Info-IBMPC@USC-ISIB.ARPA (Info-IBMPC Digest) (03/13/86)
Info-IBMPC Digest Wednesday, March 12, 1986 Volume 5 : Issue 33 This Week's Editor: Phyllis O'Neil Today's Topics: Accelerator Boards and Compaq Portable Limits on Files Open Fast Turbo Screen Writing Grammatik and Rightwriter BRIEF Arity Prolog Sector Size Newsroom Program German DOS and SORT .OBJ File Format EXEPACK Breaks Programs NEC V30 vs V20 in Compaq Deskpro Device Drivers and Sector Size MAX Availability RE: Queuing Simulation on IBM PC/XT/AT SAVVY PC Turbo Pascal Clearing the Screen Microsoft C compiler and strings Microsoft C V 3.0 and Phoenix Linker PLINK86 V 1.30 MS-C Re: Microsoft "C" Compiler Wish List Today's Queries: UCSD PASCAL on IBM-PC Query Overlay linkers Query 8087 Library for DeSmet C Compiler Query BRIEF Programmer's Editor--Macros Request What are Those Error Codes? 25 Top Utilities Request Looking for Paradox Database Users Irma Boards Query Logitech Modula-2 Compiler Review Requested Turbo Pascal Query DESQview Query HP Vectra vs IBM PC/AT Jove Query Graphics-Cursor on EGA Query AT&T 6300 Hard Disk Query Installing Xenix on Compaq Query ------------------------------------------------------------ Date: Wed, 12 Mar 86 18:41:39 EST From: Walter Bender <walter@MEDIA-LAB.MIT.EDU> To: info-ibmpc@usc-isib.ARPA Subject: RE: Accelerator Boards and Compaq Portable Cc: blue@MEDIA-LAB.MIT.EDU The PC-elevATor 286 from Applied Reasoning Corp. works in Compaq portables, DeskPros, and even PC-ATs. It comes in 8, 10, and 12.5 MHz models, with optional 5 or 8 MHz 80287 co-processor, and up to 2 MB of no wait-state RAM. Their software includes RAM Disk and EMS / Above Board support. Applied Reasoning Corporation 765 Concord Avenue Cambridge, Mass. 02138 (617) 492-0700 Yr. obt. svt., P.D.Pathe Javelin Software Corp. (blue@media-lab.mit.edu) ------------------------------ Date: Mon, 10 Mar 86 15:10:45 pst From: psivax!friesen (Stanley Friesen) Subject: Limits on Files Open Organization: Pacesetter Systems Inc., Sylmar, CA James H Coombs writes: >1) The DOS manual states that no program can have more than 20 files >open concurrently. This includes such files as stdio and stderr. Using >the C fopen() function, I bomb out on the 16th file. The manual is correct as far as *DOS* is concerned. The limit with fopen() is internal to *stdio*. The standard implementation of the stdio library only allocates 16 FILE structures for use by the library, so even though DOS permits more, stdio doesn't. Use open() or its equivalent to get at the DOS file handles instead of fopen(). Or else recompile the stdio library(if you have the source) to use 20 FILE structures. > >2) Some people have said that the limit is not 20 files CONCURRENTLY but >20 files during the execution of a program. As far as I know there is no sequential limit at all. Sarima (Stanley Friesen) UUCP: {ttidca|ihnp4|sdcrdcf|quad1|nrcvax|bellcore|logico}!psivax!friesen ARPA: ttidca!psivax!friesen@rand-unix.arpa ------------------------------ Date: Mon, 10 Mar 1986 21:06 CST From: Dean Hoyt <NU068450%NDSUVM1.BITNET@WISCVM.WISC.EDU> Subject: Fast Turbo Screen Writing To: <info-ibmpc@USC-ISIB.ARPA> >From: Bard Bloom <BARD@XX.LCS.MIT.EDU> >Subject: Generic Fast Screen I/O Query >How does one write quickly to the IBM PC's screen from Turbo Pascal? >I have a generic MSDOS version of Turbo, running on a very incompatible >Rainbow; I'm trying to write a program that will run on IBMs as well. >I also need to know how to set video attributes, at least underlined, >bold, and reverse video. > Thanks, I work on Zenith PC's that the school has. The memory for the color screen is at $8000:$F000. The memory for the mono screen is at $8000:$0000. I define an array of integer that is the size of the screen memory and then write into this array. This method bypasses both DOS and BIOS. Type a_screen = array [1..25,1..80] of integer; Var color_screen : a_screen absolute $8000:$F000; mono_screen : a_screen absolute $8000:$0000; temp_screen : a_screen; begin { code to build picture or text in temp screen } color_screen = temp_screen; { copy temp screen onto real screen } { more code } temp_screen = color_screen; { save