PATCH SUPPORT DAT D5F BF 21 F5F BF C2 e Ser com sed od The Patch Corner 101 The Patch Corner Volume I iv THE MISOSYS QUARTERLY SPRING 1987 SYS6D FIX 04 23 87 Pat
25. 08 D24 00 20 F24 00 48 D24 10 C3 F24 10 21 D24 20 08 F24 20 4A D24 30 E5 F24 30 43 D24 40 3B F24 40 44 D24 50 CD 00 00 06 C8 68 Fl 00 22 CD 1C 87 7C 00 11 44 Ic 06 4A 18 Cl Fl B7 11 2E AB AA 00 28 88 C9 22 58 97 4A 93 B5 19 2A E5 4A 00 2A 75 21 E5 C5 04 B8 49 49 39 E5 2A CD 68 44 93 E5 Fl CA 6E 48 CD 21 39 54 2A SA C1 D5 00 7C 03 03 7E E5 66 A8 88 CD Fl CD 22 OA 26 44 08 06 CD 44 54 44 21 CD 19 B5 00 00 23 E5 88 81 2A 54 22 97 54 4A 00 E5 A9 00 5C 11 44 CD SA 90 CD C4 A8 AA B6 CD 22 22 48 80 54 93 44 22 E5 CD Fl 39 B8 09 11 JE 44 42 3B 42 49 49 28 9B 58 56 44 22 44 D1 7C 54 CD 08 2A 7E CD 00 09 B8 CD Fl B7 92 05 05 13 52 44 44 7C 48 7C 7C B5 44 F8 A9 58 23 54 19 00 2A 9E Fl C5 00 AUTOIN62 FIX patch MC62 CMD D19 CE 23 F19 CE 53 DIF 72 CD DIF F4 18 D20 57 18 D20 6A C9 D20 70 BB F20 70 BB D20 80 01 F20 80 01 D20 90 CD F20 90 2A D20 A0 2A F20 A0 F1 D20 BO EF F20 BO CD D20 CO F 1 F20 C0 99 D20 DO E5 e an an a ee The Patch 46 46 Corner 05 05 46 Fl CD 2A D1 FI 00 00 46 E5 E5 E5 Fl 99 FA 2A 3D CD 86 3F Fl Fl 46 46 CD Fl CD F5 21 Fl 28 3F 64 86 A2 E5 El El 3F Fl 63 ci AB E5 ED 18 E5 23 C3 A2 Fl CD E5 E5 CD Fl 99 El 6E Cl 28 21 DD
26. 23 76 7E E5 23 45 CD 45 21 DB 00 A8 49 Ol 00 A8 49 00 AA 49 O1 00 AA 49 Fl CD 3E B5 44 B5 B5 CA 11 43 Fl 44 B6 B8 6E 19 54 B8 CD D5 00 RND 05 19 87 allow arbitrary expression as initializer of an automatic scalar variable 00 05 00 1B 46 04 00 09 46 03 00 09 46 02 00 04 46 01 00 04 46 00 05 00 4B 46 04 00 32 46 03 00 32 46 02 00 28 46 01 00 28 46 FIF 72 CD EF CD 54 01 CA AF 20 20 OA 08 Fl 2A 22 20 7C 62 6E 44 2A 54 CD 00 3B 80 32 Ic 32 OB OA 4A 00 7c 58 66 06 B5 E5 26 11 56 B8 90 69 22 68 4A 68 2A D5 2A 19 B5 44 88 CD 28 CD 00 04 44 ED 42 Fl 48 88 2A 88 48 CD 54 7E 28 22 C3 3F 05 1A E5 00 E5 5B Fl CD 44 2A 48 7C 44 87 44 FE 11 66 Ic 4A CD 43 CD 19 2A 56 Fl F23 A7 CD A8 81 22 56 44 21 01 00 22 21 66 44 B5 ES 93 11 09 2A 88 4A C3 3E FI 1A CD 54 44 CD Volume I iv F24 50 54 B8 DI CD 2E B8 7C B5 C4 42 92 FI C3 D9 45 OO FIF F4 Fl Cl EI E5 C5 F5 7C B5 C4 50 93 C9 00 00 00 00 F20 57 F1 Cl El E5 C5 F5 7C B5 C4 5B 99 F5 F5 F7 Cl 2A E5 Fl 21 E5 2A E5 CD 45 E5 E5 Cl El 2C C5 C9 66 E5 2A E5 B5 D5 21 2A E5 3F F5 21 3F E5 3F CD Cl CD 2A 2A C5 E5 E5 06 CD CD E5 38 E5 E3 3F 3F F5 CD CD 00 F7 19 cD 69 CD 7A E5 18 7C 88 Di 39 Cl 53 88 Fl 88 Fl CD QQ vs F20 6A C9 21 2A 3F E5 CD cl E3 2A 2
27. 3050 3051 3053 3054 3055 3059 305C 305F 3062 3063 3066 3067 306A 306C 306F 3043 30D6 30F4 310A 310C 310E 3115 3117 3119 311B 311C 311E 3120 3122 3124 3126 3128 3129 312B 72 2B 73 FD23 FD23 18E7 ED5B0C31 213431 013100 EDB8 EB 3E64 EF 23 DD360046 DD7501 DD 7402 210000 C9 21F430 DD 21D630 3E0C C31500 OA 43 48 4D 00051 00052 00053 00054 00055 00056 00057 THE MISOSYS QUARTERLY SPRING 1987 9 00058 00059 00060 0006 1 00062 00063 00064 00065 00066 00067 00068 00069 00070 00071 00072 00073 00074 00075 00076 00077 00078 00079 00080 00081 00082 00083 00084 00085 00086 3 t RXEND LD DEC LD INC INC JR Volume I iv HL D HL HL E IY IY RLOOP sAnd put back Loop till done Move filter LD Xi i k gt k k 9 VIASET NOMEM S IGNON NOME M VIASET 00087 180D 0000 06 50 47 50 0000 0000 2005 79 FEOC 280G DDES DD2A1531 3E 14 EF DDE 1 c9 00088 00089 00090 00091 41 57 00092 00093 kahm PAWSTRT OLDHI 53 MODDCB 00094 00095 00096 00097 00098 00099 00100 00101 00102 00103 00104 00105 00106 00107 00108 The Programmer s Corner START x Ad mes CHAIN RX01 DE OLDHL Destination address LD LD HL MODEND Last byte of module LD BC LENGTH slength of filter LDDR EX DE HL M
28. 6 3 1 2 1 1 2 They like the Dancing Demon very much add EGA A 65 30x 237 PA STOW N S W 2241 Ap ise DM LE phe EA EEE cit kad aaran i EE Letters to the Editor Volume I iv LDOS Information 80 MICRO review of LDOS 5 3 Fm MISOSYS In response to the terribly inaccurate review of LDOS 5 3 which apeared in a recent issue of 80 MICROCOMPUTING the following letter was submitted to their editor Mark Reynolds Review Editor 80 MICROCOMPUTING Rarely have I had the necessity of responding to an 80 Microcomputing review of a MISOSYS product But the June 1987 review of our LDOS 5 3 had so many inaccuracies that I am compelled to attempt a correction The statement Logical Systems Inc the company that developed LDOS has once again updated LDOS and added some new features is far from the truth LSI did originally develop LDOS however LDOS has been a MISOSYS product since March 1986 and the entire 5 3 update was designed and prepared by MISOSYS The product was announced in a press release printed in 80 Microcomputing in its April 1987 issue LSI had nothing to do with LDOS 5 3 Although author Feldman may think that the major change is the time stamping introduced into the directory Structure 5 3 was developed primarily to extend the directory dating to 1999 Time stamping was just an ancillary feature There would have been no 5 3 release generated without the need to extend directory dating past
29. 6000H can easily use any DOS library command Bit 2 says that SYSI is requesting line input That s the only time that the bit is set MINIDOS s CMD R uses this to sense if the key request is from LDOS Ready Keyboard filters which require system overlays other than SYS1 can use this bit to see if they have to force a reload of SYSI Bits 3 are adequately explained in THE MISOSYS QUARTERLY Issue Tii LDOS Information THE MISOSYS QUARTERLY SPRING 1987 26 Volume I iv 300 rpm keyboards and other matters Fm Willi E B Wald My LDOS 513 manual s index refers to a page 4 14A in the MINIDOS section but there is only 4 14 Was 4 14A in an addendum which I missed I am 100 behind you on your MS DOS route for your BUSINESS I have bought my LAST computer from Tandy because of what they are doing to me as a Model I and Model IV customer What you had to say about 80 Micro was also of great interest to me they still owe me a T shirt that was promised me with great fanfare in an editorial last year Also by SELECTIVELY quoting from my letter to the editor Eric Maloney made me sound as if I were advocating software piracy What I said was that Tandy OUGHT to give away with each computer purchase an assembler word processor and a spread sheet program If I were Tandy s marketing manager that is how Tandy COMPUTERS would be sold because the software development need to go the cost for those programs had long
30. DI I won t give an exhaustive list of all the valid combinations here but remember that they are limited For example SI DI is NOT a valid addressing mode In general SI and DI are treated as index registers and BX and BP are treated as base registers A base plus a base is not allowed nor is an index plus an index Nor are combinations of more than 2 registers allowed You may specify either an 8 or 16 bit signed offset akin to the Z80 IX IY disp8 in any of the addressing modes Note that the general 8086 opcode name for data movement is MOV as opposed to the Z80 s LD The 8086 is probably best thought of as a 16 bit processor Most instructions can deal with 16 bits at a time and the external data path of the 8086 is 16 bits wide It would appear on the surface that the 8086 is a 16 bit processor in terms of memory addressing as well After all the maximum register size is 16 bits long right Here we get into the most complicated and vexing aspect of the 8086 family This is known as segmentation Note the registers The Programmer s Corner THE MISOSYS QUARTERLY SPRING 1987 aa e Volume I iv listed above that haven t been mentioned yet CS DS ES and SS The 8086 address bus is not 16 bits but is actually 20 bits wide This means that it can address 2 20 or about a megabyte of memory This is 16 times the memory capacity of the Z80 and this means that more sophisticat
31. DRIVER TRSHD3 Let s briefly explain these parts The SYSTEM designates the DOS command The parameters are in parentheses The DRIVE parm designates the logical drive slot to be assigned the HD partition DISABLE tells the DOS to disconnect any disk driver currently assigned to the drive slot if any That s because the HD driver initialization will not allow you to assign a partition to a drive slot which is already enabled The DRIVER parm specifies the name of the hard disk driver Obviously the driver name noted above would have to be changed if you are using the TRSHD5 DCT driver Also the drive number d assigned in the above command cannot be it must be in the range 1 7 If you want to set up a partition to be your system drive drive you first have to assign it to some other drive slot then swap it later What happens next is the installation part of the driver begins asking you a lot of questions The question sequence is identical for the TRSHD5 and TRSHD6 drivers but not the TRSHD3 driver The TRSHD3 driver will ask you for the drive address If you have one hard drive connected that would be drive 1 additional HDs would be 2 3 and 4 This is actually determined by a jumper on the hard drive bubble itself Whoever installed your hard drives should have them jumpered as 1 2 3 and 4 in that order although any arrangement works if you tell the driver what is the jumpered number Yo
32. LD EOU LD CALL LD SUB CP JR LD ING LD LD CALL PUSH BIT JP PUSH POP LD LD LD LD LD INC LD CALL POP LD ADD LD JP CALL LD EQU INC CP JP NZ BADDRV A HL 30H 8H NC BADDRV DRIVEW A DRIVER A CG A GCKDRV NZ BADDRV C WPDRIV GETDCT A IY 9H DIRCYL A D A E OH READS A GATCD 3 A NZ NEWTYPE HL MSGOLD GDSPLY EXITOK D OH S 1H E 3H READS A DIRERN 2H 21H C FILES A 20H A DIRSIZE A E 2H READS HL 4 HL Z NEXT IX HL A 96H IX 12H A A 42H IX 13H A A CHANGD A CHANGD A KEY NZ ABORT HL A L A 20H L A NZ AGAIN WRTSYS A 0H 1H E E NC GET 57 SPRING 1987 Volume I iv scheck for drive 0 7 sstore drive for writes sand for reads scheck the drive in C sif drive is not available sif disk is write protected selse get DCT address in IY DCT 9 is the DIR cylinder store the DIR cylinder number sand prepare to read it sread GAT type of disk byte sbit 3 set for new date time stell that is it old type sand do not bother any more sread sector 3 of DIR cylinder ERN of DIR SYS find the number of sectors to be read sread sector 2 of DIR cyl read a sector sis the record in use first byte of PASSWORD sin place ssecond byte of PASSWORD sin place snumber of files updated sin place srestore pointer sadvance to next entry slast entry in sector si so writ
33. LS DOS 6 x environment since the early days Device filters are kind of neat they provide an opportunity to alter the behavior of a device For instance both LDOS and LS DOS come supplied with a key stroke multiplication filter KSM FLT and a printer filter PR FLT or FORMS FLT both of which alter the utility of the KI and PR devices respectively Under the LDOS 5 x environment Logical Systems and subsequently MISOSYS have made many other filters available in the FILTERI and FILTER2 program products now bundled into The Mark III Collection At one time MISOSYS offered a proportional spacing filter for the The Programmer s Corner Volume I iv DW II There also were filters as part of the MSP 02 package For LS DOS 6 x users the LS UTILITY disk still available has provided a KSMPLUS filter The PRO ESP package which we used to sell also included filters it s bundled now into The Mark IV Collection I would hazard a guess that our Compuserve forum also houses some filters THE PROGRAMMER S GUIDE TO TRSDOS 6 once again available from MISOSYS for 25 3S amp H even lists three complete filters for DOS 6 in its Appendix On the other hand more filters were available for LDOS 5 x than were made available for LS DOS 6 x Why is that Well for one thing the architecture of a filter was changed to accomodate the altered architecture of the device handler in LS DOS 6 x I changed the structure of the devi
34. LS DOS Information Volume I 1v Could still insert one but then it s their fault I m not suggesting a change that could certainly barf up something that s already out there My suggestion is that if indeed the GDSPLY routine has changed to terminate on a zero as somebody mentioned they observed it to that it should either be fixed or the documentation changed to reflect the new behavior yes I did actually put au in there but saw it and deleted it before sending lt lt grin gt gt Assuming that the doc is changed because this is an official alteration from the spec LSI should do the documenting Fm MISOSYS To my knowledge DSPLY has not been changed to terminate on NULL It s not in THE SOURCE and LSI stated that low core was not touched Fm MISOSYS Folks I just passed along to Bill today 29 May 87 a little patch which would correct the DEVICE command of 6 3 if correct is the proper word What it does is to transfer only the first character of the device name if the second character is NULL Here it is DIF D4 ED AO 7E B7 C8 F1F D4 7E 12 13 2C 7E The patch of course is to SYS6 SYS LSIDOS That s really the only place that should have any problem with one character device names if you really think that it Also DSPLY has not been changed it does not terminate on a NULL I believe what happened to the strange display from the nulled 2nd character was that the I char device name was the la
35. Record 0 byte CB hex This is the first sector of the directory cylinder In both of these locations you will find the value 51 hex 62 hex or something like that which is a record of the DOS used to format the disk 51 hex means 5 1 x 62 hex means 6 2 x and so on SET2RAM and memDISK driver Fm Gary Phillips Does the MEMDISK driver in your model 4 IDOS enhancement package work with memory expansions installed XLR8er and or AlphaTech I m planning to order the package anyway for use on my non XLR8er machine but it occurs to me to ask since I note that the LS DOS MEMDISK DCT now crashes the system under XLR8er Found it out by accident when I invoked a pre XLR8er JCL file LDOS Information THE MISOSYS QUARTERLY SPRING 1987 1 BOOT SYS Record pi Volume I iv Fm MISOSYS To begin with the MemDISK which is part of the Hardware Interface Kit only permits a 32K or 64K RAM disk using either one or both of the extra banks in a standard Model 4 banks 2 3 Second your Model 4 MemDISK should not crash at all if you have either an Alpha Tech board or an XLR8 board installed MemDISK might crash if you have Someone elses RAM drive software installed and you force Memdisk to install On the other hand some other RAM disk software should be locking out banks 2 and 3 if they are using it I suspect some other problem local to your system Fm Gary Phillips OK I IL have to look at the MemDisk th
36. SAVE command Creating a NULL length file Fm Jack Lottey I recall under LDOS 5 1 4 I could CREATE a file with records However with LDOS 5 3 I always get a bad parameter msg when I try to create a file with records I am usually doing this in one of the sections of Profile III HD version say DA2 or DA3 What is the proper proceedure in an application like this Fm MISOSYS You can t use CREATE to generate a NULL length file Fm Jack Lottey In the Profile III HD database I needed a file called database da3 Without that file the program would not open the opening screen would note that that file was missing In LDOS 5 1 4 I would enter CREATE database da3 d Records 0 and Voila I had a file that opened up the whole PROFILE database With LDOS 5 3 I have discovered that I can create a file if I specify Records 5 and I suspect any other number but the command CREATE GIFT1987 DA3 5 RECORDS 0 comes back with the response Bad parameter Fm MISOSYS It does that under LS DOS 6 3 and TRSDOS 6 2 as well We made it compatible CREATE can no longer create a NULL file which is what you were trying to accomplish You can just as easily create a NULL length file by the command BUILD GIFT1987 DA3 5 lt ENTER gt then hitting BREAK as the next key entry 2 kor Y 3 oe z Ki gt w og 204 iyl ev Sat TAA hi oN pa DSTOW A 5 matti eer cee RTI anal EDS Inform
37. SETCOM uses an identical set of parameters as the SET command which originally installed the serial driver RS232T In fact the serial driver defaults to 71E that s 7 bit word length 1 stop bit parity EVEN DTR ON a common setting for most folks with just a simple SET CL RS232T command from LDOS Ready That s lengthy Besides most users keep a few JCL files around to automate their device setups MS DOS users have found that the batch language is extremely useful It s been that way with LDOS since 1980 Now even without a hardware clock that is keeping accurate time relative time is just as important The significant reason for keeping a time stamp is to enable checking which file written today is the most recent when a file is written more than once A week s use of time stamping will turn anyone into a convert And TRS 80 time keeping is not that inaccurate Perhaps the positive statements in Feldman s review attributed to the 4 star rating On the other hand the inaccurate statements concerning risks attributed to the changes in directory structure may cause a few folks to needlessly shy away from this important release of LDOS And I can t imagine how the update was attributed to Logical Systems Perhaps the personnel shufflings going on at 80 have resulted in a green staff no pun intended to Wayne Old Passwords Fm Adam Rubin The changes in the password structure for 5 3 and 6 3 and the cha
38. START FNAME FNAME DB PAGEPAWS is the standard module header as documented for the LDOS environment It consists of an absolute branch around the data area followed by a word of storage for the previous value of HIGH to be inserted by the filter installation code followed by the one byte module name length followed by the module The Programmer s Corner THE MISOSYS QUARTERLY SPRING 1987 67 Volume I iv 3g0 if so selse restore sdo regular driver sbe sure CF reset send TOF char ssave during key input see if CR get key if not selse ready to return slength of filter code name This structure is required for all LDOS 5 x filters The guts of the filter is the following START JR C OUTP 2 PUSH AF LD A C CP OCH JR Z OUTGF POP AF OUTP2 JP OUTCF XOR A OUTP1 CALL KBD1 PUSH DE CALL KBD POP DE CP ODH JR NZ KBD1 POP AF XOR A RET PAGEPAWS is an output filter Since input is flagged when the carry flag is set the code starts out by bypassing the form feed test based on the CF To be totally correct about it the filter should also test explicitly for the control call which is flagged by a NC NZ flag state Note that the jump and call vectors denoted by were previously stuffed with the original driver vectors as obtained from the DCB of the filtered device The filter code is straightforward The flag state is saved PUSH AF to avoid altering the flag conditions in c
39. TRSDOS VERSION 6 Also to assist in the binding operation we acquired is the little with those a GBC Image Maker 2000 This machine which binds documents plastic finger thingamajigs We ve produced a few GUIDES already and I like the result The binding allows pages to turn freely lets the book lie flat when open and is not much of a burden to install So if you are still looking for a copy of this valuable book we got it The price of course has gone up from when we used to get it offset printed As I stated in TMQ I i The Blurb offset printing of The Guide was only economical when the print run was at least 500 At a print run of one per we will be selling The Guide for 25 plus 3 S amp H US Projects at MISOSYS Inc reported on the imminence of three products RATFOR M4 RATFOR 86 and DSM86 Well the first two made it out Unfortunately the free lancer doing our port of DSM4 over to MS DOS got tied up with other work forcing us to bring DSM86 in house It Last issue I The BLURB Volume I iv will be delayed since there are a few other projects going on This issue announces PRO WAM Release 2 By the time you read this it will have been in beta test for about two months getting a good shakedown Also as soon as I put the wraps on this QUARTERLY I ll start the massive rewrite to the PRO WAM manual I am quite proud of this new release it has virtually everything that our users have been reque
40. Tandy s DOS as well EnhComp does not attempt to detect what DOS it is running on so it could check to see if PR FLT is in use and then use those parms In fact if it did do that it would then be incompatible with 5 3 And then I would not be changing it either My best suggestion would be to give you zapping instructions to the SUPPORT DAT library to totally deactivate EnhComp s print formatting but I really doubt that that is what you want MISOSYS Products Tidbits 81 THE MISOSYS QUARTERLY SPRING 1987 Volume I iv Fm Bob Connors Roy I have been working on the following code for 4 days Reason it will not compile This is a stripped down version of the code all REM statements removed so that I can fit it into this message When trying to compile using the RUN command from CED invoked by S I get an error message Source file not in directory When using the command line BC RECEIVE BAS WE NO from LDOS Pass 1 displays two times and the system dies I have checked the syntax over and over and cannot find any errors Do I have a defective compiler I have applied all patches to the EnhComp package through TMQ I iii By the way with the included the system just dies the compiler is invoked My EnhComp serial number is 00027 and the errors occur under both LODS 5 3 0M and LDOS5 1 4M Thanks for your help Code follows trimmed down to the bare essentials ed REM statements no matter how
41. You can t open files without first specifying the number of buffers with ALLOCATE It does support a lot of other things that Microsoft s BASIC doesn t like a built in array sort graphics random files greater than 256 LRL strings greater than 255 characters etc You may be fighting it instead of trying to work with it I would suspect that most programs would need some degree of modification MISOSYS Products Tidbits Volume I iv Finally if you still run up against a brick wall send me a copy of a program on DISK which you cannot get to work Fm James J Chouinard I am experiencing problems with the output from a program compiled with your basic compiler The output is not consistent Variable K is supposed to be 02 and is okay when output without using PRINT USING is used Any help you can provide would be greatly appreciated Fm MISOSYS We have been the problem and also narrowed its scope to PRINT USING a certain fractional power of ten and multiples of it For instance PRINT USING HE 0 01 prints as 1 00 02 as 20 08 as 80 Because of the complexity of the floating point to ASCII conversion program under an edit mask i e USING I referred this problem back to the author of EnhComp Phil Oliver Phil has isolated the cause of the USING problem which you reported concerning our EnhComp compiler A one byte patch to the SUPPORT DAT file should correct the error This is SDAT53 FI
42. a new application which is used to establish and maintain a simple list of items to do Now I wanted to keep Brenda s todo list together with mine so I designed it with a field to indicate who does it This wammie generates a TODO DAT file which is similar to a BRINGUP file Allows priority of 1 8 a who does it value of 0 15 where a value of 0 is considered global more on that later and a text field of 29 characters The add date is added to the record Commands allow you to add an item mark an item as done which deletes it file any changes move to next or previous display page and list items A page displays the 8 items in a disk record rather than working like BRINGUP the display works like DIALER The list command allows you to print all records or the current record additionally specifying either all who does it or a particular who does it Any item tagged as who 0 will always appear on a listing Forget Hi s job jar TODO does it PRO WAM comes with an on line help facility which has been updated to include all of the features added in release 2 It can keep you from having to resort to the user manual when you forget a particular operation of PRO WAM Of course since every application has a command menu you will rarely have to resort to the manual once you have familiarized yourself with the operation of the applications you will use PRO WAM also comes with some improved utilities and so
43. again On RSCCs Fm jjkd Here s a story relating to one of the better RSCC s I ve been to It happened to be in Ontario Canada I went in on a lark as I was on a business trip and wasn t really equipped to carry anything major back with me I did walk out pleasantly surprised at the employee s attitudes and capabilities along with the neatness of the store and stock and I got a Canadian RSCC catalog That had to be one of the best RSCC s I ve ever been in I meant to get the name of their DM and write out an attaboy letter but I never did get around to it Anyway to make a long story a bit shorter I took the catalog back to work with me A friend of mine was considering buying a 3000 but had been hemming and hawing for some time Me Hey Lynn did you buy that Tandy 3000 you were looking at Him Naw I haven t really decided yet Me With a very sorrowful look on my face Gee did you hear about the new price increases Him No Computer prices go down not up Me Yeah Somethin to do with Japanese chips or the yen or stomthin like that I heard Look I just got the new Radio Shack Computer ts lo akn I showed him the Tandy 3000 page in the catalog with the catalog folded over of course There the price was but in Canadian dollars of course something like 4000 without the hard disk vs 2600 in the US His eyes just sort of glazed over and the look on his face was priceless He didn t notice t
44. and input output libraries Also FORTRAN arithmetic statements are used including mixed mode arithmetic and the exponentiation operator In addition most keywords from the C language are also part of RATFOR this is not surprising since Kernighan played a key role in both language definitions The result is that RATFOR is a fully structured language in the tradition of ALGOL PASCAL C and ADA with the power and versatility of FORTRAN RATFOR is nearly always implemented as a translator A translator differs from a compiler in that a compiler translates raw source code to hardware dependent assembly language source code Translators are compilers which translate raw source code to source code for another high level language such as C Translators are common where portability of the compilers is important and are almost universal in the early stages of development of any compiler Compiler compilers such as YACC which use data files which describe the keywords and structure of a language as input data produce compilers for the new language in C source code Compiling the result gives a compiler for the new language RATFOR translators translate raw RATFOR source code into FORTRAN object code This has exactly the same advantages as translators to C in contexts where FORTRAN is universal and largely portable This is true in the microcomputer mainframe community today This can be to advantage to the engineering workstation user
45. be checked by EVERY program using memory Would you drive off in your car with better tires without looking at how much gas you had in the tank When I worked for AT amp T I had the necessity of using a company car occassionally We were taught to use a safety check list each and every time we took out a car The few minute inspection covered gas lights horn mirrors seat belts etc and the old Lincoln test that s shoving a penny into the tire tread to make sure Lincoln s head is still visible does that work with your pennies Sure a company may need a little more double checking when there LDOS Information Volume I iv are multiple users of a vehicle however if everyone made these kinds of checks you d probably have less broken down vehicles and less accidents KI DVR was designed for a Model III The DOWN ARROW key on that keyboard and on the non clustered ARROW Model 4 keyboards is adjacent to the LEFT SHIFT key The use of LEFT SHIFT DOWN ARROW for a CTRL key stems from the mechanical design of the keyboards in use on the machines Model Is and IIIs that LDOS was designed for To change to the RIGHT SHIFT DOWN ARROW for control apply the patch PATCH KI DVR DRIVER D01 EC 4E F01 EC 46 KI4 DVR which is part of the Model 4 Hardware Interface Kit makes use of the Model 4 s CTRL key rather than a 2 key combination the Model I and III did not have a CTRL key The LCOMM keystroke interface was not designed
46. bit at least Thanks again I 1l be back with my next few dozen questions on this soon Fm LDOS Support jjkd Transferring files in a raw or binary mode as an option is indeed useful Any movement of an Q LBR or now ARC file for example What is the format of the CNV file Is it actually executable code Perhaps a translation table definition file might be easier to change even though not as flexible If the translation table approach is used an input format like MAXLATE s to allow multiple byte output for an input byte would ease tasks like CR CR LF Fm Les Mikesell Yes it is also worthwhile to write the last sector first on the TRSDOS side so all the disk space will be allocated at once instead of accessing the directory for each gran Writing dummy data will also work if the real data for that sector is not available yet Fm Adam Rubin I have a related question or three Don t I always lt grin gt I discovered that the DOS floppy disk driver uses the directory cylinder value to determine when to start write precompensation so my MS DOS TRSDOS file moving project uses 20 14H LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 middle of a page or wherever it 37 Volume I iv for the MS DOS disk s directory cylinder It uses the DOS WRSEC and its RDSEC updates the current cylinder value Since some other DCT values are altered the program sticks 0 in the directory cylind
47. colum of Friday May 29 1987 June 1 1987 as printed in THE WASHINGTON POST Dear Ann Landers Will you please help settle a friendly dispute The loser has agreed to send a 50 donation to the winner s favorite charity Question What is the first day of the 2lst century My friend insists that it is Jan 1 2000 I say it is Jan 1 2001 Who wins Hugh W in Wilmington Mass Dear Hugh I called six very bright people and guess what Three said Jan 1 the year 2000 and an equal number said Jan 1 the year 2001 Each person went back to the year Christ was born and from then on it became unbelievably complicated I am not touching Letters to the Editor THE MISOSYS QUARTERLY SPRING 1987 18 s Volume I 1v this one It s one of those tricky guestions that can make you crazy Well I have been reading the Landers colum for quite a few years who said real men don t read Ann lLanders After the lengthy discussion which appeared in the last issue of TQ I could not remain silent Taking keyboard in hand I sent the following letter to Landers Dear Ann Landers After reading your colum for years I am male age 44 married with two children I wanted to offer my insight into Hugh W s question on the calendar date of the first day of the 2lst century I am in the computer software business and publish a magazine in support of my product line Just as you have found with your readers my readers gener
48. dutifully installed within the init code where it belongs Thanks Roy appreciate your patience Fm MISOSYS That s what we re here for LS DOS Information Volume I iv The LSI LS DOS Column Hello again from LSI Again in this column I would like to point out that any material contained herein is the opinion of LSI and NOT that of MISOSYS Any comments regarding the content of this column should be directed to LSI at PO Box 55235 Grand Junction Co 81505 303 243 7070 This time I want to stress the importance of sending in your registration cards Response has been fair but note that we will have to be very strict in applying the rules in the future UNREGISTERED USERS WILL RECEIVE NO SUPPORT Tandy will probably have mailed out update notices to registered TRSDOS 6 2 owners by now and all of those users that don t read 80 Micro or TMQ or check into the LDOS forum on Compuserve will finally be finding out about LS DOS 6 3 We expect to be rather busy here at LSI filling those orders The demands on the customer service department will probably go up accordingly To assure that the Customer Service departments time is spent on the right people we will have to verify that we have received your registration card All registration cards are entered into our database computer on the day received and are available in just a few seconds since they are indexed by the Customer Service ID If you have that number ready
49. e EnhComp BASIC compiler s e MC zy C compiler gt e e 8 e o a PRO WAM Window and Application Manager The PATCH C orne r A a e e e a Ld e e e Our Compuserve Forum e s o Copyright 1987 by MISOSYS Inc All rights PO Box 239 Sterling VA 22170 0239 703 450 4181 Spring 1987 13 20 32 44 45 47 56 59 62 62 64 71 78 80 83 96 96 103 reserved Volume I iv The Blurb Speaking about The Quarterly Well folks all kidding aside this is the Spring 1987 issue of THE MISOSYS QUARTERLY I know many of you are boiling in the hot sun by the pool Some of you are stocking up on cord wood see Letters Actually it should get to you about three months since the last issue Unfortunately I haven t been able to compress the time between issues to catch up on the timing of the QUARTERLY I do hope to shorten the time until the next issue by a few weeks so you re not gethering in the pumkins while you read the Summer issue Rest assured regardless of the timing a year s worth of subscriptions will get you four issues This Spring 1987 issue marks the completion of Volume I next issue starts Volume II Sometime in the near future we will be mailing out renewal notices to those of you who have subscribed for the first issue i e if your subscription gets expired after this I iv issue Pl
50. enter char and PRMSAV This solution actually discloses about 4 bytes which could be NOP d out By doing the 16 bit load into HL to begin with instead of the 8 bit load of the high order pointer the pointer is loaded to begin with Of course NOPing is unnecessary because the redundant code is not destructive I will use your second patch without alteration You had the only correct solution and the reason behind it I might add allowance for the NULL word at the top was overlooked Incidentally that NULL word is used to store a pointer to the continuation of a KSM string prematurely stopped because of a logical ENTER character See KSMA FIX in The Patch Corner And now for a word from our sponsor Seriously though you asked about the CFLAG byte at 4758H An experienced DOS 6 x programmer should have no trouble with this one However I should provide some insight for the non 6 x ers We ll expand on this in TMQ I iv as necessary Here s a brief intro Bit l says that CMNDR is executing In other words when a program begins execution it can test this bit DIR uses it to force O N because the sort buffer is not available CMNDR is used to invoke library commands and other transients which run in the library overlay region LDOS assumes this region to extend from 5200H 5FFFH Any program which needs memory beyond 5FFFH can test the CMNDR bit to see if continued execution is feasible Thus programs ORG d at
51. even the 32 bit 80386 are currently operated in modes which essentially make them The Programmer s Corner year or two with eS ee eee ee a 1 Volume I iv fast 8086 s This will change over the next now 1987 For the time being however there are no widely accepted operating systems which utilize the full power of the new instruction sets Therefore all remarks applying to the 8086 apply to the 80286 and 80386 operating in the 8086 mode There are three major divisions of comparison between the TRS 80 line and the IBM PC First of all there s the difference between the instruction set of the Z80 and the 8086 This is essentially separate from the individual hardware the processors happen to be designed The hardware such as video output and disk drives makes up the second category Third and last is a comparison between the operating systems of the TRS 80 and the IBM PC I m going to start at the beginning with the processor instruction sets and then lead on to the higher level things such as hardware and operating systems I believe an instructive way to compare two processors is to look at the following major categories the number of registers a processor has the flexibility of its addressing modes and the amount of memory directly addressable Generally the more registers a processor possesses and the bigger number of bits they are the faster it is Similarly the addressing modes
52. fixes for this Fm MISOSYS Double check to see if your SYS6 module is stored in more than two extents If you backed up the new 5 3 to the old 5 1 4 on your hard drive it s possible that the SYS6 module went to more than 2 Also check SYS7 SYS as it must also be in no more than 2 extents That may be your DEVICE problem It s possible that DIR could do that I don t treat it as a bug RSHARD from JCL Fm Jeffrey Kline I finally got some time to sit down and work with my system in getting the new driver package installed RSHARD6 DCT The problems are as annoying as they come First off for some unknown reason I cannot install the driver via a JCL I keep getting an INVALID ENTRY FROM JCL and the JCL stops I tried every angle I could think of to make it function under the JCL configuration file I had created on the disk The second is not too bad but I was hoping to be able to partition off the drive by head only as I don t really need to be concerned with the amount of directoy slots I use DISKdisk and PDS I finally wound up with a moderate configuration of 6 logical drives and partitioned by cylinder It would also seem that maybe the problem is only with TRSDOS 6 2 but not so I also ran into the same problem under LDOS 5 3 0 I wanted to not only keep a log of the configuration in text form but also use that very same text to re configure in the event of a missing or damaged CONFIG SYS file Why can I not use a JC
53. flag which was turned on only by the SET and SYSTEM DRIVER library commands Thus the test being performed here is to insure that this filter was invoked via a command such as SET XY PAGEPAWS this same It is worth mentioning here that The Programmer s Corner a 720 os value of HIGHS and pokes it into the resident filter s module header Lines 31 56 are new for this 6 x version filter however the routine 1s ome discussed a long time ago in the LSI Journal The routine corrects all fixed address references in the resident portion of the filter to reflect the correct value after the filter is relocated to high or low memory Thus the routine won t be dissected here On the other hand the code LD IX 0 40H OR 6 LD IX 1 L LD IX 2 H LD HL 0 RET is new for DOS 6 x and is essential for DOS 6 x filter installation Remember that the job of the filter installation section is to place the filter code into memory and set up the first three bytes of the assigned DCB At this point register IX contains the address of the DCB assigned Since the filter is expecting to be associated with a device which can support output and control calls PUT CTL the DCB TYPE bit must have bits 1 and 2 turned on Also a DCB indicates a filter assignment by having bit 6 of the DCB TYPE byte turned on Here the first instruction does all of that The next two place the entry address of the resident filter into the DCB driver
54. in that powerful easy to write structured source code can be executed on any machine which has a FORTRAN compiler Since uploading and down loading ASCII text and programs is standard in engineering workstations RATFOR can be translated at the workstation the FORTRAN object file can be uploaded and the final compilation and execution can be done on the mainframe This makes best use of the workstation as a file manager and application or software facility and the mainframe is execution of large optimizing execution of applications development best used in compilers and programs The simplest way to demonstrate the simplicity of RATFOR is to show a few sample programs The one selected is a simple program used to compute and plot tradeoffs associated with the use of Dolph Chebychev windows a weighting Announcing New Products Volume I iv technique widely used in communications to control crosstalk Examination of the program reproduced below shows that a simple relationship between the algebra used and the program is evident from the listing Also Statement labels are completely absent except for FORMAT labels Use of define and other C language keywords is important to the simplicity of the program The same program in FORTRAN would be much more complex in interpretation of the source listing Program tested on TRS 80 Model 4 define PI 3 1415927 Mathematical constant define READLUN 1 1I 0 units for console
55. it will be considered a protected character Most typically this would be a reverse video character however if reverse video is not in effect any graphic or special character greater than 127 will be considered protected and not output input This will be extremely useful with our revised CARDX card file application Export import now supports a logical ENTER character which will be translated to a hard carriage return regardless of the import export state CR or NOCR Thus a text phrase can be export as multi line even though it is a continuous stream of characters without a carriage return The logical ENTER character is tested before the protected character test thus values greater than 127 are acceptable for use as a logical ENTER The default logical ENTER character is 127d 7FH which can be entered by lt CLEAR gt lt SHIFT gt lt ENTER gt We ve added an ENTER parameter to the PRO WAM installation so you can change it to one of your choice It s even configurable with DEFAULTS We ve used the expanded features in the application manager to make considerable improvement in the existing applications and added some powerful new ones Here s what s been done in the application arena Many users requested that the CALendar application flag days which have a BRINGUP activity entered Well we added a facility to display an asterisk in the leftmost position of a day cell if there is an active BRINGUP record for t
56. oh well shows that the address must be followed by a comma No mention thereof and no example Quality control anyone I am running through the SVC examples at the back of the Tech Ref Manual Example pgm B has a series of prompts and GKEYINs and they keep grabbing control of my single stepping My next I entry gets handled as if it were a response rather than a DEBUG command Is there a ready way to regain control Another way to look at the problem in DEBUG what is the easiest way to simulate keyboard entry at a prompt and continue single stepping Fm MISOSYS The best way to satisfy KEYIN response is to use the C debug command When the KEYIN is called a C will be left on the screen If you are looking at the screen that C will remind you that you are keying in If it s a KEY call then you have to slowly use the C command otherwise you will get the result you mentioned Don t forget that C works exactly like I for everything but CALLs What you really need is to have PRO DD amp T s Debug Disassembler resident so that the screen display also includes the current instruction Of course PRO DD amp T is no longer available as a separate item but is available only bundled in the MARK IV Collection the e not Fm LSI command of Virgil I d the extended aggree that debugger is explained as well as it could be Just using the following for the example would have helped Eaddres
57. on the 80x86 family in the April 1987 Computer Language He LIKES segmentation but feels that Intel s IMPLEMENTATION of segmentation on the 86 family is botched at best and incredibly stupid at worst well to be honest he didn t say he liked segmentation but he did point out the advantages of a WELL DESIGNED segmented architecture Fm Jim Kyle First and foremost keep in mind that MS redefines the language to mean what THEY want it to mean just like Humpty Dumpty did With that settled we can discuss their meaning for relocatable In the 86 MS DOS Information THE MISOSYS QUARTERLY SPRING 1987 57 Volume I iv world of segmentation an address consists of two parts the segment and the offset In any program the base value for segment 0 can actually be at any of the possible segments in the memory space that is every address in which the low four bits are 0 and in this sense you are right that both kinds of program files are relocatable However in the COM file all of the OFFSET values within the program are fixed and cannot be changed and what MS means by relocatable is changeable offset values In the EXE structure the OFFSET values can be and are manipulated by the loader In more familiar terms it s the difference between a CMD file that always loads at 3000h or 5200h and a DVR file that must relocate all internal references The BIG difference between the Z80 and the
58. other end Our Compuserve Forum Volume I iv Unsqueezing Files Fm John Bode How do you unsqueeze a file that is squeezed when you download it and then it won t run Please I need to know how to undo the things that you do Fm Michael Johnston John to unsqueeze or de arc a program you must have one of the several archive or squeeze utilities They will extract the file to its usable condition Just download the following 1 ARC4 CMD 2 XARC4 CMD 3 USQ4 CMD 4 LU4 CMD 5 ARCH4 CMD The first two are the newest and preferred files the middle two are the old style that are used rather infrequently now The last one also fits into this catagory as it has been replaced by arc4 cmd If you don t want to download all that stuff who can blame you Just send a disk in a mailer with a return address label and postage to Michael R Johnston 155 DeKalb Ave Brooklyn N Y 11217 Better send two disks and I will copy ALL the files you need onto them OK ji Fm LDOS Support jjkd It aspdads on owthe file has been made smaller The breakdown is as follows AE Ext Program needed ACH ARCH3 or ARCH4 depending on OS ARC XARC4 or ARC4 202 USQ LBR LU and possibly USQ LQR Both LU and USQ ESS PROCES not PRO CESS The above utilities are available for TRSDOS 6 in DL 6 with some new arrivials here in DL 0 nata tia t epena rm s ete yg gt lays TE i 4 Ed Bs td 4 A X 1 4
