Info-IBMPC@USC-ISIB.ARPA (Info-IBMPC Digest) (03/18/86)
Info-IBMPC Digest Tuesday 18 March 1986 Volume 5 : Issue 35 This Week's Editor: Billy Brackenridge Today's Topics: Correction to Turbo Pascal Fast Screen Write UCSD Pascal from Pecan Software High Performance w/o an AT! Turbo Pascal Screen Clearing Nansi Bug Fix & Binaries Available ClickArt Personal Publisher from T/Maker Correction to Address of AK Technology Microsoft C Bugs Directories from Turbo Pascal Directories from Turbo Pascal & Wang Compatibility Wang Compatibility MASM 4.0 OFFSET Bug (2 Msgs) Microsoft Fortran & F77 (2 Msgs) NBI to PC-DOS (and CP/M) Connections Turbo Pascal Colors XT/370 Speedup XENIX Updates Todays Queries: MS-Windows Drivers Pascal to C Translator Symphony & MSWindows DisplayWrite 3 EGA Support FORMAT Equivalent Within a Program (2 Msgs) Zenith PCs at Govt Contract Prices Xenix Device Driver Document Indexing Program BRIEF vs EPSILON ---------------------------------------------------------------------- Date: Thu, 13 Mar 1986 15:26 CST From: Dean Hoyt <NU068450%NDSUVM1.BITNET@WISCVM.WISC.EDU> Subject: Correction to Turbo Pascal Fast Screen Write In the file that I sent you about writing to the screen quickly in turbo pascal I had the address wrong. They should have been for the mono screen $B000:$0000 for the color screen $B000:$8000 Please if possiable correct the file I sent you. If not then add a notice about this correction. Dean hoyt Bitnet address NU068450@NDSUVM1 Arpanet address NU068450%NDSUVM1.BITNET@WISCVM.WISC.EDU where the wind chased the hills away. [We didn't catch this in time. The local archives have been updated. -ed] ------------------------------ Date: Sat, 15 Mar 86 22:19 EST From: Yedidyah Langsam <YDLBC%CUNYVM.BITNET@WISCVM.WISC.EDU> Subject: UCSD Pascal from Pecan Software Pecan Software has released UCSD Pascal working under a DOS Hosted P-system. The compiler can produce either native code or P-code for the IBM PC. This DOS Hosted p-system seems to be fully compatible with DOS. It is a full implementation of Pascal. Since Pecan software simply developed (or re-released it) the DOS hosted version of the P-system without tampering with the compiler, it should be bug free and be compatible with APPLE UCSD Pascal. Pecan Software advertises heavily these days in Byte, PC WORLD and PC Magazine. I have no connection with them, other than having tried the system. The only incompatibility I have found is that it will not work when Borland Lightning is in memory. At $79. for the whole system, you can try it out.... ------------------------------ Date: Sat, 15 Mar 86 22:20:09 cet From: 10409813%WSUVM1.BITNET@WISCVM.WISC.EDU Subject: High Performance w/o an AT! For those of you who might care... I put a NEC V30 in my AT&T PC 6300 this weekend. The NORTON SysInfo rating went from a mediocre 1.9 to a rather impressive 3.9! Overall performance (not just the mul/div that SI tests) is up about 20-30% based on a DHRYSTONE benchmark. Results of the benchmark went from 833 to 1086 dhrystones/sec. This is about the same performance as a 6Mhz PC/AT! Considering that the 8087 runs at 8Mhz and therefore outperforms the PC/AT on Floating Point operations I am extremely glad that I didn't spend an extra 2000 dollars or so for a "better" machine. Also get DESQview and simply love downloading with KERMIT in the background while doing real work in the foreground. This works quite well, and should promise to increase productivity significantly. Eric Schneider 10409813 @ WSUVM1 . BITNET eric @ wsu . CSNET ------------------------------ Date: Sun, 16 Mar 86 12:39 EST From: MKATZ%UMDB.BITNET@WISCVM.WISC.EDU Subject: Turbo Pascal Screen Clearing From: Jameson@UNCA-MULTICS.MAILNET Does anyone know of a patch which will stop Turbo Pascal programs from clearing the screen before they run, and possibly from leaving the cursor at the bottom of the screen afterward? Many thanks. Turbo Pascal 3.0 no longer clears the screen at the beginning (2.0 did), but it still jumps to the bottom at the end. I suppose you could force it to exit without jumping to the bottom by closing all files and then calling DOS to exit instead of exiting normally through Turbo. Manasseh Katz ------------------------------ Date: Sun, 16 Mar 86 00:36:32 PST From: kegel%Juliet.Caltech.Edu@Hamlet.Caltech.Edu Subject: Nansi Bug Fix & Binaries Available To: cc.dlnash@a20.utexas.edu, info-ibmpc@usc-isib.arpa Two notes about Nansi: 1. Peter Lyman at JPL informs me that the Nansi