current contents of screen } { display help screen } color_screen = temp_screen; { restore screen } end; As for attributes I only have experience with a color screen. As you can see the screen array is of type integer. The upper byte of an integer contains attribute bits, the lower byte contains the ascii character to display. Blink Back color Bright Fore color ASCII char ---------------------------------------------------- l x l x x x l x l x x x l x x x x x x x x l ---------------------------------------------------- Hi byte low byte I would just play around with the attribute bits and see what comes up on the screen. oh in the array it is Y,X so a gotoxy(15,10); write ('A'); would place an A in the lower byte of the screen array at location 10,15 so to get it you could use the statement, Value := color_screen[10,15]; I hope that this helps. I have found that with this type of screen I/O you can not see the writes to the screen, the picture just appears. Dean Hoyt NU068450 @ NDSUVM1 ------------------------------ Date: 11 Mar 1986 01:34:23 EST Subject: Grammatik and Rightwriter From: Norman Lyons <LYONSN@USC-ISI.ARPA> To: info-ibmpc@USC-ISIB.ARPA I use both Rightwriter and Grammatik and I thought I would comment on the capabilities of the two packages. I read about 30 Master's theses a year, and all theses are submitted to these two packages first. Working together, they handle almost all of the nit-picking editing chores. The two packages are actually complementary. Grammatik flags wordy phrases, archaic phrases, and other such problems. It has an extensive dictionary of its own, and the user can easily create his own phrase dictionary using an ASCII editor. Two passes are required with Grammatik. One pass flags problems in the document itself. Another pass prints out suggestions for changes separately from the document. Rightwriter is a much more sophisticated package. It examines sentence structure, flags passive voice, long sentences, complex sentences, and some types of grammatical errors. It also computes a fog index and a list of unusual words. It's a very useful package for streamlining prose. One has to be careful about its suggestions, however. If taken too literally, it can result in a "me Tarzan--you Jane" style of prose. It "rewards" short sentences and active voice with a low fog index. Since I use both of these programs together, I have written a pair of C routines and some BAT files that work both Grammatik and Rightwriter comments into the document in a single format. These programs can run in unattended mode. It requires about 20 minutes to do a complete check on a 100-page thesis. I can contribute these programs to INFO-IBMPC if anybody is interested. Disclaimer: I have no interest in either of these companies except as a reasonably satisfied user. ------------------------------ From: chapman%miro@BERKELEY.EDU (Brent Chapman) To: info-ibmpc@usc-isib.arpa Subject: "ENVSIZE.EXE" (included with BRIEF editor) patches COMMAND.COM Date: 10 Mar 86 23:53:44 PST (Mon) The program "ENVSIZE.EXE" that is included with the BRIEF Programmer's Editor successfully increases the environment size, but it does it by patching your COMMAND.COM file. Now, while in and of itself, this is not necessarily a bad thing, this little "feature" is not noticeably (if at all) mentioned in any of the documentation... I don't know about you, but I prefer to know in advance when something is going to go poking around in my critical files... Just thought I'd warn everyone else. Brent Chapman chapman@miro.berkeley.edu ucbvax!miro!chapman ------------------------------ Date: 11 Mar 86 09:34:00 EST From: "DESILVA, ERIC" <desilva@nbs-vms.ARPA> Subjectt: Arity Prolog To: "info-ibmpc" <info-ibmpc@usc-isib.arpa> I was catching up on my mail and happened to notice a message about Arity Prolog. We just got version 4.0 and it is no longer copy pro- tected. For the $800, you also get the Arity Prolog Compiler which runs about 10-20 times as fast as the interpreted stuff (Can Borland beat that?). There is a review of a bunch of Prolog interpreters in the Jan '86 PC Tech Journal for more info and benchmarks (LIPS Logical inferences per second !?!) ------------------------------ Date: Tue, 11 Mar 86 13:30:02 PST From: Dana Myers <bilbo.dana@LOCUS.UCLA.EDU> To: Schauble@MIT-MULTICS.ARPA CC: INFO-IBMPC@USC-ISIB.ARPA Subject: Sector Size Too Large MS-DOS determines the size of the largest sector it will have to deal with during boot time by examing the sector sizes in use by the disk driver in IO.SYS. For most PCs, this is 512 bytes. Once DOS figures this out, it creates the buffer pool which consists of a number of structures fixed in size, the size being the largest sector size plus a small amount for a header. Unfortunately, if you want a larger sector size you have to do something tricky, like modify your IO.SYS source (you do have source to IO.SYS, don't you?) or mess with IO.SYS somehow to make DOS believe a larger sector is in use (don't ask exactly how, it qualifies as black magic). You could also block and deblock the 2k sectors into 512 byte sectors... Note that having a large sector size (which means large buffers) will reduce the amount of memory available to applications for a given number of buffers. 25 buffers is a reasonable number; this is about 13k of 512 byte sectors but more like 53k of 2k byte sectors... ------------------------------ Date: Tue, 11 Mar 86 13:44:54 EST From: mr <ihnp4!inuxc!isrnix!mr@seismo.CSS.GOV> To: info-ibmpc@USC-ISIB.ARPA Subject: Newsroom Program > I would like to get some info on programs that can be used for creating newsletters or other in-house magazines. Something which uses the Hercules Graphics Card and displays the exact stuff that will be printed (What You See Is What You Get). Ideally I should be able to get a draft printout on an Epson FX printer and the final version on the Apple Laser Printer. Any information on this will be appreciated. Irfan Ahmad ECO10125 BOSTONU > You might check into Springboard's "Newsroom" program ($25). With the program, you get some 200-odd pieces of "clip art" that can be modified to your liking. (sure, the clip art is a little hokey, but you can get either volume 1 or volume 2 of their more sophisticated clip-art. add $19. volume 2 is the business clip art which allows the design and production of many, many graphs, charts, logos, piecharts, you name it.) Either volume touts over 500 pieces of clip art. With a color/graphics monitor, you're able to design your own "photos" which, if you like, can be laid out pixel by pixel; anywhere on your page. There's 6 "departments" in the newsroom: the photo lab where you design your own "photos;" the printing press; the wire service where pc's running "newsroom" can talk to one another via modem; banner department where the masthead is created; the copydesk where text is written; and the layout where it all comes together. mice/joysticks are welcome but not necessary. It is fully compatible with most dot-matrix printers (esp. with epson's) and what you see, is what gets printed (wysiwgp!). .^. michael regoli /|\ ...ihnp4!inuxc!iuvax!isrnix!mr '|!|` <mr@isrnix.UUCP> ------------------------------ Date: Tue, 11 Mar 86 13:52:32 PST From: kegel%tybalt.caltech.edu@Hamlet.Caltech.Edu To: m70b%cbebda3t.bitnet@wiscvm.wisc.edu Subject: German DOS and SORT > How do I get SORT to sort letters with umlauts correctly? In an earlier letter, I reasoned that SORT used the DOS facility for building a character sorting weight table. As it happens, the SORT that comes with DOS 3.1 (and probably DOS 3.0, as well) contains its own sorting weight table, which will have to be modified. The new sorting weight table will (like the old one) be the identity function for characters below 65 (upper-case A). Since I don't know in what order you want to put the various accents, I can't give you an exact table; but here's how to construct one: Write down the alphabet (including accented letters) in the desired order. Number each letter, starting with 65 for A and a (both cases get same number). The number next to each letter is its sorting weight table entry. Now sort the list by ASCII code number, and fill gaps (which mostly hold block graphics characters) with numbers higher than the weight generated for Z. You should have 191 entries, one for each code from 65 