59. product caused the static test program to be Using the linker variables in a option small initialized to zero so I guess it did the trick All this was started by me trying to get a version of fgrep egrep or grep up and running on the IV grep is a UNIX style pattern matching program Useful for finding strings in files None of the several versions I ve tried seem to work Don t really have the time as of right now to try and figure out why grep is a fairly complicated program The problem is not neccessarily PRO MC s I ve come across a lot of programs that take advantage of compiler feature like stack MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 on other machines 88 Volume I iv architecture or machine architecture The hard part is finding where in the code the program screws up Hey MISOSYS wanna do a source level debugger I m currently putting the finishing touches on a command line editor written with PRO MC The program is in the same sort of vein as the one that Hardin Brothers did in a column some months back but a good bit more extended Something like SCREENEDIT found on Data General s AOS VS machines Maybe I 11 upload it when I m satisfied with the way it looks It s a nice alternative for people who don t like to use SHELL s menus If anybody out there has a version of grep that compiles and runs under PRO MC I sure would appreciate it if you woul
60. rather than the result of the keyboard call and compare for ODH which would always return Z is returned back to the caller Thus if the printer became unavailable during the form feed the error status would be seen That s it I don t think that this filter conversion was excessive albeit it was a simple filter However more complex filters are just built up from simple concepts So all of you folks with copies of FILTER DISK 1 and or FILTER DISK 2 both of which were supplied with the full source code can do ports to DOS 6 x or develop some of your own filters Send your work in for a future QUARTERLY 8086 Assembly Language for Philip Oliver Z80 programmers by I was recently asked by Roy Soltoff to present a short introduction to IBM PC assembly language somewhat with the intent of targeting those of you who are already familiar with the venerable Z80 of the original TRS 80s Being the author of both ED ASM 86 for the IBM PC and ENHCOMP for the TRS 80 Model 1 3 4 I m fairly familiar with both processors The intent of course is to sspecifically asdress 80x86 code from an ED ASM 86 frame of reference So I ve agreed and here it is As a short prefix I ll note that from the viewpoint of the programmer the 8088 and the 8086 are essentially identical save for timing The difference is that the 8088 has an 8 bit external data bus the 8086 a 16 bit bus Furthermore the 80286 used in the AT and
61. s take a look at some of the features which have been added to the window manager PRO WAM now supports Model 4 reverse video detection and restoral at each level of application execution there are four levels For those of you using Visicalc or Multiplan you can now get into PRO WAM applications and return to your spreadsheet screen with reverse video intact Besides that PRO WAM applications can make use of reverse video in their screens we use it in CARDX More Model 4 folks are taking advantage of extended memory add in boards to gain additional memory banks With that have come reguests for better memory bank control in PRO WAM So we added a BANK parameter for PRO WAM installation to force PRO WAM to use your designated bank number rather than automatically search for a spare Even if you don t want to make use of that PRO WAM s automatic search will now look up to bank 30 PRO WAM is installed as a keyboard device filter Under Release 1 when you wanted to remove PRO WAM by a PRO WAM OFF command all keyboard filters were reset which required you to re filter the KI device with the other filters you had installed Under release 2 PRO WAM OFF now unfilters the WM device rather than RESET KI This leaves your KI device chain intact except for the PRO WAM removal PRO WAM now accesses application modules from a library not from standalone modules This keeps the applications from cluttering up your di
62. search for most users It s quite easy to ENABLE 2 and or 3 if you need to using the SYSTEM command I believe that LSI chose to DISKCOPY instead of the OFB for a few reasons One they licensed a QFB to Tandy in the Utilities Disk sold by Tandy under license from LSI that may be the prime reason Second diskcopy at least in Mame is the same name as a similar utility under MS DOS I personally would have opted for them to call it QFB since LSI had already bundled a QFB with LDOS 5 1 4 They had already established the function association with a name It should have remained QFB implement a more recognizable Fm LDOS Support jjkd The functions of QFB and DISKCOPY are virtually identical The reason for the name and appearance change are most likely that LSI licensed a poorly marketed by RS too version to Tandy and DISKCOPY is what the equivalent tool is called in MS DOS I know I know who cares about MS DOS but anything that reduces the sheer mass of command differences when there is no major overriding need for a difif rence ts a ptus in TN UL SRR ae ge lt ook AYP Dit jf AS 2 OL i NUD PO BOX 297 DEBUG Help PADSTOW NSW 2217 pea a Fm Ken Kane I can t find that thread on GOOD BAD docs but the Model LOW rating for its 4 manual gets a docs and examples for LS DOS Information Volume I iv DEBUG The E command is poorly explained Hunt and peck about an hour of work but
63. that routine all SYSTEM FAST SLOW has to do besides toggling the hardware is to change the value of the time slice byte One more thing taken care of in 6 x that couldn t be done in 5 x is the operation of the PAUSE routine In Model III mode GPAUSE is totally in ROM and unchangeable unless your on a 4P Since code executes twice as fast approximately when in the 4 Meg clock speed PAUSE runs twice as fast and its pausing action is subsequently cut in half The PAUSE in 6 x adjusts for FAST SLOW Okay is the confusion cleared up now CPU effect the blink rate action externalto CPU of the cursor unless the blink routine clock interrupt rate some adjusts for the SYNONYM adapted for 5 3 Fm Dave Lamkins This is a patch for one of my favorite utilities You may use it in TMQ if you wish It is a patch for Henry Melton s LDOS Information Volume I iv SYNONYM processor from LSI Journal V2n5pl4 SYNONYM hooks into the overlay loader vector and looks for the ERROR overlay number which changed from X 86 in LDOS 5 1 x to X 96 in LDOS 5 3 The patch is easily applied as a command line patch PATCH SYNONYM D01 76 96 F01 76 86 Superlog and LDOS 5 3 Fm Luis M Garcia Barrio SUPERLOG JCL is a modified INSTALL JCL for SUPERLOG to suit the new structure of SYS0 SYS Feel free to publish it in your magazine if you find it appropriate Editor s note SUPERLOG JCL will be printed in The Patch Corne
64. the mail I received in response to last issue s Letters to the Editor everybody wants to hear more about the family Gee maybe I ll try to get a family portrait in one of these issues Now although it has only been three months since I wrote the last column Stacey appears to have aged about 3 years I guess it s just due to her getting out from under our thumbs and starting to play with the other kids in the cul de sac That s a big leap in independence I did attend her graduation from 3 year old pre school however the class picnic which I was also scheduled to go on was rained out Stacey turned four earlier this month Now somewhere I read that after the first t y it si teal Ly hard to take off children are entitledtt THE MISOSYS QUARTERLY SPRING 1987 Volume I iv Speaking of Brenda get this She s going to start driving the pre school van for the morning and noon trips taking the pre school kids to and from school Brenda s a late sleeper she doesn t like to get up before 8 It s a good thing we work close by You see I can talk about her now Brenda Stacey and Stefanie are down in Florida visiting her family Brenda reports that the kids are coming along fine in the swimming pool They re even forgoing their water wings That s a big improvement over last year We were out a few times at the pool in our community over Memorial day weekend but the water wasn t too warm yet It s quite a bit warmer in
65. to zero also set M to zero must have been overlooked If you do for bizarre reason not want wrapped lines to be indented to the margin we can look this as a fortutious occurence LS DOS Information sc Volume I iv Accessing the RS232 port from BASIC Welcome to the here is the most Fm LDOS Support jjkd LDOS TRSDOS 6 Forum This experienced group of folks you ll ever find when it comes to LDOS TRSDOS 6 and the Z80 family of Tandy computers As I recall the question was RS232 output and perhaps input in BASIC For example from TRSDOS 6 before entering BASIC do a SET CL COM DVR This installs the OS support for the serial port Now from inside BASIC you can do this 10 SYSTEM SETCOM W 7 P Y EVEN DTR BREAK 0 or whatever other parameters you need to talk to the device in question You may also change these on the fly as necessary simply by repeating this line as needed with the appropriate changes To actually go ahead and produce output you can do this 20 OPEN 0 1 CL 30 PRINT 1 This is a test Remember that BASIC will provide a CR after this output but no linefeed If you want to provide output to both the display and the serial port one at a time you can do this 100 GOSUB 1100 Initalize for output 110 DEV 1 A This will go to the RS232 gosub 1000 120 DEV 2 A This will go to the display gosub 1000 1000 PRINT DEV A return 1100 OPEN O 1 A CL OPEN 0 2
66. well asa built in speaker with limited capability Details are beyond the scope of this article The Programmer s Corner THE MISOSYS QUARTERLY SPRING 1987 zije s Volume I iv As far as disk drives go the conventional 5 25 floppy disk is a 40 track double sided double density drive with a 360K capacity Hard disk sizes vary dramatically but it s now possible to get 20 meg drives with a controller for around 400 Ten meg drives are a thing of the past You can get off the shelf hard drives up to 120 megabytes and beyond IIL Comparison of Operating Systems The PC as well as the TRS 80 has a more or less standard boot rom or BIOS as it is called which contains various useful low level routines for doing I O For example although it possible to directly control the video hardware yourself and there are definite speed gains in doing so the short and simple way to print things out is to use a BIOS call to do it for you Essentially a BIOS routine is called with a useful short 8086 instruction called INT The INT instruction is analogous to the Z80 s RST instruction although the 8086 supports 256 possible INTs compared to Z80 s meager 8 INT is basically a short position independent form of the inter segment far CALL The lower numbered INTs are reserved for BIOS call usage For example INT 10H note that H after a number means hexadecimal supports the video adaptors For example to do a si
67. will get you a boot disk this configuration given a disk in the first floppy drive Keyboard Filters Fm Bob Haynes I feel very foolish bringing this up as I bet I m overlooking something stupid Nevertheless here goes 7 I m experimenting with a small piece of code which will eventually be used to filter KI under 6 2 6 3 Init code memory header are properly established I m trying to intercept keyboard chars under certain conditions and pass them to the CL device as well as back through the KI chain Here s the code SVC MACRO NUM LD A NUM RST 40 ENDM LS DOS Information Volume I 1v THE MISOSYS QUARTERLY SPRING 1987 Ga Volume I 1v s preserve DCB link if present j 7 fup a relocation routine for the hard START PUSH IX gt amp faddresses Then I THINK I m done cept for get DCB pointer from header RX LD IX MODDCB call the KI driver SVC CHNIO restore link POP IX exit if no char RET NZ abbrev d for benefit of this PUSH 4HL DE BC AF AF get CL DCB into DE LD DE LC SVC GTDCB EX DE HL s get char into C send to CL POP AF LD C A SVC GPUT restore regs again abbrev d POP AF BC DE HL RET Of course there s no error trapping yet and the GTDCB isn t needed for every loop thru the chain but as I say this IS skeleton It works for all chars except lt ENTER gt and lt BREAK gt When either of those are pressed it crashes I suspect the KFLAGS scann
68. with TRSDOS 6 2 the one that came with LS DOS 6 3 and one of my own generation which is a true copy of the ROMs in a model 3 I have experienced the system in BASIC The reason I ask is that consistent lock ups of while running with the MEMDISK installed as a system disk I install SET2RAM and MEMDISK backup all the SYS overlays onto the MEMDISK then kill SYSO SYS5 and SYS9 to make room for BACKUP CMD as well Then I do SYSTEM SYSTEM 2 and a SYSTEM SWAP 1 DRIVE 2 followed by SYSTEM DRIVE 0 WP YES DOS commands and utilities work fine in this environment but BASIC seems to freeze after running for awhile keyboard is dead the screen unchanging no response from BREAK and only a hardware RESET will recover From the times at which this happens my guess is that it strikes during garbage collection I am using the BASIC that is supplied with LDOS E PE and have high memory modules above approx x F400 all with appropriate headers Any guess what might be going on Which ROM image did you use for testing on your 4P Or was development mainly done on the regular 4 Fm MISOSYS TMQ I iii has a two byte patch to SET2RAM which corrects a problem if you use BASIC s TIMES The patch is as follows PATCH SET2RAM D00 50 00 38 F00 50 FF 37 some versions of the not operate without that patch I should think that the style of 4P would be irrelevant The MODELA III file should likewise be irrelevant once that pa
69. won t work Nor will any but the very simplest versions of PRINT USING But what about printing variables My programs need to print several variables of all different formats on the same line This has to be in a loop processing lots of different numerical values and strings Fm MISOSYS Nothing in our advertising or in our catalog states that PRO EnhComp will compile a program written to run under Microsoft s BASIC Interpreter used on the Model 4 There is no commonality amongst BASICs each implementation may include more or less than some other manufacturer s interpreter or compiler Your third paragraph suggested that EnhComp should compile your existing BASIC programs Quite the contrary the review in 80 MICRO which you reference to be your reason for MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 89 Volume I iv purchasing EnhComp states While EnhComp bears a strong relationship to Basic the word from MISOSYS is that there is no attempt to make a Basic compatible compiler Besides differences among Basic interpreters for the Model I III 4 would make the chances of compatibility practically nil The term Basic as used by the reviewer represented the Microsoft BASIC interpreter as distributed with TRSDOS 6 under license to Tandy Corp Your criticism of the EnhComp reference manual may well be a personal one There are literally hundreds of books covering the BASIC language Every inte
70. you use Fm H Brothers I m using BRIEF for programming and something called EDIX for word processing on the MS DOS machine LeScript on the Model 4 BRIEF isn t it is great PC DOS vs MS DOS Fm Adam Rubin Probable dumb question here but that s never stopped me yet What does it mean when a program works under PC DOS but won t run properly under MS DOS Did someone use something undocumented somewhere Have official MS DOS guidelines been ignored How frequently does something like this show up in commercial programs The program in question is Webster s New World Writer and was tried under both DOSes on an XT clone Many thanks for any enlightening comments Fm LDOS Support jjkd Several possibilities MS DOS Information Volume I iv 1 The program relies on undocumented internals of the OS This will usually tie a program to a particular release of DOS or at least those that it knows about in addition to eliminating MS DOS compatibility I wouldn t expect a program like that to work under say the just released DOS 3 3 for example There is generally no excuse for this In many cases it is caused by the copy protection scheme if any I wouldn t keep such a program as I wouldn t trust it to keep on working for any length of time It would be returned immediately 2 What you have is a barfy botched up version of MS DOS When MicroSoft licenses a version of MS DOS to a vendor the vendo
71. 0 but is there not a read problem the disk drives I understand they format differently Also who is the commercial vendor you mentioned Is it possible to modify the board to accept 256k chips as a non tech or must this be done by a computer technician Thanks again for the information Fm LDOS Support No be very careful The opposite is true the majority of IBM programs will not run due to the use of direct video access and or hardware access Virtually anything that pops up for example is immediately ruled out WordPerfect is one of the lucky exceptions The T2K will read and write either 360K or 720K floppies though some care needs to be taken when writing 360K disks MS DOS Information 47 THE MISOSYS QUARTERLY SPRING 1987 Volume I iv The vendor I m thinking of is Envision Designs you can find references to them over in the TRS 80 Pro Forum go PCS21 in either DL 9 or DL 10 I forget which They also used to advertise in 80micro I haven t looked recently Most of the boards I ve heard modified by non techs have been problems it isn t something I d recommend to somebody who isn t at least reasonably familiar with digital logic anti static techniques and soldering on multi layer boards Fm MISOSYS There is a newsletter called Tandy 2000 Orphans which may be of assistance The address is 387 Main St Westport CT 06880 Envision Design mentioned by Joe is at 1539 West Pearl St Pasco W
72. 00 0065 00011 FLAGSS EQU 101 3000 00012 ORG 3000H 3000 D5 00013 BEGIN PUSH DE sMove DCB pointer to IX 3001 DDEI1 00014 POP IX 3003 216F30 00015 LD HL SIGNON 3006 3EOA 00016 LD A GDSPLY 3008 EF 00017 RST 40 3009 35865 00018 LD A GFLAGS Get flags pointer 300B EF 00019 RST 40 300C FDCBO25E 00020 BIT 3 1Y C A System request 3010 CA6330 00021 JP Z VIASET 00022 00023 install new HIGHS and move filter code 00024 3013 210000 00025 LD HL 0 sget current HIGHS 3016 45 00026 LD B L Set B 0 3017 3E64 00027 LD A HIGHS 3019 EF 00028 RST 40 301A 204B 00029 JR NZ NOMEM Error if no memory 301c 220C31 00030 LD OLDHI HL sput in filter header 00031 00032 Relocate internal references in driver 00033 301F FD213B31 00034 LD LY RELTAB sPoint to relocation tbl 3023 113431 00035 LD DE MODEND i 3026 B7 00036 OR A Clear carry flag 3027 ED52 00037 SBC HL DE 3029 44 00038 LD B H Move to BC 302A 4D 00039 LD C yi 302B FD6E00 00040 RLOOP LD L IY Get address to change 302E FD6601 00041 LD H 1Y 1 3031 7C 00042 LD A H 3032 B5 00043 OR L 3033 280F 00044 JR Z RXEND 3035 5E 00045 LD E HL P U address 3036 23 00046 INC HL OO 3037 56 00047 LD D HL 3038 EB 00048 EX DE HL Offset it 3039 09 00049 ADD HL BC 303A EB 00050 EX DE HL The Programmer s Corner 68 The Programmer s Corner Volume I iv 3038 303C 303D 303E 3040 3042 3044 3048 304B 304E
73. 1987 r 3 2 21 9 k4 k X 9 s 3 i X MHI SHI DEVNIL NOTOUT ERROR 3 NILMSG OUTMSG OLDHI FNAME 57 53 START LD LD LD LD POP BIT JR BIT JR BIT JR Volume I iv DE 428AH logot MHI 1 HL sinstall HIGHS SHI 1 HL ERROR 1 DE sinstall logot AF sdevice type byte 3 A sdevice set NIL NZ DEVNIL 4 A Routed NZ DEVROUT 1 A Output device Z NOTOUT ok to filter device POP LD LD LD LD install LD LD LD XOR SBC INC LD LD EX LD LDIR JP LD JR JR LD CALL JE DB DB DB JR DW DB DB JR PUSH CP IX sget dcb L 1X 1 H 1X 2 sget old address OUTP 2 1 HL AAA OUTP1 1 HL new HIGHS and move filter code HL 0000H sget current HIGH OLDHI HL sput in filter header BC LENGTH slen of filter A HL BC figure new HIGHS 0000H HL sstore HIGHS HL IX 1 L IX 2 H sinstall addr into dcb DE HL DE to new address HL PAWSTRT SHL to filter code srelocate filter GEXIT sall done HL NILMSG ERROR HL ROUTMSG ERROR HL OUTMSG 0000H scall LOGOT GABORT 31 PAGEPAWS Line Printer Pause Filter Copyright c 1981 by Logical Device not active ODH Device is routed ODH Not an output device ODH START START FNAME PAGEPAWS sHIGHS before filtering C OUTP 2 AF A C OCH stof char 66 The Programmer s Corner Volume I iv 5321 2804 00080 JR Z
74. 1987 Feldman goes on to state that Unfortunately the new directory structure poses a potential danger to your LDOS 5 1 4 disks The problem is that the individual directory entry being longer puts your old disks at risk That s just not true The directory is exactly the same size 32 bytes per directory entry Besides there are no dangers posed to 5 1 4 disks Those disks can be read from or written to by LDOS 5 3 without any corruption whatsoever to the files themselves or the directory Of course after December 31 1987 the directory date of a 5 1 4 disk file would be incorrect if the file was modified while running 5 3 but then that s the whole reason for updating MISOSYS went to considerable lengths to design a date extension environment Which was as transparent as possible You can read write from to 5 1 4 disks with 5 3 When a 5 1 4 disk file is updated while under 5 3 the extended date and time fields are not LDOS Information THE MISOSYS QUARTERLY SPRING 1987 ar JO s Volume I iv touched on quite safe the 5 1 4 directory It s really The DATECONV utility is only needed when you want to convert a 5 1 4 diskette to the 5 3 extended dating mode Certainly after December 31 1987 this issue will be moot if you are using LDOS You ll need 5 3 Feldman s final statements concerning the lengthy routine to set those parameters when referring to the serial port and forms kind of left me puzzled
75. 1l have to duplicate the routine to find the DCT swap in bank 0 and so on It looks like it would be easier to implement new functions as DCINIT GDCRES etc On the other hand if new functions are implemented as subfunctions of for example DCINIT then any software which expects DCINIT to initialize the end up writing to the disk my new function with predictably unpredictable results Doesn t sound too good either disk controller may In this particular case I m tinkering with a floppy disk driver which suggests two possibilities One extend the existing RDSEC and WRSEC to perform new related functions depending on bits 6 and 7 of the track and sector values Two use GHDFMT for my new write functions and abort if it s actually a HD Any thoughts on these Oh and can I count on the NMI routine staying essentially the same in future versions i e POP return address RET or should I stick my own routine into 66H 68H and restore it each time Fm MISOSYS My suggestion would be to utilize the RDHDR SVC You could use HL as a pointer to a data region to pick up any set of values you need useful since you are deciding your own protocol RDHDR has the advantage that no software could expect the DOS to suppport that function since it doesn t Any utility package actually needing a RDHDR function needs to write their own disk driver to support it As an aside I had hoped to add an SVC to perform a d
76. 2 for US 3 Canada and Mexico 6 elsewhere If you purchase the corresponding DISK NOTES with the coupon which accompanies this TMQ issue you can save 2 50 the cost then being only 87 50 S amp H Here s what s on Disk NOTES 8 LSIFIX TXT Fixes for LS DOS 6 3 INSTRUCT TXT See APPLICATIONS LDOSINFO DAT ditto LDOSINFO TXT ditto LIBENTRY BAS ditto LIBENTRY TBA ditto LIBREAD BAS ditto LIBREAD TBA ditto UNDATE CMD opposite of DATECONV PGPAWS5 ASM source to PAGEPAWS v5 UNDATE SRC source to UNDATE CMD PGPAWS6 ASM source to PAGEPAWS v6 UNDATE DOC docs for UNDATE MPATCH FIX Model 4 SS patch SUPERLOG JCL Just what it says STRIP BAS utility to strip 1 byte CODE TXT source to code program SOUND BAS SOUND for PRO EnhComp FIXES8 TXT The Patch Corner QUARTERLY Coupon Don t forget the coupon which accompanies this issue It qualifies you for a 2 50 savings on DISK NOTES 8 We have assemblers on sale EDAS PRO CREATE ED ASM 86 We have data base managers on sale LB LB86 And here s a great price on The Basic Answer Remember you must return the coupon to qualify Please select only one sale item per coupon and gosh don t forget to include appropriate S amp H fees Some folks just never read The BLURB Sh hg ee Volume I iv Family Update Here s where the folks who are not interested in my family can stop reading and proceed to the NEW PRODUCTS section Judging by
77. 286 386 I was referring to protected mode Operation in regards to the linear addressing mode Linear addressing is not available in real mode with or without the new instructions Of course the 386 memory management also allows you to operate in virtual 86 mode to simulate the kludge Fm LDOS Support to Nate Salsbury Entirely a matter of semantics To me a program is relocatable if the code is entirely based on relative jumps or if locations in memory are calculated at run time not compile or assemble time of course offsets must be calculated then Real 6809 and 680x0 relocatable code works this way due to the linear address space and the fact that almost every instruction has a relative form In both COM and EXE files there are absolute offsets all over the place The COM file works only because the segment registers are set up to magic values that represent the location the program segment only one allowed was loaded at In EXE files since more than one segment is allowed any offsets requiring reference to an absolute location are hot patched on the fly when the program is loaded Note that according to this definition the Z80 code used in LDOS filters and drivers is not relocatable code but instead is code that is relocated A fine difference indeed Fm Nate Salsbury I think the distinction re LDOS filters between code that is relocated vs relocatable code is NOT a fine difference
78. 3 Send Ascii FILE PRINT 4 Send from KEYBOARD PRINT 5 Send RANDOM sequences PRINT 6 Send QUIZ PRINT PRINT The lt gt key will return you to this menu PRINT PRINT Your selection please LINE INPUT MM MM VAL MM IF MM lt 1 OR MM gt 6 THEN 530 CLS ON MM GOTO 510 470 350 180 450 670 CLS PRINT You will be sent a character you must respond by typing that character PRINT lt SPACE gt will RESEND the character counts as 1 wrong PRINT lt gt will PASS the character counts as 1 wrong PRINT lt F1 gt will TOGGLE VIDEO display of dit dah PRINT lt gt will TERMINATE quiz and display score PRINT PA 0 CA 0 RT 0 Q 0 WA 0 GOTO 850 B INKEY IF B THEN 740 rem IF B CHR 129 THEN V V 1 GOTO 740 _ E Peay Ares IF BS gt G THEN BS CHR ASC B AND SHSE ari IF B Q THEN 820 Mot li IF B THEN GOSUB 300 RT RT 1 GOTO 740 PADSTOW NS wr go IF B G THEN 930 a ad IF B THEN WA WA 1 PA PA 1 GOTO 830 WA WA 1 GOTO 740 CA CA 1 PRINT C 0 840 FOR L 1 TO 2000 NEXT L 850 IF Q 8 INT Q 8 THEN PRINT 860 0 0 1 870 880 890 900 910 920 930 940 950 960 970 980 990 GO RND 47 43 Q CHRS GQ IF Q gt AND Q lt Z THEN CODE TBL GQ 64 GOTO 910 IF Q gt AND Q lt THEN CODE TBL GQ 43 26 GOTO 910 GOTO 870 IF V lt 0 THEN PRINT CODES AS CODE GOSUB 300 GOTO 740 Q Q 1 PRINT PRINT PRINT Characters sent
79. 3 X 32D1 D03 2D 33 F03 2D 78 GAT set DOS version to 63 set bit 3 of byte OCDH X 376C D07 DE 63 00 8B F07 DE 62 00 83 Boot sector 2 DOS version X 39C1 DOA 3B 63 FOA 3B 62 The Patch Corner 100 The Patch Corner Volume I iv THE MISOSYS QUARTERLY SPRING 1987 Volume I iv NAME63 for NAME 19 Ju1 84 Change check for year lt 88 ti year lt 100 X 30C8 DOO FD 64 F00 FD 58 Change 6 x to 6 3 X 31E7 DO2 20 33 F02 20 78 Since TRSDOS 6 2 Boot sector 2 bytes x 18 x lF are no longer used for date storage but rather Levelxx system disks it would be better to prevent NAME from changing this area Change JR Z to JR x 3165 DO1 9E 18 FO1 9E 28 MRS512 FIX 05 19 87 Corrects DW expansions with mixed modes Apply via PATCH MRAS MRS512 DO2 46 7C 8F F02 46 75 8F D1C 03 75 8F F1C 03 C4 62 X 8F75 AF 32 9B 65 C3 C4 62 MRS612 FIX 05 15 87 Corrects DW expansions with mixed modes Apply via PATCH MRAS MRS512 DO2 2A 67 62 F02 2A 60 62 D1B C6 60 62 F1B C6 BB 35 X 6260 AF 32 92 38 C3 BB 35 PDS553 FIX Patch to PaDS s BUILD module for 5 3 Apply via PATCH PDS PDS PDS553 D15 D1 70 3D 52 45 29 OD 00 00 00 F15 D1 61 3D 2C 70 3D 52 45 29 OD PRO PaDS PATCH PDS PDS D20 55 20 30 F20 55 05 20 RSH61 FIX 05 27 87 Patch to RSHARD6 DCT Apply via PATCH RSHARD6 DCT RSH61 enable JCL D00 C1 C3 C1 68 F00 C1 B7 20 27 D08 ED B7 C2 DF 60 C3 BB 60 F08 ED 00 00 00 00
80. 5 n asdb Compute beam bin crossover point if sdb lt 0 return Exit now if weights are not to be computed Compute frequency response nback n l of fset 5 nback n2 offset x0 n dcf n q 0 Normalization factor x 1 1 n DC term is trivial do k 2 n2 f k 1 n x k 2 dcf n q x0 Factor of 2 here instead of in series do j l n2 Use DFT to compute weight from frequency response s lt Ju 1 d 2 PI offset j oe B Dr be a sum x 1 ha do k 2 n2 I p en s 0 sum sum x k cos k 1 df j z h j sum jo Ja i nback j E LP 3 2 h i sum EAN EM ET Netti 4 return w end function dcf n q fa Dolph Chebychev frequency response n Number of weights Normalized frequency fraction of sample rate Returns T n 1 q cos pi f where T n 1 x cos n 1 acos x ifa fa 0 5 Limit f to lt 0 5 f fa ifa argl g cos PI Compute fundamental argument switch n case l tn l T0 x 1 case 2 tn argl T1 x x default Compute rest by recursion lt tn2 1 tnl argl tx 2 argl do 1 3 n tn tx tnl tn2 tn2 tnl tnl tn Announcing New Products Volume I iv return tn end function qfroms db n Computes Dolph Chebychev parameter q from sidelobe ratio s Sidelobe ratio dB 20 alogl0 s n Number of weights Equation is q cosh acosh s n 1 double precision ddb s argl dval ddb db Convert to double precision s 10 05 abs db Compute voltage ratio from dB
81. 62 Volume I iv SCRIPSIT previewing output Here s a little tip for using Model 4 SCRIPSIT at MISOSYS for some specifically letters booklets We have those of you still We still use it here word processing jobs and small documentation always wanted to conserve paper thus we have always looked for better ways to proofread rough drafts of text for form content and outline Back in July of 1984 we put together a series of patches for Tandy s SCRIPSIT program These patches were called MPATCH and they were published in NOTES FROM MISOSYS Issue III the FIX files are also on DISK NOTES III The primary enhancement to SCRIPSIT added by these patches was the 0 command This command allows you to invoke any DOS library command or other program running totally in the library region of the DOS from within SCRIPSIT For instance one can easily invoke a DIR command or a REMOVE or ROUTE for that matter The Q command makes it very easy for us to preview the formatted text output of SCRIPSIT without wasting any paper This operation makes use of three DOS commands ROUTE RESET and LIST The idea is to redirect the printer device to a disk file with a command of the form ROUTE PR PRINT TXT 2 Since the DOS needs some high memory to set up a buffer and file control block space for such a ROUTE command I issue this command before entering SCRIPSIT When I am in SCRIPSIT ready to print since the PR device i
82. 6D B6 C9 4B E5 OF E5 F4 E5 E5 OC B5 C5 97 21 6B E5 6B CD E5 D5 21 2A 28 Fl FI 06 CD CD E5 1D E5 CD C8 OF 6B OF 11 C9 Fl 00 36 FE CD 95 CD F20 B0 6D C5 F1 C9 21 OF 6B E5 CD 90 9A FI 4C 00 F27 29 F1 Cl EL E5 C5 F5 7C B5 C4 90 A7 00 00 F27 8C F1 Cl EL E5 C5 F5 7C B5 C4 24 AE F27 B7 18 76 2A 97 45 7C B5 28 05 F5 E5 D5 CD Fl El E5 CD F4 7F DC 2A 8B Fi FI FI 8D 45 E5 Fl FI El E5 45 18 OC 2A Cl El E5 C5 CD 89 AE Fl 98 The Patch Corner Volume I iv F27 EO F5 D27 F0 C9 F27 FO CD D28 00 B6 F28 00 CD D28 10 36 F28 10 E5 D28 20 09 F28 20 E5 7C 21 D7 CA 23 71 CD 29 CD B5 8B 7A 31 DC Fl 89 E5 89 28 45 Fl 92 2A 21 AE CD AE 12 E5 Fl Fl 8B 16 FI AA Fl 21 CD 18 21 45 29 CD 71 18 8B CO 38 09 E5 E5 36 C3 08 THE MISOSYS QUARTERLY SPRING 1987 45 7D Fl 29 CD DD AE SF 21 E5 Fl El E5 89 E5 21 4C 8B Fl C9 E5 E5 AE E5 8B 00 45 Cl 21 F5 E5 Fl CD 45 00 E5 El 06 E5 CD 18 03 23 00 CD E5 00 Cl 15 22 78 23 00 CO C5 39 21 5C 2A Fl CD 00 7D F5 7E C7 Fl 8B FI El 00 Fl AUTOIN61 FIX patch MC61 CMD RND 05 19 87 allow arbitrary expression as initializer of an automatic scalar variable DIE DA 04 FIE DA 04 D23 A7 21 D23 B0 21 F23 B0 22 D23 C0 97 F23 C0 00 D23 D0 88 F23 DO E5 D23 EO CA F 23 EO CD D23 FO F1 F23 F0
83. 8 FI FI 23 2A CD Fl 21 Fl 21 7A 2A il C9 FI B6 3F 7C CD ED C9 66 Fl 3F 21 21 c9 CA E5 68 63 28 21 F20 D0 2A 3F E5 CD AB 6E Fl The Patch Corner Volume I iv THE MISOSYS QUARTERLY SPRING 1987 Volume I iv BC53 FIX 05 19 87 Patch to EnhComp BC CMD Corrects error trap of IF exp THEN missing clause Apply via PATCH BC BC53 D2C 9A 17 5B F2C 9A 98 63 D05 2F B7 C2 98 63 C3 BD 66 F05 2F 00 00 00 00 00 00 00 BC63 FIX 05 19 87 Patch to PRO EnhComp BC CMD Corrects error trap of IF exp THEN missing clause Apply via PATCH BC BC63 D2C AD 11 2F F2C AD A6 37 D05 29 B7 C2 A6 37 C3 CF 3A F05 29 00 00 00 00 00 00 00 INSTRU51 FIX patch MC51 CMD to correct initialization of nested structures RND 05 24 87 D22 2B 21 08 00 39 6E 26 00 7D B7 C4 71 AD E5 F22 2B F5 21 00 00 39 E5 CD 71 AD D1 CD BI E4 INSTRU52 FIX patch MC52 CMD to correct initialization of nested structures RND 05 24 87 D1C A6 21 08 00 39 6E 26 00 7D B7 C4 B4 AB E5 F1C A6 F5 21 00 00 39 E5 CD B4 AB D1 CD F9 ES INSTRU60 FIX patch MC60 CMD to correct initialization of nested structures RND 05 24 87 D23 9E 21 08 00 39 6E 26 00 7D B7 C4 E9 91 E5 F23 9E F5 21 00 00 39 E5 CD E9 91 D1 CD E6 DB INSTRU61 FIX patch MC61 CMD to correct initialization of nested structures RND 05 24 87 D22 1A 21 08 00 39 6E 26 00 7D B7 C4 60 81 E5 F22 1A F5 21 00 00 39 E5 CD 60 81 D1 CD 61 B8 INSTRU62 FIX patch MC62 CMD to corr
84. A 99301 509 547 1139 Clear to end of frame Fm Nate Salsbury Does anyone know a command in GW BASIC that will partially clear the screen from the present cursor position to the EOS lower RH corner In M4 BASIC the command PRINT CHRS 31 would clear from where the cursor was at that moment to the EOS In GW BASIC I find PRINT CHR 12 which clears the WHOLE screen same as CLS Not what I want Fm LDOS Support jjkd I don t think there is one A kludge would be 1000 print string 80 P0S 0 32 but that type of thing can give you troubles with wrapping the right margin and scrolling the screen unless you always stop one short which can in turn cause problems with not clearing the last column I think that you d get an error back from the string function if the cursor happened to be in the last colum If you were not working in BASIC ANSI SYS but BASIC ignores driver I d say uSe the ANSI SYS Fm JIM OWENS Joe What do you mean I only have a vague understanding of what the ansi driver does In fact I know only two things about it 1 Some programs need it and 2 it slows down execution if it is not needed MS DOS Information Volume I iv So Basic does not use it any way Well darn it Joe don t you see what you ve done You have opened up a whole new to ask grin Fm LDOS Support jjkd The ANSI driver is an attempt to provide a Standard interface for programs to use This is the equi
85. C this address is conceptually used as a pointer to a structure of type dct thus that s the reason for the cast MC and include variations THE MISOSYS QUARTERLY SPRING 1987 some CPU s reverse the high Volume I iv lt file h gt and include file h the UNIX environment when used in As I m sure you re aware K amp R indicate the quoted include searches first in the directory of the source file before searching the sequence of standard places while the bracketed include does not search the directory of the source file Standard places usually meaning usr include While this means nothing under LS DOS implementations I could impact the systems developed under LS DOS MC once migrated to UNIX or other environments Fm MISOSYS Since MC is not used in a UNIX environment why should it describe the differences between lt stdio h gt and other forms Fm Byron P Peebles It s my tumble lt rightfully so gt opinion that since you suggest that programs developed under LS DOS with MC will be compilable under UNIX you should point out areas where the handling of things like forms of include differs I would have kept quiet on the entire issue except your documentation which does appear to be excellent uses several forms of include for the same declaration on the same page Just seems it might cause confusion that s all Fm MISOSYS That s to show that in MC all of those forms are equi
86. D 22 1B 26 70 39 06 CD 70 65 70 21 CD 19 B5 B9 BB B6 E5 B4 AD 2A 65 22 AB 65 76 00 E5 D5 00 AC 11 70 CD 6B AI CD C4 75 13 28 CD 22 22 59 AC 65 BF 70 22 E5 CD Fl 39 E4 09 11 EE 70 6E 8B 56 AC 69 67 70 22 70 D1 7C 65 CD 39 2A 7E cD 00 09 E4 cD Fl E3 BE AUTOIN52 FIX patch MC52 CMD allow arbitrary expression as initializer of an automatic scalar variable 71 03 00 EE 71 02 00 E9 71 OL 00 E9 71 72 03 00 17 72 02 00 OD 72 01 00 OD 72 D19 B3 08 F19 B3 38 D1F 57 CD D1F D9 18 D20 3C 18 D20 4F C9 F20 4F C9 D20 60 A6 F20 60 E5 D20 70 6B F20 70 21 D20 80 01 F20 80 01 D20 90 39 F20 90 E6 D20 A0 7E F20 A0 6D D20 B0 F1 72 72 09 09 09 21 21 Fl CD CD OF 02 02 7E 2A Fl C5 FI 05 05 12 Fl CD OF OF Cl C8 FA 6B 00 00 23 OF CD FI 21 00 05 00 00 72 04 00 05 00 30 72 04 F1F 57 CD D4 AB CD 6E 6B 6B 21 A6 71 E5 72 72 B6 6B 61 CD DO 6E CE E5 E5 4B Fl CD Fl 21 C9 CA E5 94 48 54 CE Fl CD CD 6B Fl 48 Cl OF Fl D4 CD Fl C5 E5 Fl C9 AO AO CD Fl C5 El 6B El AB 6D 21 21 CD C3 00 9C 9C 36 Cl 2A E5 E5 E5 Fl C5 DD OF 22 BE 00 Di Fl E6 El 11 C5 CD F5 21 Fl 54 6B 90 AUTOIN60 FIX patch MC60 CMD allow arbitrary expression as initializer of an automatic scalar variable 00 3F 4D 03 00 3F 4D 02 00 3A 4D 01 00 3A 4D 00 6D 4
87. D 03 00 6D 4D 02 00 63 4D 01 00 63 4D D20 47 59 F20 47 8F D26 70 CD D27 29 18 D27 8c 18 D27 B7 C9 D27 CO 4D F27 CO CD D27 DO C4 F27 DO F5 D27 E0 8B The Patch 4D 4D SA 09 09 00 21 BA 8A E5 45 05 05 4D Fl CD 2A 8B 4D Fl 21 CD 00 05 00 51 4D 04 00 05 00 87 4D 04 CD 2C 97 45 18 Cl 8B 4B Corner 2C BA 45 E5 6A 21 45 4D BA Fl 7C CD 21 C7 E5 CD Fl C9 B5 F4 8B 45 CD 36 F26 70 CD 31 92 C3 00 C2 TF 45 CD C4 AE F2 00 BA D1 E5 23 8A 2A TE 70 70 7C 59 7C 7C B5 70 09 D5 69 23 A4 19 00 2A EE Fl C5 00 RND 05 19 87 00 00 71 00 Fl Fl 2A E5 6B F5 90 E5 DO 18 E5 23 C3 RND 05 19 87 Fl CD 3E B5 70 B5 B5 CA 11 70 Fl 70 B6 E4 6E 19 65 E4 CD D5 00 EE 17 00 00 El El OF C5 E5 E5 9A Cl 54 21 DD 23 5B AC AC 32 2D 32 OB OA 76 00 TC 69 71 06 B5 E5 26 11 67 E4 Al 95 22 79 76 79 2A D5 2A 19 B5 70 B4 CD 28 CD 00 04 70 ED 6E Fl 59 B4 2A B4 59 CD 65 TE 28 22 C3 50 05 2B E5 00 E5 5B Fl Volume I iv F23 B8 CD B9 AD 22 67 70 21 01 CD 70 2A 59 7C 70 9B 70 FE 11 77 2D 76 CD 6F CD 19 2A 67 Fl F24 60 CD A4 E4 DI CD 7E E4 7C B5 C4 56 BE F1F D9 F1 Cl El E5 C5 F5 7C B5 C4 35 BF F20 3C F1 Cl EI E5 C5 F5 7C B5 C4 40 C5 E5 E5 6B F5 CD CD Fl 21 ES 2A E5 CD Ti F5 F5 18 7C
88. ED ASM 86 manages segments in the following manner If all the program code and data can live in the confines of one segment or 64k then life is easy ED ASM 86 will handle the overhead for you transparently Otherwise you will have to define multiple segments in the following manner segname SEGMENT PARA class your code or data definitions go here segname ENDS You will define one segment per segment you need Usually for multiply segmented programs there are at least three segments A code segment a data segment and a stack segment MSDOS defines two executable file types COM and EXE The former are limited to 64k though not necessarily one segment ED ASM 86 can actually generate multiply segmented programs and output a COM file unlike MASM The latter the EXE file is designed for larger programs and can support the 640k limit imposed by MSDOS The Programmer s Corner THE MISOSYS QUARTERLY SPRING 1987 ee A Volume I iv Segmentation affects all aspects of data moves and program branching statements For example there are two basic types of jumps and subroutine calls in the 8086 inter segment and intra segment As the names imply inter segment jumps and calls jump between segments They are so called far references These require 32 bits of information the segment and offset of the target routine Since these are long and slow programmers try to avoid them Far more common is the intra segme
89. ERLY page 48 Winter 1987 in reference to the incompatibility created by the new time date stamp of LDOS 5 3 Since I am not a CIS subscriber I have not a way of sending it to him UNDATE CMD For LDOS only c 1987 Luis M Garcia Barrio Free distribution Not for sale Syntax UNDATE drive The new time date stamp used by LDOS 5 3 and LS DOS 6 3 has created a minor incompatibility with previous releases of those DOSs In the new format bytes 18 19 of each directory entry are used by the date time stamp Under the old format those bytes were used by the THE MISOSYS QUARTERLY SPRING 1987 Volume I iv mentioned in THE MISOSYS QUARTERLY Winter 1987 As a result a disk formatted or DATECONVerted with LDOS 5 3 OR LS DOS 6 3 will appear to have all its files with ACCESS password protection when used under LDOS 5 1 or TRSDOS 6 2 There are two ways to solve the problem 1 Format all disks with LDOS 5 1 or TRSDOS 6 2 The new versions of those DOSs will not attempt to stamp the files with the date time 2 UNDATE the disks by resetting bit 3 of GAT CDH and giving files an ACCESS password of 8 spaces LDOS 5 3 LS DOS 6 3 will access them as if formatted with LDOS 5 1 TRSDOS 6 2 and these two will access them as their own In short UNDATE reverses the process of DATECONV The problem I can foresee is a file trying to access another expecting it to have an ACCESS password In those rare cases use ATTRIB to chang
90. Editor s note it was the Flotsam and Jetsam article on page 138 of the April 1987 issue PRO MC and Variable Initialization Fm Mark Harris Is there an easy way to have MC initialize all variables to contain zeroes when they are created A fair amount of C programs assume zero initialization If I ve missed something in the manual I offer my apologies By the way after several months with PRO MC I remain impressed It s a very solid package Fm MISOSYS Check out the z Y option in MLINK It will generate hex zeroes for DS regions in DSEGs If you want that it MISOSYS Products Tidbits Volume I iv will make the linking a little slower and take up a lot more CMD file space since all your declared buffers will be zeroes in the output file instead of a DS null space If you happen to have a char buffer 5000 in your program your CMD file will be 5000 bytes bigger That what you want If you do then add the Z Y switch to the JCL line which invokes MLINK immediately following MLINK Fm LDOS Support jjkd No C program should ever assume that a variable not explicitly initalized contains anything of significance I can t even think of a single C compiler that this would work with Can you give any examples Fm Mark Harris to MISOSYS Perfect Now don t get me wrong Programs written in any language that as sume some default initialization of variables make me ill But reality being wha
91. FIX S PATCH DISKCOPY CMD UTILITY D using DCOPY3 FIX S PATCH BOOT SYS LSIDOS D using BOOT3 FIX S PATCH SYS6 SYS LSIDOS D using DIR1 FIX S RRAREKKREKRRRRERERRERERRRRRRERRERRRRRRERRRRRREREERRRRERRERREERE Your LS DOS is now a version 6 3 0 Level K RAR AIR RRR k kiok kkk kk kkk RRR II AA kk END Here are the current patches that are applied by the above JCL SYSO1 FIX 02 20 87 For Level J SYS0 SYS LSIDOS PATCH 011 This patch will cause the day of week byte at DATE 4 to be set to 0 during the load of SYSO correcting the state of the LEAP BIT This patch to be applied to SYS0 SYS LSIDOS File dated before 07 01 87 X 0036 00 TED3 FIX 05 22 87 For Level J TED CMD UTILITY PATCH 012 This patch corrects a problem that occurred when another file is loaded immediately after deleting one character with lt CTL gt D This patch to be applied to TED CMD UTILITY File dated before 07 01 87 D04 B1 AC F04 B1 8A DEVICE1 FIX 05 29 87 For Level J SYS6 SYS LSIDOS PATCH 013 This patch enables DEVICE to function correctly when fed a one character device name which is forbidden in the manual but not actually prevented by the operating system This patch to be applied to SYS6 SYS LSIDOS File dated before 07 01 87 D1F D4 ED AO 7E B7 C8 FIF D4 7E 12 13 2C 7E DCOPY3 FIX 06 05 87 For Level J DISKCOPY CMD UTILITY
92. Florida especially down in Miami I would have liked to go to Florida too except then you wouldn t have gotten this sere aiad for at least another few weeks Besides had other work which had to get Wh n you run a small company like we do The week in birthdays which are usually ak ab ski st wi 11 be the first week I have ever immediate family have at their party as many friends my eee MISO nals ass L age According to that rule we shoul had four other kids Unfortunately difficult to narrow down commitments Since both Brenda s family and my family live far from here birthday parties for the kids have meant inviting some friends we have known for years They too have children of Stacey s and Stefanie s age 5 in all That would put us over the top But we still had to invite a few of Stacey s classmates So with an ensemble of about 10 kids and say 8 or so adults Stacey celebrated her fourth At least I now know why parents start letting their kids stay at birthday parties by themselves when they get to be four We all had fun but the pace was non stop Stefanie has made some big advances in her selection of foods She still won t even taste orange juice and Brenda growing up in Florida But Stef has begun accepting meat I have gotten Stacey turned on to soups She ll even scoop up the juice left over from the vegetable serving bowl calling that her soup Stacey even tried to get Stef to try s