source file nansi.asm had a TAB character in the device name field instead of five spaces; this makes the device name field the wrong length. The device loaded, but its name was CON\t%$#** or some such garbage; I think that INT 29 was installed properly, so it probably worked on single-character writes. the device chain for the device name, and would have fallen thru to the default CONsole driver! Talk about mysterious symptoms... The tab was inserted by an editor that thought it was doing me a favor, I suppose; anyway, this is a serious problem, and anyone who got the sources should change the line in nansi.asm that reads db 'CON ' ; (8 chars) to read db 'CON' ; (3 chars... db 5 dup (20h) ; ... and 5 blanks) to avoid future screwups. Somebody at ISIB should change their copy, too. (Congratulations to Peter for cracking this problem.) 2. I am posting a binary version of Nansi on my machine for people who don't have MASM. The three files are as follows: nansi_doc.cat - Documentation, examples, and Intel hex version of Nansi.sys. nansi_src.cat - Source code. nansi_demo.cat - Documentation and Intel hex code for view.com, a vi-like file viewer which demonstrates the full speed of nansi.sys. In case it matters, the files are on the local directory USERS:[ANONYMOUS]. You should log in using FTP to hamlet.caltech.edu with username anonymous and password ftp. (Hamlet and Juliet share file systems, but only Hamlet is supposed to be visible from Arpanet.) Finals-weekly yours, Daniel Kegel [NANSI.CAT has been updated. The tab character was there! -ed] ------------------------------ Date: Sat, 15 Mar 86 19:54:26 pst From: adobe!wff@decwrl (Ware and Freidenrich) Subject: ClickArt Personal Publisher from T/Maker A significantly more sophisticated (but not that much more expensive) program than Newsroom is ClickArt Personal Publisher from T/Maker Company in Mountain View, California. T/Maker is at 2115 Landings Drive, Mountain View, California 94043, (415) 962-0195. Personal Publisher is a WYSIWYG desktop publication tool that works with dot matrix, HP LaserJet, and Apple LaserWriter printers. T/Maker's popular ClickArt clip art files which have been reproduced on 5 1/4" media and can be read by Personal Publisher into a document. Carey Heckman ------------------------------ Date: Sun, 16 Mar 86 13:17:14 est From: Mark D. Freeman <mdf@ohio-state.ARPA> Message-Id: <8603161817.AA09854@ohio-state.ARPA> Subject: Correction to Address of AK Technology Corrected version: AK Technologies Corporation Columbus, Ohio (614) 764-0737 I do have financial ties to this company, but not in regard to CXI products. [AK Tech makes the IRMA replacement board and software for IRMA cards. -ed] ------------------------------ Date: Sun, 16 Mar 86 12:44:27 PST From: Matthew J Weinstein <matt@LOCUS.UCLA.EDU> Subject: Microsoft C Bugs It appears to be unsafe in general to declare a single 'register' variable in MS C large model in the presence of structure-valued functions. The following type of code will break for some values of STREXP[12]: struct any {...}; int IntValued(struct any); proc() { register i; ... if (IntValued(STREXP1) < IntValued(STREXP2)) { } } where STREXP[12] are structure valued expressions (such as StrucFunc() or (*(struct any *)"data")). Generally speaking, replacing the register declaration with an integer declaration fixes the problem. Microsoft has indicated that there will be no interim fix. - Matt ------------------------------ Date: Sun, 16 Mar 86 12:44 EST From: MKATZ%UMDB.BITNET@WISCVM.WISC.EDU Subject: Directories from Turbo Pascal To: CHADM1%UCONNVM.BITNET@WISCVM.WISC.EDU There is a Turbo Pascal kermit available from KERMSRV@CUVMA which includes a DIR routine for MS-DOS and for CPM-80 (I am working on one for CPM-86). The pertinent programs are in QKKER.PAS, look for SYSFUNC.PASMS in that. Manasseh Katz ------------------------------ Date: 16 Mar 1986 21:37-EST Sender: ABN.ISCAMS@USC-ISID.ARPA Subject: Directories from Turbo Pascal & Wang Compatibility From: ABN.ISCAMS@USC-ISID.ARPA Could someone tell me how to get a directory listing from a running (compiled) TURBO PASCAL program on the IBM PC? Is there an assembler routine that can be imbedded in the source code? Is there a DOS routine one can hook to? Thanks. Carl David In MARKET's TURBOPAS directory: MSLIB.PAS -- several routines that use DOS interrupts, etc. Two include directory-related routines. (Since Carl David has an address outside the ARPA, I'll forward this code