to 255. Finally, convert the weights into hexadecimal (ugh). Here's how to use DEBUG to find and change SORT's weight table (this will only be understandable if you already know DEBUG): 1. Make a copy of SORT.EXE under a different extension, e.g. copy sort.exe sort.bin 2. Locate the case map table within sort.bin: DEBUG SORT.BIN -RCX ; How long is the file? 0680 ; Now we know how long to search -S 100 L 0680 0 1 2 3 4 5 ; Search for start of case map table xxxx:0592 ; Now we know offset of table -H 592 41 ; Find where letter A starts sum=05d3 difference=0551 ; it starts at 5d3 -Q ; back to DOS (The above numbers are examples) 3. Make a file called Script with your 192-entry hex weight table as follows. (Let's say the first 4 entries of your table are 41h, 48h, 49h, 50h, and the last 4 entries are a3h, a4h, a5h, a6h.) f 5D3 L F 41 48 49 50 etc.... f 5E2 L 10 ... ... f 682 L 10 .... a3 a4 a5 a6 w q 4. Run debug on sort.bin with Script: debug sort.bin < Script 5. Copy sort.bin to newsort.exe, and try it out. 6. If it does not sort correctly, change the file Script, and go back to step 4. Again, good luck. - Daniel Kegel ------------------------------ Date: Tue, 11 Mar 86 14:12:44 PST From: kegel%Juliet.Caltech.Edu@Hamlet.Caltech.Edu Subject: .OBJ File Format To: joe.newcomer@a.sei.cmu.edu, info-ibmpc@usc-isib.arpa The Microsoft .obj file format is pretty obscure, and is slightly different from the Intel .obj file format (a clean subset, I think). I was able to write a .obj file writer by intensive study of the issue of IBM PC Tech Ref Journal (?October 1985?) that explained the format, and by examinging the output of very, very simple MASM runs. The only thing the article doesnt mention is that, although indices usually are coded into 1 byte, the high bit is a "next byte is low 8 bits" flag, allowing more than 255 external symbols (for instance). There is also a document available mail-order from Microsoft, the MS-DOS Programmer's Reference, which (i am told) also describes the format. - Daniel Kegel ------------------------------ Date: Tue, 11 Mar 86 16:31:58 PST From: Matthew J Weinstein <matt@LOCUS.UCLA.EDU> To: info-ibmpc@isib Subject: EXEPACK Breaks Programs EXEPACK v 4.00 and previous are known to break programs. No specific causes or solutions. EXEPACK broke a 220k program in obscure ways the other day... MS has no new version at this point; no word if they are working on it. The /E option of LINK is known to break programs too. I don't use it. - Matt ------- UUCP: {ucbvax,ihnp4,randvax,trwrb!trwspp,ism780}!ucla-cs!matt ARPA: matt@LOCUS.UCLA.EDU ------------------------------ Date: Tue, 11-Mar-86 08:34:56 PST From: scgvaxd!pelican!pete@csvax.caltech.edu (Pete Carah) Subject: NEC V30 vs V20 in Compaq Deskpro To: marc%uofm-uts.cdn%ubc.csnet@csnet-relay.arpa A friend and I have put a V30 in a deskpro, and find that the time to compile and assemble a large ms-pascal/assembly program was reduced from 25 to 19 minutes, where the time in a PC went from 38 to 37 minutes when a V20 was installed. In other words, the V30 improves a deskpro more than a V20 improves a PC, at least for MS-pascal and the assembler. -- Pete [I suspect the NEC V20 is choked by the 8 bit data path and spends more time waiting for memory -wab] {scgvaxd, ihnp4, vortex}!pelican!pete ------------------------------ Date: Tue, 11-Mar-86 08:58:06 PST From: scgvaxd!pelican!pete@csvax.caltech.edu (Pete Carah) Subject: Device Drivers and Sector Size To: schauble@mit-multics.arpa Cc: info-ibmpc@usc-isib.arpa DOS gets its idea of sector size for buffer allocation from the BPB of the first block device driver loaded. Since this is the one that comes with DOS, you are kind of stuck without a patch. Several people manage to get around this to increase the allowable disk size over 32MB. It should be possible to sneak around in the DOS.COM data area and fix the problem at the time of loading a driver, but I don't know what side effects there might be, and code to do it would not work on another version of DOS. One side effect of having a buffer size over 512 bytes is that the buffer area uses a lot of memory and (in your case) 3/4 of it wouldn't get used