93. Fm Gary Phillips You need Machine Language Disk I O and Other Mysteries by Michael Haguer published by IJG if you want to learn to program disk controller operations at the assembly language level It IS possible to read IBM PC diskettes in a model 4 or even a model 3 Commercial software to do this is available from HyperSoft and or PowerSoft at a very reasonable price 50 see 80 Micro ads IBM diskette formats use 512 byte sectors while TRSDOS LS DOS LDOS uses 256 byte sectors so not all the disk i o primitives will work directly I once read PC DOS diskettes on a model 3 using only my bare hands and DEBUG but a small zap had to be made to the DOS so it would read the full 512 bytes Now I use SuperCross and I m glad I do GRDTRK is not very effective with the Western Digital controllers and standard floppy disk formats Although there is a hardware command code that tells the controller to read an entire track due to the way synchronization is handled at the hardware level the only guaranteed correct data you will retrieve is the sector address marks The rest is usually ohase shifted and appears as garbage Such a function is useful for determining the structure of an unknown alien diskette but useless for actually reading the data located thereon I would have expected RDSEC to work if you have coded your program correctly Fm Adam Rubin In this case the two topics adding disk driver functions reading 512 b
94. Fm MISOSYS You guessed wrong The reason that HITAPE or other utility from LDOS 5 3 shows up as a DATA member of PaDS has nothing to do with passwords or the lack thereof in 5 3 What it has to do with is a test that PaDS makes to ensure that the appended member is a CMD file It rejects as being data any file which does not start with either a OIH load record or a 05H comment In LDOS 5 3 the header suppressed to save disk space just in case a file went to a granule boundary without the header this makes 5 3 s files start with a IF record Thus PaDS assumes those are not programs comment record was Here s a little fix you can make to PaDS which will make it more lenient it will accept any char less than 20H as valid PATCH PDS PDS D20 75 20 30 F20 75 05 20 MISOSYS Products Tidbits Volume I iv While you re about it add the PDS553 FIX patch to the PDS BUILD module That revises PDS BUILD to the password convention of 5 3 A similar patch for PRO PaDS was in TMQ I1 i111 That should fix you up A similar patch for making PRO PaDS more lenient is PATCH PDS PDS D20 55 20 30 F20 55 05 20 PRO WAM Window Application Manager Fm record of ADDRESS DAT only affects the PSORT sorting Changing it has absolutely no effect on the Search for gt key I m looking to search by Company name the third field 20 characters long seems like ADDRESS is hard coded to use the last amp first name field
95. GTH EQU S PAWSTRT slength of filter code 313B 2431 00122 RELTAB DW RX01 RX02 0 2D31 0000 3000 00123 END BEGIN Before we dissect this filter let s flag convention was added to LDOS 5 3 remember one thing about DOS 6 x A filter is installed into memory by the SET Lines 22 30 are similar to the 5 x filter command This command passes the address of counterpart the code obtains the current a spare DCB which has been assigned to the filter The FILTER command also inserts into the DCB the device name you have designated and passes the address in register DE of the DCB assigned to your filter The action which the installation code must take is to relocate the resident portion of the filter to high or low memory then setup the first three bytes of the DCB The code in the DOS 6 x filter from lines 1 through 16 are pretty obvious This next piece pertains to 6 x LD A GFLAGSS RST 40 BIT 3 Iy c A JP Z VIASET user could confront nothing restricted a One of the problems a under LDOS 5 1 was that filter from being invoked as if it were a command program Some folks for instance tried to install the LDOS keyboard driver simply by typing KI DVR This of course wouldn t work but the user would get no indication of failure as the driver really had no way of knowing that the entry conditions to the driver register DE were incorrect without some fancy checking This problem was eliminated in DOS 6 x by a system
96. H IIT 49h 520E 2808 00017 JR Z M3 5210 214940 00018 LD HL 4049H Mod I HIGHS 5213 117B44 00019 LD DE 447 BH Glogot 5216 1806 00020 JR LDHIGH 5218 211144 00021 M3 LD HL 441l1H Mod III HIGHS The Programmer s Corner 65 The Programmer s Corner THE MISOSYS QUARTERLY SPRING 1987 Volume I iv The development of DOS 6 0 permitted me to clean up that mess and design a reasonable device handler I think that got accomplished Anyone who has read my discussion of device handling which appeared in the book mentioned above should be able to bear that out Since most of the filters I mentioned above were written by LSI perhaps there just wasn t sufficient time to adapt all of them to 6 x On the other hand they may have felt that not enough people purchased the 5 x versions which resulted in no justification to spend any further time on them What follows then is a look at one of the filters from the filter pack PAGEPAWS This module is designed to filter the PR device and pause awaiting a keystroke when a form feed is received That allows you to use single sheets to feed a printer when the program is expecting continuous forms Let s take a look at the LDOS 5 x version of this filter But first some information is needed Let s look at a command such as FILTER XY PAGEPAWS The DOS s FILTER library command will search the device table for the XY device then load the PAGEPAWS module assuming XY was found W
97. I amp MISOSYS about helping people who don t pay their dues I want to say that we have purchased the appropriate software with every piece of hardware we use and we will continue to support those who are willing to support us when we need help Enclosed please find a check for First Class handling of your MISOSYS Quarterly but please keep in mind How can we successfully convert LDOS 5 1 4 Hard Drive to 5 3 0 with the following configuration 15 Megabyte Hard Drive 0 1 2 3 and Floppy 4 5 Fm MISOSYS Perhaps you read too much between the lines You are correct in assuming that MISOSYS cannot afford to provide you with 50 worth of assistance when you bought only a 25 product We would have gone broke a long time ago if we did that On the other hand I believe I should be able to point you in the right direction without you having to take a course in hard drives What follows is some information which should have been provided to you by Radio Shack with your hard disk package and from information provided in both the LDOS user manual and the upgrade documentation LDOS Information z Volume I iv First there is no standard configuration of a hard drive Even a 4 head 5 Megabyte Tandy hard drive can be arranged in one of about eight different configurations The most typical configuration of a5 Meg drive is as four partitions each partition being one head If your 5 Meg drive is already configured that
98. It s all a matter of point of view I guess Both COM and EXE files won t work without some form of management by the DOS For COM it is as simple as setting all the segment registers to 0010 0000 less than the code start address For EXE files the loader has its work cut out for it All internal segment references must be adjusted to account for the offset and the segment registers set to probably various different values What makes both of these not relocatable code in my book is that you can t move let s say half the code in a segment relative to the first half at run time and have the end result work Of course to have this work even on a machine with truly relocatable code like the 68000 each piece must be self contained and not reference the other In an EXE file they could then also be put in different code segments MS DOS Information THE MISOSYS QUARTERLY SPRING 1987 es ae Volume I 1v As for loading no it doesn t work that way If you load an EXE file it doesn t load into this code segment it always loads into its own code segment Those could of course overlap Fm Les Mikesell Remember that in COM files the segment sizes cannot exceed 64K that is the limit of the relocation that can be done by setting the segment pointers at run time Fm LDOS Support to Jim Beard I call it a clever kludge because it works at all in the general case for non trivial programs RE the 80
99. L to install them Fm MISOSYS Because there was a bug in RSHARD The fix files are in The Patch Corner Murphy strikes again ON ET Ka ee ene eam E i d a y yee OSS i F 4 PET OLID PSA WEEE RT GES eee i ks PAU F G BOX 287 a i PASSTOW N S ww 2211 RA hs va AALS Pt mee Sedge Weenies el tak A ma LDOS Information riieteta 00 E 2 S Stevenson hard coded and unchangeable Volume I iv LDOS on Model 4 and cursor speed Fm MISOSYS There appears to be a large contingent of active TRS 80 users in Australia I believe they number fourth in size based on our customer base US CANADA GREAT BRITAIN AUSTRALIA The Adelaide Micro User News is one of a few club publications sent to us The June 1987 issue carried a review of LDOS 5 3 written by Rod One paragraph of Rod s review expressed puzzlement over clock speed and its relationship to cursor speed Since I am in a prime position to clear up that puzzlement Rod this one s for you Before Tandy had finalized their implementation of the Model 4 hardware one of the few requests for a design change submitted by LSI and actually accepted by Tandy was for a faster interrupt rate The Model III interrupt occurs at arate of 30 per second Since a Model 4 has to be able to emulate a Model III exactly its interrupt rate is also 30 per second when the machine is switched to the 2 Meg clock rate When the Model 4 is switched to a 4 Meg clock rate t
100. LD EQU LD OR RET LD BIT RET JP LD LD LD INC DJNZ PUSH DEC LD CALL ADD LD DEC LD OR JR POP PUSH LD BIT JR PUSH PUSH PUSH RES LD CALL BIT JR CALL OR JR POP POP POP POP RET LD LD E OH READS A GATCD OF7H GATCD A WRTSYS HL CHANGD DE MSG1 DE MSG1NO HL LOGOT HL 0000H SP 0000H 2H A H SFLAGS NU Pn re GABORT B 5H A 20H DE A DE MSGINI DE DE A OAH DIV A 30H DE A DE A H L NZ MSGIN2 DE HL HL KFLAG 0 HL Z KEY3 AF BC DE 0 HL BC OBOOH GPAUSE 0 HL NZ KEY1 GKBD A NZ KEY2 DE BC AF HL HL BUFSEC C OH 58 Volume I iv if so read GAT sector sreset bit 3 of GAT CDH sin place write new GAT snumber of changes to sprepare exit message sput number in message srestore SP and prepare to end sput of files in message first 5 spaces in place sdo it backwards HEX to DEC sto ASCII sin place still done with the five skb check for lt BREAK gt write system sector Applications for the User Volume I iv 530D 530E 5311 5313 5316 5318 531A 531B 531E 531F 5320 5323 5324 5326 5327 5329 532C 532E 5331 5332 5335 5338 533B 533E 5340 5342 5344 5356 537B 537C 539A 53B3 53B5 53D7 53E8 53EA 5414 5415 542D 542E 5444 5445 546F 5470 5500 55CD 5514 5200 CD6847 2014 CD7247 FE 06 200D c9 210055 0E00 CD7747 c8 FE06
101. Laboratories at the same time RATFOR became a standard UNIX utility The West coast UNIX community adopted RATFOR as a systems language in lieu of C for a UNIX like shell for use on all minicomputers not using UNIX In addition the Berkeley UNIX organization has continued to develop and market RATFOR as a major enhancement of UNIX 4 2 A major extension of RATFOR called EFL Extended FORTRAN Language is also included with UNIX 4 2 UNIX 4 2 is marketed commercially by Mt Xinu UNIX spelled backwards Although alive and well in the UNIX community RATFOR is largely unknown elsewhere An article I wrote for Computer Language Magazine for their February 1986 issue was printed as their Exotic Language of the Month page 75 Therefore a brief description of RATFOR is in order The RATFOR language is defined in terms of tokens Tokens are alphanumeric identifiers numeric constants and non alphanumeric characters Statements are made up of tokens and are delimited by carriage returns or semicolons The language is free field Meaning that there is no particular significance to any particular colum Lines are usually limited to 80 columns so that unformatted listings may appear on standard 80 Announcing New Products THE MISOSYS QUARTERLY SPRING 1987 6 Volume I iv column terminals and printed pages Nearly all FORTRAN keywords are carried over including the extensive and powerful standard FORTRAN function
102. OS s APPEND command the target file of a need to go with the using the STRIP parameter file to search for a way to strip for purposes My first into here Applications for the User when the source file is NULL My next attack was to write the STRIP BAS program in BASIC which follows This can be used to remove the last character of a file by copying all but the last character to another file but APPEND aborts 99 Applications for the User Volume 10 20 30 40 50 60 70 90 100 Ae Eee ee AE cea ee eee ea A eee em ee ree E ce ce ge re re ee ee ee ce ee ee ee re ee ee ee ee ee ee EN E MORSE Fm MISOSYS Here s one I found on one of the SOUND routine The address is LI iv THE MISOSYS QUARTERLY SPRING 1987 DEFINT I N INPUT Input filespec INFILE OPEN R 1 INFILES INPUT Output filespec OUTFILE OPEN O 2 OUTFILES LC VARPTR 1 GET 1 IPTR 0 OD GOSUB 80 ITYPE ASC CH GOSUB 80 NLEN ASC CHS PRINT ITYPE NLEN IF NLEN lt 3 THEN IF ITYPE lt gt 2 THEN NLEN NLEN 256 Allow lengths to 258 bytes FOR I l TO NLEN GOSUB 80 NEXT I Feed through block IF ITYPE 2 THEN PRINT 2 0D CLOSE END ELSE GOTO 40 IF IPTR 255 THEN 100 l AN CHS CHRS PEEK LC IPTR IPTR IPTR 1 OD 0D CH RETURN CH CHR PEEK LC IPTR PRINT 2 0D CH OD GET 1 IPTR 0 RETURN Volume I iv Code from a Model 4 4 I note that LINE 100 pokes 18H into low memory This makes a slight alter
103. Only the segment registers ever need to be changed An EXE file is more like dynamically relocating a driver or filter under LDOS A table of things that need to be fixed or relocated is maintained and then once you know where the code is going to go you can re align the stuff in memory The quote you mention is not strictly true the segments contained in the program must be aligned to some degree For executable code that is typically a paragraph every sixteen bytes boundary That s because the segment registers can t be adjusted any closer that that MS DOS Information THE MISOSYS QUARTERLY SPRING 1987 5 51 13 Volume I iv 2 For a standalone machine language program that uses only interrupts no interface to BASIC true 3 The HIGH option places a flag into the header code so that the system EXE file loader will shove it up to the top when loading Your DOS tech ref manual if IBM s documents the structure of the header for an EXE file not that you really need to know it for this 4 Sure The more memory you have the higher the top of high memory is right Remember the 32K Model 3 vs the 48K Model 3 Forcing your code to load at a specified address is even more foolhardy than doing so in the LDOS TRSDOS 6 Z80 world The correct way to do this is documented in Appendix B of the IBM BASIC Manual for version 3 Do you have access to this or do you want more details You really need real
104. PATCH 014 This patch will correct DISKCOPY to properly deal with software write protected source drives This patch to be applied to DISKCOPY CMD UTILITY File dated before 07 01 87 D03 A8 78 F03 A8 F8 BOOT3 FIX 06 09 87 For Level I BOOT SYS LSIDOS PATCH 015 This patch will cause the display LEVEL of 6 3 0 to change from I to K This patch to be applied to BOOT SYS LSIDOS File dated 12 01 86 D02 1E 4B F02 1E 49 BOOT3 FIX 06 09 87 For Level J BOOT SYS LSIDOS PATCH 015 This patch will cause the display LEVEL of 6 3 0 to change from J to K This patch to be applied to BOOT SYS LSIDOS File dated 12 01 86 D02 LE 4B F02 1E 4A qen ee y ee g SE ere D A G a eee ele ee ree ig Se D AA HAUS van ee SUD O D e eee See eee See a ete TUND ee ere eee See eee eee D eee a eee Se D eee eee eee See ee ee ee eee ee See ee eee eee eee ee eee eee eee ere ere ee ey e ee o p im ee mio annan mein Ti M RGA SEND m mm vorme Vitis V eti M RA m as E a iaa ni _ Tova Veldi met vahti MOVE MU oe oe ee es DIRI FIX 07 01 87 For Level J SYS6 SYS LSIDOS PATCH 016 This patch will correct the handling of a very rare exception by DIR This patch to be applied to SYS6 SYS LSIDOS file dated 02 10 87 D09 C9 FF F09 C9 00 Y ar awn m tt A a een A Be PLE Y yoy cP hy A i a z kd 4 8 ns a as ae 20 SO Ta van 28 Qe OO Re eae a cy 3 97 h eo BUA
105. Provides stucture and greater portability to FORTRAN programs Fully documented with tutorial user manual Requires FORTRAN compiler 99 95 5 S amp H Note Mode ttl products may be available on request MISOSYS Inc PO Box 233 Sterling VA 22170 0239 703 450 4181 MC VISA CHOICE Orders Only 800 MISOSYS 1P 5P EST Monday Friday VA residents add 4 2 sales tax S amp H Canada add 1 Foreign use S amp H times 3 Applications for the User Volume I iv Volume I iv The Programmer s Corner Using I O redirection with PRO MC for debugging purposes by Michael R Johnston CIS 73240 1777 05 06 87 Users of MISOSYS MC or PRO MC compilers may or may not know of the special I O redirection inherent in the pre processor MCP and compiler MC that are part of the package Using these features it becomes possible to Save a great deal of time and some money In this article I will attempt to give some helpful instructions on using this feature made before we Some special notes need to be proceed First the standard output file stdout is normally the CRT This file is used to display all copyright notices and informative messages such as Including STDIO H Second the standard error file stderr is used by the pre processor and compiler to output ERROR messages Normally stderr is sent along with stdout to the CRT when one begins to play with the redirection feature of these two programs th
106. RLY SPRING 1987 2 Volume I iv time I have to spend working up articles and the more time I can spend on programming Now speaking of TMQ we felt that was so valuable we started paying corrections If you re a First subscriber you may have noticed Forwarding and Address Correction the content for address Class Mail the phrase Requested If you have moved you ll still get your QUARTERLY but then we have to pay 0 40 to find out that you moved For Bulk Mail subscribers the phrase Address Correction Reguested Forwarding and Return Postage Guaranteed will cost us a lot more To begin with forwarding is handled according to First Class rates and we still have to pay the address correction fee Now here is the kicker Although your change of address filed with the Post Office is supposed to be good for one year we have been informed that the COA is valid only for THREE MONTHS for magazines We have gotten a few back already noting Forwarding time expired This hits on us for anywhere from 1 18 to over 3 00 which Brenda refused to pay So please if you move don t forget your subscriptions and get us a change of address The PROGRAMMER S GUIDE now available We recently acquired a two color high quality copier This gives us the capability of producing in house small g antities of printed matter We have decided to use this facility to make available onae again THE PROGRAMMER S GUIDE TO LDOS
107. RSDOS 6 2 PRO WAM APPLICATION MODULES ADDRESS Mailing Labels and Rolodex MCards e BRINGUP Tickler File and Appointments e CALENDAR Any Month From 1582 to 4902 CALCULATOR Four Function Floating Point e RPN CALC Seven Function in Bin Oct Dec Hex e CARD 480 Character 3x5 Cards for Notes and Data e CHARSET Display All Video Characters e DIALER Telephone Number List and Auto Dialer e DOSAVE Save Entire Screen to Disk e TERM A Really Small Terminal Program TYPER Line Buffered Typing to Your Printer PRO WAM cccccccceceees 74 95 5 S amp H plus 7 new applications in release 2 Here s Mr ED More powerful applications for your PRO WAM This pac includes an editor for your every need You get e DED Edit disk sectors On any drive FED Edit file records in any disk file MED Edit any page of memory in any bank o TED A pop up full screen text editor VED Edit the video screen gives cut and paste e CARDFORM Populate CARD with a form DOLOAD Brings DOSAVEd files back to the video e REGENBU Shrinks BRINGUP DAT file Mr ED aata 59 95 3 S amp H What You don t have PRO WAM yet 800 647 6797 MISOSYS Inc PO Box 239 Sterling VA 22170 0239 703 450 4181 MC VISA CHOICE Orders Only 800 MISOSYS 1P SPEST M F VA residents add 412 sales tax S amp H Canada add 1 Foreign use S amp Htimes3 Announcing New Products Volume I iv Letters t
108. S not DS as the segment This is due to the fact that BP is designed to access local procedure variables allocated on the stack ES is a spare The Programmer s Corner Volume I iv It is generally possible for MOV instructions to override the default segment value In this case the SEG segreg instruction is used or the segment name can be directly given as with MOV DS BP DI 12H SI Rather than using the stack segment by default the DS or data segment would be used instead Note that segment overrides only last for the following instruction amusing implications of the fact that there are 64K exactly the same physical memory location using 64K different combinations of segments and offsets arithmetic overflow allowing One of the Segmentation is ways of specifying Another implication is that is it difficult time consuming and annoying to attempt to deal with more than 64k of memory at a time with the 8086 due to segmentation This is because there are no arithmetic instructions which operate directly on the segment registers and no way to add a 20 bit address directly to a segment register offset register combination The main reason for this mickey mouse scheme is that Intel designers basically didn t want to have to deal with the issue of extending an 8 bit processor the 8080 and so just slapped in a few extra registers with minimal instructions for dealing with them
109. SOSYS QUARTERLY SPRING 1987 Tom Gallaudet The Sort key in the first 96 Volume 1 1v Fm MISOSYS The only thing you would lose by using a function key as the PRO WAM activation hot key is the ability to get the PRO WAM menu displayed once you got into an application Some folks forget what they have assigned to a key If you are already in one application pressing the hot key still brings up the menu unless the hot key happens to be a function key then you would invoke the application assigned to that function key Fm Dirk Vandenbossche There is a problem concerning the file WINDOW CCC also included in the PRO WAM packet When compiling this program with MC I came across error messages In the manual you describe on page 2 12 the pointer operations that can be used The operations lt lt gt gt 8 and seem not possible Can you please indicate how you did compile this program Fm MISOSYS Your problem with the WINDOW CCC functions is because they were written for our LC compiler long before MC was released MC being a full compiler is tighter about your use of identifier types MC flags an error when you try to use a pointer in a non pointer Operation even though on the particular implementation an integer and a pointer are of the same bit length It just makes good sense for portability to insist that pointer operations use pointers and normal arithmetic operations using pointers except tho
110. Sd a i b r Lail ae SES i PO E Sig i VV 3 OT TATA Pag TS in t Tt If you wish to use a bank other than bank 4 replace the 4 s with whatever bank number in hex up to OF you wish to use Note in the 7th line that the number is 84 id NOT 04 DOL 1A 04 FO1 1A 02 D01 9E 04 F01 9E 02 D02 1F 04 F02 1F 02 D02 28 04 F02 28 02 D03 83 04 F03 83 02 D03 F2 04 F03 F2 02 D07 EC 84 F07 EC 82 DO8 32 04 F08 32 02 D08 50 04 F 08 50 02 The Patch Corner 97 The Patch Corner Volume I iv THE MISOSYS QUARTERLY SPRING 1987 AUTOIN51 FIX patch MC51 CMD RND 05 19 87 allow arbitrary expression as initializer of an automatic scalar variable 05 00 B9 75 01 00 B9 75 05 00 BB 75 01 00 BB 75 D1E EB 04 F1E EB 04 D23 B8 21 D23 C0 13 F23 C0 00 D23 D0 36 F23 D0 21 D23 E0 77 F23 E0 70 D23 F0 B5 F23 FO E5 D24 00 BF F24 00 11 D24 10 09 F24 10 2A D24 20 B4 F24 20 76 D24 30 C3 F 24 30 50 D24 40 F1 F24 40 2B D24 50 CD F 24 50 65 D24 60 70 00 00 06 21 22 C3 00 B4 E5 CA CD Fl 08 20 59 C3 21 19 76 E5 6F 8B 70 CD B9 BB 00 D7 79 Fl 00 22 CD 2D 9B 7C 00 11 70 2D 06 76 18 Cl Fl E3 Li 7E 75 75 39 54 B4 C9 22 69 AB 76 BF B5 19 2A E5 76 00 2A 75 21 E5 C5 04 E4 03 03 TE E5 2A CD 79 70 BF E5 Fl CA 6E 59 CD 21 39 65 2A 6B Cl D5 00 7C 00 00 23 E5 71 B9 B4 CD Fl C
111. T ELSE SOUND TONE DASH FOR LL 1 TO DELAY 2 NEXT NEXT FOR L 1 TO DELAY 9 NEXT RETURN PRINT PRINT Enter Filespec LINE INPUT F OPEN R 1 FS5 1 FIELD 1 1 AS C FOR F 1 TO LOF 1 GET 1 F A C GOSUB 210 AS INKEYS IF A THEN CLOSE GOTo 530 NEXT F CLOSE GOTO 530 DATA oTa eee o Teo s 8 ge 5908 s Sgt eegee 96777 ge gue ogg 5 393 Te s TTT DATA poe sry ce my eee my emmy mee T mem mee DATA 6 emmy es esT eT ey weg ng my em ee meee em ye eees DATA eees Toe Te oe meme me AS INKEY IF AS THEN 530 A RND 79 43 AS CHRS A GOSUB 210 GOTO 450 PRINT PRINT 5 4 WPM 4 5 WPM 3 7 WPM 2 10 WPM 1 16 WPM O 26 WPM ations for the User 60 Applications for the User a ta n s Volume 480 490 500 510 520 530 540 550 560 570 580 590 600 610 620 630 640 650 660 670 680 690 700 710 720 730 740 750 760 770 780 790 800 810 820 830 IL iv THE MISOSYS OUARTERLY SPRING 1987 Volume I iv PRINT PRINT Set Rate 0 7 INPUT A IF A lt O OR A gt 7 THEN 480 RATE A GOTO 140 PRINT PRINT Set TONE 0 7 3 INPUT TONE IF TONE lt O OR TONE gt 7 THEN 510 M CLS PRINT PRINT TAB 13 MO RSE CODE LEARNING SYSTEM PRINT TAB 13 Version 1 0 9 Feb 1985 by Logical Systems Inc PRINT TAB 21 This program is Public Domain PRINT PRINT PRINT PRINT TAB 28 MA IN MENU PRINT PRINT PRINT 1 Set TONE PRINT 2 Set RATE PRINT
112. Terminals and Printers Handbook might be useful too I have a copy of the VT 100 User Guide if you have any specific questions Requests for CIS files on disk Fm jjkd The original author retains all distribution rights for materials uploaded to CIS So if at their behest an article is routed to somebody else for republication there is no problem If the author limits further distribution or the author is not consulted does not give permission there is a problem Wholesale reproduction of a group of files from CIS can be viewed as violating CIS Compilation copyright on them all as a collective work Upgrading MISOSYS Products Fm Bob Connors Can you tell me how to and how much is an upgrade to the products you carry I would like to upgrade my SAID from 1 0 to 1 1 since I have noted that that is the version currently being patched somehow the fact that the version number changed Letters to the Editor THE MISOSYS QUARTERLY SPRING 1987 2 15 ae Volume I iv escaped me Do I send back the master disk or is my registration card I mailed in sufficient to request an upgrade I guess mailing the disk back makes more sense I think I have seen this policy stated before just can t remember where One final thing do you think that it would be possible to include the current version number of your products in your price lists Would sure help to let us busy types know when an upgrade not patchable is nee
113. X and SDAT62 FIX in The Patch Corner able to reproduce Fm MISOSYS In TMQ I iii Mike Harrow asked about the Data Table length values issued during an EnhComp compile Here s what the tables are for in order from 0 17 variable names label names label addresses variable types variable addresses data forward references line numbers and resolved addresses IF forward references COMMAND names COMMAND addresses FUNCTION names FUNCTION addresses FUNCTION types ERROR list reserved reserved reserved COMMAND FUNCTION local variable pointer list For what its worth there it is MC C Compiler changes for x 3 Fm Donald P Vincent Jr Are there any changes required to PRO MC to support LS DOS 6 3 Fm MISOSYS We expect to have an upgrade to MC later this year when I get unburied We expect to tweak the time functions to make use of the time field and extended date for stat amp fstat Stat is a newie When we have something to announce we ll announce it MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 83 vi Volume I iv MC and GREP speed Fm mark harris By the way grep compiles and works under PRO MC Assuming that GREP4 CMD was compiled under the AZTEC C PRO MC optimized CMD file is about 3K smaller but seems to run a tad slower Speed is only an issue on certain expressions Ex foo takes a lot longer to process than foo Fm David Huelsmann One explanation tha
114. _ptr count 0 edo something continue with function In this case the C function will not go to the do something but will continue on past the end of the if statement Yeah I know have you checked your logic and punctuation YES YES YES Nothing is wrong except this small idea that C doesn t like to use it s passing variables as working variables Les Mikesell The variables function act exactly like conveniently initialized local variables Your function looks reasonable but of course I can t tell if the if buf ptr count o statement should be true or not without knowing what is passed to the function I usually add a printf to display the variables for debugging when something doesn t work the way I expect It is pretty easy to have an off by one error working with 9 based arrays Fm passed to a Fm that s what I wanted to hear Shane Dawalt Actually I didn t like the idea of creating variables inside a function for variables passed to the function Now what s wrong with that dumb thing Fm jeff brenton Well the function looks ok but did your calling function know that buf ptr was a pointer and not an int Is it possible you passed it as amp buf ptr which would send a pointer to the pointer to your buffer rather than the pointer itself As for the variables themselves you CAN use them within the function without copying them aroun
115. acility in ADDRESS and CARD BUP also incorporates the REMOVE command which used to be in BRINGUP I keep an ADDRESS file in the computer on my desk Since I like to use it to capture address headings when writing letters I got tired of exporting the address record as it appeared in the window which required editing in my text editor That s why I designed HEAD This new application accesses the ADDRESS DAT file and presents an address record in the format Company gn rt esseed First Last gy ie r ifm i ARG addressl z i S an ern fe Z f address2 E PADS 2219 City ST ZIP i MO titan If the Company field is blank then it won t appear in the window Likewise if Address2 is blank it won t take up a line I even coded HEAD to drop trailing spaces from the fields so for instance there s no big gap between First and Last HEAD provides commands just like ADDRESS search next prev first last An lt ENTER gt will automatically export the record The application is useful for extracting address information for use asa heading in a letter It sure saves me some time and eliminates needless editing We ve modified ADDRESS APP to pick up the sort key and key length so that the search command now uses the external key information Announcing New Products Volume I iv specified in the data file header record rather than hard coding the search to the last first fields This means you can alter t
116. across an interesting fact It appears that when using the FORMS command with PR FLT installed if I specify only one parameter on the command line FORMS resets all the other parameters to their DEFAULT and no DEFAUT is not specified on the command line Is this the way it should work Could it be that the default for DEFAULT in FORMS is set to ON for the MAX 80 Also why isn t EnhComp set up to use the PR FLT I have PR FLT set up as follows P 66 L 57 M 5 C 75 T F While EnhComp will use the margin setting and I assume number of characters on a line it does not abide by the lines per page or the hard top of form code Instead it uses apparently its own internal printer routine Line feeds for TOF are most annoying Is there anyway to get EnhComp to use the system configuration at least as far as the printer routine is concerned Fm MISOSYS That s the way it works I didn t realize that TRSDOS 1 3 has a PR FLT That s what it would take to have EnhComp use it Although I would certainly like to see LDOS as the standard DOS for the Model III it just ain t so Thus EnhComp was not written to use the LDOS PR FLT formatting That s the real world I don t think there is a way to MISOSYS Products Tidbits Volume I iv get EnhComp to use the LDOS PR parameters short of writing a library function in Z80 MODE to accomplish that Perhaps this message will tickle me to post the internal to EnhComp storage names w
117. add 5 to the c register Memory transfers are also oriented primarily for use with HL IX and IY The latter two registers IX and IY provide the most versatile memory addressing modes of the Z80 Any 8 bit value or register can be loaded or stored to memory with an optional offset from the base register value supplied For the most part the Z80 can be described as an 8 bit processor Although certain 16 bit operations are supported such as addition of a 16 bit register pair to HL or the loading and storing of 16 bit registers to from memory most instructions of the Z80 deal with 8 bits at a time Furthermore there is an upper limit of 64K of memory directly addressable to the Z80 This is due to its 16 bit address bus This imposes a restrictive limit on the sophistication and size of the programs which may readily be run on a Z80 based computer Now to look at the registers of the 8086 and compare them with the Z80 First of all the 8086 has the following 8 bit registers AL AH BL BH CL CH DL DH Editor s note notice the Similarity in quantity and correspondance of 8 bit registers between the there are the 16 bit registers BP DI SI SP CS DS ES SS As with the 2Z80 certain pairs of 8 bit registers may be treated by some instructions as a single 16 bit register Thus AL and AH combine to form AX and similarly with the other 8 bit registers to form BX CX and DX There is no convenient way howeve
118. al drivers filters options you desire Then type SYSTEM SYSGEN to save this configuration SUPERLOG will be operational each time you reBOOT EXIT The Patch Corner 102 Volume I 1v The Patch Corner Volume I iv Editor Note Contrary to The Blurb I found room to squeeze in this section Our Compuserve Forum PCS49 FORUM commands The following commands may be entered at the main forum FUNCTIONS menu To find out more about a particular command enter the command at the prompt below If you d like a listing of the entire online HELP displayed at your terminal enter three periods without the quotes at the next prompt The listing is approximately 5 pages long Remember to use CONTROL A to stop the output from scrolling and a CONTROL Q to start the output For general information about using a forum type INstructions at the main forum FUNCTIONS menu Please note that H may be entered at most Forum menus and prompts to obtain help specific to that menu or prompt More information is available for INstructions Leave Read Scanning BROwse COnference DLibrary Bulletins MDirectory SSubtopic OPtions Help EXit USTatus SENd NEW Information Userlog SN ames LNames VERSion HIgh RF orward RReverse RThread RSearch RMarked RIndividual SThread WHO DAY TIME Forum policy on deleting messages Fm LDOS Support jjkd 1 Please delete all P rivate messages after reading A P rivate message
119. ally come up with some good advice Our subject in guestion was not which day was the first day of the 2lst century but rather was the year 2000 a leap year The responses also shed some light on the answer to your readers question In the interest of providing advance discussion on a topic which will certainly rise in significance as we approach 2000 I offer you a copy of the dialogue which was printed in my Winter 1987 issue for what its worth I ve highlighted the paragraph pertaining to your question 7 Glad to be of assistance but not a contender for your previous job in that other paper Compliments and complaints Fm Lloyd Davidson Hello folks Have been absent from the forum for some time as I have been busy with other computers languages and a coupla BBSs The first thing I did when I logged on then was to download and read off line all of the current messages from all categories A task that is much easier and less expensive since CIS decided to give us a break on 2400 baud Would like to comment on a few of the messages for no other reason than to get my two cents in LS DOS 6 3 From an end users point of view this is a dandy system Does what it is supposed to and does it well I have a number of disks configured in many different ways and have not had a hitch from 6 3 One disk I use for word processing has both SuperScripsit with dictionary and Powerscript on it Even threw in TED and have Pronto
120. amming under LDOS or LS DOS thus my thrust will be to expose these readers to the kinds of things which can be done under MS DOS which would be very similar to what you are already used to Thus it is expected to serve as a bridge 8086 coding is confusing enough From my check just now as I write this The TRSHD6 DCT driver is 330 bytes long when resident RSHARD6 DCT is 335 bytes long 5 bytes longer than the Tandy supplied driver Mine is longer due to the additional coding overhead necessary to allow partitioning by both cylinder and head and on supporting physical cylinder sizes up to 1024 cylinders Fm Michael Johnston Does anyone know if the EXMEM and BANKER programs have been uploaded to Compuserve I have the TMQ but I hate to type them all in I also need the update fixes to the MC libs to include the EXMEM functions Fm MISOSYS The EXMEM and BANKER program are part of the Copyright which MISOSYS has on its THE MISOSYS QUARTERLY The only thing we upload onto the board presently are the fixes File copies of any programs which appear in TMQ can be obtained from DISK NOTES unless the original author of the program is not MISOSYS and chooses to distribute the his her program in another manner I have not made any decision on uploading other parts of THE MISOSYS QUARTERLY other than the fixes On Family news Fm JACK LOTTEY I just received my TMQ I iii as I returned from my vacation Ref the p
121. anced CARD file It supports up to ten CARDx DAT files 0 9 rather than the single one under Release 1 It makes use of reverse video to designate protected screen positions which cannot be edited nor exported CARDX will actually skip over any screen position which is in reverse video CARDXF APP is used to construct and edit a CARDFORM record for use with CARDX This wammie allows you to create new data files and add edit a cardx record identified as CARDFORM Reverse video which is used for protected fields in CARDX APP is toggled via lt CTRL R gt The screen indicates the current state of reverse video CARDXF also allows you to populate the CARDX file with any designated quantity of records initialized with the CARDFORM form Announcing New Products THE MISOSYS QUARTERLY SPRING 1987 11 Volume I iv It s easy to use the cardxf editor to paint your form onto the text area of the window By using reverse video for every position which you want protected from entry while using CARDX you can create a form to be filled in later When using CARDX the cursor will be automatically moved to the unprotected positions This combination of CARDX CARDXF can really provide you data base functions Around here we always seem to forget to do this or forget to do that Sometimes BRINGUP doesn t work when you just want to keep a list of items to be done Remember that todo pad That s why I designed TODO APP
122. and that the directory entries under this release are no longer than they were in the past You can read and write diskettes from previous versions of LDOS and TRSDOS 6 while running LDOS 5 3 with absolutely no hazard to the data or directories on those diskettes The only problem is that the directory entries will be maintained in a manner compatible with the previous versions no timestamp and no dates after December 31 1987 You cannot inadvertently destroy your disk if you do not follow instructions and that is why the manual does not tell you that you can Unless of course Mr Feldman is the sort that uses FORMAT when he means BACKUP or switches the drive specifiers on a BACKUP command or does any of the numerous things that can destroy disks in any DOS system No manual can save you from yourself LDOS 5 3 is indeed downward compatible with all previous releases of LDOS and of TRSDOS 6 Perhaps what Mr Feldman really means to say is that previous releases are not upward compatible with 5 3 Under certain limited circumstances involving password protection Letters to the Editor THE MISOSYS QUARTERLY SPRING 1987 218 Volume I 1v LDOS 5 1 x will be unable to written on a diskette formatted by version 5 3 This can be prevented by not using password protection on files that must be read in this particular and unusual manner access files In reply to his complaint about the lengthy routine ne
123. any caveat about THAT problem before e g having to MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 85 Volume I1 1v include a struct template in every file using Said structure As I read all the comments a struct is a type of variable and variables can be declared as extern and presto all the problems would be cleared up at link time This omission is not just in the MC docs I have never seen a hint about it in ANYthing I have read Well NOW I know one more thing about this weird C language I wonder if 1 11 ever get enough knowledge to feel as comfy with C as I do with B S C Les I think I ve got it Thanks for that lucid explanation One reason for my confusion is that I have NEVER seen the illuminating sentence you wrote to wit Preprocessor defines structure declarations and typedefs must be repeated in each file I would have intuitively decided on the defines but have NEVER seen even a HINT that structures and typedefs needed repeating as you suggested with the include technique Your guess was exactly right the structure definition was in another Separately compiled modu le When I added it to the one in question all smiles until the next mystery comes along Fm LDOS Support jjkd See the new Dr Dobbs not the Baby Duck issue the next one for a clever macro solution so that you can use a single external file to handle both situations
124. ar we will be making public a set of patches to accomodate Model I LDOS users These patches will de activate the Directory file dating and extend the DATES dating to accept dates through 1999 Surplus hardware for sale Brenda asked me not long ago going to get rid of all that extra hardware stored in the back room I m not sure she used the term hardware When MISOSYS picked up the operation LSC used to run out in Milwaukee we acguired what can be best called an excess of computer eguipment if you get my drift not to mention an excess of office furniture George Carlin refers to this as stuff Brenda s right it s time to see if anyone has a need for some of this stuff at a price of course Here s a list of what we sorted out so far as extra Prices shown include ground shipment via UPS Sorry but this stuff can only be shipped to continental US only Everything is used and sold as is All equipment was in working order when removed from service Qty Item Price 4 TEC FB 202 40T floppy 25 3 Texas Peripherals floppy 25 l Shugart Assoc 400 floppy 25 2 BASF 6106 2 3 ht floppy 25 l Percom 20ms floppy jaws 25 2 MPI B 51 floppy 25 7 floppy case amp power supply 10 1 MAX 80 128K keyboard 200 1 BMC Monitor BM 12A 40 1 16K Model I complete 1 c 100 1 32K Model I E I 50 All floppies noted above are bare drive We have other stuff around here Like a Line Printer III an MX 100 a DP 800 some 5 m
125. ared as static in another file which might be linked into the same program Global variables defined outside of functions without being declared static are accessable by functions in all files but they must be declared extern in all files where they are used except the one where they are defined Most compilers will resolve multiple definitions of the same variable into a single global but don t count on it Fm LDOS Support jjkd Yes external refers to defined allocated storage initalized external to this source code file The tag extern is not optional Static is defined allocated storage and possibly initalized in this source code file but not within any function The tag static is optional Fm Les Mikesell Static varibles may also be declared within functions or inside any block to provide permanent variables that are only accessable within the block MC Structures and bit fields Fm Nate Salsbury I ve been reading the referenced file in order to learn more about structures and bit fields Two questions 1 It seems to me that the part labelled DCT byte 4 flags is really going to be reading what the Tech Manual calls DCT 3 and the part labelled DCT byte 3 flags will really be overlaying what the Tech Manual calls DCT 4 It also seems to me that the two sections DCT byte 8 and DCT byte 7 will actually overlay DCT 7 and DCT 8 respectively Is that an error in the DCT CCC