while floppy or normal hard disk sectors were bufferred. Another way to handle the problem is to translate all requests to/from "virtual" 512-byte sectors in your driver. That isn't too hard but it would require an internal buffer to handle partial sector requests, and if the cluster size were smaller than 2K you would get some thrashing. This scheme has the disadvantage of not increasing the allowable device size over 32MB, though you can always allow more than one DOS partition (my disk driver does, and allows up to 128 MB in a drive, in 4 virtual devices). -- Pete {scgvaxd, ihnp4, vortex}!pelican!pete ------------------------------ Date: Tue, 11 Mar 86 23:56:17 PST From: larry@Jpl-VLSI.ARPA Subject: MAX Availability To: info-ibmpc@isib.arpa In answer to several queries, I got my copy of MAX (a microEMACS) from UTEXAS-20.ARPA from <CC.GALVIN.PUBLIC>MAX.LBR... Larry @ jpl-vlsi ------------------------------ To: info-ibmpc@usc-isib.ARPA Cc: tektronix!stever%tektronix.csnet@CSNET-RELAY.ARPA Subject: RE: Queuing Simulation on IBM PC/XT/AT Date: 10 Mar 86 19:08:52 PST (Mon) From: stever%tektronix.csnet@CSNET-RELAY.ARPA consider evaluating SIMAN, it is a cross between GPSS and GASP/SLAM. Systems MOdeling Corporation Calder Square P.O. Box 10074 State College, Pennsylvania 16805-0074 (814)-238-5919 ------------------------------ Date: Wed, 12 Mar 86 08:24:05 est From: Scott Guthery <sguthery%slb-doll.csnet@CSNET-RELAY.ARPA> To: info-ibmpc@usc-isib.ARPA Subject: SAVVY PC I use SAVVY PC for a coupld of little databases I ride herd on and find it quite useful. It has a funky model of computation (tasks kinda) and the AI component is very A but all in all for $49 I'm a happy customer. I also wonder why you'd spend more. There's nothing I've wanted to do that SAVVY PC can't handle (probably a comment on my imagination or needs). It allows full-fledged programming which means you can simulate a Turning machine if you have to. I recommend you pay the extra $30 and get the hardcopy documentation cause there's lots of it and you'd waste at least $0 worth of your own time and paper printing it all out. ------------------------------ Date: Tue 11 Mar 86 21:14:14-EST From: Doctor I & III & V <WESALUM.A-LIAO-85@KLA.WESLYN> Subject: Turbo Pascal To: INFO-IBMPC%USC-ISIB.ARPA@VAX.WESLYN About Turbo Pascal: As far as I know the 64K program space limitation is still in effect on Version 3.xx. This is not true, however, for dynamically allocating variables off the heap. He and I managed to allocate something close to 200K on his machine (he has a 256K IBM PC). Warning to all, however, about using Turbo's version of linked lists: If you're not careful you, referencing a NIL pointer can be REAL dangerous. It turns out that, at best, reading data via a NIL pointer will only return garbage. However, writing to a "variable" will hang your machine (and possibly worse). It turns out that Turbo sets the NIL pointer as an address in REAL LOW Memory (i.e. the DOS area). So writing thru a TURBO NIL pointer down in that area isn't really a recommended thing to do. In fact, Turbo really doesn't tell you if you try to dereference a NIL pointer. Fortunately, I talked to Borland and they said they finally fixed the problem (I'll believe it when I see it). "NIL", they said, "doesn't do that nonsense anymore". This is as of 6 weeks ago. So I assume a new version of Turbo Pascal is out on the market. Hope this info helps.... - drew liao ------------------------------ Date: Tue 11 Mar 86 21:25:37-EST From: Doctor I & III & V <WESALUM.A-LIAO-85@KLA.WESLYN> Subject: Clearing the Screen To: INFO-IBMPC%USC-ISIB.ARPA@VAX.WESLYN RE: To GALVIN@R20.UTEXAS,JAZBO%BROWN The following is the code to clear the screen and leave the cursor in the top left-hand corner: ;The following was written to interface with MS-Pascal. I wrote this ;code for an IBM monocrhome monitor. It makes use of the DOS 2.00 ;BIOS routines. PLEASE NOTE THE DOS Version! ; ;HOMCLR Clears the screen of garbage and moves cursor to top left ; corner of the screen, and leaves the cursor there. ; Called from MSPascal as: PROCEDURE HomClr;EXTERN; HOMCLR PROC FAR PUSH BP ;Set for frame pointer MOV BP,SP