126. ase the filter should operate transparently during this character output The character under test is obtained from the C register If it is not a form feed OCH the original flag state is restored and the filter jumps to the original DCB driver address This action was then totally transparent to any downstream modules The Programmer s Corner Volume I iv If a form feed is detected the flag state is forced to the state proper for output by the XOR A instruction Z NC of course that was the state after the CP OCH but then I didn t write PAGEPAWS At OUTP1 the form feed character is passed downstream by a CALL instruction this means the filter regains control after the formfeed is sent to the device assumed printer THE MISOSYS QUARTERLY SPRING 1987 Volume I iv The next short piece polls the keyboard taking care to save register DE as that is the only register whose contents are corrupted by an KBD call Once an lt ENTER gt is pressed the filter issues a POP to clean up the stack It then clears the flag register and returns to what called the module That s it a short introduction to filtering Now let s take a look at the same filter rewritten for DOS 6 x 00001 PGPAWS FLT Version 6 x 000A 00002 LF EQU 10 000D 00003 CR EOU 13 0014 00004 CHNIO EQU 20 0015 00005 GABORT EQU 21 000A 00006 DSPLY EQU 10 000C 00007 LOGOT EQU 12 0016 00008 GEXIT EQU 22 0008 00009 KBD EQU 8 0064 00010 HIGHS EQU 1
127. at the two begin to go their separate ways It is only Turning to page 1 8 of the MC manual one will notice on the bottom of the page a general example of how stderr might be redirected MC TESTLIB PR This example has compiler outputs astute observer will example has used all error messages that the going to the printer The note that although this a printer as the target for the output this does not necessarily mean that it could not be directed elsewhere In fact using the wonderful facilities of DEVICE INDEPENDENCE that were so thoughtfully included in TRSDOS one can redirect this error mesSage output to ANY device or FILE Since there are no other examples of this remarkable feature given in the manual it is left to the user to uncover and utilize them where stderr is A practical example is one redirected to a file for reference when debugging If you are using SAID the text the MRAS package you listing to aid in your debugging process In fact the error messages will tell you which line that the problem occurred It is then a simple matter to use the GO command provided in the Beene to locate the line for you editor supplied with could use this error The Programmer s Corner THE MISOSYS QUARTERLY SPRING 1987 long run it GA s Volume I iv The ideal setup is one where you load the C source file in one bank and have the error file in another All one has to do then is to swi
128. ation ee otsani Volume I iv Fm Ray Pelzer An easy method is the old stand by BASIC The statement OPEN R 1 GIFT1987 DA3 5 CLOSE creates the empty file s directory entry with roughly the same number of keystrokes as the CREATE command including the execution of BASIC LS DOS Information Fix for Tapel00 Fm Paul Jaeger I don t use tapel00 very often but occasionally I want to move a bunch of files from my model 4 to a modl00 cassette under batch control When I switched to 6 3 tapel00 stopped working I noticed that the sound of the data was much higher in pitch coming from the 4 compared to the data from the 100 I can t figure out why 6 3 would screw it up but here is the patch I arrived at hours of experimentation to adjust the cassette write timing in tapel00 tape1001 fix Patch to adjust write timing for 6 3 adjusts the wait timing between bit set reset to cassette port D06 0D 1A 16 18 03 11 32 2E F06 0D 17 12 18 03 11 2F 2B Patch to remove Cassette ready ENTER DOO 8D 00 00 00 F00 8D CD 3B 2A Hop The ratio of the two changed words seems to be important so keep that in mind if you want to experiment Fm LDOS Support jjkd I m not surprised that you needed to alter the timing to make TAPE100 work What is surprising is that it worked before and you needed to change it because of the LS DOS 6 3 upgrade TAPE100 has problems because it is sensitive to CPU speed It must be
129. ation to the consistent hard drives which we acquired when we acquired across 6 2 and 6 3 On the other hand the tail operation of Logical Systems last original value of 28H is not restored when you Judging from some other files Iwas terminate the program the menu doesn t really able to discern that this Morse code program give you an option to STOP Anyway for what was derived from one written for the IBM PC it s worth to you aspiring Morse code folks Apparently LSI reworked it for use on a Model here s CODE BAS the re year 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 410 420 430 440 450 460 470 Applic POKE amp H3AF amp H18 CLS CLEAR 5000 DEFINT A Z DIM TBL 48 FOR L 1 TO 42 READ TBL L NEXT RATE 2 TONE 5 V 1 RANDOM DELAY 100 RATE IF DELAY 0 THEN DELAY 50 DOT RATE DASH RATE 4 IF DASH 0 THEN DASH 2 GOTO 530 AS INKEYS IF A THEN 180 IF A THEN 530 GOSUB 210 GOTO 180 IF AS gt THEN AS CHRS ASC AS AND amp H5F PRINT A ELSE 230 CODE TBL ASC A 64 GOTO 290 PRINT AS 0 IF AS lt OR ASD THEN 260 CODE TBL ASC A 43 26 GOTO 290 IF AS OR A THEN CODES GOTO 290 IF A THEN CODE GOTO 290 PRINT FOR L 1 TO DELAY 20 NEXT RETURN A CODE PRINT A FOR L 1 TO LEN A IF MID A L 1 THEN SOUND TONE DO
130. aven t been sorry for my decision on buying PRO MC because it is useful I don t buy what I don t need and I don t expect others to do so even if it might be a 15 00 subscription to a magazine People who are ticked off at you should be ticked off with themselves I believe you are quite correct to discontinue support if sales figures don t warrant it I can hear the nashing of teeth already I think the Model I people should be wondering where all THEIR programmers went rather than placing the blame on you Finally portability is a problem with the Model I III 4 I have a strict rule when I develop software if it uses undocumented addresses PEEKs POKEs and whatall it should not be sold or distributed by information services or BBSes Too bloody hard to Support never know what people are doing on their systems great TMQ BIW are TMQs gettin larger or do I need new glasses NN Fm MISOSYS I hope I can quote you on that The last issue I iii was way too big It got that way because it was about 6 weeks late thus it had a greater time interval for input I did cut out a few things as I noted I expect the next to compensate so that the average is not above 100 pages Can t afford the shipping at that weight The Source for 6 2 Fm Kevin R Parris I understand that The Source was published by LSI but I bought my copy from you I also have read the disclaimer in the front of the book questio
131. ays evaluate to zero I wonder why it is used in lieu of the constant value zero I own PRO CREATE Ver IV and would appreciate an I have noticed as listings the expression explanation for this practice Fm MISOSYS It has become common practice to use the S nomenclature when coding a word Space which is to be subsequently altered during runtime Thus a quickie look at the source code gives you the knowledge of the dynamics of the code if you use the technique consistently DED86 MS DOS based disk editor Fm Jim Gaffney I received DED86 today and I m impressed It s all that I anticipated it might be and more I particularly like the documentation and having the foldout pages to illustrate the displays while you read the docs sans a computer is a masterstroke I assume from the docs that you would have no problem with the DEMODED EXE being uploaded to the bulletin board circuit Not the DED EXE file DSM Disk Sort and Merge Earl Jamison gave us a brief report that DSM4 aborts a Job Control Language file at the conclusion of processing That s the Model 4 MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 78 Volume I 1v version Bill Norman gave us a similar report concerning DSM51 Model III version Bill s report was that 1f you invoke a DSM51 MAP run from LBASIC 5 1 4 via a CMD DO when it gets back to LBASIC the program which invoked the CMD DO aborts Be
132. been recovered and the incremental cost of producing an extra copy of the same program would come out of the advertising budget I am beginning to learn how to use SAID I was am apalled to find that SAIDINS completely ignores the lt CTRL gt lt F1 gt lt F2 gt lt CAPS gt and lt F3 gt keys For a while I thought of building a clamp for the lt LEFT SHIFT gt key to enter text in upper case but then TMQ3 came to the rescue and I found out about the lt SHIFT gt lt gt combination That sure helps and I have to learn how to use LDOS I It is only now that use my Model 4 mostly in the Model 3 mode because 64 chars per line are easier on 58 year old eyes If your Model 4 hardware interface kit is half as good as I expect it to be it is the one thing I have been waiting for You are 100 right in NOT using the 80 x 24 chars in Model 3 mode I completely fail to see your point in calling 300 000 RPM a disaster in reference to Jack Decker he no longer publishes Northern Bytes I can see that such a degree of accuracy is technical overkill but it does NOT constitute a disaster To me it is more of a disaster having to initiate KI DVR and JKL of Model I and NEWDOS fame to be able to use the screen print function which is already present in ROM Now please don t get mad at me I merely want you to make your software more USERFRIENDLY How I hate that word but nothing else will do In the interest of user
133. ble for the order of sector access jumping around in the buffer as necessary This would give a speed up during file access as well as the format and verify You might kick around the idea of making a resident driver that would let TRSDOS read write directly to MSDOS formatted disks This could be done by faking the i o to the TRSDOS dir cylinder while actually writing the MSDOS directory I don t know if it could be done in a reasonable amount of memory though approach for make up your own Fm Adam Rubin I never thought of that before but I IL do it Let s see four revolutions per track times forty tracks is 32 seconds which sounds fast enough to me An interleave table would definitely be an improvement and I ll use it in the format routine One of the first things I considered was that many files need some sort of conversion to be usable when transferred such as CR to CR LF I decided to pass the file through a filter as it s moved so CONV TEXT on the command line would load TEXT CNV and call it for each byte read With this byte at a time approach I don t know how much faster multiple sector I O would be for file access Also I m not sure how useful direct access to unconverted MS DOS files would be What do you think Fm Les Mikesell If you are at all concerned with speed use full sector I O on the TRSDOS disk and read at least IK a track would be better at at time Then do any necessary convers
134. brings up the issue of typed symbols or variables It is possible to define say a byte variable in the following way XYZ DB 0 or a word WVAR DW 0 Obviously DB denotes a define byte operation DW a define word Unlike Z80 symbols when a symbol is defined immediately prior to a DB or DW it is typed to that definition XYZ in the above example then is typed as a byte variable and WVAR is typed as a word Essentially this information can be used if the symbol has been previously defined on pass 1 to specify the type of The Programmer s Corner THE MISOSYS QUARTERLY SPRING 1987 5 Volume I iv instruction generated So MOV AX WVAR would if XYZ had been defined as above generate the instruction to move the contents of WVAR into AX Otherwise according to ED ASM 86 convention it would move the offset of WVAR into AX Note that MOV AX OFFSET WVAR would unconditionally generate a direct load of the offset of WVAR into AX regardless of the type of WAR with ED ASM 86 Similarly MOV AX WVAR would unconditionally generate an indirect load of the contents of WAR into AX regardless of the type of WVAR again using ED ASM 86 conventions which are that bracket characters always specify an indirect reference Note however that if WVAR has been typed then it must match the size of the register For example MOV AL WVAR would generate an error because WAR was typed as a wor
135. c of Don Scarberry from the December 1982 80 U S I m not a DOS lover or a DOS lover s son But 1 11 use LDOS til a better DOS comes Now if somebody can just figure out what rhymes with SOLTOFF Report from France Fm Michel Houde Please excuse this familiarity but it is a real pleasure to meet you every quarter in TMQ It is also a pleasure to deal with you Your order servcing is FAST Ten days is the usual time I have to wait when the Customs don t open the packet a few days more when they do It s actually faster than getting shirts and trousers for the kids from most mail order firms here in FRANCE Further I usually have the feeling that I get more than what I paid for The latest instance is the RSHARD package which included Model 4 software although no mention was made of it on the flyer included with LDOS 5 3 By the way I found 4 bugs in RSHARD6 DCT While I am Letters to the Editor THE MISOSYS QUARTERLY SPRING 1987 19 Volume I iv on bug reporting I also found one in MAPPER from PRO MACH2 but the cure was easy I am pleased to tell you that SET2RAM works perfectly with MODELF III which is the French version of MODELA III the Germans have a MODELD III It is actually a modified version because the original like your network ROM did not allow LDOS 5 1 4 to boot Of course I had to remove those mods that made the clock run correctly with 5 1 4 I had to write KI4F DVR to a
136. c8 F6C0 CD0944 180A 212554 DD 211554 CD8A42 21EFFF C3BF532 3E20 18E5 OA OA 55 4C OA 28 20 OA 54 20 OA 20 OD 44 OD 57 OD 54 OD 0000 THE MISOSYS QUARTERLY SPRING 1987 DRIVEW READS DRIVER ERROR WPDRIV ABORT EXITOK BADDRV TITLE MSG1 MSGABR MSGWP MSGOLD CHANGD BUFSEC GATCD DIRERN EQU CALL JR CALL CP JR RET LD LD EQU CALL RET CP RET Volume I iv S 1H WRPROT NZ ERROR VERSEC 6H NZ ERROR drive number for writes HL BUFSEC read disk sector C OH 1H RDSECT Z 6H Z OCOH GERROR EXITOK HL MSGWP ODDH HL MSGABR GLOGOT Sate ce HL OFFFFH Lou EXIT Dg E A 20H ke ERROR sanS OAH OAH ie enon sdrive number for reads schanges next to LD IX xxxx UNDATE CMD v1 0 LDOS 5 3 to LDOS 5 1 Date Conversion OAH c 1987 Luis M Garcia Barrio 8 N 1 4 215 848 5728 OAH OAH The ACCESS passwords for all files will be removed OAH ODH file s updated conversion complete ODH Date conversion aborted ODH Write protected drive ODH This disk does not use the time date stamp ODH OOH 5500H BUFSEC OCDH BUFSEC 14H START a w a A A PS cs pee rm SSS r prey Se ech Se er Sm me ms a S i i AO a ao te es ee ey ee te ee ee ee ee ee ee es ee ee ee ee ee d an STRIP BAS by James Beard Ph D I had occasion off the last byte that I don t attack was to try the appending a NULL file to
137. cause of two similar reports I thought it best to go over the documentation and evaluate the two reports In my DSM4 test I sorted a 1200 record file of 128 byte records The JCL file was this is a test of dsm4 via jcl dsm4 usermap jcl s n userclub idx 2 a 2 this is the jcl continued I experienced absolutely no problems with this procedure I anyone has difficulty using DSM4 from JCL you must send me complete details I will need a copy of your data file I will also need your MAP file your JCL file and specific details concerning all of the responses you made to the DSM4 module which generated that MAP file Please provide all of the files on a floppy disk I also investigated the report of DSM51 aborting the running BASIC program which invoked a sort job via Job Control Language and BASIC s CMD DO commandstring In my test I sorted a 1200 record file of 128 byte records My test BASIC program was 10 PRINT Here before invoking DSM 20 CMD DO TESTDSM 30 PRINT Here after sorting 40 END The JCL file was this is a test of dsm5l via jel dsm51 userclub map 4 jel userclub idx 4 ei 4 this is the jcl continued I experienced absolutely no problems with this procedure If anyone has difficulty using DSM51 with JCL from BASIC you must send me complete details I will need a copy of your data file a copy of your BASIC test program which demonstrates your DO invocation and abort af
138. ccomodate my French keyboard but that s not a real problem as I ve been doing it since Model I time Of course I did it for LS DOS 6 3 THE SOURCE was a good help I ve done my home work to enable ALTDISK NAME UNREMOVE all from PRO ESP to work with LS DOS 6 3 I have even improved UNREMOVE to make it display mod time as well as mod date when more than one REMOVEd files are found By now you may wonder who s that guy out there in France doing all those things Isn t he giving sharing selling software I have never earned a penny nor a Franc from computer activity I am a teacher researcher in Chemical Engineering at Compiegne University I use a Model 4P for word processing and data processing we do a lot of modelization in Chemical Engineering I prefer using BASIC on my Model 4P than FORTRAN on the VAX I am the one and only here in Compiegne with a TRS 80 Model 4 But I do not feel too lonely as I am a long time subscriber to 80 Micro and to TMQ of course I know my computer intimately and thanks to you I can do what I want with it Please allow me to add a few words about my family My wife is named Sylviane PULVIN she is also a teacher researcher at Compiegne University Biochemistry Like artists or journalists female researchers usually keep their name when they get married I wouldn t mind She does not really like computing The Mac is OK for her We have three kids one girl and two boys aged
139. ce handler in 6 0 for many important reasons Here s a few LSI received many complaints over the inability to RESET devices and conveniently re install a filter Numerous kludges were implemented to various releases of 5 x s filters to provide a limited form of re installation The LDOS device handler also did not permit filtering a ROUTEd device Linking also took up high memory with a link module for each LINK command It was also necessary for a 5 x filter to pick up vector values from the filtered device s DCB and poke in its own Also when filtering the KI device the filter installation code had to account for the effect of JCL operation on the KI DCB s vector All in all it was very messy Now the ground rules for the device handler were parameters options etc Now the LDOS 5 x restricted by the ROM resident device I O filter routines specifically GET PUT and CTL handlers were in ROM unchangeable 00001 PAGEPAWS FLT Version 5 1 fee acta eeaites ede 4030 00002 ABORT EQU 4030H lamas 4467 00003 GDSPLY EQU 4467H adikiteti t3 402D 00004 GEXIT EQU 402DH an 80x137 0028 00005 KBD EQU 002BH oansSTOW NSW 2211 5200 00006 ORG 5200H harane manean 5200 D5 00007 PUSH DE 5201 1A 00008 LD A DE stype byte 5202 F5 00009 PUSH AF 5203 217552 00010 LD HL SIGNON 5206 CD6744 00011 CALL DSPLY 00012 00013 Mod I or III determination 00014 5209 3A2501 00015 LD A 125H sMod I or III 520C FE49 00016 CP 49
140. cessary to set up the RS232 driver or FORMS filter I urge Mr Feldman to read the documentation on the SYSGEN command which is not a new feature One only has to install such drivers once if they are made permanent by the SYSGEN command Thereafter they are automatically set up each time the system is booted You are doing disservice when your readers a terrible you print misleading reviews and incomplete information The advice given in Mercedes Silver s Feedback Loop is frequently incomplete or dangerous Likewise many of the so called hints that appear in the Reader Forum need to be screened by a competent person I have seen advice given there that is potentially disastrous and could indeed destroy diskettes or even damage hardware I was astonished to see that you paid someone ten dollars for instructions on uSing the REMOVE command of TRSDOS 6 2 that are fully explained right in the manual for that system Bill Boggess should look at his manual if he thinks he discovered something new when he began deleting more than one file with a single command line Jerry Semler s suggested abbreviation for controlling passwords from Basic is potentially dangerous he is overlaying all 8 bits of a system control flag in order to change just one of them Mike Zarowitz made the correct allowance for this in his original recommendation by using OR or AND to change only the desired bit Altering the others could produce unexpe
141. ch to LDOS 5 3 ROUTE Command Corrects re use of previous buffer FCB Apply via PATCH SYS6 SYS SYSTEM D36 31 6F F36 31 67 SYS6E FIX 06 18 87 Patch to DIR Corrects obscure problem in space calc Apply via PATCH SYS6 SYS SYSTEM SYS6E DOA 4D FF FOA 4D 00 TED5B FIX 05 19 87 Patch to LDOS TED Fixes text deletion on LOAD if DELETE mode is active Apply via PATCH TED UTILITY TED5B D04 CF DB F04 CF B9 TED6C FIX 05 19 87 Patch to LS DOS TED Fixes text deletion on LOAD if DELETE mode is active Apply via PATCH TED UTILITY TED6C D04 Bl1 AC F04 B1 8A INSTALL3 JCL Thomas L Keister Jr 15 Jun 83 LDOS 5 3 change by Luis M Garcia Barrio 7 May 87 i 8 N 1 4 215 848 5728 Following a Global RESET this JCL file will install SUPERLOG into a minimal floppy based system including KI DVR Type ahead and JKL screen print Valid for MISOSYS LDOS 5 3 For LDOS 5 1 3 restore the original PATCH listed below SUPERLOG series L PAUSE Type lt ENTER gt to proceed lt BREAK gt to exit Installing SUPERLOG Screen Storage to MEMORY SUPERLOG LOAD MEM Configuring System SET KI TO KI DVR TYPE JKL VERIFY OFF e Disable shift lt BREAK gt option of LDOS MEMORY A X 44DD B x D4 The LDOS 5 1 3 patch was PATCH SYSO SYS SYSTEM 0 D03 37 D4 The new LDOS 5 3 patch follows PATCH SYSO SYS SYSTEM 0 D03 2F D4 F03 2F CD Now add any additon
142. course are NOT in my REAL source code extern char lbrname extern char nulls extern int num slots define MAXFILES 256 find memb fname char fname int slot if get directory 1 error Fatal error program terminated n slot find fname if slot 1 lt printf s is not in the library n fname return 1 return slot get directory unsigned size int 1 dircrc auedeaed cksum 0 extern struct ludir ludir lt MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 modu le 84 Volume I iv extern struct ludir ldir lt lseek lbr OL 0 if read Ibr Sid sel0 seizeot indir sizeof ludir First record return 1 num slots 1dir 0 len 128 sizeof ludir The pre processor MCP doesn t have any trouble with this but when MC starts up I am only trying to get the REL module to put into a LBR it promptly stops with the following error message get_directory 9 error 79 Unknown struct union if read lbr amp ldir 0 a eb ir and if I let it continue it barfs on amp ldir O next I tried this first with the extern struct etc declarations at the very top of the file along with the other extern declarations Same result in fact that s why I moved them into the body of get directory The comment in the MC manual for Error 79 is The structure union object being dec
143. cted or unwanted results MSDOS and IBM copycats are dominating the microcomputer market today I take it you intend to enhance this domination by spreading misinformation about the alternatives Perhaps it is time you put a warning on your publication Let the reader beware Recovering REMOVED files Fm Jim Owens Where can I find info on recovering removed files I just had a major oops this morning on the 4p Any help will truley be appreciated Letters to the Editor Volume I iv files are probably invoked any writes Fm Bob Haynes Your recoverable IF you haven t to the disk in the meantime The process involves changing certain data in the directory sectors There are articles dealing with this subject in the 10 85 and 7 86 issues of 80 Micro magazine which will help you If you are not a subscriber any large public library will have them The second alternative is to use a MISOSYS product called UNREMOVE which was part of their PRO ESP package discontinued but now available within a larger package called MARK Iv 99 A bit steep just to fix a disk I ll admit but the many other useful tools included may make it worth considering TMQ Literary Section to further assure I propose the basis of a for THE In order LDOS urgent poetry section Fm Dave Hardy the immortality of establishment on an literary and epic MISOSYS QUARTERLY To start the section following is the immortal epi
144. d Those values plus the ones defined as local to the function you are writing are available to the function block MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 Fm 90 Volume L iv Last thing are you SURE there is a XO in the bu ffer Fm Shane Dawalt In writing that message I was assuming that the calling function was correctly written and assumed the proper things As for my program everything is fine Your idea about NO has already been thought of and I have placed an initialization statement at startup to take care of this Actually it was declared extern and public to all modules functions which should have cleared the space to O automatically Or am I wrong to assume this Fm LDOS Support jjkd The variable should be declared as a static in exactly one module and as an external in all others You may initalize it when declaring it as static if desired If it were external in all modules there would never be any storage allocated to it Fm Shane Dawalt Perhaps I m not using the correct language to explain how I have dec lared the variable The main module has the variable declared BEFORE the main function The variable is not preceeded by extern It is declared with it s data type then the variable name I have the impression right or wrong I don t know that a variable declared at the very BEGINNING of the program is automatically assumed static Of course I could
145. d and AL is a byte sized register This could actually be overridden if required by MOV AL OFFSET WVAR The Z80 contains I O ports which are separate from the main memory So does the 8086 which has 65536 addressable ports versus the Z80 s 256 The general form of the 8086 port output instruction is OUT DX AL or OUT DX AX which uses the contents of the DX register to specify the port For port addresses less than 256 a direct form may be used OUT portnum AL or AX And of course inputs from a port are allowed as well The syntax is the reverse of OUT s II Comparison of Hardware amp Peripherals Both the IBM PC and the TRS 80 have certain The Programmer s Corner Volume I iv hardware standards such as of video adaptors etc types There is far more add on hardware available for the IBM PC than was ever available for the TRS 80 Aside from the popularity of the IBM PC the reason for this is the expandibility of the PC with its internal slots Therefore it would be impossible to fully cover the possible options However there are certain more or less standard hardware configurations The TRS 80 has supported only one type of video adaptor primarily because all TRS 80s have had them built into the standard hardware although there have been some high resolution add ons available The PC has at least 3 standard types of video output all addressed in a standard way These are the monoch
146. d t a eee a is 2 i caAnSyTOW M S A 224 M anata eel ma mr ae TTT are The LSI LS DOS Column 46 The LSI LS DOS Column Volume I iv MS DOS Information Tandy 2000 general Fm Steve Jensen I have recently acquired a Tandy 2000 with 256K My experience is with IBM PC s and my software is all in that format My questions Can I run any of it on the 2000 How can I expand the RAM without paying a fortune Are there third party hard drives that are not similarly prohibitively expensive I appreciate any information you can give me about the 2000 Fm LDOS Support jjkd Well behaved IBM applications will run fine on the T2K This means screen output via DOS or BIOS not directly to the hardware Any other fancy program that uses anything other than BIOS and DOS resources iS going to be a problem WordPerfect works fine for example but the IBM version of MS Word wouldn t have a chance Extra memory from Tandy is not bad at all around 170 for the board You can save 20 by going mail order That comes with 128K of chips you can add two IBM PC 64K sets to bring you to 512K The board can be modified to use 256K chips giving you 768K A commercial board for 896K total is available but expensive at around 500 The same vendor also has a 20 meg HD kit for a bit less than Tandy s 10 meg unit Fm Steve Jensen Thank you for the quick response It is good to know that IBM software will run on the 200
147. d upload it SIDEWAYS with MC Fm Jerry Wagers Have you been wanting to run SIDEWAYS on your Model III 4 4P If you ve seen this interesting program running you had to be impressed Now a public domain version of it has just been uploaded to the database called OFFSID ACH OFFSIDE will allow you to print any ASCII file SIDEWAYS with your EPSON compatible or GEMINI printer with a line length of up to 255 characters and 80 lines per page It also permits multiple pages so virtually any ASCII file can be printed All ASCII characters are in the character table and are well formed and properly spaced No more funny looking printouts Even better OFFSIDE is written in C and was compiled on Misosys PRO MC and the source code is included in the archive for you hackers If you want to run it in model III mode just recompile it as is with M instead of PRO MC Full instructions are also included in the package although the program is extremely simple to use Additional instructions are included on changing the character pitch of the printout if desired PRO MC and LINE __ Fm Michael Johnston Roy I am trying to implement some debugging macro s for my system and I m having a few problems First can I use a macro from within a macro What I am trying to do is something like this MISOSYS Products Tidbits Volume I iv define TRON printf Line 4u n LINE 1 and then sprinkle TRON statements in the p
148. ded I just never even realized my SAID was not the current version until I went to patch it Since I got this version of SAID with my EDAS 4 3 will the EDAS be upgraded also along with MAS MED et al Fm MISOSYS The SAID 1 0 gt 1 1 upgrade costs 5 2 S amp H It was covered in TMQ I ii Ah carrying a version number would not work since SAID is no longer a standalone product And yes we would reduplicate the entire disk on the SAID update gosh you d get all the patches You need to return your disk in a protective mailer TMQ kudos Fm Marc Nowell Received TMQ I iii today and I m greatly impressed It is a HUGE issue and will give me plenty of reading for the next few weeks Also it s nice having my names first and last spelled right in the text where I am quoted It is appreciated and I m glad that I could contribute Fm MISOSYS Jim Gaffney wasn t so lucky I mispelled his in the TOC Keyboard must have had a sticky key Now don t think that the next issue is gonna be that big The weight drives my foreign costs way above subscription costs due to the cost of mailing That sucker cost 5 49 to mail to Australia Actually the next issue should stay within 100 pages it has to Also as I fine tune the look and feel of the QUARTERLY I have decided to include the full names unless specifically requested not to This includes public messages which have been extracted from our Compuserve forum
149. dn t allow anything other than two characters In my book I d call this a bug even though it probably has always worked this way Do you have a copy of 6 2 handy to try Fm Shane Dawalt I don t have a copy of 6 2 1 but I do have a copy of 6 2 0 isn t that strange I attempted the same device setup on the 6 2 0 system and the same results were found I wrote a program which did some fetching of device specs from the DCB I never considered the fact that only 1 character could be in a devspec since I didn t think that could occur The program stuffs the two characters inside of a message string before displaying the message with the DSPLY SVC I noticed that when the program comes to a1 character devspec the message up to and including the first character of the devspec is displayed but subsequent characters of the line are not The cursor is positioned at the end of the line as if an ETX had been encountered but actually a 00 h is encountered But that raises another question doesn t DSPLY terminate the displaying of a line when either a CR or ETX is found This implies any other control character is displayed without termination of the print line Fm LDOS Support jjkd As I suspected this is not a problem created by 6 3 0 as it has been there all along most likely since 6 0 0 without anybody finding it I imagine that Virgil will put it on the list for future attention but I certainly wouldn t consider this some
150. documentation Get the BASIC manual mentioned in my previous reply Then get every book written by Peter Norton There are better references Peter has a few minor errors and tends to use decimal when he should use hex but few better tutorials You can graduate from there to Advanced MS DOS by Ray Duncan and MS DOS Developer s guide by Angelmeyer sp and somebody else Both are great references and sources of code ideas but little or nothing on BASIC I find it fascinating that Intel and the IBM 8086 8088 line are based entirely upon an architecture that most sophisticated programmers and hardware hackers would have considered a clever kludge at best Of course the 80286 and up do have linear address models unfortunately it is far too late for that to do much good Fm Jim Beard What do you mean clever It s just a kludge The extended addresses of the 80286 80386 are in new instructions portability with the 8088 8086 80186 machines demands that these instructions be avoided Microsoft FORTRAN 77 V 4 0 and probably other languages have compiler options for use of 186 286 386 instructions but I doubt that these options have much effect except to prevent portability I 11 find out soon enough amp say as I run RATFOR FORTRAN 77 V 4 0 on an AT at work With the same 8087 programs it is a dead heat with my Tandy 1000A with 8087 MS DOS Information Volume I iv though I would have to compare programs on the Sam
151. does anyone realize that the Model 4P ties port FF and port 90H together for WRITE That s why a 4P will use the sound port for programs strobing the cassette port for sound output to a speaker amplifier We found that one when doing GOBBLER LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 get now probably a dumb 33 Volume I iv Disk driver function codes Fm Adam Rubin Serious interlude here I have two questions about the proper method of interfacing with LS DOS 1 How can a disk driver accept requests for additional functions I e other than the documented functions 0 15 I ve considered either using an unused register to pass the information or using function values 16 255 but is there a better official way 2 The module headers of the system device drivers have useful data that s not obtainable anywhere else I know that they shouldn t be accessed as absolute addresses but is it acceptable to use GIMOD and find the modules data that way Obviously there s no problem with either question under the Radio Shack School of Programming but I d rather use the correct method if there is one Fm LSI Virgil Why not use the four unsupported functions DCINIT DCRES RDHDR and RDTRK that were provided in case utility software needs these requests for communications with custom drivers Quote is from The Programmers Guide to LDOS TRSDOS Version 6 by Roy Soltoff Do
152. dshaw I have no problem sending a CHR 6 to FORMS using PRCTRL If you wish to reset the printer to top of form use the PRCTRL keyword SETTOF this will reset the printers counter A chr 6 will set the FORMS internal counter and all will be in sync Fm LDOS Support jjkd Your problem is that you are manually advancing the paper to tear it off Don t Your BASIC program can advance the paper via the command LPRINT CHR 12 and then you tear it off This will keep all five the printer the paper the FORMS filter your program and you happy and well synchronized If you insist on manually advancing the paper you can put a LPRINT CHR 6 in your program wherever you want the FORMS line counter reset to the top of the page Then however it is the responsibility of the operator to check and confirm the paper position whenever this is used if you are using a printer capability FORMS with FFHARD this is not sufficient You need to send a LPRINT CHRS 12 to the printer it will advance to where it thinks the top is and then prompt the user to manually advance the paper to the proper position to force synchronization As a final gotcha with hardware formfeed The MARGIN parameter requires the CHARACTERS per line parameter in an attempt for the filter to know when to apply the margin to lines that wrap the right margin This taken that This some also upon to its logical conclusion would require setting C
153. e impossible to collect all static uninitialized vars into one segment for run time initialization purposes We had considered using a COMMON but that would have been much too dificult I guess that is where the GROUP operator came from in MASM Thus since the linking tool we had to use did not support a method we chose to deal with it as we did We also recognized that the machine environment was better off without autof tically filling zeroes I can t stand to see those files generated by L80 with sector after sector of wasted garbage or zeroes At least with MRAS you have the capability of keeping a smallet OBJ module MISOSYS Products Tidbits Volume I iv Fm Byron P Peebles Static and external variables that are not initialized are guaranteed to start off as zero Automatic and register variables that are not initialized are guaranteed to start off as garbage on 3B2 UNIX running AT amp T C LDOS Support jjkd That certainly is bad I don t know of any micro that zeros memory I can t implementation of lint not about that kind of Fm coding practice imp lementation imagine any screaming bloody murder usage K amp R says pp 37 Automatic variables for which there is no explicit initalizer have undefined i e garbage values External and static values are initalized to zero by default but it is good style to state the initalization anyway Fm Mark Harris You re right Lint will complain if
154. e machine with and without the 286 instructions Fm Nate Salsbury One of my problems seems to be with the word relocatable Both COM files and EXE files can be loaded into ANY segment To me that is the meaning of relocatable The fact that COM files always start at 100H of the segment they find themselves Tiris that is a minor distinction I would still call both type relocatable l Joe if I write a little COM ML program to print the famous hello message I have to put that hello into a DB place with a label and then someplace in my code I have to POINT at that label e g when the program is actually running it has to be told AN ABSOLUTE ADDRESS in memory where it can find hello in order to reveal it on the monitor Since COM programs can load at ANY segment this means that my hello text could be almost ANYPLACE in memory the almost is allowing for the constraint on 16 byte boundaries for segments Since the program is finding an absolute reference SOMEPLACE INSIDE ITSELF my education from Z80 says that this is what is called a RELOCATABLE program The fact that is must start at 100H in its present segment is trivial because that absolute Start address can be changed in 16 byte increments by merely changing the CS setting EXE COM files So to me BOTH and relocatable are Fm jeff brenton to LDOS Support You should see P J Plauger s editorial comments
155. e the new sector sdone with all DIR sectors Applications for the User Volume I iv 529E 52A0 52A3 52A6 52A8 52AB 52 AE 52 Bl 52 B4 52 B5 52 B8 52B9 52 BC 52 BF 52C0 52C2 52C3 52C4 52C5 52C8 52CA 52CB 52CE 52D0 52D2 52D3 52D4 52D6 52D7 52D8 52DA 52DD 52 DF 52E0 52E1 52E2 52E3 52E5 52E6 52E7 52E8 52EB 52ED 52EF 52F0 52F1 52F2 52F4 52F7 52FA 52FC 52FE 5301 5302 5304 5305 5306 5307 5308 5309 530C 1E00 CD1B53 3ACD55 E6F7 32CD55 CD0953 247054 I1EA53 D5 CDCE 52 El CD8A42 210000 310000 7C B5 c8 3A2B44 CB6F c8 C33040 0605 3E 20 12 13 lOFC D5 1B 3E0A CD5144 C630 12 1B 7C B5 20F3 D1 c9 E5 219F42 CB46 2818 F5 C5 D5 CB86 010008 CD6000 CB46 20F4 CD2B00 B7 20FA DI Cl Fl El c9 210055 0E 00 00096 00097 00098 00099 00100 00101 00102 00103 00104 00105 00106 00107 00108 00109 00110 00111 00112 00113 00114 00115 00116 00117 00118 00119 00120 00121 00122 00123 00124 00125 00126 00127 00128 00129 00130 00131 00132 00133 00134 00135 00136 00137 00138 00139 00140 00141 00142 00143 00144 00145 00146 00147 00148 00149 00150 00151 00152 00153 00154 00155 00156 THE MISOSYS QUARTERLY SPRING 1987 EXIT SPSAVE MSG 1 NO MSG1N1 MSGIN2 KEY KEY 1 KEY2 KEY 3 WRTSYS Applications for the User LD CALL LD AND LD CALL LD LD PUSH CALL POP CALL LD
156. e the password For bug reports suggestions etc leave a note to the sysop at 8 N 1 4 215 848 5728 ACCESS password This incompatibility was 00001 00002 UNDATE CMD by Luis M Garcia Barrio 8 N 1 4 215 848 5728 00003 For free distribution Not for sale 00004 Based in LDOS 5 3 DATECONV CMD 00005 Allows conversion of LDOS 5 3 LS DOS 6 3 disks with the 00006 new date time stamp to the LDOS 5 1 standard format 00007 The ACCESS password installed is 8 spaces no password 00008 4030 00009 GABORT EQU 4030H 4 209 00010 CKDRV EQU 4209H 4451 00011 GDIV EQU 4451H 4467 00012 DSPLY EQU 4467H 4409 00013 ERROR EQU 4409H 002B 00014 GKBD EQU 002BH 428A 00015 LOGOT EQU 428 AH 0060 00016 PAUSE QU 0060H 478F 00017 GETDCT EQU 478FH 4777 00018 RDSECT EQU 447774 4772 00019 VERSEC EQU 4772H 4768 00020 WRPROT EQU 4768H 429F 00021 KFLAG EQU 429FH 442B 00022 SFLAG EQU 442BH 5200 00023 ORG 5200H 5200 ED73C052 00024 START LD SPSAVE SP save for the end 5204 E5 00025 PUSH HL save parameters 5205 CDE752 00026 CALL KEY 5208 214253 00027 LD HL TITLE say who 520B CD6744 00028 CALL DSPLY 520E El 00029 POP HL parameters in HL 520F 7E 00030 PRAMS LD A HL 5210 FE20 00031 CP 20H bypass the spaces 5212 23 00032 INC HL 5213 28FA 00033 JR Z PRAMS 5215 FE3A 00034 CP 3AH slook for Applications for the User
157. ease wait for your statement we are going to process renewals in only a few inputs to our order entry system That s to avoid wasting Brenda s time All we really have to get accomplished is to alter your expiration field in the data base and reflect the total dollar value of renewal transactions in the invoice system monthly revenue figures here are processed electronically As I write this Blurb June 18th I i s left we have 15 I ii s left and 105 I iii s left Since our last print run was 800 that gives us about 700 subscribers The actual count in the data base is 717 out of 18066 this means we have over eighteen thousand names in our database 717 are subscribers of THE MISOSYS QUARTERLY TMQ I iii was mailed to 658 subscribers thus we have picked up about 59 new readers since our last issue The rate of renewals will let us know if we re doing a good job with TMQ we have no This is another full issue In fact again we had to defer some material I had eight pages of The Hardware Corner which must wait until next issue I also had to forgo Our Compuserve Forum Rich s stuff on stat and rm are likewise postponed Some of the correspondance which I had selected for QUARTERLY insertion also is postponed But that s good it means I have a partial box full of material for starting Volume II 1 11 still welcome your input though The more contributions I get from readers the less The BLURB THE MISOSYS QUARTE
158. ect initialization of nested structures RND 05 24 87 Dic cl 21 08 00 39 6E 26 00 7D B7 C4 CF 7F E5 F1C C1 F5 21 00 00 39 E5 CD CF 7F D1 CD BA Cl KSMA FIX 03 30 87 Patch to LDOS 5 3 KSM FLT Apply via PATCH KSM FLT FILTER KSMA DOO 60 2A FC 4D 01 47 F00 60 FD 6E 02 01 51 D00 BD 59 F00 BD 57 PATCH to correct MAPPER 24 Oct 84 from PRO MACH2 M Houde MAPPER displays 4 granules in a row which means that when disk allocation is 5 or more granules cylinder 2 lines cyl are used and 9 cylinders are displayed on the screen whereas 18 cylinders are displayed when only one line cyl is needed filling the screen in each case Unfortunately when allocation is exactly 4 grans cyl only 9 cylinders are showed leaving half of the screen blank Cure is easy as the culprit is an erroneous CP 4 JR C n which should be CP 5 JR C n X 335B X 335C D03 74 05 F03 74 04 UNREM63 for UNREMOVE 12 May 84 M Houde This patch enables UNREMOVE CMD version 1 0a from PRO ESP to operate on 6 3 and 5 3 disks as well as 6 x and 5 1 disks This patch must be added as an X type patch as no room is avd ila inside the file x 311D cD C0 35 was E E6 07 X 35C0 3A CD 36 CB 5F 20 04 7E E6 07 C9 7D C6 11 6F 7E X 35D0 E6 1F 00 C6 50 CD 2E 31 OE 20 3E 02 EF 3E 02 EF X 35E0 2B 7E E6 F8 OF OF OF CD 2E 31 OE 3A 3E 02 EF 7E X 35F0 E6 07 23 4E CB 21 17 CB 21 17 CB 21 17 D6 50 C9 ALTDSK63 for ALTDISK 7 Feb 85 change DOS7 6 x to 6
159. ed programs or at least bigger ones can be run on the 8086 The natural question at this point is how this somewhat odd number not an even power of 2 is supported by the addressing modes of the 8086 The answer is relatively simple The 8086 actually forms all physical memory addresses by the following formula 16 segment register offset Segment register means any of CS DS ES or SS Offset is what is conventionally thought of as the memory address and is 16 bits wide In hexadecimal this can more easily be visualized as AAAAA 20 bit physical address The convention for specifying a full address is SEGMENT OFFSET That is the segment value followed by a colon followed by the offset The 8086 decides which segment registers to use based on the context of the instruction When fetching instructions for example the CS or code segment clever acronym eh is used in conjunction with the internal conventional PC which represents the offset Thus CS PC would denote the full 20 bit address of the currently executing instruction Data accesses usually use the DS or data segment register implicitly Thus the example given earlier MOV AL BX DTI actually means load AL from the memory location specified by DS BX DI or 16 times the DS register value plus BX plus DI The SS register combines with SP to form the stack address Note that addressing modes involving BP use S