XOR CX,CX ;Top left corner of screen MOV DX,24*256+79 ;Bottom right corner of screen XOR AL,AL ;We want all lines moved up MOV BH,7 ;Want white on black attribute MOV AH,6 ;Select scroll screen up (CLR) INT 10H ;Go do it XOR DX,DX ;Where do we want cursor? XOR BH,BH ;Current active screen MOV AH,2 ;Select move cursor function (HOM) INT 10H ;Go do it POP BP ;Restore original base ptr RET ;Return to MSPASCAL program address HOMCLR ENDP Hope it helps. - drew liao ------- ------------------------------ Date: Tue, 11 Mar 86 13:18:39 PST From: Dana Myers <bilbo.dana@LOCUS.UCLA.EDU> To: Ya-akov_Miles%UBC.MAILNET@MIT-MULTICS.ARPA CC: INFO-IBMPC@USC-ISIB.ARPA Subject: Microsoft C compiler and strings On page 181 of "The C Programming Language" by K and R, in section 2.5, it is stated that "All strings, even when written identically, are distinct.". KR offered the only standard for C compilers for a long time and a lot of programs exist which make assumptions about the uniqueness of strings. Even though ANSI is hard at work on a standard for C compilers, etc., I don't think this can change. AT one time I used Lattice C a lot, and it would perform constant folding of strings, which made some programs non-portable to Lattice C. I don't know if the newest versions of lattice C still do this. Dana ------------------------------ Date: Tue, 11 Mar 1986 18:59 EST From: LENOIL@XX.LCS.MIT.EDU To: IU%DACTH51.BITNET@WISCVM.WISC.EDU Cc: info-ibmpc@USC-ISIB.ARPA Subject: Microsoft C V 3.0 and Phoenix Linker PLINK86 V 1.30 > Two friends of mine are working on a comprehensive word processor using the above C compiler from Microsoft. Due to the size they have to split the program into overlays and thus have to use PLINK V 1.30. > I'm fairly certain that the newest LINK from Microsoft supports overlays. ------------------------------ Date: 12 MAR 86 11:39-AST From: IU%DACTH51.BITNET@WISCVM.WISC.EDU To: INFO-IBMPC@USC-ISIB.ARPA Subject: MS-C Re: My question about the Microsoft C compiler 3.0 and the Phoenix Linker 1.30 in INFO-IBMPC 5 31 I forgot to mention that the reason to use PLINK86 is the need for hierarchical overlays, which the Microsoft linker doesn't support. Sorry, el (Eberhard W. Lisse) ------------------------------ Date: Wed, 12 Mar 86 17:03:14 PST From: Jim Carter <jimc@LOCUS.UCLA.EDU> To: info-ibmpc@usc-isib.arpa Subject: Re: Microsoft "C" Compiler Wish List In Info-IBMPC Digest v5 vol32, Ed Nather writes about folding string literals. The current draft standard for "C", X3J11/85-138, says in sec. C.1.4, that different copies of the same string literal need not be distinct, and that string literals may not be modified. James F. Carter (213) 206-1306 UCLA-SEASnet; 2567 Boelter Hall; 405 Hilgard Ave.; Los Angeles, CA 90024 UUCP:...!{ihnp4,ucbvax,{hao!cepu}}!ucla-cs!jimc ARPA:jimc@locus.UCLA.EDU ------------------------------ Date: 10 Mar 86 21:09:56 +0100 From: XBR1YD22%DDATHD21.BITNET@WISCVM.WISC.EDU (YD22@BR1.THDNET) Subject: UCSD PASCAL on IBM-PC Query To: info-ibmpc@usc-isib Hi folks, I have an application program written in UCSD-PASCAL and currently running on an Apple ][ computer. I'd like to transfer the program to an IBM-PC or compatible computer. So I am looking for a PASCAL compiler that is similar to the UCSD PASCAL and runs on the PC. I'm not very keen on 100 % compa- tibility, but I'd like to get e.g. the string built in functions/procedures and the like. Please respond to me, so we don't clog up the net. If there's any interest, I'll summarize to the network. Thank you very much Ralf Bayer Computing Center @ the Technical University of Darmstadt, West Germany Arpanet address: xbr1yd22%ddathd21.BITNET@wiscvm.wisc.EDU (personal mail) BITNET address: xbr1yd22 @ ddathd21 (no NETDATA format) #d22 @ ddathd21 ( NETDATA format, beware of the Number sign, it's part of the USERID) ------------------------------ Date: Mon, 10 Mar 86 22:10:52 EST From: James H. Coombs <JAZBO%BROWNVM.BITNET@WISCVM.WISC.EDU> To: INFO-IBMPC@USC-ISIB.ARPA Subject: Overlay linkers Query Could someone give a summary (or a reference to one) of overlay linkers: what they are, why use them, and which ones are good/bad? I have an application that could consist of 10+ separate programs, each loaded at the appropriate time by a system call. Unfortunately, four of the routines attempt to load as much data into RAM as possible and then have to "page" out to disk. If I can avoid giving up 60K or so to the driver, then I will have that much more room for data. Also, one of the programs uses many of the same routines as the driver, and I would hate to give up RAM to duplicate versions of some fairly complex routines. The answer seems to be to use an overlay linker. Are there linkers reliable enough to sacrifice the modularity of separate programs, or am I better off giving up the RAM? Thanks. --Jim ------------------------------ Date: Mon 10 Mar 86 22:23:56-PST From: D.LEWIS%SCU%PANDA@SUMEX-AIM.ARPA Subject: 8087 Library for DeSmet C Compiler Query To: info-ibmpc@USC-ISIB.ARPA I've noticed that the 8087 run-time library for the DeSmet C compiler occassionally places *TWO* (as opposed to one) WAIT instruction directly in front of an 8087 instruction, even some that only reference the 8087 internal stack. Does anyone know why more than 1 WAIT might be needed? Also, does anyone know the code fragment that will determine if an 8087 is installed? Thanks in advance! Dan Lewis Dept. of EECS Univ. of Santa Clara Santa Clara, CA 95053 ------------------------------ From: chapman%miro@BERKELEY.EDU (Brent Chapman) To: info-ibmpc@usc-isib.arpa Subject: BRIEF Programmer's Editor--Macros Request Date: 10 Mar 86 23:44:54 PST (Mon) I've just started using BRIEF lately, and absolutely love it. I was wondering if any other users of the system might have some useful macros they'd be willing to share? If you'll send them to me, after the smoke clears, I'll summarize them all in one post to Info-IBMPC. Send anything that YOU find useful; you never know when someone else might actually agree with you... Brent Chapman chapman@miro.berkeley.edu ucbvax!miro!chapman ------------------------------ Date: 11 Mar 86 07:53 GMT From: meaders @ KOREA-EMH Subject: What are Those Error Codes? To: info-ibmpc @ usc-isib.arpa Can someone fill me in on what a parity check 68000 (s) is? Same 60000, 48000, 64000, 48000 etc. Thanks in advance Mark Meaders "From the Land of the Morning Calm..... and Mid Afternoon Madness" ------------------------------ Date: Mon, 10 Mar 86 22:44:16 est From: ucdavis!lll-crg!seismo!cmcl2!phri!greenber@ucbvax.berkeley.edu (Ross Greenberg) Subject: 25 Top Utilities Request Organization: Public Health Research Inst. (NY, NY) As part of a book I've been asked to write, I'm looking for the top 25 Utility programs for the IBM-PC. Everybody as their favorites, I'm sure, and so I'd like to ask you to forward to me a list of *your* favorites. Both Public Domain, and Commercial programs are fine. If you could give a (very) brief description of what they do, and where I can get them, I'd appreciate it immensely. Please mail them to me, and I'll eventually summarize, analyze and whateverels-irize. Thanks! ------ ross m. greenberg ihnp4!allegra!phri!sysdes!greenber [phri rarely makes a guest-account user a spokesperson. Especially not me.] ------------------------------ Date: Tuesday, 11 March 1986 10:25:07 EST From: Chuck.Weinstock@a.sei.cmu.edu To: info-ibmpc@usc-isib.arpa Subject: Looking for Paradox Database Users Are there any other users of Paradox out there? This is the best database system that I've found for the PC. I'd like to talk to other users to share experiences. Chuck ------------------------------ Date: Tue Mar 11 11:53:39 1986 From: jperry@sri-unix To: info-ibmpc@usc-isib Subject: Irma Boards Query Does anyone out there have experience using Irma boards to emulate IBM terminals when connected to an IBM mainframe? Does anyone out there know what the known software/hardware problems are with this board and whether the error rates are high/low/medium? Any and all experiences/ rumors/innuendos about this product will be appreciated. I am doing research for a group which may purchase many such boards and since they are about $1200 a pop, they want to know that they are spending their money wisely. John Perry SRI International ------------------------------ Date: 11 Mar 86 