160. ed KSM you better start looking at it it can save you a great deal of repetitive typing Here are the KSM strings I use which are pertinent to this modified version of SCRIPSIT q list print 2 q reset pr q route pr print txt 2 r Sincerely Roy Soltoff This is in response to your letter of The semicolons are important KSM allows you to specify a logical lt ENTER gt character which defaults to the semicolon In the g commands noted above you will find them terminated by two semicolons one designates an lt ENTER gt to complete the DOS command while the second designates an lt ENTER gt to satisfy the Q command s prompt after completing the requested command Finally I normally assign the LIST string to the L key the RESET string to the Q key and the ROUTE command to the R key although you can make those assignments to suit yourselves If you use Model 4 SCRIPSIT without the MPATCH you are losing out on some very handy enhancements DISK NOTES 3 are still available for 10 plus S amp H 2 in the US Applications for the User 63 THE MISOSYS QUARTERLY SPRING 1987 Le v ee i Your Model 4 computer may not speak XZ M S But with MISOSYS language products she will speak ASSEMBLER BASIC C and RATFOR PRO CREATE The standard macro assembler used by professionals and novices alike Nested macros nestea includes nested conditionals Full screen editor cross
161. ed does indeed Ss Ghee Volume I iv interrupt timer routine rate doubles because per second is called the blink there are now 60 counts Any program which does not use the ROM blinking cursor routine such as a word processor or text editor which is performing its own overstrike blinking will be just as dependent on CPU clock speed since their code which controls blinking is invariably based on the speed with which their code executes at the 4 Meg clock speed their code should execute twice as fast depending on CPU wait states LDOS 5 1 4 booted up at a 4 Meg clock speed on a Model 4 because LSI had hard patched enabling the 4 Meg clock rate into the startup code On the other hand LDOS 5 1 4 had nothing added to it which would correct the doubling of the interrupt rate In fact the low and high speed slots of the task processor would be serviced at twice the rate expected That s why the machine s CLOCK became worthless it gained 8 12 hours every 12 hours Okay what was done to 6 0 to make the interrupts uniform regardless of the system CPU clock speed A time slice routine was added at the front end of the task processor This little routine rotated a byte value of or OFFH at 2Meg and serviced all but the highest task only when the rotation placed a one bit into the carry flag That s a very fast and short way to divide by two reset the counter and test the result all in one instruction RRC HL Using
162. ed to take the bull by the horns Note that my function as forum manager gives me no right to download files en masse and begin distributing them elsewhere I also have no intentions to selectively obtain files and then make them available elsewhere MISOSYS has no time to perform a service of distributing public domain software files On the other hand We could probably bend to organize disks of PD software to be made available for a fee similar to what Montezuma Micro is doing for CP M programs Considering these issues we will start to build a library of public domain software for the TRS 80 The programs must be submitted via floppy disk directly to us by mail from the original author we will not take submissions via Compuserve Please use LDOS or LS DOS formats 40 cylinder DDEN 1 or 2 sides We expect pricing per disk to be similar to DISK NOTES although we will listen to your input on pricing Our next issue will have the results of your input No Model I LDOS 5 3 Our last issue of TMQ anticipated the decision recently made that there WILL NOT be a 5 3 release of LDOS for the Model I We are extremely disappointed with the sales volume of the Model III LDOS 5 3 Upgrade Kit Extrapolating potential Model I sales results based on the ratio of actual Model III sales to the entire Model III 4 marketplace precludes any justification for a Model I release of 5 3 However before the end of the The BLURB Volume I iv ye
163. editor for ASCII files LCOMM now gives you access to LDOS library commands while in terminal mode e PATCH supports D amp F patch lines with REMOVE capabilities e DATECONV has been added to convert older disks to the new date convention BASIC Enhancements Improvement to line editing with the addition of line COPY and MOVE Very flexible INPUT added for screen fielded input e We ve added a CMD V to dump a list of active variables with values including arrays For 34 95 S amp H the LDOS 5 3 upgrade kit includes a DOS disk and documentation covering the enhancements Specify Model 3 4 or MAX 80 P S Don t return you old disk MISOSYS inc PO Box 239 Sterling VA 22170 0239 703 450 4181 MC VISA CHOICE Orders Only 800 MISOSYS 1P 5P EST Monday Friday AN AA VA residents add sales tax S amp H US 2 Canada 3 Foreign 6 Our Compuserve Forum Volume I iv S rumma ee eee ee a rt a a a a ama re M 5S Ne gt 4 lt eee n
164. eg hard drive bubbles a Lobo WIN5 5meg drive etc We also have a ton of 8 floppies used MISOSYS closed in August Here s an important announcement No we won t be closed the month of August This business doesn t pay THAT well It just happens to be time for the Soltoffs to enjoy a week off down at the beach Actually I should say a week at the shore since I come from Philly and that s what they say there We will be closed from August 8th through August 16th That s the second full week in August including surrounding weekends We expect to be somewhere in Virginia Beach depending on our guess as to what constitutes an ideal spot for The BLURB THE MISOSYS QUARTERLY SPRING 1987 When are you a short patch the 4 Volume I iv the family No we won t be there the whole time but don t count on us answering our phone And don t worry if mail orders are delayed we ll catch up when we get back Summertime has been a little slow anyway Contents of DISK NOTES 8 THE MISOSYS OUARTERLY contains Each issue of program listings patch listings and other references to files we have placed on DISK NOTES Some people enjoy typing in long listings Sometimes you may have need for only If you want to obtain all of the patches and all of the listings you may conveniently purchase a copy of DISK NOTES There is a cost involved DISK NOTES is priced at 10 PLUS S amp H The S amp H charges are
165. eing accessed In fact WRSEC and some other requests are passed on to FD Anyway I don t have anything that would use RDHDR Including the driver directly in the program which I ve sort of done I suppose might be the best way and I ll certainly have to think about it Do you or anyone else have any comments on anything here Fm LDOS Support jjkd GRDTRK would not be what you want It does not work very well in the WD FDC controllers especially in double density Nothing in the OS needs it so it would have been excess baggage that wouldn t work anyway to code a replacement for RDSEC as noted in the other replies Note that MS DOS sectors are numbered from one to nine where LDOS LS DOS sectors are numbered from zero to seventeen Trying to read sector zero on an MS DOS disk won t get you very far at all You basically need LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 36 2 Volume I iv There is no interleave on many MS DOS disks because real IBM DOS on a real IBM PC or close enough can keep up with 1 1 interleave on a multiple sector I O request One trick you can use to eliminate the problem is to write and read verify the sectors in the following sequence 1 3 5 7 9 2 4 6 8 instead of 1 2 3 4 5 6 7 8 9 You d be doing the interleave in software instead of by the format as in some releases of CP M Fm Les Mikesell The best multi sector i o would be to interleave ta
166. emember again what you typed you ll only have to remember the name of the JCL file The JCL EXIT gotcha Mendenhall I have run into what appears to be a bug in LDOS 5 3 for the Lobo MAX80 Using BASIC when I start a BASIC program using a JCL file to automatically load and run the program the computer displays JOB DONE and branches out of the BASIC program back to LDOS whenever it receives an INPUT statement Would you please send me a patch to correct this bug Fm C M Fm MISOSYS No patch reguired suspect that you ended your either nothing or EXIT terminate a JCL file when you within a program which uses does BASIC s INPUT is to use STOP That terminates the executing JCL and returns control to the program Omitting both STOP and EXIT is interpreted as EXIT so that the LDOS Ready prompt will appear rather than having a blind line input request hanging without a visible prompt When all else fails read the manual no bug I JCL file with The correct way to want to remain line input as the JCL macro TED Problems Fm Arthur R Wynott The text editor TED is not functioning properly I cannot exit TED with lt CLEAR SHIFT gt lt CONTROL U gt sends the cursor to the beginning of text not the previous page Left and right arrows do not work up arrow gives me a left bracket Please repair LDOS Information Volume I iv Fm MISOSYS Here s some answers to your TED problem
167. enia The COM files are CP M 86 compatible files which load at 100H in the lowest available segment EXE files are relocatable and I haven t been able to get full documentation on them myself yet Roy tells me that there is an old article in PC WORLD on REL and EXE files but I haven t looked it up Apparently headers on EXE files can give a preference for location but I am very hazy on that Good luck Editor s note the article referred to was in PC TECH JOURNAL October 1985 Vol 3 no 10 and it referred to the structure of OBJ files The Microsoft standard is a subset of the Intel standard which is documented in 8086 Relocatable Object Module Formats An Intel Technical Specification Fm H Brothers Nate Answers to your first two questions 1 Yes DOS decides where to load the next program by using the first free paragraph given up by the parent process normally Command Com That is a program will load at the lowest free address in memory in the normal run of events 2 Yes As long as you aren t trying to use DOS 3 x calls on a DOS 2 x system your program should be directly portable Fm LDOS Support jjkd 1 Part one True Part two Not True Yes a COM program will end up running anywhere in memory This is not relocatable code however The segmentation structure of the Intel processors uses the CS code segment to make the program think it knows where it is executing from beforehand
168. ent of the Tandy users run no more than three programs a word processor a Spreadsheet and a data base LDOS Information THE MISOSYS QUARTERLY SPRING 1987 In addition to the The probably don t 5 98 Volume I iv Installing 5 3 on hard drives Fm John P Lyman Recently we purchased LS DOS 6 3 from LSI and LDOS 5 3 from MISOSYS We appreciate the fact that for the very reasonable price we paid that you can not give individual assistance However I have tried to take your advice of read the manual and don t bother us and after spending several hours in that mode I feel I must bother you TRSDOS system which I have successfully converted to LS DOS 6 3 their instructions were workable I have two LDOS Hard Drive systems to upgrade 5 Meg LDOS 5 1 3 and 15 Meg LDOS 5 1 4 and a new Hard Drive to be 15 Meg LDOS 5 1 4 They are standard Radio Shack systems HD 0 1 2 3 and Floppy 4 5 With Radio Shack turning their backs on us TRS Model III 4 users there is no one to turn to other than yourselves I have had no success trying to convert the new LDOS 5 1 4 system using your very brief and inadequate instructions page 3 1 5 87 and I need help I simply do not have the time to take a crash course in hard drive hardware to learn what your people already know and I don t dare go near the existing systems with conversion in mind One other point Given the between the lines attitude LS
169. entally how Should budding authors submit query letters message U S Mail or some other way Fm MISOSYS Query letters If you mean an article please submit on disk If you mean a query about submitting an article then either a note on our Compuserve Forum PCS 49 don t use EZP or a letter or a call More on TMQ I iii Fm Shane Dawalt I just received my Winter TMQ and was just a bit surprised with the Letters to the Editor The person who said the material in TMQ was over his head has boggled my mind There certainly isn t anything that profound And if it is it is for hardcore programmers which understand its implications perhaps on the second or third reading if they are rusty programmers If the information doesn t make sense that particular person is missing background Letters to the Editor THE MISOSYS QUARTERLY SPRING 1987 Anyway 2 16 Volume I iv information on the should take a couple steps back subject and I agree that user should support companies who are supporting them but the comment you made and I m paraphrasing user who don t support the people who cater to their machines don t need the support in the first place If a magazine is talking about BASIC I m not going to buy it since I don t use BASIC all that often It is not that people don t want to buy the product they must weigh its usefullness I bought PRO MC from you because I was tired of pampering BASIC and I h
170. eps the CPU interrupts enabled until actual data transfer The result is that a disk drive locked on to 300 rpm could take a very long period of time to perform the data transfer because 300 rpm is synchronous with the CPU system interrupt clock When all of a sudden users who have installed those PLL disk drives become infuriated with LDOS because data transfer is inordinately slow compared to the other DOSs yes I call that a fiasco You also don t have to initiate the KI DVR and JKL to be able to use the screen print function Where ever did you get that idea You do have to specify the JKL parameter if you CHOOSE to use the enhanced keyboard driver provided by KI DVR however the choice is yours If you want to make use of lt CLEAR gt functions such as KSM a full ASCII keyboard and type ahead then yes you do need to install the LDOS keyboard driver Since the screen print detection is an inherent part of the keyboard driver and because LDOS s screen print passes graphics values if you have SYSTEM GRAPHIC on the ROM screen print function is worthless But if you don t install KI DVR then you still can use it The customizing of a DOS is not a trivial manner No I don t think that an INSTALL command or commands would be feasible to make it more friendly Sorry but Tandy didn t put the printer parameters in low memory The code for SMOOTH happens to be within the floppy driver in the resident OS HIGH is supposed to
171. er but I m not familiar with what is involved at the Task level Thought KFLAGS scanner would be transparent here since all it does is update the KFLAGS bits which I don t care about Since there s no need to sense and act on BREAK ENTER other than to pass the char s value via PUT Any suggestions Fm MISOSYS Since GTDCB does disk I O may need a system overlay it would be better to isolate that into the installation code to extract the address of the CL DCB then stuff the address into your resident filter code That just makes good programming sense when the KI static and resident Fm Bob Haynes My concern is filter remains relatively in configuration but the CL chain is more dynamic filtering etc If filtering or redirection was applied to CL AFTER my filter was applied to KI I believe the KI filter would access the wrong CL vector with your suggestion What I ve done is set up the code to trap certain command characters and if appro priate perform a GTDCB then pull SYS1 back in via your technique outlined on pg 181 of your Programmer s Guide Tis almost finished just need to adjust a few error traps and set LS DOS Information ta AG s i the debugging of course Would appreciate any further comments you might have Fm MISOSYS Absolutely wrong I said use GTDCB to get the address of the CL device control block Once CL is installed into the device table it s DCB address nev
172. er value when it s done to force logging the next time the drive s accessed Questions Is cylinder 20 a good place to Start write precompensation on an MS DOS disk Will what I am doing cause any problems Is it all necessary Is there a better way FORMS questions Fm Dick Houston In using the FORMS command with say a BASIC program it works OK for the first page but merely starts out in the left off on the last run when I run it again Does anybody know how to reset the line counter Fm Ray Pelzer I may have missed a discussion of this once before but maybe Someone can tell me if the following was an intentional design or a missed bug When the FORMS filter is installed and you attempt to set the MARGIN you will get a Parameter Error if the CHARS setting is OFF I was messing with a file I wanted to print at 12 pitch and loaded FORMS to shift the printout over 10 characters rather than messing with physically shifting my forms in the printer Got the error message upon issuing FORMS M 10 but FORMS C 132 M 10 worked just fine The REAL kicker is that AFTER loading up the margin I could go back and do a FORMS C OFF and the margin stayed loaded properly Am I losing my mind or is it already long gone 3 Fm MISOSYS That s how its supposed to work You must have a chars per line value before setting a left margin Fm Alan H Pesetsky I m using the forms command to set the number of lines pr
173. er gets changed You used the GTDCB in your filter code after it was installed That s wrong Use the GTDCB SVC in your installation code and plug the address returned by it into your resident filter code as if that address was just returned by GTDCB Fm LDOS Support jjkd You ve got it and I think that you have the fix too no Your code can check the currently resident system overlay Any device I O that causes a resident SYSI to be overwritten by anything else is responsible for reloading SYS1 before releasing control That basically means any file I O open or close along with GTMOD GTDCB CHKDSK RAMDIR and so on Note that this applies to resident drivers and filters anything resident hooked into keyboard byte I O And finally even though not related to this problem I d like to mention that RTC tasks can t do any disk I O but can hook through KITSK Fm Les Mikesell If you want to output to the unfiltered com dvr you could use gtmod to find CL It takes a little more set up to call the driver directly instead of using put but it would certainly be more efficient Fm Bob Haynes And now here I stand my hat in my hand Eating some humble pie Your system well planned the DOS support grand How silly I feel said I Re your comments on needing GGTDCB withing a filter module you are of course correct as usual I was confusing the vector s VALUE with its ADDRESS The GGTDCB call is now
174. er into memory under LS DOS 6 3 but I messed up on the devspec Instead of typing SS I entered xs The DOS didn t complain so I didn t investigate it I filtered the KI using S and this too worked fine However when I executed the DEVICE command the device map was screwed up This is what it showed Note the DOS was freshly booted no other programs had been executed prior to this SETting FILTERing KI lt PR gt SI lt SO lt gt JL gt S S DO lt gt X 0B88 X OEOL KI DO NIL The filter module set to S worked perfectly It filtered the KI device as it should have The device table just doesn t show it correctly notice S has no route link information behind it Also notice the S DO Looks as if the KI data line terminated without a x OD before listing DO I was able to duplicate this problem by routing a device PR to another 1 character device name 2Z Z had been previously routed to a file the filename wasn t displayed in the DEVICE listing Again the route worked fine According to the operations manual the devspec must be two characters Why doesn t LS DOS fuss when given a one character devspec LS DOS Information Oe a C ST eee EES A 7 7 TE i a y A 7 Ri A p e i Volume I iv Fm LDOS Support jjkd If the docs say two characters required and DEVICE barfs ona single character devicespec then it does indeed make sense that SET shoul
175. erations The less complex the bootup procedure can be made the better since I d like the system to be able to reboot automatically after a power outage Fm MISOSYS The ROM image on the 4P IS a file that s the MODELA III file SET2RAM will work on a 4P however since it doesn t know it s a 4P it still goes through the gyrations of copying the ROM image to higher RAM Switching to the RAM model III mode then copying back the ROM image which in the case of the 4P was still in low RAM However the memory configuration is maintained in a byte that needs correction Even if you would try to dump off a new ROM image which has the banking and memdisk handlers you WOULD have to go through the motions because the new BANK and EXMEM vectors were used for the BOOT initialization The reason is that the BANK and EXMEM handlers use the memory space and jump vectors which have to do with the ROM boot code That also happens to be why you cannot best not do a software BOOT command when SET2RAM is installed Suggest you use a startup jcl to have your system re boot after a power outrage ooops outage LDOS Information THE MISOSYS QUARTERLY SPRING 1987 Volume I iv Fm Gary Phillips I am having some unexpected difficulty with SET2RAM and MEMDISK in my environment I am using a non gate array 4P Does it matter which version of the ROM image I load before invoking SET2RAM I have at my disposal the one that came
176. ersonal news in or out of THE BLURB I find it helpful to get to know a little more about you and your family I too am a Philadelphian from the Northeast Frankfor High but left after college for a life scattered between NY and Central America I am still not much more than an end user but Letters to the Editor Volume I iv feel reading the Forum correspondence has given me a few insights into the LDOS and Model III world Thanks for everything Fm MISOSYS We were a little further NE Actually 0 10 was in Feltonville C amp Wyoming and 10 20 was in Lexington Park near Cottman and Blvd I think most folks are not upset about the family news 80 Micro lt grins gt Fm Gary Phillips Roy letter to 80 Micro today I sent the following Dear Editors Your June issue would have left me laughing except that the inaccuracies were so severe as to be painful of LDOS 5 3 is an If Jack Feldman s review example of what we can expect from your present editorial policies and staff then I suspect users of Z80 equipment might as well quit reading your magazine now Mr Feldman makes statements about LDOS that amount to pure libel I can only wonder whether he actually used the DOS or read lt the documentation before writing his article As a technically competent and experienced Z80 user I am a professional systems programmer I can assure Mr Feldman that LDOS 5 3 in no way puts your old disks at risk
177. es up your second problem We have revised scan which is used by the higher level scan functions to agree with the System V handbook on termination conditions It was a major revision however in so doing we have cleared up a number of bugs in that function We even added code in scan to properly terminate float fields The FIXBUFS option did Corrections had to be reported in TMQ I iii not work correctly made to fopen as unsigned long to double problem down to an obscure bug in the UL2DST routine This routine in LIBA converts an unsigned long number on the stack to a double on the stack Fifty percent of the time it would work correctly That s because the error would occur if the return address of the UL2DST caller was an even number I tracked your Fm Richard Deglin I came across a bug which was interested when I fixed some other problem associated with structure initialization Here are patches which correct a problem encountered when initializing nested structures They are in The Patch Corner as INSTRU FIX MISOSYS Products Tidbits Volume I iv I also noted the discussion in TMQ I iii which concerned itself with the initialization of automatic variables This is something I overlooked when implementing MC Here are some patches which extend the syntax of automatic scalar variable initialization to include any arbitrary expression as the initializer In doing this however the code generated
178. f a symbol has been defined then an indirect reference will be assembled whenever the symbol is referenced This confusing to me system is what MASM does as shown in Hardin s listing In the third from last line of code in the same module I had to revise it to mov attrib bl and the next line had to be changed to inc byte ptr ers row before I could get a good assembly Oh also in the set curs procedure I had to use mov dh curs row Are these truly differences in design between MASM and ED ASM 86 From the sentence in your manual which I quoted earlier I had expected ED ASM 86 to work the SAME way as MASM Fm MISOSYS Nate in investigating your report I found that your quote from the ED ASM 86 manual was missing a VERY IMPORTANT WORD You should have quoted If a symbol has been previously defined immediately before the DB DW then an indirect reference will be assembled unless overridden You left out the word previously Now in Hardin s listing 3 the symbols CRS ROW and ATTRIB are defined in a DATA segment which is placed after the CODE segment in the source Since that is not previously I must assume that the assembler did not treat the non bracketed references as indirect Just for the heck of it I moved that DATA segment block to precede the CODE oe 5 MISOSYS Products Tidbits Volume I iv segment and re assembled Didn t get any errors I m not sure what MASM does in handling reference
179. fer to LDOS 514 About 24 25ths of the time this is OK but when the above HIGHS value occurs value in A and value LDOS Information THE MISOSYS QUARTERLY SPRING 1987 1915 ai Volume I iv in IY 3 differ by one The result is that HL points one page too high If you think that scenario was rare you will have to agree that bug two is a winner in most unlikely contest If you try to replace a KSM data set with a longer set the installation checks the lengths and aborts right not always If the new data is only one or two bytes longer ZAP Try this build two KSM files that differ in length by one byte Install the short one and then attempt to replace it with the longer one no error message Replacement done then the first use leaves smoke coming out of high RAM I went through all of this on the 5 3 version and ran off a bunch of appropriate screen prints and fixes I am sure you will have no trouble in duplicating the test Now I could use a small favor Could you offer some quick documentation on the use of the byte at 4758H I pretty much get the gist but I would value your reply Fm MISOSYS Can t believe category Believe it If no one writes or calls about a bug especially an obscure one rarely encountered it may never get fixed Since we acquired LDOS in March of last year no one has given a bug report on 5 1 4 s KSM facility I can t say whether that was the case with either LSI or LSC ho
180. file or in my understanding of what a structure is does MISOSYS Products Tidbits 91 THE MISOSYS QUARTERLY SPRING 1987 and possibly Volume I 1v 2 In main there is a line of code dct struct dct regs IY Do I read that to say that the value of regs IY is being cast to be a pointer to a structure of the type dct and then the result is assigned to dct which was created as such a pointer earlier Fm MISOSYS the README Careful reading of file supplied with MC should clear up the confusion Let me cite the applicable text All bit fields are treated as unsigned short integers there are no signed bit fields Bit fields are ordered from high most significant bit to low least significant bit in a machine word 16 bits The ordering is such that the first declared bit field will start in bit 15 of a double Z80 register e g bit 15 of HL or bit 7 of H and so on Note that when a word containing bit fields is fetched from or stored into memory the high and low bytes are reversed This reversal must be taken into account when creating a set of bit fields to describe an externally imposed format e g a bit structure maintained by the operating system The reversal is of no importance if the structure is wholly contained within the C program and not used outside of it Now let s take a look at that det structure as used in DCT CCC abbreviated i Ler struct dct uchar dct inst
181. for menu based commands If you care to utilize a user friendly read menu driven communications program there are about four still on the market at costs ranging from 40 to 200 LCOMM is bundled as part of your DOS You are not forced to use it it is there if you choose to learn how to use it There is no economic justification for totally revamping the LCOMM user interface to be userfriendly We are not selling a communications program with the DOS thrown in for the heck of it A DOS is being sold with a usable communications program thrown in for the heck of it Actually it only takes a few moments to learn how to use LCOMM Just as you have found that there are many other debuggers around you should have realized that a DOS can t be all things to all people Most of the features ancillary to the specific function of a DOS file management are really not going to be elaborate The economic justification is missing the disk Space is missing the documentation space is missing Ninety percent or more of the LDOS users don t write in assembler and thus have absolutely no need for DEBUG 99 of Tandy s customers would have absolutely no need for having an assembler bundled with their computers and those same customers would then consume 50 additional customer support trying to figure out what the assembler is for That same percentage of users have modems and likewise don t use LCOMM I would venture a guess that 90 perc
182. friendliness would you consider issuing zaps or INSTALL CMD LDOS Information Volume I iv programs that would allow me maximum flexibility in customizing LDOS And I do not mean SYSGENed High drivers Such basics as JKL belong in LOW core memory where Tandy put them ditto the printer parameters etc Perhaps I have this NIGHTMARE about a hugh collection of essential drivers occupying HIGH RAM where they will be violated by every program which does not constantly check High SMOOTH in low core perhaps I do not wish to re invent the wheel I am looking for better tires In that spirit I am using POWERSCRIPT While it greatly improves SCRIPSIT it is not as good as it easily could have been no ASCII characters below 20 hex in their translation tables and that is where the Model 4 has the European characters And again the extra keys are also ignored It is enough to make this Tandy orphan weep On the one occassion when I used COMM I found its command syntax bizarre to put it politely Even if 5 3 Interface does not clear up 4 8 key sequence commands excluding filenames I am still game And if some multiple keystroke commands still insist on lt ARROW gt lt LEFT SHIFT gt my keyboard will then sprout a small toggle switch which allows me either 2 left shift or 1 right and 1 left shift keys As you can se I am certainly not a Software engineer but I am determined to enjoy the use of my TRS 80 Even as a hardwa
183. g I met the same error messages Do you see a possible explanation solution for this Fm MISOSYS There s an easy explanation for the problems you were having with assembling PRSET ASM using MRAS Your solution is to MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 205 Volume I iv directly generate the APP a command such as core image file by MRAS PRSET 0 PRSET APP GC CI a REL file the operand experienced can be 2 15 of the That s because trying to generate imposes certain constraints on fields The errors you explained by simply reading page MRAS manual A statement such as DC HIGH lt 8 256 0 has at least three operations not permitted in the expression when assembling a relocatable object module It is permitted when assembling an executable command file I believe that Hardin Brothers used our PRO CREATE assembler MRAS would behave similarly when the GCc switch is used PaDS Partitioned Data Set Utility Fm Irwin B Burton I have tried to use PaDS with LDOS 5 3 I found that by eliminating the PASSWORD protection I can make up a PaDS of my utilities However I have trouble adding any 5 3 utilities to this file As an example HITAPE from 5 1 4 loads in and runs fine under 5 3 I m sure this has to do with the PASSWORD DATE TIME changes in the directory but I m not enough of a hacker to figure out just what is happening I would appreciate any help you can give
184. ges that make even switching from version 3 1 to 3 2 of IBM DOS risky MS DOS Information THE MISOSYS QUARTERLY SPRING 1987 49 Volume I iv Fm Adam Rubin Thanks for your reply I d guess it s 3 BIOS Hardware incompati bilities If it ll help clarify anything here s some specifics on everything It s actually my father s system I m just visiting This is an Epson Eguity II PC XT clone with a V30 6 4 77 7 16 MHz 640K RAM and 2 360K floppy drives MS DOS 3 10 was included with it The program in guestion is Webster s New World Writer v1 01 which appears to have been compiled with Lattice C v2 1 It s not copy protected The one problem I ve noticed when running Webster s under the Epson MS DOS 3 11 occurs when the program checks to see if a disk has been inserted e g right after Insert Spelling Checker in A and press ENTER The result is an apparently endless sequence of Drive not ready Abort Retry Ignore messages When it doesn t check for a disk e g Save Document there s no problem I tried the program under PC DOS 2 00 2 10 and 3 10 and that problem doesn t occur In fact my father called the customer support number and was told that the program is supported under PC DOS only not MS DOS That sound as weird as I think it does Does all that explain anything My knowledge of MS DOS is admittedly pretty near zero at least compared to the level of knowledge in th
185. ginally in the DCB of the device being filtered Here we may take up more code Space but we gain a consistent interface It is necessary to preserve the state of register IX Also the DOS CHNIO service call sets up the entry flag state for the next filter driver based on the value in register B which we have not disturbed That s why I don t need to PUSH AF POP AF here The key point is that when we want to pass control to the device being filtered we use the protocol of this chain routine The RXOI1 label is used by the relocation routine mentioned above Since the CALL CHAIN instruction generates an absolute address reference it must be corrected to the new r ference needed when the routine is relocated in memory PUSH AF PUSH DE KBD1 LD A KBD The Programmer s Corner THE MISOSYS QUARTERLY SPRING 1987 aTi lt Volume I iv RST 40 CP ODH JR NZ KBD1 POP DE POP AF MODEND RET Finally we get to the code which polls the keyboard awaiting an lt ENTER gt before returning to the caller Again we preserve register DE as that register is corrupted by the GKBD service call Under DOS 6 x all devices utilize return codes The convention is that a z flag state indicates a successful operation NZ indicates an unsuccessful operation This filter preserves the state of the return code coming back from the CHNIO call which passed the form feed character down the device chain This return code state
186. hat day CAL scans the entire and maintains a table of flags indicative of activity per day for an entire year The file is only re scanned when you change years if it found a BRINGUP file initially We also added a bringup hot key in the CAL command list so that you can easily position to the day showing an asterisk and BRINGUP DAT file Announcing New Products THE MISOSYS QUARTERLY SPRING 1987 10 Volume I 1v hit one key to bring up the list of activities you have scheduled for that day That facility has become indispensible here at MISOSYS Speaking of BRINGUP we altered the structure of the data file so that the year spans 16 years from 1984 to 1999 The date field storage was also reorganized so that is can be sorted in ascending order Since sorting by PSORT requires a data file header record that was added too BRINGUP data files can now be sorted by date and time using PSORT This requires a convert utility for your existing files which is included with the release Now sorting the bringup data file only makes sense when you have a facility for printing activities Enter the new BUP APP application This wammie which can be invoked directly from BRINGUP supports printing of bringup activities by date range using a syntax exactly like that in your DOS for BACKUP and DIR date parameters We also added as a shorthand for today Standard user input print formatting is supported similar to that f
187. hat you use to pass the parameters to the SVC when using the USRIL access feature in BASIC on 6 3 The name of the array can not be more that two characters long There isn t much need for a longer name for this purpose anyway so it is no big deal unless you have spent all day trying to make a longer name work It is all a consequence of the way VARPTR works Remember that the syntax is USR11 VARPTR ARRAY 0 Well VARPTR returns the address of the beginning of the data not the beginning of the header and the length of the header depends on the length of the name and the number of dimensions and the dimensions and the data type The problem is that the fields in the header that give the information needed to calculate how long the header is are at the OTHER end of the header USR11 does as it should check that the type of the array is integer but to know where the type information is it has to assume that the array name is less than three characters long and that the number of dimensions is one It is just too cumbersome to do the checking any other way The other choice would be to do no checking at all I do have a patch that disables the checking for those of you that think a longer array name is worth the effort Write to the Customer Service department if you want it requests for LSI to disks on the Model 4 hardware differences it would require that a completely new driver be written to support these drives LSI does
188. have my wires crossed again MISOSYS See page manual through 2 17 2 14 of the M users This defines scope of varlables and storage classes True that variables defined outside of any function are Static main is a function Fm Shane Dawalt Well what I call external is what should be called static Like I said the language I use is not what should be used I m having a rather hard time in programming in C The language barier I m building around me isn t helping as you can see Perhaps my explanation above wasn t a good one I call variables external when they are declared outside of a function I don t see variables as static unless they have an absolute declaration of STATIC This is wrong I ll have to take Roy s advice and try to reread the pages he outlined MISOSYS Products Tidbits Volume I iv Fm Les Mikesell No no Static means something else in C A variable declared as static has permanent storage but limited scope That is it is only known within the A static variable will retain its block where it is declared declared within a function value between calls to the function but cannot be accessed outside of that function and will not conflict with another variable of the same name inside another function A static variable declared outside of any function may be accessed by any function within the same file but would not conflict with another variable of the same name decl
189. he listing for MS DOS available in French for 99 or so thank goodness I let him off the hook about five minutes later Two days later he bought a 3000 through one of the mail order places Letters to the Editor Volume I iv ANSI standards where to get them Fm Adam Rubin ANSI standards can be purchased from where else Sales Department American National Standards Institute Inc 1430 Broadway New York NY 10018 Unless they ve moved that is Also some libraries have reference copies of these I ve found them at both the Chicago Public Library and the Engineering Library here at U M for example Incidentally I believe the full name is X3 64 1979 where X is the committee 3 is the subcommittee 64 is this particular standard and 1979 is the year of the last revision of this standard which may be later by now X3 64 covers most of the escape sequences including a few interesting sounding ones Operating System Control Message Waiting etc You ll probably also want X3 41 which covers other control sequences Shift In and Shift Out Select Character Set and so on and X3 4 which defines printable and control characters and their functions I believe this last is the definition of the ASCII character set If you ll be using these to write a VI 100 emulator for KERMIT I d strongly recommend Digital s VT 100 User Guide which explains which functions are implemented and what they do Their
190. he interrupt rate is doubled to 60 per second There is nothing magic about that the rate changes when the clock speed changes Now a Model 4 can be operated under DOS 6 x considered to be Model 4 mode or under a Model III OS such as LDOS considered to be then in Model III mode When operating in the Model III mode the first 14K of the machine s address space is taken up by ROMs which duplicate the Model III operating mode Thus when you boot a DOS such as LDOS you are using the Model III ROMs in ROM which are BASIC interpreter There are interrupt based routines which keep the timer and time values as well as blink the cursor The timer counts down from 30 this is The cursor blink The Model III has routines more than just the rate is also predetermined by a hard coded value Anyone wanting to look at that code need only disassemble the ROM from 3529H through 3590H The routine is accessed every time an interrupt occurs unless something external to the routine traps interrupts and takes a different action Since the video cursor counter counts down from 7 to 0 the cursor should change state once every 8 interrupts or blink once every 16 interrupts which is approximately every half second or two blinks per second When the clock speed is changed to 60 per second without any external adjustment to the frequency with which the LDOS Information THE MISOSYS QUARTERLY SPRING 1987 55H at 4Meg clock spe
191. he key location and length fields in the ADDRESS DAT file to order search by ZIP Company or whatever This was to satisfy a few requests from folks wanting ADDRESS to sort and search on the Company name field I fail to understand why the MS DOS world clamors for key stroke multiplication facilities such as SUPERKEY POWERKEY and whatnot yet the TRS 80 users rarely explore the capabilities of their own KSM facility which has been available in LDOS and TRSDOS 6 for years Nevertheless we have added a new application called PHRASE This wammie accesses a PHRASE TXT file consisting of any number of phrases You can scroll through the phrases or enter a phrase mnemonic An lt ENTER gt will export the phrase selected Your phrases can make use of the logical ENTER character of export This is akin to a more or less infinite KSM The syntax of a phrase is a 2 character mnemonic followed by 1 78 text characters and terminated by an lt ENTER gt The PHRASE TXT file is composed of any number of phrases It is terminated by a period following the last lt ENTER gt Use any ASCII text editor to input and edit the PHRASE TXT file use TED for instance The phrase facility can work wonders when you want to prepare letters drafts etc using canned phrases You can also use PHRASE to store strings for application print formats How about DOS command strings It boggles the mind CARDX APP is a new application which is an enh
192. heir new machines IBM has essentially said to hell with existing hardware and eventually existing software This move seems to be motivated primarily by a desire to squelch competition from the clone makers This is leading to divergent pathways On the one path will be the millions of IBM PC clones which will almost surely continue to be manufactured like popcorn with the attendent low prices On the other will be IBM aiming at the corporate markets who want 3270 terminal capabilities and the Blue label What this holds in store as far as the operating system goes is anyone s guess The 386 based machines will become more NY R SRR NAN lt a NS an s The Programmer s Corner THE MISOSYS QUARTERLY SPRING 1987 SS Volume I iv popular and cheaper But here there are no set standards either hardware or software Microsoft has indicated that a 386 operating system is at least two years away probably more considering their usual schedules What will the market be There are indications that UNIX will become a popular OS for the 386 I do intend to start doing development with the 386 If so then ED ASM 86 will as it has for the past 2 3 years continually evolve to a new and better lifeform In any case regardless of the way the hardware and operating systems work out 8086 assembly language will be a common denominator of everything from IBM PC clones to 386 powerhouses At least until a t