14:50:00 EST From: "INFO1::ELDER" <elder%info1.decnet@wpafb-info2.ARPA> Subject: Logitech Modula-2 Compiler Review Requested To: "info-ibmpc" <info-ibmpc@usc-isib.ARPA> Could anyone who has used the Logitech Modula-2 compiler provide a review. Is it a complete implementation of Modula-2? I heard the compiler is now selling for $89 so I am interested. Thanks in advance. Greg Elder elder@wpafb-info1 ------ ------------------------------ From: "JAMES P FERRILL" <ferrill@eglin-vax> Subject: Turbo Pascal Query To: "info-ibmpc" <info-ibmpc@usc-isib> I have been working on a program that writes output to the printer using the standard Writeln routine. In working with this program which does a good bit of screen formatting I discovered that if the printer was off-line when the Writeln(Lst,foobar) statement came along I got the familiar DOS message Printer Fault. Abort, Retry, Ignore or something similar. I then tried to use the {$I-} Writeln(Lst,foobar) {$I+}; construct but still get the DOS error message on the screen. Is there a way to turn this message off or at least catch it? It sure does mess up my screen that has been formatted for screen I/O. Any help would be appreciated. Paul Ferrill ferrill@eglin-vax.arpa ------ ------------------------------ Date: Mon, 10-Mar-86 19:09:13 EDT From: David Farber <farber%pcpond.pc.udel.edu@Louie.UDEL.EDU> Subject: DESQview Query To: ud-ibmpc%pcpond.pc.udel.edu@Louie.UDEL.EDU, info-ibmpc%usc-isib.arpa@louie.udel.EDU CC: lauren%rand-unix.arpa@louie.udel.EDU I have been having a ball running DEQview on my At with UULINK in the background and everything else (well almost) in the forground. Only serious pain I have seen and it may be just the sparse documentation is that when it starts up a partition it puts the path as it existed before DESQview got control but NOT the sets that were made. That is a BIG pain. Anyone know a trick. Much nicer than Double DOS ------------------------------ Date: Tue, 11 Mar 86 19:36:40 pst From: consult%ucbopal@BERKELEY.EDU (CS Consulting) To: info-ibmpc@usc-isib.arpa Subject: HP Vectra vs IBM PC/AT Cc: arch@renoir.berkeley.edu Arch Turner <arch%renoir@Berkeley.EDU> would like an answer to the following: #1- Do you know anything about the HP Vectra as an alternative to the IBM PC AT? Someone in Mexico City called, asking the question, and in a moment of weakness I promised to call him back. Thanks, Arch Turner, CSSG Staff 467 Evans, 2-1319 arch@renoir Bill Wells consult%opal@Berkeley.EDU ------------------------------ Date: Wed, 12 Mar 86 01:13:57 cst From: cmkuo@im4u.utexas.edu (Chin-Ming Kuo) To: info-ibmpc@usc-isib.arpa Cc: cmkuo@im4u.utexas.edu Subject: Jove Query Could you tell em where can I get a complete set of JOVE running on msdos machine ? I think object code is prefered cause I don't have too many tools. Thanks . C.M. Kuo ARPA:cmkuo@im4u.utexas.edu UUCP:....!ut-sally!im4u!cmkuo Ics.cmkuo@r20.utexas.edu ------------------------------ Date: 12 Mar 86 10:29:07 +0100 From: XBR1D36K%DDATHD21.BITNET@WISCVM.WISC.EDU (D36K@BR1.THDNET) Subject: Graphics-Cursor on EGA Query To: info-ibmpc@USC-ISIB.ARPA Does anybody know any way to produce a Graphics-Cursor on the EGA? What I need is a FORTAN-callable Routine. Is there any DOS-Interrupt, to handle this purpose? Gunther Hartmann Technical Univ. Darmstadt West-Germany Bitnet: D86V@DDATHD21.BITNET ------------------------------ ------------------------------ Date: Wed, 12 Mar 86 07:56:53 pst From: Gerry Key <key%tetra@nosc.ARPA> To: info-ibmpc@usc-isib.ARPA Subject: AT&T 6300 Hard Disk Query Anyone have any recommendations for a 30 Mb external hard disk (< 40 ms), controller, and streaming tape combo for either an AT&T 6300 or Xerox PC? --Gerry Key key@nosc.arpa ------------------------------ Date: Wed, 12 Mar 86 16:58:19 CST From: frank@db.wisc.edu (Dan Frank) Received: by db.wisc.edu; Wed, 12 Mar 86 16:58:19 CST To: info-ibmpc@usc-isib Subject: Installing Xenix on Compaq Query IBM Xenix has the nifty limitation that it only installs on 20 MB IBM hard disks. Does anyone know a way to install it on non-standard hard disks such as the 30 Mb. disk used in the Compaq? -- Dan Frank (frank@db.wisc.edu) -------