193. hen control is passed to the filter program register DE contains the address of the DCB for the device designated Register HL points to the next character on the command line so the filter module could check for Volume I iv 521B 521E 5221 5224 5228 2229 522B 522D 522F 5231 5233 5235 5237 523A 523D 5240 5243 5246 5249 524C 524D 524F 5252 5253 5256 5259 525A 525D 525F 5262 5265 5267 526A 526C 526F 5272 5275 118A42 224452 225052 ED537052 Fl CB5F 2035 CB67 2036 CB4F 2837 DDE1 DD6E01 DD6602 222353 222953 240000 221053 012900 AF ED42 220000 23 DD7501 DD7402 EB 210E53 EDBO C32D40 21D652 1808 21E852 1803 21F952 CD0000 C33040 lF 00022 00023 LDHIGH 00024 00025 00026 00027 00028 00029 00030 00031 00032 00033 00034 00035 00036 00037 00038 00039 00040 00041 00042 00043 00044 00045 00046 00047 00048 00049 00050 00051 00052 00053 00054 00055 00056 00057 00058 00059 DEVROUT 00060 00061 00062 00063 00064 00065 SIGNON Version 5 1 QOAH 52A9 Systems 2CH 52D6 52E8 52F9 530E 5310 5312 5313 531B 531D 531E 531F 43 44 44 4E 180B 0000 08 50 4l 47 45 3807 F5 79 FEOC 00066 Inc 0AH ODH 00067 00068 ROUTMSG 00069 00070 00071 PAWSTRT 00072 00073 00074 50 41 00075 00076 00077 00078 00079 The Programmer s Corner THE MISOSYS QUARTERLY SPRING
194. hich pertain to the print formating I assume you are referring to compiled stuff Here goes for the internal variables PAGELEN page length LINESPAGE printed lines per page GGLMARGIN left hand margin GGRMARGIN right hand margin GLINESP internal line counter Now the left right margins can be set with BASIC commands LMARGIN and RMARGIN LINESPAGE will set the printed lines per page and PAGELEN changes the first value This means that you really don t need to twiddle with any of those variables at the assembly language level unless you want to poke the number of lines already printed Fm Bob Connors What I was referring to was the use of the H command of CED I have the PR FLI parameters set and CED does not recognize them at all Instead of getting the hard formfeed after 57 lines I get that then I also get the formfeed aka line feeds from CED when it apparently prints 60 lines This tends to throw off my listings pretty much However if it cannot be fixed since it ain t broke then I guess I will have to live with it Wish there was someway to change the number of printed lines per page though similar to the way EDAS does it as well as the margin although CED does honor the PR FLT margin setting Fm MISOSYS CED does not honor the PR FLT margin setting since it uses PRT the PR FLT still is doing its thing I made reference to TRSDOS 1 3 since EnhComp that same version works with
195. ing again But I thought what happened to me was that if the XLR8er was set to fast speed and FIXBANK was installed then invoking SYSTEM DRIVE 2 DRIVER MEMDISK produced a crash In any case I need KI4 DVR so my order will be forthcoming when the tax refund arrives No I m not trying to use LS DOS commands under LDOS just referring to two different issues in the same paragraph When you install a MemDisk driver under LDOS where does it go in storage Without a low memory driver area I assume it goes in high memory Then it pages the alternate banks into the low memory area and buffers data back and forth FM MISOSYS It goes in low memory That s discussed in the docs which you get with the package The package re uses some of the area of ROMC which of course is RAM after the SET2RAM switchover The cassette routines and the bootup code memory area is reused Fm Alan Kaplan Just a couple of points about the interface kit On page 5 of the documentation when invoking the memdisk memdisk should be in quotations i e system drive d driver memdisk Also when I invoke set2ram and later boot the system via the boot command I get a sys error and then the Idos prompt Othertimes I get to the debug screen After recovering from either of these events if I boot the system again the computer hangs and I have to push the orange button to recover I have a model 4d Fm MISOSYS Yes the memdisk installation
196. inted per page If I interrupt the printout and then reset my paper to the top of the page manually or by the printer Form Feed button how do I tell Forms flt to reset to line 1 for the next print out I m not using basic printing out the PROWAM CARD DAT file Hope I just didn t miss that bit of information in the manual LS DOS Information Volume I iv Fm Paul Bradshaw Sending a CHR 6 will reset FORMS s internal page counter To do this easily since you mentioned that you have PRO WAM use the program PRCTRL APP that you can find in DL2 here in LDOS Or use PRCTRL CMD Also if you get The Misosys Quarterly there was a program in the first issue and duplicated in the second issue that allowed you to send characters to your printer by specifing their ASCII codes Fm MISOSYS You don t use the FORMS command to reset the line counter The command LPRINT CHR 6 will reset the system line counter to top of page Fm LDOS Support jjkd If you use KSMPlus you can advance the paper to the next perf by using lt clear gt lt shift gt lt T gt instead of manually or at the printer That will keep the printer and FORMS in sync If you can get out to DOS or to someplace you can issue a FORMS command re issuing a FORMS command with the same number of lines should also re set the counter Might be easier than getting to where you can do a CHRS 6 Fm Les Mikesell Printing a CHR 6 will reset the FORMS line coun
197. ion Volume I iv ME that is relocatable code it will run almost anywhere in REAL memory without CHANGING THE CODE As I read statements from you and others I think you are saying that EXE files are more relocatable because THEY can play at ANY address in a segment e g they are not tied to that 100H offset BUT are they not fixed on the fly by DOS as they are being loaded to ADJUST addresses for where the EXE file will be in ACTUAL memory more or less the same scheme used to relocate a filter just under HIGH in you know what Right If so then they are not relocatable by MY definition Put it anywhere in memory and it will run Question SUPPOSE I manage to load an EXE file high in the current CS space and the EXE file extends beyond the end of this CS segment NOW what Does it work Does my computer BEEP Does my mono screen display colors Or will DOS refuse to load it this way Joe I BLUSH at not realizing that a program fixed to load near 640K would differ from one fixed for 256k I knew that in my Z80 32K Model III vs the 64K Model 4 Must have been blinded by the BIG NUMBERS I can STILL recall my original Model I complete with 4K and a cassette I have a lot of good reference books both P Norton AT amp T s BASIC manual getting DOS manuals from Bob Haynes and have Adv MS DOS on order STAND BACK MORE questions are SURE to follow Thanks for the current info Fm LDOS Support jjkd
198. ion of relocatable vis a vis the Intel architecture The byte format for a conditional jump and a SHORT jump is opcode displacement and the operation is ip lt ip displacement In both cases the displacement is a signed 8 bit quantity just like a 2Z 80 JR and is sign extended Isn t that relocatable by your definition For a NEAR jump the byte format is E9 dispL dispH and again the operation is ip lt ip disp As far as I can See there s nothing in any of the jump instructions that makes them tied to the CS Only indirect jumps and jumps out of the current segment are non relocatable The same is true for calls The byte format for a NEAR call is E8 dispL dispH and again the effect is ip lt ip disp Indirect calls and calls out of the segment are non relocatable but calls within a segment and that stay within the segment after relocation should be able to move without problem I think the 80x86 family is the 68000 has it but I looked closely at set is a relative instruction which I think while since The missing and it s been a the 68K instruction indirect jump or call which would have the effect of ip lt ip REGI6 or ip lt ip MEM16 But if a programmer stays away from things like jump tables and other indirect addressing on an Intel chip a COM program or any portion of it should be relocatable within the current segment no Fm Jim Kyle to H B
199. ion package also There is nothing specific about each implementation except the supplied libraries and where the executables run Although I had intended to make the Model III libraries available for some fee there was never MISOSYS Products Tidbits Volume I iv interest sufficient to even justify grooming a disk and documentation to explain how to do it The few folks needing Model III libraries have purchased an MC package to go along with their PRO MC package it s really not that expensive Fm Mark Harris Did you optimize the PRO MC code and then compare it I didn t notice a factor of 2 difference in speed but I didn t research very much either In any case PRO MC seems fast enough Fm Les Mikesell Yes I did optimize the PRO MC code To see the difference you must make sure the search time is significant compared to the time taken to load the programs Try searching all the files ona couple of disks MC and Separate compilation Fm Nate Salsbury I have been attempting to take apart LU CCC and compile parts of it separately to use in my LRUN project I like undoubtedly many others am having problems in getting the proper SYNTAX of all the various things the compiler looks for My present problem stems from what is probably an incorrect concept of the extern principle Here is a snippet of code mostly taken directly from LU CCC The things that I have added are noted with a lt which of
200. ions in memory before writing out to the MSDOS disk using the interleave table It would probably also be worthwhile to buffer as much as possible in memory even multiple files and allow swapping disks in the same drive Remember that most TRS580 s only have two floppies As for the file conversions we LS DOS Information PY oe We Er AS tee t i Volume I iv use the LDOS version of Wordstar and often need binary conversions between these disks and MSDOS since the file format is identical Most other files need the CR LF conversions a aes a I OEP ac i Ce 8 6 2 AE SEP g PgR EP CE bis t l Srt i ks as lech around and timed them gt and youare absalutely right I m already using a 1 sector buffer for both the source and destination so expanding both to an entire track should be fairly simple Swapping disks in a single drive would be useful too but something like COPY s X parameter with appropriate overlays SYSRESed would serve the same purpose and probably be easier to implement o r G Biwi Zy Fm Adam Rubin Well picE moved lt a few files It turns out there s something else that s taking up almost half the time of an MS DOS to TRSDOS transfer allocation Programmer s Guide has an example of that and I suppose the value in the MS DOS directory would be a good guess for its length so that doesn t sound too hard either So that oughta keep me busy for a little
201. is DCT 4 similarly the ninth bit being fielded starts referencing the low order register and thusly references DCT 3 the fourth dct byte A lot of the confusion comes from not recognizing that you are really declaring up to 16 bits of a machine word once you start a bit field whether you use the entire 16 bits or not The DCT CCC example program separates the 16 bits into two 8 bit byte bit fields of declarations to follow along with the documentation for the DCT itself The second bit field you asked about follows the same reasoning It s important to understand the thrust of the the final words cited from the README file Not all processors reverse a short integer in memory Ever wonder what swab was designed for The storage of double words is even more of consternation and low word and then reverse again the individual storage of each word 80x86 for instance Some just reverse the high and low words while others don t reverse a double word at all Then there s long words You are right on about your question concerning det struct dct regs 1IY Let s take a look at what it is doing regs IY is declared in the Z80REGS H header file as an unsigned short integer MC is rather tight on permitting pointer operations only on variables declared as pointers The CALL GTDCT amp regs statement issues an GTDCT service call which loads register IY with the DCT address of the drive being referenced Within
202. is denoted by the characters P in the header P rivate messages don t help the general membership and do take a valuable message slot 2 Delete all Thanks types of messages them This will free up a message slot for another message If any information is in there along with the thanks go ahead and leave the message It may help other readers single line or quickie after you read 3 Of course if you are not satisfied with the appearance or content of one of your messages after sending it go ahead and delete it when posting another copy Our Compuserve Forum THE MISOSYS QUARTERLY SPRING 1987 103 Volume I iv Other than these situations not be deleted Thanks messages should SIG paging Fm LDOS Support jjkd You may temporarily set paging off via the command SET PAG OFF at the main Forum prompt To turn paging off permanently GO CIS 9 and set your terminal parameters to have paging off While you are there you can set the proper number of lines per page and your terminal type if desired Garbage Messages Fm LDOS Support jjkd You can delete a message that is to or from yourself by using the Delete command from the main Forum prompt or menu You will then be prompted for a message number to delete Supply the correct number and away she goes You will not be allowed to delete a message if you are not the sender or recipient SNe weet ja qa E sr Ps SAT Meer fs t me H g
203. is forum Well the manual says Webster s runs with PC DOS versions 2 0 and above though PC DOS isn t ment ioned in the trademark acknowledgements Yep I ve run into the rules for MS DOS programming I saw the 1000 page book Microsoft finally published last summer Anyway thanks for your comments A few of the gory technical details are in my previous message to jjkd if you re interested and masochistic HELP with EXE files Fm Nate Salsbury I am in serious need of help I m doing things in MS DOS using MASM and LINK that I don t understand Consequently when the results are not as expected I have no idea of how where to fix things I m not trying to shirk my responsibility to read the manual I have read it MANY times plus several explanatory books and I am still completely in the dark MS DOS Information Volume I iv l In MISOSYS TMQ Roy says No MSDOS file has absolute loading instructions A COM file loads relative to the next paragraph An EXE file has a header record listing a chain of file locations needing segment fixup This reads to me like a description of what I know as a relocatable program in the 2Z80 world i e it will work run play ANYPLACE in memory My exguisitely BADLY written DOS manual says this the output file from MS LINK run or EXE is not bound to specific memory addresses therefore it can be loaded and executed at ANY CONVENIENT ADDRESS my
204. is to have the compiler do it for you The AZTEC compiler does it at run time when you use their linker but generates all the zeros in the CMD file if you use the M80 format Seems like it shouldn t be too hard to to generate labels for the start and end of the area and add the code to zero it at start up Of course you then have to keep variables initialized to other values separate Fm MISOSYS Aztec does it when you use their linker because they support a third segment type code uninitialized data initialized data I wanted to stay compatible with Microsoft M80 REL format supports only one CSEG and one DSEG Rich and I looked at handling initialized data in a COMMON but that just wasn t feasible My easy solution was to add the Z Y statement in the linker command line and that fixes it What s wrong with that solution Fm Les Mikesell The zZ Y option is adequate in that it makes the generated code conform to the definition of the language However it would seem more intuitive to follow the language definition by default and require the option for the nonstandard optimization Or better yet find some way to initialize at run time This might be impossible to link with L80 but if MLINK were link all Microsoft REL files it wouldn t matter much Fm MISOSYS Yes it would have been impossible with M80 and we had imposed the specification that MC was required to work with M80 That was a given It then becam
205. isk function call using the value contained in the B register but it was determined to be fluff overhead since most folks didn t need that facility The only LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 a 1906 as Volume I iv modules in the DOS which set up their own disk driver interface are BACKUP and FORMAT They could have used the un implimented SVC but why use up another four bytes memory is always at a premium The existing disk driver interface architecture has served us well for a few years now IBM PC Diskettes Fm Kevin R Parris I have what I think is a rather simple application I want to read a sector from an IBM PC single sided diskette 180K PC DOS 2 1 using the factory original diskette drives in my TRS 80 Model 4 Portable and a program running under 6 x TRSDOS btw jjkd there is one possible solution to your concern about trailing periods in the text at the end of a sentence put the version number in front of the name I have THE SOURCE and THE PROGRAMMER S GUIDE and the MISOSYS PRO MRAS assembler system all purchased from MISOSYS and a First Class Postage Subscription to TMQ now the GOTCHA I also have a great deal of ignorance as to the workings of Z80 processors and companion diskette drive controllers I understand the IBM 370 processors and MVS Operating System well enough but this is relatively new to me I have tried the to me at least obvious things RDSEC
206. istributed with 6 2 aren t they If so how big are they Is there a size savings over the ones I m using now Also about Thanks again for a superb issue of TMQ I have only one comment suggestion In your new section covering assembler on MSDOS could you back off just a little bit and fill in necessary information such as how many registers there are what are their names how do you address memory what are some instruction names how do you interface with the DOS etc Your first example was nice and fairly easy to follow but never having seen the structure of the 8088 chip or its brethren I had to make assumptions regarding the nameing of the registers AH and DX etc I m VERY interested in seeing this column continue Fm MISOSYS The new SVC 108 is semi official That means that I have received approval from LSI that the EXMEM can use that SVC number as it is more a system service Letters to the Editor THE MISOSYS QUARTERLY SPRING 1987 ON se Volume I iv function It is not part of the 6 3 release as it was not done by LSI I would have no comment concerning any future release of 6 x No comment does not imply any between the lines assumptions I have asked Phil Oliver author of ED ASM 86 to address 8086 programming from the standpoint of exactly what you have suggested I expect to take a different angle My supposition is that TMQ has a lot of readers already familiar with 2Z80 progr
207. it can positively determine that a variable is not initialized before usage Still next time you re on a UNIX machine that has sources run Lint on a couple of the programs You will find several programs that do not initialize sources mix pointer types etc wc for example doesn t initialize some of the counters to Byron P Peebles I thought I remembered reading it in K amp R so I looked it up again today It on page 80 something If you think about it you don t want automatic local variables initialized to zero Too much overhead everytime a function is called Anyhow Fm Mark Harris include lt stdio h gt static int i main lt printf Zd n i 12699 when run on my machine Certainly not initialized to zero Granted not a major earthshaking issue but since I was mildly flamed for asking the original question I thought I ought to support my case printed out Truth be known I really do not like to program in C I earn a living supporting an Ada compiler I find the language well thought out in most cases I wait for the day when one can obtain an Ada compiler for a PC that 1 doesn t cost thousands I want a validated MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 87 2 Volume I iv compiler 2 doesn t need massive amounts of memory 3 doesn t take minutes to compile the simplest of programs and 4 is available for the model IV Fm mark harris Another thing
208. k and BC with the move count Then LDIR or LDDR performs the move LD HL SOURCE LD DE DEST LD BC COUNT The Programmer s Corner Volume I iv LDIR or LDDR The 8086 operation is very similar the SI register is loaded with the source offset DI with the destination offset The CX register contains the move count The segment register used for the source is DS the destination segment is contained in ES In other words source DS SI destination ES DI Thus appropriate setup of DS and ES values MOV SI OFFSET SOURCE MOV DI OFFSET DEST MOV BX COUNT CLD or STD for decrement REP MOVSB Actually REP and MOVSB are two separate 8086 instructions MOVSB specifies a single byte transfer between DS SI and ES DI incrementing or decrementing SI and DI depending on the transfer flag value The REP prefix indicates that the operation is to be repeated CX times You can also move 16 bits or a word of data at a time with the MOVSW instruction This ability to move either 8 or 16 bits is common to most of the MOV instructions Note the use of the OFFSET operator This means that the offset value what in Z80 terms would be the PC value of the symbol is used rather than the contents of the memory location specified by the symbol The MASM convention is to assume the reference is indirect the ED ASM 86 convention assumes the reference is direct that is OFFSET is implicitly specified This
209. lared or referenced has not had a template defined for it I had expected that the template would be the external one at the beginning of LU CCC which is going to be part of my main I did notice one strange thing when I examined my TOK file with LSFEDII In the phrase extern struct ludir ludir the extern and struct keywords were replaced by tokens the ludir and ludir were ASCII characters as expected but the SPACE between _ludir and ludir was replaced by a token BO I also looked at the source file and that space is REALLY a space it s 20H just like it should be There is undoubtedly something I m doing wrong vis a vis the extern concept Can you make me any smarter on this arcane subject I really need to be made smarter on THIS point Fm H Brothers extern means the SPACE for this object is allocated in another and will be accessible at link time necessarily allocate any which refers to the copy of the template so to the structure members struct does not Space Each module structure must have a it knows how to get Only one module should have the actual allocation of space A handy way to do all MISOSYS Products Tidbits Volume I iv this is to put your global structure declarations in a project h file and simply include them in all modules for the program Just don t use them to allocate space in the header or things will really be weird Fm Les Mikesell The
210. lume I iv for the absence of an action once it detected the THEN Use of the THEN required an action because the compiler didn t trap the error It would be wrong to just forget the syntax error I could have done it that way in the patch it would be wrong since the omission of an action after the THEN violated the two forms of the IF Got it Fm Richard C Baker I recently purchased your PRO EnhComp BASIC compiler So far I have found it totally misrepresented I ordered it on the basis of a glowing review in Micro 80 magazine The biggest criticism I have is your manual It explains very little I am not an advanced BASIC programmer but I can get practically nothing reading your manual The examples are uselessly simplistic These should show all major possible variations of function uses as the TRS BASIC manual does What I would most like to understand is just what is the procedure for compiling existing programs written in TRS Model 4 BASIC I have tried several of my programs which run very well with the interpreter but will not compile by EnhComp without a multitude of errors And I can not discern what the errors are Most of them are listed as 138 and 139 But try as I might I can not figure out what is wrong or how to fix them to work Also I would like examples showing the use of the USING function in a working program I have tried every way I can but this function will just not work Even your examples
211. machines together at the same time BASIC amp SOUND Fm Paul Jaeger I have always felt that the sound command in basic and the sound service was fairly useless becau e it was so00000 loooooong Here is a patch for LS DOS 6 3 to shorten the tones Sys00l fix to shorten the sound in 6 3 0 patch sys0 sys lsidos using sys00l fix D10 33 00 00 00 00 00 00 F10 83 CB 24 CB 25 CB 21 eop The code is the same in 6 0 6 1 and 6 2 although not in the same place so you can apply the patch with a file editor just by locating the code CB 24 CB 25 CB 21 Don t forget that the patch will not become active until you re boot since sys0 is only loaded at boot time I have been using this patch since the days of 6 0 so I think it is bugless Special Characters Paul Bradshaw Using CHR 28 will reset for example To or space Fm many things inverse video test whether the special characters compression codes are in effect do a print chr 28 chr 194 Then test the POS 0 function of BASIC If POS 0 returns a 1 then LS DOS Information Volume I iv the special characters are in effect otherwise the space compression codes are in effect Fm Pete Betz Ureka you ve got it I blush to say that I never thought to look in cousin Roy s tome Thanks to all Paul s method is pretty foxy too But not to be an ingrate but I m still only halfway there After much experimenting I don t see any cha
212. me new ones to boot With all the talk about sorting BRINGUP files by date and time you may have realized that PSORT needed some work Well PSORT CMD has been revised to accept an expanded data definition of up to two key sort fields Not only that but the fields can be either character or integer Announcing New Products Volume I iv Because of the switch to application libraries we ve included WAMLIB a utility to build and maintain application libraries WAMLIB has functions to create a library as well as the functions add a module delete a module replace a module extract a module list module directory WAMLIB is menu driven and easy to use Since applications are now invoked from libraries PRUN has been enhanced to load applications from libraries too However since some folks may have a need to invoke an individual application file PRUN has been enhanced with a parameter APP abbreviated A PRUN will normally load the application from the WAMO APL library or other library when the module name is prefixed with the library number If you want to force PRUN to load a standalone application i e for testing before adding it to a library specify the APP parameter as in PRUN application APP Well that s it PRO WAM release 2 includes the window application manager 18 different applications in a library a library manager the PSORT utility the PRUN facility a help Facility and a DEFAULTS utility to cu
213. modified depending if you are running on a M4 with zero one or two wait states on Ml Fm Shane Dawalt Why couldn t a simple but small program be written into TAPE100 to check the wait states of the hardware TAPE100 is being run under I downloaded a program from one of the DLs which checked the wait states and reported its findings I m sure a simple ML program could be written to do the same After this program determined the of wait states in the system it could insert the appropriate timing values into the counter LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 32 Volume I iv storage locations Hum now I know why I couldn t get TAPE100 to work Fm LDOS Support jjkd That certainly is possible A more elegant method in machine language would be to test how much work you can do in a fixed number of timer ticks which will always occur at a fixed rate regardless of CPU wait states This info can then be used to choose timer values that result in proper operation This problem has existed in TAPE100 since day one of TRSDOS 6 0 0 and Tandy never caught it or complained to LSI We had about two direct inquiries in about two years and thus never really considered it a big enough problem to gt EIX Transferring data from a Modl100 is done far better via RS232 though the tape method could be useful for sending info from a Modl00 to a Mod4 via the mail or otherwise when you don t have both
214. mple teletype like character print on the screen MOV AH 14 MOV AL char MOV BL foreground color MOV BH display page INT 10H And there are various other routines which set up serial communications send a byte to a printer do disk I O and so forth All of these are completely standardized in terms of the INT involved and the registers used for parameters with the possible exception of certain limited extensions for non standard hardware The IBM BIOS technical reference manual describes all of them in detail as do many other reference sources The TRS 80 has had a long disk operating systems One of NEWDOS and of course LDOS line of varying the first was Somewhat strangely there has never been a real competitor to MSDOS It is the standard operating system for the IBM PC Essentially MSDOS has its roots in a cheap CP M clone The Programmer s Corner Volume I iv hacked up at some Microsoft and cleaned time and purchased by up a bit As with the BIOS MSDOS contains its own set of system calls which are accessible to application programs Most of these calls are accessed via the INT 21H instruction the AH register is loaded with a function code along with other pertinent registers as entry parameters Various utilities such as file access are supported by MSDOS IV Conclusion The future of the IBM PC world is hazy at this point as is the rest of the future With the introduction of t
215. n 5 3 whereas on 5 1 they were LBASIC This means that after the above BACKUP command the old LBASIC files should be PURGEd Also note that the upgrade documentation advises you at this point to install any other drivers or filters you normally use such as the LDOS keyboard driver That can be done with a SET KI KI TYPE JKL At this point you then can switch the system drive to be the hard drive system partition with a command such as two are shown SYSTEM DRIVE 0 SWAP 4 or SYSTEM SYSTEM 4 Since you drive l to be your issuing the which switches drive 0 and drive 4 assigned a hard drive partition to you now need to assign drive 5 second floppy This is done by command SYSTEM DRIVE 5 DRIVER MOD3 LDOS Information THE MISOSYS QUARTERLY SPRING 1987 22430 Volume I iv for the drive I O second drive is 2 so The prompt will ask you address 1 4 Your respond with a 2 the configuration Your is still in your first is now numbered 4 because Issue the command Now you get to Save booting floppy disk floppy drive which of the above SWAP SYSTEM DRIVE 4 SYSGEN The configuration will be saved booting floppy onto your As my final remark I will have expected you to write down in Sequence the answer to each question prompted during the installation That s to make sure you can do it again Type the commands and answers into a JCL file and you will never have to r
216. nary minus sign is a Special case in expression analysis Needless to say MRS612 FIX Model 4 and MRS512 FIX Model III are a short patch which will correct the bug As for your first note it is true that the CI flag of MRAS must be entered in addition to the GC switch We do feel that the workaround precludes the necessity of attempting a program patch Although MLINK does I y flag and generates a core image file the filespec generation doesn t switch to CIM from the CMD default Again I don t feel it necesssary to bother with a patch MLINK has been patched enough We do have a plan for a possible re assembly later on this year if time permits It should be correctable then Again thanks for making our job easier Point two is well taken indeed accept the Fm Dirk Vandenbossche I ve been using your software packets PRO WAM PRO MRAS and PRO MC for some time now and they all seem to work well Yet I ve met some problems In the 80 MICRO language program your PRO VWAM much about and see out an assembly works with issue of May PRSET APP that is explained As I don t know assembly language I ve typed in compiled the program with MRAS As you can in the listing added the assembler points several error messages As I considered the program to be correct I assumed something had to be wrong with MRAS So I tried out the TYPER program you enclosed in your PRO WAM packet When compilin
217. nge in DO when changing between SPECIAL characters and ALTERNATE characters nor is this status included in the above mentioned documentation Life will not be complete until I can distinguish between these modes too Any Ideas Fm Paul Bradshaw I m not sure about this so I ll look it up later and correct this if I m wrong or there s a better way to do it but port x EC contains the bit that controls the SPECIAL ALTERNATE character set switch This can be retreived via an INP 255 as BIT 3 l alternate 0 non alternate This port image in maintained in one of the GFLAG bytes MFLAGS I believe It is also bit three of that flag Fm Pete Betz Looks like you ve done it again The INP 255 approach is giving me the desired information I think you had the significance of the bit reversed I m getting bit 3 SET for special characters and RESET for alternate characters As one would expect the setting holds even when you drop out into Space compression Very nice What I don t question is why am I getting this information from an INP 255 Both you and the Tech Manual refer to port x EC which translates to 236 not 255 In fact I tried INP 236 first and got a big nothing Why 255 Fm MISOSYS Probably because an input from port X EC resets the interrupt latch You really don t want to extraneously reset that The Tech Manual on the Model 4 notes that an IN from ports FC FF are a mirror of port EC Of course
218. nges in 5 3 s passwords inspired some investigation Under 6 2 and earlier there was a user password for the SYS files Those files had access of none so the user password made no difference anyway but I recently figured out that it s was LSINC LDOS Information m a ee ET ae ee Volume I iv For 5 1 s passwords the significance of GSLID was relatively obvious but does anyone know what RRW3 stands for Was LSI founded by two guys named R and three named W Fm LDOS Support jjkd I hate to admit it but RRW3 predates me I m sure it s a head slapper Roy Fm MISOSYS RRW3 came from Roy Roger and William There were three of us We were the original founders of LSI Now who s Roger What s the version Fm Theodore Masterton Ok I have patched up my 5 3 sent in my cards and am about ready to convert ALL my disks to the two new DOS s One problem remains I have never been careful about noting which DOS 5 1 4 or 6 2 1 I have used to format a disk assumming they were identical Now I find out that I can only DATECONV the disks using the same DOS they were formatted with But I can t remember Is there an easy way to tell whether a disk was formatted with 5 1 or 6 2 Ido have a track and sector type disk utility but I am not a hacker type myself Help please Fm Adam Rubin There s at least two places on every 5 x 6 x disk 2 byte 0 This is always track 0 sector 2 2 DIR SYS
219. not have any plans to write a driver for 3 5 There have been several support 720K 3 5 Because of several drives It is hard to see much chance to break even let alone make a profit on that project But if anyone out there tackles the project we wish you luck The physical configuration for 3 5 drives doesn t seem to be standardized yet a driver that might work well for one manufacturer s drive may not work at all on another drive Just going to 720K probably doesn t justify the expense of new drives but if they start getting 4 or 5 meg on each disk reliably and inexpensively then they will have something HYPERSOFT S announced new Model 4 emulator PC Four for PC compatible machines is intriguing At the time of writing this I haven t seen or heard anything about it except The LSI LS DOS Column A TS NC ahh a Si ta a gar TT a g NME a ae res em a car Fj z 7 Volume I iv the ad in July 80 MICRO LSI considered doing something like it but never thought a software only approach would work very well The 8088 just isn t faster than a Z80 such that it could emulate a Z80 and take care of the MS DOS environment without a significant loss of speed There seems to be some demand for something like this and LSI is not going to fulfill it so we are glad to see that HYPERSOFT has moved into this area I will look into this product and discuss it at length in my next column Where do we go from here LSI has had freq
220. ns will not be answered But I am willing to take the risk that your response will simply be the disclaimer is correct on the chance that you might answer this question on page seven of Letters to the Editor Volume I iv volume one in the listing for Bootstrap Loader in the section after label SAL there are two lines where object code is shown and comments but no instruction mnemonics appear This is at locations 02E4 and 02E6 and also further down that page this happens again as well as a few other places I have noticed in the book Why is this Thank you for your fine work The new TMQ arrived Thursday the heavy plastic wrapper is a very good idea MISOSYS The source code in question relates to access of Tandy hardware Proprietary agreements between LSI and Tandy precluded LSI from publishing the source to those particular statements Fm New EXMEM SVC etc Fm Paul Bradshaw Regarding the newest issue of TMQ I m curious of something Is this new SVC 108 EXMEM official I see you are including it with the Model 4 interface kit for LDOS Is it now a part of the DOS Will the next version upgrade include it Why wasn t something like this included in the 5 3 6 3 upgrade Why not in 6 0 for that matter It is an absolutely WONDERFUL addition and was sorely missed your new RSHARD hard disk drivers These are different from the TRSHARD that I got with my Hard Drive and that was d
221. nt or near jump or call These require from 8 to 16 bits of displacement information It is interesting to note that inter segment references specify an absolute address whereas the intra segment branches or calls are PC relative This makes it impossible to directly search for instructions jumping or calling a specific address The Z80 has one useful feature lacking in the 8086 16 bit conditional jumps For example JP NZ target branches to target a 16 bit address on the non zero condition Amazingly there is no conditional 16 bit jump in the 8086 only conditional 8 bit jumps or branches Furthermore the 8086 lacks a conditional CALL To synthesize these extremely useful instructions I developed a set of macros included with ED ASM 86 to simulate them using the 8 bit conditional branches The Z80 supports block move and block byte search with the instructions LDIR LDDR and CPIR CPDR The 8086 has similar instructions REP MOVSB and REP CMPSB In the 280 there are separate instructions to perform incremental block moves LDIR and compares CPIR and decremental LDDR and CPIR in the 8086 a processor flag the direction flag indicates the direction of the move compare The CLD instruction clears the flag and specifies a decrement STD sets the flag and specifies increment To set up a block move with the 280 you load the HL register with the address of the source block DE with the destination bloc
222. ntents I m just writing now because it took me this long to figure out how to reliably reproduce the phenomenon It occurs equally reliably with the version of TED that comes with LS DOS 6 3 I lost the buffer contents with both the 5 3 and 6 3 DOS versions of TED when I did the following 1 Load TED and type in or load in some text 2 Delete one character in the text with lt CTRL D gt 3 Press lt CTRL L gt to chain in another file from disk 4 Type in filespec at prompt 5 When the Press lt ENTER gt to confirm message appears my understanding is that that message should NOT be appearing in conjunction with the lt CTRL L gt command press lt ENTER gt Pressing enter causes the text in the buffer to be deleted The filespec you typed in on the command line appears in its place followed by a drivespec LDOS Information THE MISOSYS QUARTERLY SPRING 1987 31 E Volume I 1v Fm MISOSYS Thanks very much for bringing that TED bug to my attention You narrowed down the problem very well which enabled me to find the program bug guite easily I have developed a patch for both the Model III LDOS version and the Model 4 LS DOS version They are both only a one byte change the fixes are in The Patch Corner as TED5B FIX and TED6C FIX I have forwarded the Model 4 LS DOS patch to Logical Systems Inc for their consideration Incidentally this problem should only have prevailed with the LOAD command not the
223. o test that the program is being run from a terminal isatty 0 before clearing EOF Otherwise the proper response to EOF is to exit gracefully 2 Unless you are parsing computer generated output predictable it is better to avoid using fscanf directly for input Instead buffer then sscanf to is fairly difficult to use fgets into a parse the buffer It recover from an error in fscanf and its concept of newlines are whitespace seldom matches reality Fm MISOSYS I assume that you really specified TIOCGETP and not tiocgetp there is a difference you know MC and C variables Fm Shane Dawalt I m puzzled on the operation of C when variables are passed to a function Specifically what do I do after they are passed Consider the following hypothetical function void myfunc buf_ptr count length int count length char buf ptr the function coding goes here return MISOSYS Products Tidbits Volume I iv Can I use the variables defined outside of the lt gt of myfunc or must I define a new set of variables inside the and set them equal to the variables which contained the passed information I ask this because at times PRO MC acts as if it cannot access those variables passed to the function unless they are assigned to variables defined within the function This happens most of the time in a Situation such as using the variable definitions of myfunc if buf
224. o the Editor Leap years a la 2000 revisited Continuing from TMQ I iii here s some more on the subject Fm Adam Rubin 71320 1052 Sorry but you ll have to take that up with Parliament Assuming the Statute of Limitations hasn t expired of course i Actually the Act included provisions for contracts and other dated events which were extended eleven days beyond any old style date agreed upon For example George Washington was born on February 11 1732 old style but was not twenty one years old until February 22 1753 new style For the benefit of anyone else if there IS anyone who s following this rather disjointed thread the library finally found a booklet I d requested The full title is Act and Bull or Fixed Anniversaries a Paper submitted to the Numismatic and Antiquarian Society of Philadelphia Nov 4 1880 by Lewis A Scott with an Appendix containing the Bull of Gregory XIII translated and the body of the Act of Parliament and it was published by the Society around 1881 or 1882 The question at hand was the proper date for the bicentennial of William Penn s landing on November 8 1682 old style it contains an interesting discussion as well as the other documents mentioned in the subtitle I found it very informative and recommend it to anyone else interested in the subject Here s what must be the final remark concerning the next centennial The following appeared in the Ann Landers
225. o to some fraction times two to the minus 127th That s why you got a large negative result from sinh 0 0 calculated as _ddv2 exp x exp Note to our readers corrected versions of the appropriate library modules were put on DISK NOTES 7 as suggested in TMQ I iii Fm Jeffrey McLean I have found more bugs in MC External and static variables are not initialized to zero K amp R states on page 198 Static and external variables which are not initialized are guaranteed to start off as 9 Scanf and fscanf stdin do not echo characters that are typed in to the screen Scanf and fscanf do not handle non white space non conversion characters correctly The program TEST CCC demonstrates this problem The FIXBUFS option does not work One last bug occurs in expressions where UNSIGNED LONG values are converted to DOUBLE see enclosed test program MISOSYS Here are the solutions As far as guaranteeing non initialized statics and external variables to be zero that essentially requires them to be initialized to Zero The Aztec assembler does this by collecting all uninitialized statics et al into a third segment type supported by their assembler code data uninitialized data Their runtime support can then zap that region with zeroes By doing so they don t take up disk file space in the object program for the zeroed region We decided to support Microsoft s REL bit stream which restricted us from implementing such a me
226. ome real soup but Stef s not quite up to that Stefanie finished up her fun for two year olds class In the Fall you know when I get out the Summer QUARTERLY Stefanie will start into the same class that Stacey just left Both of my girls will be in school on Tuesday Wednesday and Thursday mornings What will Brenda and I ever do The BLURB et 086 during the summer months since SS a a full time job That s been since In fact the only week taken off has always been the week between Christmas and New Years Now when I worked for AT amp T I had three weeks vacation a week of personal days and a week of Management days I never knew what to do with all of that time It s different now The most amazing thing about parenthood which has really struck home to me is the strong perception that I am reacting just like my parents reacted I m talking read as yelling like my father talked I m thinking like my father thought Perhaps I am more aware of that because I have the maturity which comes from being 44 years of age and I didn t have children of my own until I was 40 Or perhaps this is just a natural phenomenon Doing some vegetable gardening again this year I m trying some new crops having planted some peas green beans lima beans cukes cantelopes pumkins and the proverbial tomatoes Don t misunderstand it s just a small plot But it was fun starting everything from seeds Also started a bunch of zinnia
227. or number and the length of the sector 128 256 512 or 1024 bytes First let s look at how WRSEC works and why it will work on an MS DOS diskette Once the floppy disk controller FDC has found the requested sector it also knows that sector s length It then asks the CPU for the data a byte at at time until it s been given all the data for that sector based on the length value in the sector header Once the FDC has been given all the data it sends out a signal to tell the computer that it s done In the M4 this signal is a non maskable interrupt NMI Anyway this i e WRSEC works properly for any valid sector length What RDSEC does in 6 x is assume that every sector is 256 bytes long so the CPU asks the FDC for data 256 times and then assumes the read is done However if the sector is longer than 256 bytes the FDC realizes that the CPU hasn t asked for all of the sector s data and correctly signals a lost data error 6 x is designed to retry a lost data error indefinitely and of course each time is another lost data error Thus it keeps trying until you press RESET I don t know why RDSEC is implemented this way I once asked jjkd and he said it was needed because of the interrupt structure but I don t see why that would matter for reading and not for writing If RDSEC were implemented in the same manner as WRSEC adding read sector header and read track would have been fairly straightfo
228. ove new HIGHS to HL LD A GHIGH Set new HIGHS into the system RST 40 B 0 from above INC HL Bump to filter entry LD 1X 0 40H OR 6 Stuff TYPE byte IX 1 L LD 1X 2 H install addr into dcb LD HL 0 Successful RET er Error message handling pe ery hits LD HL VIASETS md a DB ODDH pare WS yy ee LD HL NOMEMS papSTON ae LD A LOGOT Rr JP GABORT DB LF PAGEPAWS Line Printer Pause Filter Version 6 x 10 DB Copyright 1987 MISOSYS Inc All rights reserved LF CR DB High memory is not available CR DB Must install via SET CR The actual filter JR DW DB DW DW JR LD CP JR Pass to PUSH EOU LD RST POP RET START HIGHS before filtering 6 PGPAWS Length of name and name sLoaded with DCB pointer 0 NZ CHAIN Go if not output A C OCH stof char Z OUTCF sgo if so next filter driver IX Save current pointer IX MODDCB P u this module s DCB 2 A CHNIO Chain to the next 40 IX 69 The Programmer s Corner Volume I iv THE MISOSYS QUARTERLY SPRING 1987 Volume I iv 312C CD2031 00109 OUTCF CALL CHAIN ssend TOF char 312D 00110 RX02 EQU 2 312F F5 00111 PUSH AF Save RET code 3130 D5 00112 PUSH DE save during key input 3131 3E08 00113 KBD1 LD A GKBD 3133 EF 00114 RST 40 3134 FEOD 00115 CP ODH see if CR 3136 20F9 00116 JR NZ KBD1 sget key if not 3138 DI 00117 POP DE 3139 Fl 00118 POP AF 313A C9 00119 MODEND RET 00120 0031 00121 LEN
229. place outside of a comment line where you could find the letter oh used is if instead of showing the patch bytes as a series of hexadecimal pairs it was depicted as a string A string could be used if one was patching a string of displayable ASCII characters For instance the patch D03 14 This is a new string F03 14 extra space for what would replace the string extra space for what with the string This is a new string Strings are shown within double quotes That s the only place where a letter G through Z could be used Also even though TRSDOS supports the colon notation to put more than one patch line on the command line e g PATCH TEST D01 27 56 F01 27 65 it does not support the colon separator when used in a FIX file it does support a semicolon which is used under LDOS to signify a trailing comment LDOS 5 3 supports a colon separator both in a command line patch and a fix file patch In order to conserve space in THE MISOSYS QUARTERLY we may logically print more than one FIX line on a printed line HOWEVER ALWAYS USE A HARD lt ENTER gt FOR THE COLON WHEN TYPING IN A FIX FILE for TRSDOS 6 If you use the FIXES TXT file from the DISK DNOTES corresponding to this issue please Separate out the individual fixes which you need by use of any text editor you find convenient to use wp RA 7 p ist SEIT 7 A n REA 42 a 2 es j m 8 A 2 A BFE 24 a 5 k i z PI Pr 3 ar ge F FF 2H EES Bn 3 Apes
230. pport 5 15 Meg drives and ask the same questions Your first job to install a new DOS version into your currently configured system is to make a few backup copies of the new DOS Since you may need some free Space on the system diskette you may have to delete one or two files Do this only with one of the backups you will be using to create your boot disk Next make sure you know what the current configuration is That means you need to know what heads and how many are currently assigned to each drive partition You should also know the driver currently being used If you don t know how can you tell The DEVICE command can be used to identify the starting head number of a rigid drive when using the abovementioned drivers For the R S driver LDOS Information THE MISOSYS QUARTERLY SPRING 1987 90 se Volume I 1v the number which appears after the word rigid in the display is the starting head number zero based i e a 2 indicates head 3 The FREE command also can be used to display the number of heads Write this information down and keep it Our MEMDIR utility can be used to display the names of resident high memory modules such as the hard disk drivers But you may know what driver you are using A hard drive can be divided up into pieces known as partitions A partition is assigned to a logical drive slot 0 7 by means of the SYSTEM command The exact syntax would be SYSTEM DRIVE d DISABLE
231. r KSM bug for years Fm Jim Hawes I am beginning the job of looking at LDOS 5 3 Most of what I see is valuable I appreciate that someone is still interested in LDOS and the Z80 in spite of the IBM pressure and the industry sub standard mania I truely hope the financial rewards are there for you Hey that is pure selfish if MISOSYS doesn t do well I know I lose access to some good programs and maybe some good bugs as well This has got to be in the I can t believe it category Do you know or suspect anything about bugs in the KSM FLT I had truely meant to write you a note on this about 6 months ago Please accept my apology I just got busy with some other stuff and it got forgotten These go all the way back to 1981 I see that you rewrote some parts of KSM FLT and rearranged some stuff but the bugs are still there They require highly improbable circumstances to show up Both involve replacement of KSM data Still after all this time someone should have reported it Bug one happens only when HIGHS after installation of KSM is within 10 bytes of a page end On re installation KSM looks for the proper place to put the CR replacement token The IY register points to an image of OLD and NEWHI values obtained from the KSM module The value at 4DFC is a pointer to this address group Now the A register has just picked up the contents at 4DFD to test If A is NZ you have LD H A and LD L IY 2 These load addresses re
232. r M P oi te met ha a 2 eye char det driver addr i m 9 DCT byte 4 flags ae bit field declaring 8 bits i z DCT byte 3 flags oe bit field declaring 8 bits E Se uchar dct curr i uchar dct high DCT byte 8 allocation data 7 unsigned dct gpc 3 unsigned det _spg 5 DCT byte 7 allocation data unsigned dct head 3 unsigned dct spc 5 uchar det dir gt dct FO saad DSTOW N S W 2244 The uchar dct instr declaration references a single byte which happens to be the first byte of the gt DCT DCT 0 The char dct driver addr declaration references the driver vector address which is DCT l and DCT 2 note that this is declared as a pointer to char which takes up two bytes MISOSYS Products Tidbits Volume I iv Now the fun begins You are confused over why the structure says that the flags at DCT 4 are declared next followed by a declaration of the flags at DCT 3 The cited reference stated that bit fields are unsigned short integers 16 bits and that the first declared bit field will be starting in bit 15 of a double Z80 register say HL Now that s the high order bit of register H Since Z80 words are stored in low high order it means that the first bit field declared will wind up in the high order storage location when placed into or read from memory Thus the first bit field declared must be referencing the fifth byte of the DCT which
233. r to break apart the other registers such as SI into upper and lower 8 bit pairs As with the Z80 addressing modes on the 8086 are not symmetrical However they are generally more flexible than the Z80 For example there is less bias in using the accumulator which is the AX register in the 8086 over the other registers For example SUB CL 5 is a valid instruction which subtracts 5 from the CL register Note at this point that the The Programmer s Corner Volume I iv 8086 follows the same convention as the Z80 is terms of the position of source and destination operands The familar Z80 arithmetic instructions which specify only one operand implicitly specifying the accumulator as the other operand are not supported in the 8086 For example the Z80 instruction ADD 7 would be analogous to the 8086 instruction ADD AL 7 where you must explicitly specify AL since BL CL etc can be specified as well Note that the use of AL and AX is generally more efficient than using other registers one byte rather than two is needed for the opcode The 8086 does support a wider range of addressing modes than the Z80 The BP BX DI and SI registers may be combined with restrictions to form a memory address for most instructions that access memory For example MOV AL BX DI would load the AL 8 bit register or the lower half of AX from the byte located at the address formed by BX
234. r whatever happens to be resident in memory conflicts in fights over hotkeys etc is an entirely different question DEDs6 1M won t leave you puzzled A powerful sector oriented disk file editor and a page oriented memory editor Search your entire disk while you tag sectors or clusters to be kept direct DED86 to write the sec _ tors as a disk file Great for file re covery without fussing over FATs DED86 gives you important features Position by cylinder head sector Position by logical sector or cluster Jump about subdirectories Keep sectors clustersimemory e Egit bytes in hexadecimal or ASCII Search strings in ASCII or hexadecimal DOS subshell available for your use Hmm I wonder if an analog of WHERE would DEDS6 M 86 020 59 95 5 S amp H have some use Especially in programs that use the absolute indexed jumps and calls MISOSYS Inc PO Box 239 Sterling VA 22170 0239 703 450 4181 MC VISA CHOICE to any and all code since you can always add some delta to a hard coded immediate address MISOSYS the first thing I ever saw Gary Kildall s VA residents add 412 sales tax S amp H Canada add 1 VX Foreign use S amp H times 3 AN A Dd ie MS DOS Information 55 MS DOS Information Volume I iv Applications for the User UNDATE by Luis M Garcia Barro I wrote UNDATE in answer to the query from Kenyon F Karl and to your guidelines THE MISOSYS QUART
235. r is responsible for final assembly of the OS along with mastering and duplication The vendor is also usually responsible for writing or at least assembling a number of the utilities that come with Some just never can get it right Who s version of MS DOS is this 3 Possible BIOS Hardware incompatibilities Because part of DOS directly modifies the low memory regions normally maintained containing and controlled by the BIOS interrupts a not fully 100 compatible machine BIOS could become a problem An example of this is popping up with the use of some 3 5 inch drive options on some clone PCs They will only format to 360K with many releases of MS DOS but will usually work fine with real IBM DOS except for BASIC of course Often a BIOS change will fix the problem Fm jeff brenton Well technically since PC DOS is a trademark of Digital Research Corp and refers to their concurent CP M 86 which now supports most MS DOS stuff from what I hear I would be surprised if ANYTHING that requires PC DOS would work under generic MS DOS grin but seriously MS DOS is bigger than IBM DOS due to trade secret problems plus IBM essentially re writes MS DOS anyway there ARE differences but anything that follows the rules for MS DOS programming you know the ones written by Microsoft should run on either Howsomeever some people decide that by the rules isn t good enough and do implementation specific klud
236. re are a couple of rules you have to observe when compiling modules to be linked together The variables that are global to more than one module must be defined in only one file and declared extern in all the others Preprocessor defines structure declarations and typedefs must be repeated in each file This is typically done by creating a header file with the necessary common information and including it in each dependent source file Note the difference between declaring a structure and defining one struct junk char line 100 struct junk next junk 2 is a declaration and does not allocate any memory for Storage However sizeof junk will be known to to compiler and B struct junk pile instance of the would then create an structure c struct char line 100 struct junk next junk junk is a definition and allocates the memory for storage Normally you would put a declaration like A in the header file and create the structs using syntax B where needed in each module You probably only have syntax C in one module with nothing to identify the structure in the other Fm Nate Salsbury Hardin Your comments regarding extern and struct made eminent sense and even better were right The module compiled without a hitch as soon as it found the template Now I have read a REAL LOT about C even including my MC manual if anyone is willing to believe THAT and I have NEVER seen
237. re hacker of some 45 years standing I blanch at the thought of doing more brain surgery on this computer The expansion to 128K was easily done on the dining room table but wholesale butchery of internal components and grafting of unmarked sanded off chips with doubtful results is not MY idea of adventure I will wait for the arrival of an external 500K RAM drive One last remark I have never been too fond of any DEBUG they seem deliberately crippled My monitors of choice are TASMON MICROMIND and MONITOR 5 I still have T BUG My next project has to be the study of various keybaord scans Keep up the good work and don t change TMQ Fm MISOSYS The 4 14A must have been a misprint There is no Appendix which could have been referenced The correct number should be 4 14 A Model III computer keyboard does not have the lt CTRL gt lt Function keys gt nor lt CAPS gt key thus SAID which is a Model III program would not expect to have them available and LDOS Information THE MISOSYS QUARTERLY SPRING 1987 1 97 Volume I iv thus does not use them From your Statement about SHIFT I can almost guess that you do not have an LDOS User Manual That kind of information would be in it in the section on the KI DVR It would also be in a Radio Shack Model III manual I believe the term which I used ala the 300 rpm issue was fiasco In order to adequately Support LDOS keyboard type ahead the floppy disk driver ke
238. rectory PRO WAM initially loads from the WAMO APL library Thereafter the UNIVERSAL function request for Application is now satisfied from that library on the drive where the library was found unless overridden by the application module specification mspec entry Even though a library can hold 32 applications more than most folks will have use for UNIVERSAL can still specify the library to access or default to WAMO APL Specify modname to access WAMO APL for modname Specify modname to access WAM APL for modname The new drive will be again used in subsequent UNIVERSAL requests until changed by explicit entries in a mspec either UNIVERSAL or programmed execution You can have up to ten application libraries designated WAMO APL through WAM9 APL Announcing New Products Volume I iv We ve expanded the capabilities of the window supervisor call with two significant enhancements programmed control of export and programmed invocation of an application module We use the former to automate the PHRASE application and the latter to enable the bringup hot key in CAL Of course both facilities are available to every programmer as well To expand on the concept of protected fields we ve added the concept of protected characters to the behavior of export import Any character within the defined rectangular export import area which has a value greater than 127D i e bit 7 set will not be output input i e
239. roducts Tidbits Fm Volume I 1v LDOS Support jjkd to Mark Harris Many examples does not a good programming technigue make As you indicated this is referring to using automatics not statics or externals As Les and yourself pointed out there is a big difference in the way these two categories are handled Fm LDOS Support jjkd to H Brothers Thank you as Les indicated I didn t point out the difference between the handling of statics and externs vs autos Gosh you just never know where that nasty ol stack has been grin gt gt Fm LDOS Support jjkd thanks for emphasizing handling of statics to automatics to Les Mikesell Yep the difference in the and externals as compared Fm mark harris Huh without beating a minor point into the ground this all started out with the question of how to have the compiler initialize variables to zero since plenty of public domain C code depends on this I never claimed that this was a good programming technique It obviously is not Furthermore K amp R I do not know about the upcoming ANSI standard Although I should I am an alternate on the committee specifies that statics and externals are initialized to zero PRO MC does not follow this No big deal It merely makes porting C code a somewhat tedious procedure Anyhow thanks for the help and the answers to my question It s nice to be able to get a quick intelligent response to a query on such a solid
240. rogram I tried this and the pre processor expanded the TRON statement but not the LINE macro like it says in the manual Then I figured that I might have to run through the pre processor twice to get the full expansion performed When I do this and get to the compile stage and I get errors on the lines with the LINE_ thingy in it I figured that the second pass would expand it to the line but no go I then tried searching all the files on both MC disks with GREP for LINE but I came up empty handed Fm MISOSYS Page 2 40 states that macro text replacement may be nested to a depth of 20 however I am not sure if the predefined macros such as LINE fall into that class I do note that your example only showed LINE_ which is not the predefined macro LINE Do you catch the difference Or did you just drop one of the underlines on each end of LINE in your message Page 2 39 clearly shows TWO underlines required on either side of the name Fm Michael Johnston Roy your right I was only using LINE instead of _ LINE Thats a major distinction anyone could miss at a guick glance I ll try it again and see if it works PRO MC and EOFs Fm Shane Dawalt I wrote a program which uses the FSCANF function in PRO MC to obtain an integer value from stdin I have set O KBECHO using the OPTION function so that stdin will echo to stdout i e keyboard input will be displayed Here s the kicker when I press lt BREAK g
241. rome adaptor the Color Graphics Adaptor or CGA and the Enhanced Graphics Adaptor or EGA The monochrome and CGA adaptors are controlled by the MC6845 video controller chip Editor s note The MC6845 is the same type of video controller chip found in your TRS 80 Model IILI and 4 Various display parameters are controllable by on chip registers which are addressed at standard I O port locations Similarly the display memory resides at standard locations The start of monochrome display memory is 08B000 0 and CGA memory is at 0B800 0 The standard setup is an 80 by 25 line screen with two bytes specifying one displayed character One byte is the ASCII code of the character to be displayed the other is a so called attribute byte Depending on the mode enabled this can specify whether the character is to be blinking dim etc The adaptors conventionally have 16k on board it is possible to have 4 separate screens in memory at once in alphanumeric mode The CGA can also support graphics The conventionally supported modes are 320x200 in either B amp W or 4 colors and 640x200 B amp W The EGA is much more complicated although it can emulate CGA mode In one enhanced mode 640x350 resolution in 16 colors is possible ssk There are standard I O port addresses for accessing serial and parallel ports the standard serial I O chip is the INS8250 The PC also has DMA Direct Memory Addressing controller chips and timers as
242. rothers Hardin You realize of course that there s a column crying to be writtef here To me what Joe is calling relocatable is position independent code which will operate the same way no matter where in memory it s located without any changes at all being made to it My definition for relocatable is based on the MS DOS Information Volume I iv THE MISOSYS QUARTERLY SPRING 1987 l Volume I iv derivation of the word capABLE of being N Wi RELOCATed and in the general sense applies Y byline on was a method for doing just that to SOFTWARE FOR YOUR PC 8080 code in an early DDJ In practice though I follow the G E mainframe convention which gave me my first exposure to the idea and use it to refer to code which is modified at LOAD time to account for its memory location Order Now 800 MISOSYS 647 6797 Interesting things come from this when considered as a single entity a COM program is position independent because the loader only sets the segment registers and does no modification at all to the actual image An EXE file on the other hand is relocatable IF and I think only if it makes any references to the SEG operator of MASM i e requires segment fix ups Yet the nature of the relocation done by the loader is not very different from the non relocation it does for i COM files i se AN Laas 1M a flexible data manager Easily used by anyone for managing their data I
243. rpreter or compiler does not need a detailed tutorial on the language since each BASIC verb and function should operate similarly You will find that true with EnhComp Besides you already have a tutorial in your existing BASIC Interpreter reference although I would question the term tutorial applied to that manual I don t understand your problem with the USING function It should work exactly like interpretive BASIC in PRINT USING strS var as well as a more powerful string function as in var USING str var Thus the formatted result can be manipulated further in the assigned string This you cannot do with Microsoft s Basic Besides the examples do work just like they are documented as modified by the note in README TXT For instance the first example of USING dH HF 1 5555 2 6666 3 9999 could have been written as USING Hee HE svarl var2 var3 or USING dk Ahi Ak ik o svarl var2 var3 etc I don t understand your confusion Perhaps our stating that USING is a function that can assign the result to a string may have led you to believe that you couldn t use it normally For instance consider the analagous pair of statements PRINT TIMES versus T2S TIMES PRINT T2 See the similarity Compile time errors usually result from either invalid syntax or use of verbs not supported by EnhComp like COMMON CHAIN EQV IMP ERASE CALL WHILE WEND WAIT DEFUSR NAME or SOUND
244. ruly nice processor like the 68020 is widely accepted grin to be one attitude towards specifically on acquainted with 8086 if when there is readers concentrate My next article depending on the this one will using ED ASM 86 to get assembly language I intend to use this article as a base for the next presuming a little bit of familiarity next time these 80x86 Editor s note You may address and ED ASM 86 isses with Phil on our Compuserve forum PCS 49 Philo s PPN is 72437 2057 A f mae tha a Ct TS TTT A prn 3 xe gt 41 iy a M ND ONE a The Programmer s Corner Volume I iv Product Tidbits ALTRES part of PRO ESP Fm Paul Bradshaw After a little digging I have found the patch areas that allow the ALTRES CMD program from your PRO ESP package to be moved around from Bank 2 This specific patch modifies ALTRES to use bank 4 but it may be modified to select any bank available This makes it possible to move ALTRES out of the way for those who have the Alpha Tech or XLR8er boards leaving banks 1 and 2 free for applications that MUST use one of those banks See ALTBANK FIX by Paul M Bradshaw in The Patch Corner Assemblers PRO CREATE Fm Clay S Scott While reading throught THE PROGRAMMER S GUIDE TO TRSDOS6 and other source the frequent use of operand in source statements as LD HL DW S etc Since it appears that this expression will alw
245. run on an MS or PC DOS machine Some like FED86 require a true compatible DED86 includes a version for the T2K Compatibility in the 8086 instruction sets Do you mean compatible across them or compatible with Z80 DED86 is for everybody Just like everybody using a TRS 80 needs a good disk zapper Obviously ED ASM 86 is only for a programmer It s architecture is perfectly suited for those folks wanting to learn 80x86 assembly language Since it combines the editor assembler linker debugger and disassembler into one memory resident file Fm most likely not ED ASM 86 questions 1 Fm Nate Salsbury Roy this is sent to you from the bottom of my learning curve To gain some experience I tried reproducing Hardin s tutorial Listing 3 in the June 87 issue of 80 Micro page 44 MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 TO Volume I iv After a LOT program that without error took the longest of cut and try I ended with a would assemble into memory My stickiest problem e g it to cure by brute force NOT reason was when the wrt l1n procedure was assembled It produced Symbol value difference between passes for wrt ln and wrt_l as well as mov curs and wait key This problem went away by changing the second line of code in wrt ln from mov bl attrib to mov bl attrib As an old Z80 man that was what I was inclined to do but your manual says on p 47 I
246. rward However as I learned from a discussion here RDTRK isn t very useful for reading sector data Nothing to do with TRS 80s it s part of the FDC s design The FDC interprets several valid data values as sync bytes is that the right name assumes that it s at the beginning of a sector and the rest of that sector s data is garbled Remember that data is stored on disk as a stream of bits not bytes Fm Les Mikesell The Trsdos 6 floppy rdsec is hard coded to stop reading after 256 bytes so if you want to read 512 byte msdos disks LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 oe ae Volume I 1v you will have to write a substitute RDTRK is not generally considered reliable in DDEN and is not needed by the 05 so it is not implimented The infinite retry on lost data errors is to allow interrupts to be kept on between finding the sector header and the start of the data This makes it fairly likely that the read will be interrupted by the heartbeat clock and since the disk revolution speed is an even multiple of the interrupt rate if it does it once it will take a while to recover I suspect the 256 byte limit was imposed by someone thinking about the chance of accidentally reading a 1024 byte sector into the DOS 256 byte buffer However the lockup from the retries is not much better The write code lets the FDC interrupt out when the command completes because it must also handle wrtrk
247. s The TED on your LDOS disk worked perfectly for us The three key combination lt CLEAR gt lt SHIFT gt lt gt worked for us to exit TED The lt CTRL gt lt U gt key moved up a page in text don t forget that a page is 22 lines Also the LEFT and RIGHT arrow keys worked properly You gave me a clue when you said that the UP ARROW key gave you a left bracket I don t recollect if you related that on the phone however it is a strong clue TED uses the LDOS keyboard driver s Extended Cursor Mode ECM ECM can be invoked in one of two ways There is a bit in the KFLAGS which if set establishes ECM that s what TED does The ECM mode can also be toggled by depressing the three key combination lt CLEAR gt lt SHIFT gt lt SPACE gt The up arrow would generate a left bracket 5BH if ECM were not engaged Also the arrow keys would not function for cursor movement if ECM were not engaged Thus I suspect that you had inadvertantly depressd that key combination and disengaged ECM On the other hand even if ECM were not engaged CTRL U and the exit combo would still work Better double check your procedures Fm Jane A Layman I very much enjoyed receiving the latest issue of THE MISOSYS QUARTERLY with the very welcome patches provided for both LDOS 5 3 and the LDOS 5 3 Model 4 Interface Kit I have found however that even with the patch to TED CMD my cavalier editing habits can still cause me to lose the TED buffer co
248. s and marigolds which are now in the ground With luck I 1l be harvesting my pumkins while I get out the next QUARTERLY lt grin gt I am really thinking about planting a walnut tree although they take a good ten years to develop they are said to be great climbing trees and Stacey really likes to climb trees We ll see The BLURB Volume I iv Announcing New Products RATFOR 86 and RATFOR M4 This product has been a long time coming But MISOSYS is finally shipping both the MS DOS compatible RATFOR 86 and the TRS 80 Model 4 compatible RATFOR M4 Both products were developed by James K Beard Ph D Currently James is a Senior Scientist working for the Hughes Aircraft Company To get you started in understanding a little about the RATFOR language here is a short introduction to RATFOR and its usage written by James K Beard RATFOR is RATional FORtran Although sometimes described as a preprocessor RATFOR is actually a language unto itself Conceived in 1973 by Brian Kernighan of Bell Laboratories RATFOR was originally imp Lemented in collaboration with P J Plauger now of Whitesmiths Inc in 1975 The original version on nine track tape for microcomputers such as the DEC PDP 11 is still available from Addison Wesley Publishing of Reading MA The authoritative book at that time was Software Tools by Kernighan and Plauger The C language and UNIX were under development by Kernighan and Ritchie at Bell
249. s ACTUAL addresses when called in by DOS Or asking the question differently is there a way to fool DOS into thinking that MY machine has only 256K of RAM so that what I do HERE would work on my customer s 256K machine also I m probably asking for too much information via CIS I would be happy to just get back some recommended reading that would explain the REALITY of what is going on the structure of the files being created why they act the way they do this stuff MUST make sense to Lotus and how to take real CONTROL of the processes that create and use these files The manual that came with my XT prepared by Compaq is absolutely devoid of ANY information along these lines e g There is a program called EXE2BIN which Changes an EXE file to a COM file Well that is TRUE but to me no help WHY does one do that HOW do you tell if you CAN do it to any specific EXE file WHAT changes are made to the file itself These are just a couple of completely unmentioned unexplained items I would like to be able to read something that described the entire system and the rationale behind it Thanks for reading this far If you can shed some light in these murky quarters I cannot think of adequate words to thank you further MS DOS Information ee Volume I iv Fm Jim Beard Welcome to the wonderful world of MS DOS The 8086 uses a segmented architecture which means that it is basically a 64K chip with schizophr
250. s byte BTW lt spacebar gt works as well as lt comma gt for the delimiter between the address and the new contents Cursor strangeness Fm LSI Virgil There was a bug in the previous versions of BASIC that has been fixed Trouble is that it was such nice bug and a lot of people miss it When execution of a program in basic is ended and you are returned back to the BASIC Ready prompt BASIC is supposed to tidy a few details up just in case the program left a mess One of those details is to turn the cursor on So when you LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 40 Volume I iv do a PRINT CHRS 15 from the BASIC Ready prompt it looks like nothing happens Indeed the net result is that you are back where you started But put exactly the same statement inside a program followed by a statement that should present a cursor INPUT A for example and the cursor is gone Then comes the end of the program and the BASIC Ready prompt complete with cursor The only case that I can think of that is still a problem is if the break key is hit interrupting the program for some reason and the program is resumed with the CONT command CONT does not and should not turn the cursor off even if it was off before the BREAK Of course the programs that have to survive a BREAK could easily turn the cursor off more often just in case One character device names Fm Shane Dawalt I was SETting a filt
251. s for searching the header key field s are only for PSORT I changed this in PRO WAM Release 2 so that ADDRESS can pick up the information on the first key and use it If you go for the upgrade you ll have what you were asking for and more Fm MISOSYS It Fm Alan H Pesetsky Been fiddling with a copy of Pro NTO now WAM that I bought from a RS store some time ago 00549 I seem to have trouble using Shift fl amp Shift f2 8 shift f3 when I have KSM Plus installed fl amp f2 8 f3 work however KSM Plus allows the use of both sets and I do Is there a way around this conflict if it is a conflict Also are there patches or updates I should have Fm LDOS Support jjkd If you have KSMPlus installed with definitions for the function keys they can t be used to control Pro WAM You must undefine those keys before Pro WAM can see them Fm Bob Haynes Here s a mini tip for anyone using Roy s PRONTO PROWAM program Would you prefer not to use a multi key control P or whatever command to activate your PROWAM system Just found out that you can specify an Fx key as the activation character even though the docs seem to imply you cannot Works just fine apparently since PROWAM does not take over those keys under AFTER it activates When doing assembly programming a quick double tap on Fl takes me directly to my SVC APP screen once for PRONTO once for the APP I love it MISOSYS Products Tidbits THE MI
252. s redirected to a disk file SCRIPSIT generates a listing file instead of actually generating the output to the printer it doesn t know that the DOS is sending the printer output to a disk file nor does it care Using the Q command once the printing has stopped I can close the listing file and restore the PR device by simply invoking a RESET PR DOS command again from within SCRIPSIT using the Q command added by the MPATCH Now I am ready to look at the file listing For this I use the DOS command invoked via the Q command LIST PRINT Applications for the User Volume I iv I have the convenience of the PAUSE facility SHIFT to slow down the listing while I read it This gives me the advantage of making revisions to the source text without having to waste paper ribbons and typewheels And printing to a disk file is quick especially if its a hard disk or a RAM disk When I am ready to generate another listing I can easily invoke the ROUTE command from within SCRIPSIT That s because the DOS will re use the high memory previously allocated during the first invocation of ROUTE PR PRINT TXT Of course I have to add the REWIND parameter to that ROUTE command to ensure that subsequent printings are not appended to PRINT TXT but rather written anew Now another DOS facility further eases the burden of issuing these Q commands that facility is Key Stroke Multiplication KSM If you haven t us
253. s to symbols not yet defined If MASM doesn t generate an error then I guess it defaults to always assuming an indirect refference unless OFFSET symbol is specified That may be arbitrary to an assember s implementation My suggestion is to always use the brackets to indicate indirection to the observer of the source code In that way you know what is the case EnhComp BASIC Compiler Fm Clyde W Preble Received PRO EnhComp in good condition and have converted several of my BASIC files However when I first tried to convert the Russian Roulette game I got error conditions in lines 280 and 305 280 PRINT CLICK SOUND 0 0 and 305 SOUND 0 2 When I deleted SOUND 0 0 in line 280 and SOUND 0 2 in line 305 the program converted correctly and works all okay except the SOUND of course is absent When I tried to transfer EnhComp to TRSDOS 1 3 I got errors which showed that the files were not converted and were listed as protected files which as far as I can determine are not protected The disk I formatted for the conversion formatted correctly and the four EnhComp files transferred correctly before I tried to convert the disk to TRSDOS 1 3 and the TRSDOS 1 3 disk had ample granules available Fm MISOSYS Here s some answers for your EnhComp questions EnhComp does not support any SOUND statement It s not listed in the manual On the other hand a little COMMAND addition in Z80 MODE should allow you to adap
254. says data record not found RDSSC turns on the drive light and never comes back only way out is the RESET switch RDTRK true to the documentation does not read a track I do not want to invest in a commercial program right now I do not need PC Disk directory handling or multiple TRS80 Operating System format support or BASIC token conversion I just want to read a selected PC Disk sector so I can put the data from it in a TRSDOS file in the other drive I have not studied it very closely yet in The Source but I do not see why RDSEC locks up I do know that the PC uses 9 512 sectors while TRSDOS uses 18 256 and suspect that has something to do with the problem A way to implement support for RDTRK would be OK I could take the sectors apart I think without too much trouble BIW NOT a Complaint just curious what were the reasons for not including RDTRK support in 6 x TRSDOS And if it were included would it help me with this project Any assistance will be greatly appreciated you may be sure I confess I have made little contribution here and have asked several questions most of them answered very well but with so much help from so many LS DOS Information r Volume I iv people one day I will learn enough to do something really useful Thanks Everybody in advance Fm Adam Rubin As you may know each sector on the floppy disk is preceded by a header which gives among other things the sect
255. se permitted for pointers be disallowed I have attached a sheet showing the two functions which need updating by the addition of a cast to unsigned int Edit your WINDOW CCC file to include these casts and you will have no further problems These are the revised lines 11 12 15 and 28 of WINDOW CCC bu unsigned int regs BC gt gt 8 0 if unsigned int regs AF amp 1 0 return unsigned int regs BC amp 0xff 1 return unsigned int regs AF gt gt 8 MISOSYS Products Tidbits Volume I iv The Patch Corner General Information The following information should be read before you type into a file any of the patches noted in THE MISOSYS QUARTERLY It is unfortunate that our printer prints the letter O and the number 0 almost identically Unless we utilize a filter to slash the number zero the two are difficult to distinguish However when it comes to patches all is not lost In an LDOS 5 or TRSDOS 6 direct patch the letter oh is not used in the patch code it may appear in comments which are lines beginning with a dot The direct patch format of TRSDOS 6 which we use in our patches is Drr bb xx XX XX XX XK XKX oe Frr bb xx XX XX XK XKX KX ee The patch is usually a pair of lines The first line begins with the capital letter dee This is immediately followed by the rr field which stands for record The rr field is always two hexadecimal digits Actually i
256. short and could easily be done from the command line but they are presented here in JCL format for the benefit of Disk Notes and this same file will be put into Data Library 0 and or DL6 of the LDOS forum on CompuServe Make that five new patches plus the level change Roy just called with a bug that he just found and he fixed Under some very unusual circumstances DIR refuses to read in an extended directory record because it thinks it already has it when it doesn t To install the patches included here you must first create each PATCH file exactly as shown as an ASCII file Then save it with the patch name specified in the patch Each physical line shown must be terminated with a lt ENTER gt Each individual patch has been separated by a line of signs The JCL file that follows will properly install all these patches After all files have been created do a DO MAKE63K D d S s lt ENTER gt where the s in the example is replaced by the Source where the FIX files are and d is replaced by the Destination where a COPY of your Master 6 3 disk is AK PV kd ae Seppe se ane 47 NV ks OA mY MvaDS ota w A g t id As see ESE a v Mee ee we i a ee ga a n 2 The LSI LS DOS Colum Volume I iv THE MISOSYS QUARTERLY SPRING 1987 Volume I iv PATCH SYSO SYS LSIDOS D using SYSO1 FIX S PATCH TED CMD UTILITY D using TED3 F1X S PATCH SYS6 SYS LSIDOS D using DEVICE1
257. st thing on the line The following CR was then picked up by the DSP driver to designate as displaying rather than as control Hard Drive Problems Fm LDOS Support jjkd Let me mention that if you currently have the physical floppy drive zero in as logical drive zero there are only two methods of getting the floppy addressed elsewhere First off if you have the system files installed on some logical partition of the you can use SYSTEM SYSTEM n of the logical drive to hard drive where n is the number LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 is a problem 42 e Volume I 1v be used as the system drive Typically you ll want to set up the hard drive bent slightly so that it comes out straight after doing this command As an example here s a drive setup before Physical Floppy zero HD partition HD partition HD partition HD partition HD partition HD partition Logical two three four five Six one NUE amp be with system files on the HD partition one After a SYSTEM SYSTEM 6 Logical Physical 0 HD partition one l HD partition two 2 HD partition three 3 HD partition four 4 HD partition five 5 HD partition six 6 floppy zero To bring in logical drive seven as your second floppy you can then execute a SYSTEM DRIVE 7 DRIVER FLOPPY DISABLE and answer the 1 4 prompt with 2 Finally i SYSGEN D 6 that will produce bootable system
258. station which does not have its own compiler at all The Model 4 RATFOR translator is provided as a relocatable object file which needs to be linked with FORLIB REL A Job Control Language file is included which performs the linking process using your FORLIB REL FORTRAN library and L80 linker The Model 4 version RATFOR M4 also includes a customized version of the LED text editor for source code maintenance Each version comes with a batch or JCL file useful for performing Announcing New Products n ET tein yim maa Volume I iv the RATFOR translation FORTRAN compile and subseguent linking to generate an executable file from the RATFOR source These RATFOR translators are each currently priced at 99 95 5 S amp H US the price includes installation and reference manuals disk based software and binder Order M 21 073 for Model 4 or M 86 073 for the MS DOS version PRO WAM Release 2 Since the release of LS DOS 6 3 we have had many requests from PRO WAM nee PRO NTO users asking for revisions to the BRINGUP application to extend its date support beyond 1987 Well here is an announcement that should please every PRO WAM user and make many other folks run out to get PRO WAM Release 2 In the 80 Micro review of PRO NTO Release 1 back in November 1985 Hardin Brothers said PRO NTO is one of the most useful products for the Model 4 I ve seen Well we have supercharged this new release to include just about every feat
259. sting plus lots more Of course due to the state of the 8 bit marketplace I think this will be the last release of PRO WAM That s why we wanted to put so much into it With the release of new 8 bit chips the Hitachi HD64180 and Zilog Z280 we want to explore the feasibility of revamping MRAS to support the extended instructions of these two processors We may be able to report something next issue We also want to spend some time later this year grooming Little Brother LB and LB86 I have a personal commitment to support a co existance of LB and PRO WAM Currently LB cannot run with the reduced memory available when PRO WAM is loaded That may be due to the file buffering requirements imposed by the Aztec C compiler used to compile the Model 4 version of LB We hope to explore a redo of the LB source to compile with MC From time to time we get program submissions for consideration as potential MISOSYS products Currently under evaluation is a new interactive game targeted for the Model 4 This MegAdventure package titled LAIR OF THE DRAGON has been submitted by David Goben Preliminary feedback from my evaluators reveals great potential for this package I ll probably have more to report in TMQ II 1i as some decisions will be made during the next few months If we go ahead with this package an MS DOS version will be developed as well I am looking over a fast backup and restore utility for MS DOS Even though the marke
260. stomize your PRO WAM configuration all on disk PRO WAM Release 2 includes a totally revised reference manual housed in a 5 1 2 by 8 1 2 3 ring binder If you don t have PRO WAM yet order M 51 025 for 74 95 5 S amp H US If you are already a PRO WAM or PRO NTO user and want to trade up to release 2 send in your old master diskette and order M 51 125 the complete PRO WAM Release 2 package for 25 5 S amp H note if you have purchased PRO WAM release 1 since April 1 1987 you can trade up to release 2 for 20 5 S amp H with the return of your old disk The trade in offers are valid only until 03 31 88 so hurry up with your old PRO WAM tradein to Release 2 Announcing New Products 12 THE MISOSYS QUARTERLY SPRING 1987 Volume Liv ual Release 2 ia ha MAAD Window controller and Applications Manager PRO WAM supplied applications can turn your 128K Model 4 4P TRS 80 into a sophisticated business or personal machine rivaling the best of them That s because PRO WAM comes with many useful and powerful menu driven time savers and work organizers PRO WAM includes eleven applications a complete HELP facility a data file sort program a 99 page user manual and is easily installed While you operate other programs you can request its services with a single keystroke PRO WAM saves you typing with its EXPORT and IMPORT functions which allow you to move data across windows between Programs Reguires 128K T
261. t at the FSCANF function the program locks up in an infinite loop redisplaying the prompt then an error message over and over again The routine is designed so that integers less than 1 and greater than 1000 generate an error message after which the routine loops back and redisplays the prompt I know that an EOF returned by PRO MC is OxFFFF I assume that FSCANF is returning OxFFFF as the result My routine sees it and displays the error message then goes back for more But why doesn t the EOF condition remove itself after the next FSCANF Also how can the EOF be changed from lt BREAK gt to the lt SHIFT gt lt whatever gt lt whatever gt sequence I tried the IOCTL function with the sgtty header included of course but I kept getting an MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 89 Volume I iv Illegal logical record number The exact IOCTL syntax I used was static struct sgttyb sg 0 0 other statements which have nothing to do with ioctl if ioctl stdin tiocgetp amp sg perror IOCTL error exit 1 and the program continues Am I missing a simple point in the above code e e Or am I completely wrong Fm Les Mikesell You have to use an explicit clearerr stream to reset the EOF condition A couple of points 1 If there is any chance the program will be run with redirected input isn t there always you should be sure t
262. t might account for the speed issue is that AZTEC C uses a 1 K buffer for file 1 0 Fm mark harris Could be but target search file was on memdisk Therefore disk I O speed shouldn t figure in too much I should point out that I didn t run exhaustive tests and the difference in speed was slight For example something like AZTEC s 10 seconds to PRO MC 11 seconds Fm David Huelsmann Even on memdisk you would see some impact and the slight impact you described would be consistant I believe Fm Les Mikesell Yes I also compiled grep with PRO MC to test it The odd looking asm code at the end that watches for the break key press appeared to be the only way to interface to both compilers You can reduce the size a bit more by using PRO MC s wild card expander instead of the coded one but then you give up the ability to use followed by an alpha character in the filenames In the very limited testing I did with the two versions the AZTEC version ran nearly twice as fast I expected this on floppies since the AZTEC compiler uses buffered access reads in IK chunks but I was surprised that the difference remained consistent even on a memdisk The other point in favor of AZTEC is that I can link to a mod 3 version of their library modified to also work with mod 1 LDOS while running in the Mod 4 mode Fm MISOSYS You could do the same thing with PRO MC if you have the MC model III also Model I compatible vers
263. t that Russian Roulette program very easily Here s a little code which adds a SOUND statement and demonstrates its use DEFINT A Z FOR D 0 TO 4 FOR T 0 TO 7 SOUND T D NEXT T D STOP COMMAND SOUND TONE DUR 280 MODE LD A 8 TONE AND 7 LD B A LD A 7 SUB B LD B A LD A amp DUR AND 31 RLCA RLCA RLCA OR B LD B A LD A 104 RST 40 RET MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 35 track disk parameters 80 Volume I iv HIGH MODE ENDCOM You would need to include the code verbatim from COMMAND SOUND TONE DUR through to ENDCOM A statement such as the fourth line will beep the requested sound Although this syntax SOUND tone duration is a little different from SOUND tone duration in Microsoft s BASIC it s a straightforward conversion is another example assembly language into your BASIC or packed strings For our other readers this of the ease in which an routine can be integrated programs without USR CALL None of the EnhComp files are protected The Model III EnhComp comes supplied on a stripped down LDOS 5 1 4 disk To transfer the files over to TRSDOS 1 3 if you really want to FORMAT a 35 track Single Density disk using LDOS Then copy the BC CED S and REF CMD files and the SUPPORT DAT file over to that You should have no problem in using TRSDOS 1 3 s CONVERT utility Fm Bob Connors Was just using EnhComp with LDOS 5 3 0M and came
264. t can be a 4 hexadecimal digit number if the file to be patched has more than 256 sectors Hex digits use nothing but the numbers zero through nine and the first six letters of the alphabet A B C D E F or a b c d e f The record number is immediately followed by the bb field which stands for byte The byte field is also two hexadecimal digits just like the record field This is immediately followed by an equal sign The equal sign is immediately followed by the first patch byte the xx shown above The patch byte is again two hexadecimal digits Where more than one patch byte is included on a line it is separated from its predecessor by a single SPACE The line is terminated with an ENTER TRSDOS 6 and LDOS 5 3 patch formats use a find line record This is used to verify that the file being patched is actually the file you want patched All of the bytes ALTBANK FIX by Paul M Bradshaw 03 22 87 Patch to cause ALTRES to use bank number 4 THE MISOSYS QUARTERLY SPRING 1987 Volume I iv noted in the F line or lines must be matched in the file before any of the D patches will be utilized The second line of the pair begins with the letter F which stands for FIND The next six positions are identical to the preceding D line Following the equal sign on the FIND line are pairs of hexadecimal digits which should align themselves with the preceding line So far the letter oh is not used The only
265. t is crowded with such utilities there s room for more When you move into a crowded market the trick is to make your product a little bit better than the rest Which brings up the topic of program submissions Yes MISOSYS does publish products which have originated from independent authors like many of our readers are After all when MISOSYS started out nine years ago we Were an independent and we still are we re just better known due to The BLURB 3 THE MISOSYS QUARTERLY SPRING 1987 Volume I iv perseverance and a quest for doing the right thing We are always open to program submissions But please I don t think we need any more Model III software Even for a Model 4 I have my doubts as to the viability of marketing a new Model 4 program On the other hand a program running on a 4 which is easily ported to MS DOS is another issue At this point we have not considered addressing the Apple Atari or Amiga markets nor UNIX either So products for those markets won t be viable here If you have something you think we should see either call or write first no need to send your package Our primary consideration is for products in the MS DOS arena Focus on Public Domain We continue to get requests from folks primarily overseas wanting to obtain files which have been placed into the download section of our Compuserve forum Since we have had no takers for our plea of assistance in this regard I am forc
266. t it is I have come across a lot of public domain C programs that assume all pointers and ints are initialized to zero As you might imagine it s a bit of a drag to go through 200 lines of declarations and initialize all variables Fm Mark Harris to LDOS Support Well actually I ve never used a C compiler that didn t have all variables initialized to zero But you see I ve always used C on mainframes minis and the OS of course zeros out all pages when your process initially receives them Security y know So while the compiler itself is not initializing the data area the end result is that everything starts out as zeros And the fact remains that a lot of programs take advantage of that Examples pec on Vax 3B2 Perkin Elmer NCR Tower C on DG MVs IBM 370s etc etc To be honest I don t know how the seemingly thousands of C compilers available on the PC clones behave Fm Les Mikesell My K amp R is always in the wrong place when I want to look things up but The C Programmer s Handbook from AT amp T says Any permanent variable is initialized to 0 NULL unless otherwise specified I m pretty sure K amp R says roughly the same ANSI might not guarantee this since there are some machines where the same bit pattern does not represent zero for all data types MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 able to 4196 e Volume I iv Fm Les Mikesell to MISOSYS Or the correct way Which
267. t o o 7 n t Pa VA wot s 5 45 a At A AF Ale tart A y Aga IN gt A j t 4 t T 3 cH Ep EES a Sa Maudi PO eC lt A Saray v NL vy Fm LDOS Support jjkd kasuks that you are using the exact same equipment to call our Compuserve forum as a BBS you are having problems with then there are three possible sources of problem Problems accessing a BBS 1 Phone lines from you to them Since your dialed route to CompuServe is likely not the same as to the BBS there easily could be a problem from you to them If long distance try another carrier AT T SPRINT etc If not try taking your equipment to a friend s house who lives more than five miles or so from you and see if that makes a difference 2 Baud rate parity stop bits modem mode Make sure that you match what the other end is expecting Most common is seven bit word even parity and one stop bit You are limited to 300 baud and should be in originate mode 3 Make sure that computer with a the other end modem When I really is a called the number you gave insert number here I got a message indicating that the number was disconnected Always try calling with a regular phone in addition to a modem to make sure that a person or message isn t answering Repeatedly calling a normal residential phone number mistakenly thought to be or published as a BBS number can cause untold grief at the
268. t s menu driven for ease of use absolutely no programming needed Why settle for expensive over kill with data managers you have to program when economically priced LB86 wili do mailing labels customer lists registration data Siore up to 65 534 records per data base Us to 1 024 characters per record Uo to 64 fields per record Un to 254 characters per field Nine fieid types for flexibility Select and sort on up to 8 fields Keen 5 different indexes for data access 10 inout update screens per data base C crintout formats per data base Extensive on line help always available N And of course as you point out all 8086 88 i 5 2 conditional jumps are position independent ae U like the Z80 JR or the 68xxx series EONA AA teu a entirely right Now I only have to figure out Fm LDOS Support to H Brothers You are j P e A a what I was thinking of LB86 L86510 74 95 5 S amp H Fm H Brothers to Jim Kyle Jim I agree that this is fodder for a future column was thinking that to myself when Nate first brought the subject up There s a bunch of experimenting I have to do first to see exactly what happens with an EXE file don t think I ve written many COM files at all lately BIW when LINK is done with it and as the loader gets ahold of it Of course the outcome from a user s point of view is that almost all programs are relocatable since the user doesn t have to worry about addresses no
269. tch is installed Could the use of TIMES be the root of your problem It s possible that MODELA III file may think this is more but BOOT does not For some time now I Fm Jim Hawes I documentation than code work when SET2RAM is in have had several Mod3 programs which either use a complex bank switch deal to get a doubler or copy the ROMs to RAM and go full RAM in the Mod3 mode I have always found it necessary to hit the reset switch rather than issue the BOOT command It s hardly worth a lot of worry and I am sure that anyone having LDOS Information Volume I iv run in to the situation will find the orange button fairly quickly Fm MISOSYS The reason why the BOOT command does not re boot your Model 4 in III mode while operating under LDOS 5 3 and the SET2RAM auxiliary utility is due to the following The BOOT command executes a RST 0 instruction This of course transfers control to address 0 The code in the ROM image at address 0 turns off interrupts and then branches to the bootstrap code through vector 3015H The technical specifications supplied with our Hardware Interface Kit mention that SET2RAM uses the address range specifically associated with the boot strap code In fact address 3015H is reused for the EXMEM handler All is not lost What needs to be done to reclaim the BOOT handler is to change the code at address 0 after SET2RAM is installed Here s the old and proposed new old
270. tch between banks when referencing error messages By redirecting stderr to a file you also avoid another problem THE PAPER BLOB RRR Yes that disgusting animal that proliferates beside your desk when debugging will no longer have a fighting chance And there is yet another advantage to this method too Your wife won t throw your printer out the window because it is keeping her awake It may also be possible to write a filter that will read the error file and make a note of what line numbers are in error The filter could then feed SAID with the GO commands and LINE numbers to position the cursor to the proper place in the file for correction Of course this would require someone to actually WRITE a filter for this specific purpose Which is something that most people are probably not willing to do However in the would prove to be worth it in terms of the amount of time saved looking for the line numbers in the listing and then issuing the proper GO command Just a pipe dream of course but possibly worth looking into if your pro gramming in C serious about on the mod 4 Writing FILTERS for LS DOS 6 x by Roy Soltoff Because of the size of TMQ I iii a few articles programs had to be deferred One was a discussion of filters for DOS 6 x I ll pick that up in this issue Rich s stat function and rm utility will have to still be deferred po g Device filters have been an inherent part of the LDOS 5 x and
271. ter but if you are able to do that why not just send a form feed in the first place instead of moving the paper manually Fm Alan H Pesetsky Using PRCTRL CMD I tried sending 6 to the printer to reset the internal FORMS page counter didn t seem to work I have been using PRCTRL CMD never got around to getting the app version to send a form feed and then would set the paper to the top of the page manually Was wondering how to do it the other way round I assume that LPRINT CHRS 6 only works from basic Was looking for a solution from DOS KSMPLUS is a good idea also re issuing my forms set up sound even more convenient as its in a jcl file This problem came up when I was printing out my Pronto card dat file in various ways I usually have KSMPLUS disabled then because my current KSM file using the function keys which conflict with PRONTO Thanks to all for your suggestions LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 39 Volume I iv When I think of it that s what I do But sometimes I interrupt the printing move the paper up to tear something off then reset the paper etc Just looking for a way to make sure that the computer knows that it s the top of the page now 2Fm Gary Phillips 0 From DOS level you can always try COPY KI PR followed by lt CTRL F gt lt SHIFT CTRL gt That should achieve the same results as the LPRINT CHR 6 Clumsy but it works Fm Paul Bra
272. ter return please slim down your program to the bare essentials neccessary to reconstruct the error I will also need your MISOSYS Products Tidbits Volume I iv MAP file your JCL file and specific details concerning all of the responses you made to the DSM51 module which generated that MAP file Please provide all of the files on a floppy disk ED ASM 86 80x86 assembler Fm MISOSYS I have made arrangements for Phil Oliver the author of our ED ASM 86 assembler product to check into our Compuserve forum more frequently than he has in the past So if you have any questions concerning ED ASM 86 or 80x86 assembly language please leave a message for our expert Phil s PPN is 72437 2057 Fm Bob Haynes Will be getting into MSDOS assembly in the near future and am interested in your ED ASM 86 and DED86 packages Forgive my ignorance of 16 bitters but here s a couple of questions Will a DSMBLR 86 also become available Do these packages run on any MSDOS system What sort of compatiblity exists in the 8086 8088 80x86 instruction sets and how do your packages fit in as a broad range development package I d really like to have the same kind of professional software under that environment as on the Z80 MISOSYS DSMBLR 86 will be done ED ASM 86 includes a memory disassembler which generates labels Considering that the debug portion can load an EXE file what more do you need All of the 86 packages
273. they DON T MEAN ANYTHING to me because they seem SO contradictory In one place I m told that EXE files can load ANYPLACE in memory here I m told that that LINK is selecting a SPECIAL part of memory with the HIGH switch MS DOS Information THE MISOSYS QUARTERLY SPRING 1987 50 Volume I iv 4 So now I have my EXE file composed with the HIGH switch of LINK Using DEBUG I load it into memory and find that the DOS has assigned it to Code Segment CS register 9F 94H I am supposed to use THAT address to merge this file into a BASIC program with THAT address being set by using a BASIC statement to DEF SEG amp H9F94 I do all that is outlined and it works just fine HOWEVER when I move the program and BIN file that I created on my 640K XT over to a 256K Tandy 1000 HD the thing bombs After some experimentation I discover that when I load my EXE file created on my XT using DEBUG on the Tandy 1000 it NOW loads at CS 3BEOH and recreate the BIN file needed using THIS address This means that my BASIC program is NOT PORTABLE TO OTHER GW BASIC MACHINES because the EXE file used with it loads at 9F94H on MY machine 3BEOH on ONE Tandy machine and WHO knows where on any other MS DOS computer Is this progress Is there ANY way to use MASM and LINK to create an EXE file that will ALWAYS load into some predetermined address Even COM files with an ORG 100H will still be actually loaded into variou
274. thing bad enough to force a new release As far as DSPLY goes yes the documented termination characters are ETX and CR and anything else below X 20 should be output for display I can see taking a zero as a terminator also it is a useful construct It was most likely just never documented Even though you can think of a zero as both X 00 and as I don t consider it as a real control character Far too many things treat a binary zero as a special case For example it does have a special meaning to the DO device that I m not sure is documented anywhere Who all knows what it is LS DOS Information THE MISOSYS QUARTERLY SPRING 1987 2 AT s Volume I 1v Fm MISOSYS Taking a zero as a terminator is against the use of a zero to imply that the next character has a value of 0 31 and should be displayed instead of being treated as a control character Of course the bug with that would be that the DSPLY handler won t trap and pass the next character if it turns out to be a ETX or CR Fm LDOS Support jjkd Allowing a X 00 to terminate an DSPLY string output would not preclude the ability to use a X 00 with DSP or PUT of course And of course it should either be fixed or documented depending on whether it is considered broke or mis documented right Virgil Fm MISOSYS I doubt it immensely to see LSI change DSPLY so that a NULL will terminate the string One reason is that historically only t
275. thod The alternative was to either ALWAYS initialize such data to explicitly or not initialize i e leave initialization subject to whatever was in memory by using DEFS for declaration Fm We decided that to force the compiler to generate zeroes would not necessarily be best for the target environment a Z80 with 180K drives That s why we went to such great extent in our linker to provide an ability to keep DS regions declared in data segments On MISOSYS Products Tidbits THE MISOSYS QUARTERLY SPRING 1987 E Volume I iv the other hand we provided a linker switch Z Y to allow the programmer control over the generation of zeroed space You can edit the MC JCL file to add that switch to the MLINK command line You can also make a change in the MCMACS file The macro VAR is used to declare space for statics local or global The space is defined by the statement DS SIZE If you want to have those regions explicitly initialized to zero change the statement to DC SIZE 0 You could even get fancier by adding conditional code centered around a option We may discuss that in TMQ I iv That s the best we can do with uninitialized data The echoing of characters during stream input from the keyboard is governed by the KBECHO option There was a bug in fopen which effected KBECHO as reported in TMQ I iii If you add the statement option 0 KBECHO 1 to your code you will get echo That fix
276. u next get asked the quantity of heads and the starting head for the partition assigned to the drive number identified in the SYSTEM command The response to these questions should now be obvious from the data you collected with the DEVICE and FREE commands LDOS Information Volume I iv or already knew That takes care of the first partition For each succeeding partition you repeat the SYSTEM command noted above designating a different drive slot i e er ae ask you about The TRSHD5 6 drivers will also the drive step rate mostly always zero or the smallest number given in the message prompt the number of heads the 5 Megger usually has 4 the 15 Megger has 6 and the number of tracks per surface the 4 head 5 Meg has 153 whereas the 15 has 306 When you are through assigning each partition to a distinct drive slot you are almost home From your descriptions of your current configurations I suggest you assign the partitions to these drive slots 4 1 2 3 Drive 4 is where your current hard disk SYSTEM drive is You then BACKUP 5 3 to your hard drive SYSTEM partition via a command such as BACKUP 0 4 S I If you have read the instructions in the upgrade documentation you will note that I have not told you anything different here it just takes a little thought as to what your current configuration is Here s a few more tips Note from our documentation that the BASIC files are named BASIC o
277. uent requests to make a good DOS for PC hardware The trouble with that is the limitations of the PC hardware which is already approaching its own obsolescence and a need to be MSDOS compatible which would defeat the whole purpose The PC came out BEFORE the Model 4 remember The trick is to write a really good DOS for the machine that hasn t been built yet Although a better DOS could easily be done for the IBM world for One or Two Million dollars we don t feel that it would be economically feasible So for now we wait Now where did I put that crystal ball I am working on a major new project it is for the IBM world of users though and it is only for users who invest in mutual funds The design is almost complete now and I have started the coding It should be ready in a year or so If any of you are involved with mutual funds and would be interested in a system that completely handles and monitors your holdings drop me a letter or give me a call I ll tell you all about it We knew it would happen There are some patches out there on some BBS that claim to be able to fix the date handling in TRSDOS 6 2 A word of caution If it could be fixed adequately with patches we would have done it that way The patches that we have seen just alter the base date from 1980 and deal with only about half of the other things that need to be done to keep all of the operating system working correctly And the range of dates remains only 8
278. und without concern as long as we can get to the entry point and back from the exit point Presto Changeo we now have relocatable code On the 8088 non long jumps are always absolute within the current code segment but are treated relative to the value of CS Thus if we move an entire code segment no problen everything is hunky dorey Since the CS register has been changed all the offsets within the segment are still OK One can consider this the same as the way Z80 absolute jumps work the only difference being that the 64K doesn t have to start at zero in the physical memory map that s what the CS segment register is for Now let s move just part of the code in that code segment just the second half at run time not at assemble time Since the assembler calculated all the displacements from the start of the code segment and the MS DOS Information THE MISOSYS QUARTERLY SPRING 1987 54 Volume I 1v distance from the Start code in that second half has changed all the will stop working On the 68000 and 6809 relative jumps can be coded relative to the current IP not some other landmark So as long as we keep track of the entry points and where we came from properly generated code can be moved willy nilly without fixing anything on the part of the DOS the loader or anything else In my book only that is truly relocatable code Fm H Brothers Joe Now I m confused about your definit
279. ure requested by our PRO WAM users and added other significant enhancements that should make PRO WAM even more useful to just about every owner of a 128K Model 4 To those Model 4 owners uneducated about PRO WAM here s a little PRO WAM overview The WAM stands for window application manager PRO WAM is a manager of applications which pop up in windows on your video screen You can pop up a PRO WAM application whenever you are running any other DOS behaved program The PRO WAM application can capture anything on the video screen by means of a function called import Similarly a function called export can be used to pass back to the interrupted program anything which is in the PWO WAM application s window There are essentially two parts to PRO WAM 1 a resident module which supervises and controls the operation of the video windowing and application execution and 2 a set of application programs some of which may be made to reside in memory PRO WAM uses about 2500 bytes of your machine s high memory and one memory bank That s why it requires a 128K machine Application programs provided with PRO WAM include a calender a bringup and tickler file an address rolodex file a 3x5 card file a modem aautodialer a 4 function calculator a disk based key stroke Announcing New Products 9 THE MISOSYS QUARTERLY SPRING 1987 Volume I iv multiplier a todo list manager a terminal program and more Let
280. valent That s for the folks trying to port stuff from other systems MC and bug reports fixes Fm Roger A Sabin While using the PRO MC compiler I came across a problem Some of the float routines do not work properly Specifically the following routines do not work fraise fexp and fsqr Also the sinh function returns HUGE when the argument is zero Fm MISOSYS We had a number of little problems pending with MC that we have corrected I located the root of the problem you were having with fexp and fsqr The problem also carried over to fraise of course fsqr uses fraise which uses fexp The latter was where the problem lay Apparently one line of data from the table of Fm Byron P Peeb les I think the constants associated with the polynomial documentation should adda brief description expansion was corrupted Fixing fexp cleared of the differences between include up the problem with all three MISOSYS Products Tidbits 92 MISOSYS Products Tidbits Volume I iv I also located the bug in sinh 0 0 The problem was actually with the ddv2 function it didn t test for a zero value before doing its thing That function divides by two a float or double by decrimenting the exponent byte which is kept in excess 128 notation Since a zero exponent is used to designate a zero value float an arbitrary decrement from OOH to OFFH would arbitrarily change the float from a value of zer
281. valent of using the SVCs and the DO device in the Model 4 environment Unfortunately Microsoft is their own worst enemy and they generally refuse to follow their own guidelines when producing their products The result Virtually all MS DOS BASICs ignore ANSI SYS and go directly to the display hardware Fm Les Mikesell Did you try using ANSI SYS then opening CON and printing the control Sequences to it Just a thought I haven t tried that either If it does work BASIC will probably then forget the current cursor location Fm Nate Salsbury I solved MY problem about clearing to the end of screen with a ML program which I can merge into any GWBASIC program I just put the details into DL 0 with the name KLEER ASM Perchance it might be of use to you with slight modification I found that printing a string of blanks was fast enough for my purposes to just clear one line LDOS MSDOS Utility Fm Ed Boudrie Is it possible to transfer files from LDOS 5 1 4 into a MS DOS system I keep a large mailing list on Profile III HD over 4000 names and would somehow like to install that data into a MSDOS database Any suggestions Fm LDOS Support jjkd Yes Ed as a matter of fact there are numerous methods The problem breaks down into two stages 1 Physically transferring the data from one machine to the other and 2 Altering the data format so that it is useable by software availalble on the new machine
282. vector storage area After that it s simply necessary to clear register HL which tells the DOS that the program successfully executed The RET transfers control back to DOS PAWSTRT JR START OLDHI DW DB 6 PGPAWS The Programmer s Corner Volume I iv MODDCB DW DW 0 This section of code begins the resident portion of the filter The first three instructions are the same as under LDOS 5 x they represent the module header However under DOS 6 x the module header has been expanded The two additional word fields are used by the DOS for linking between DCB s The MODDCB field is loaded by the filter installation code with the DCB assigned to the filter by the DOS Actually the second word field is not yet used by the DOS but remains reserved for future use START JR NZ CHAIN LD A C CP OCH JR Z OUTCF CHAIN PUSH IX LD IX MODDCB RX01 EOU 2 LD A CHNIO RST 40 POP IX RET The first four instructions here are similar to that for LDOS 5 x Here I corrected the entry state test Since GPUT represented by Z NC is the only condition I want to examine I test for this simply by branching directly to CHAIN on an NZ flag condition This single jump is necessary and sufficient since CTL and GET calls each produce and NZ entry state The CHAIN routine is unique to 6 x Under the LDOS 5 x convention we chain downstream to the filtered device by CALLing or JumPing to the address vector which was ori
283. way then the LDOS 5 3 master disk contains a Job Control Language file which will install the new 5 3 release onto your existing drive The name of the file is INIT54 JCL and assumes the use of the TRSHD3 DCT driver If you are using some other driver you will have to make any changes in the JCL responses to the questions asked by your driver The INIT54 JCL file is mentioned in the README TXT file There is no way I can discern the partitioning of your 15 Meg drive into 4 partitions how do you divide 6 heads into 4 pieces Even though your little printout shows the four partitions as 5Meg 5Meg 2 1 2Meg and 2 1 2Meg must I assume that s heads 1 2 3 4 5 and 6 When it comes to re doing a configuration one better not make any assumptions one better know Thus I must say that you have to figure that one out yourself You had to do it for LS DOS 6 3 if you had to redo the configuration The procedure for LDOS 5 3 should be quite similar on the other hand we provide instructions for doing that whereas Logical Systems had absolutely none Now here are some general words that may help you A Radio Shack hard drive is installed using a driver The driver may be named TRSHD3 DCT an early R S driver supporting only the 5 Meg drive TRSHD5 DCT a later Model III mode driver supporting the 5 15 meg drives or TRSHD6 DCT a driver for the Model 4 mode The TRSHD6 driver is equivalent to the TRSHD5 driver they both su
284. wever the code they sent us 5 1 3 source and the 5 1 4 master did not provide any indication of a bug report Can t fix what we don t know about I agree perfectly with your analysis of the two KSM bugs which have existed in KSM for more years than I have fingers on one hand Sure we revised some of KSM for 5 3 But the revisions were strictly related to the CFLAG checking and copyright messages On the other hand I don t agree with your fix I don t like X patches when I can introduce a D patch Your analysis of the first bug was correct The DEVIBL stores a pointer to the PRMSAV region which is exactly 10 bytes higher than the new HIGHS pointer stored at PRMSAV 2 The initialization code picked up the high order byte of the DEVTBL pointer and used the low order byte of PRMSAV in an attempt to point to one byte before the KSM module in Memory It then added an offset calculated from the distance between the logical enter character and the module origin This is using apples and oranges as you pointed out You LDOS Information Volume I iv were also correct in demonstrating a correct solution to be the use of both the low and high order bytes from PRMSAV However your solution of changing the LD A H with a LD H 1Y 3 wouldn t fit into the existing code space The better solution would be to replace the LD L IY 2 with a LD HL 2 6 DEVTBLS and then change the offset to use the distance between the
285. when you call the customer service department we can deal with your problem quickly with no wasted time Some people are having difficulty finding their Customer Service ID Actually it is quite simple but the instructions just don t get the message across to everyone Just type in ID lt ENTER gt from the LS DOS Ready prompt I am amazed at the number of people that send in the registration card with the serial number instead of the Customer Service ID They are very different and serve different purposes The Customer Service ID is a must on ALL communications with the LSI customer service department Now with 20 20 hindsight I wish we had used CS instead of ID for the command to get the Customer Service ID That may have prevented some of the confusion While I am covering some of the basics let me answer one more common question Why is the TRSDOS62 or DISK NAME that appears after the drive designator in a directory listing of so many disks not changed in the LS DOS 6 3 update process That field is of course a user maintained and created field It can be you want with the ATTRIB be tampered with by an changed to whatever command and should not update utility The LSI LS DOS Colum THE MISOSYS QUARTERLY SPRING 1987 44 Volume I iv An interesting tidbit about Basic has been brought up by Hardin Brothers author for 80 Micro It turns out that there is a restriction on the name of the array t
286. while it does in fact produce a core image file does not use the CIM extension it continues to use the CMD extension as documented I hope you are able to it is rather nasty since the listing appears just fine it is only the CMD file generated with MLINK which shows the error not too many people are likely to look there for such a thing The CI and I y bugs are perhaps really just documentation changes track down this DW bug Fm MISOSYS I greatly appreciate the detail of your bug report and the extent to which you have narrowed down the problem It certainly made it easy for me to find and rectify the problem Actually the bug was more obscure than you even thought It of course was isolated to specifically the case of multiple operand DW s Next you had to have a relative value in any operand position immediately followed by an absolute value which was negative particularly one in which the negativity was indicated by the presence of a MISOSYS Products Tidbits Volume I iv minus sign or another relative value of different mode For instance if in your case you coded the first line rather than the second DW DW DUMLAB OFFFFH 0 DUMLAB 1 0 the correct values would be generated The bug was that the mode of reference was not being initialized to absolute for the 2nd through nth operands of the DW statement since that was being done only preceding the parsing of each statement The u
287. will often be less efficient than the old way nevertheless the patch should stand because it brings MC into conformance with standard C They are in The Patch Corner as AUTOIN FIX MRAS assembler Fm Michael R Johnston I received the fall TMQ and it s great I can t say enough about it I sure hope that subscriptions pick up I d hate to lose this one Your comments about spending more time with the family really hit a soft spot here All too often in the height of writing this program or debugging that program we forget about the ones we love If more programmers would remember this maybe there wouldn t be so many of us divorced Conclusion Keep the copmments about your family in TMQ it makes it much more of a friendly publication y P I would like to congratulate you on PRO MRAS I never thought that writing in assembly could be so much fun I used to use ALDS but I got tired of it s limitations i e DS s of 255 bytes or less per line among other things I have a quick BUG REPORT for you on PRO MRAS version 1 0a The following statements do NOT report an error ETX EQU 03 UNDRLN DC 80 ETX The code generated will be 80 underlines followed by a zero byte even though ETX is defined as 3 Using this with DSPLY really had me going I couldn t figure out why I was getting garbage after it printed the string Fm MISOSYS It is probably wrong of MRAS not to complain about the syntax of your statement
288. with FORMAT label Announcing New Products THE MISOSYS QUARTERLY SPRING 1987 5 8 Volume I 1v 100 format 15a1 C Resulting FORTRAN code 23003 continue ierr l write MDL for range Sanivation 0 to end read APAJA ierr 2 2 continue 23004 if not ierr eq 2 g0te 23003 23005 continue Gea 23002 continue i1err l write Input file name read 100 err 3 fname lerr 2 3 continue 23003 if not ierr eq ESE 23002 23004 continue 100 format 15al1 Macros can call other macros or even themselves Macros which call themselves are called recursive macros These can result in very powerful new keywords which are limited only by the imagination of the user MISOSYS publishes two versions of RATFOR One executes on MS DOS machines and one executes on the TRS 80 Model 4 under LS DOS TRSDOS Version 6 The two versions share a RATFOR manual of about 86 _ pages including an extensive index An implementation specific installation manual concerning hardware and workstation issues is available with each version the advantages and disadvantages of the FORTRAN compilers available are also discussed candidly in the installation manuals All features of RATFOR are described in stand alone reference form in the manual A shorter narrative of RATFOR features is also provided there The MS DOS version RATFOR translator is provided as an EXE file so that it can be used with all MS DOS FORTRAN compilers or in a work
289. wo values have terminated the string CR and ETX Although passing a NULL actually does something on the Model 4 implementation I believe that that operation is machine dependent i e a Model 4 feature and not part of the DOS per se Another reason is that any change such as that would require additional memory for its implementation and thus would be out of the question Personally I don t think that DSPLY should ignore the next character in the string after a NULL I also don t think anyone is taking advantage of the lower 31 characters 32 on a Model 4P since a NULL NULL will print a character as there is a character in the 00 position of a 4P s character generator for display purposes through dsply Maybe through GDSP though So who is supposed to document it Tandy or LSI Fm Bryan Headley Tandy did document it in the Tech Manual Over in the appendices which deal with the character set it addresses the business about character 9 Machine dependency Although I still laugh at the circumstance that brought the question to the head Fm LDOS Support jjkd In my opinion the only problem is that SET accepts a single character devicespec If you fix that the fact that DEVICE hoses up on a single character spec may or may not be considered necessary to fix because you ll plug the only documented hole that could result in a single character devicespec Of course a program that mucks with the device table directly
290. years For example to get 1988 you lose 1980 Please don t get involved LSI patches to LS DOS 6 3 MAKE63K JCL JCL Procedure to create a LS DOS 6 3 0 Level K from a level I or J NOTE there are two version of BOOT3 FIX Use the correct one for the FA level that you have NOTE Destination to be patched and Source of patch file must be given on JCL command line ie DO MAKE63K JCL D d S d lt ENTER gt Replace the d in example with correct drive numbers The LSI LS DOS Column THE MISOSYS QUARTERLY SPRING 1987 45 Volume I iv with this mess if you don t have to There was an error in the patches in our last column Don t panic I think everyone just fixed it on the fly without giving it a second thought What we did was to give the same name to two of the patch files Both patch 002 and 009 were given DCOPY1 FIX and 009 should have been DCOPY2 FIX as it was referenced in MAKE63I JCL Sorry about the oversight If you didn t notice it at the time you should go back and verify that patches 002 and 009 are both installed There are four new patches All are very minor but a new level has been assigned to keep things orderly The fifth patch is just to change the level designator in the boot banner Remember that different level letters are assigned according to whether you put in the patches yourself or LSI does it by recutting your master All of these patches are very
291. you need more that four additional functions As far as I know offhand the other four bits of B are not used for anything but without having Bill here to ask I don t have a practical way to verify this He told me something about that once but I have forgotten It seems like there was some reason for not using the top four bits but I can t imagine what it would be now GTMOD looks like it was designed for just that kind of stuff Wouldn t it have been nice if Basic s VARPTR returned two pointers like that Of course the very nature of a function precluded that Fm LDOS Support jjkd 1 Virgil s suggestion is a good one though I would probably try to use subfunctions of a single function perhaps DCINIT rather than grabbing all remaining functions 2 Access to module header data via GIMOD is certainly kosher The only problem is finding the documented structure of the header Most LS DOS Information Volume I iv are documented in Roy s book and can also be found in THE SOURCE Fm Les Mikesell Are you planning to write more than one program that would use the new disk functions If not why not just imbed the functions in the program instead of using an SVC linkage What kind of new functions do you have in mind Obviously nothing in the DOS is going to use them Fm Adam Rubin Thanks for the information Let s see if I use function values above 15 that means there s no SVC for them and so I
292. yte sectors are related since the program LS DOS Information Volume I iv I m working on a looong term project will copy files to from MS DOS disks To read 512 byte sectors I followed the method LS DOS uses for writing sectors i e loop until NMI and haven t noticed any problems That s why I wonder about the DOS s method I admit reading 1024 byte sectors into a 256 byte buffer isn t a good idea Once I had RDSEC working well enough the only additions needed for RDHDR and RDTRK were the appropriate FDC commands in place of read sector In the format MS DOS disk section I m using RDHDR to determine if the disk is already formatted and thought that the multiple sector read and write commands would be a quick way to fill all the sectors with OF6H and then verify them I can t write OF6H using WRTRK and figured that the sector layout of SS MS DOS disks every track is 1 2 3 9 is that no track skew no interleave would otherwise require 18 revolutions to read and verify all the sectors on a track I realize the multiple sector commands lose bytes but that doesn t matter in this case does it else in mind or RDTRK but figured it would be easier to implement the whole thing as a disk driver in case I ever wanted to use them in the future What exists now is the MSfile program installs and removes my driver each time the program is run so my program can use RDSEC regardless of which drive is b

TMQ Vol I.iv - tim

Contents

Download Pdf Manuals

Related Search

Related Contents