separately.) RDIR.ARC -- Haven't downloaded this, but it's directory related and should have appropriate procedures for MS-DOS. Does anyone know why the WANG PC doesn't run "normal" packages like MS DOS 3.1, Turbo Pascal, Wordstar, The IBM Personal Editor (just to name a few packages I tried). I always get an error "Invalid/Unrecognized Interrupt" or the PC just freezes up. Any suggestions are welcome. Marc Chang Sing Pang Compadre, Wangs were never claimed or sold as PC-compatible. For one thing, they don't implement Int 10H (all the PC video stuff)! You can't even just stuff bytes in video memory and expect something to appear .. gotta hit another interrupt to force a refresh! Heck, they don't even provide DEBUG with their usual menu systems. It also fires up with Interrupt error messages, but appears to run (kind of). I wouldn't even claim MS-DOS 2.0 "compatibility" for the Wangs since I've personally tried to run programs I've checked for pure MS-DOS 2.0 functions only .. and had them fail! I purchased a "generic" MS-DOS version of Turbo Pascal 3.0, and (thanks to some pointers from a kind NetLandian) managed to configure it to a peculiar "almost ANSI" terminal that was supposed to work with Wang PCs. I can now compile and run Turbo Pascal programs on a Wang PC (and still have them fairly presentable (if limited) on PC clones. But dumb dumb dumb Wang .. thought they'd learn from TI it doesn't pay to try a "closed architecture." David Kirschbaum Toad Hall ABN.ISCAMS@USC-ISID.ARPA ------------------------------ Date: Sun, 16 Mar 86 10:52:26 EST From: Chris Schmandt <geek@MEDIA-LAB.MIT.EDU> Subject: Wang Compatibility Cc: singpang%hlerul5.bitnet@wiscvm.arpa The Wang is fairly DOS level compatible with the IBM, but definitely *not* BIOS compatible. Most likely your problem with "invalid interrupt" is that the software is using a Bios call like INT 10H to write directly to the screen. I've often considered writing my own version of the Bios screen driver for the Wang, not a formidable task, but never have the time. There is some Wang product, that I think includes a new display board, that allows full PC emulation. ------------------------------ Date: Mon, 17 Mar 86 10:50:39 est From: dewar@nyu-acf2.arpa Subject: MASM 4.0 OFFSET Bug In a recent message, I noted that OFFSET does not always work in MASM 4.0. The editor noted that you could use LEA instead. Three quick comments on this suggestion: First, least serious, LEA is slower than MOV immediate Secoond, in the example I had, LEA would not have worked, since I had BP indexing present. LEA would have included this indexing, OFFSET should have excluded it. Third, most subtle, at least in the Intel semantics, OFFSET and LEA are different in the presence of GROUPS. LEA will use the current ASSUME's to obtain the segment for the relocation target. OFFSET will always use the segment as the target relocation, unless you provide an explicit GROUP: override. I have no idea what microsoft does here, but beware! ------------------------------ Date: 17 Mar 1986 10:08:31 PST Subject: MASM 4.0 OFFSET Bug From: Richard Gillmann <GILLMANN@USC-ISIB.ARPA> To: dewar@NYU-ACF2.ARPA I certainly agree that it would be desirable to use OFFSET if it assembled correctly all the time. Using LEAs is just a work around, but it's never failed me yet. If someone could formulate rules for when it's safe to use OFFSET, that would be a big help. The way LEA works in the presence of GROUPs is consistent with the way Microsoft Pascal and C work, which is my main use of GROUPs. [The Microsoft assembler/linker has gotten better over the years. See some of the earliest INFO-IBMPC digests for horror stories. The OFFSET bugs have been particularly difficult to trace. Sometimes the linker catches an error other times it just generates wrong code. Dick had some modules that were stable for years, and when other code was added to the system suddenly the modules would go bad and always the OFFSET operator was at fault. Of course it never failed in simple test cases only when large multi segment links were involved. I gave up using macros back with MASM 1.0 and even though MASM 4.0 is much improved I still don't trust the assembler and try to keep my assembly programs short and simple. -wab] ------------------------------ Date: Mon, 17 Mar 86 13:20:29 cst From: pool@anl-mcs.ARPA (Pool) Subject: Microsoft Fortran & F77 The most current release of Microsoft Fortran is a rather poorly defined subset of F77. It is not easy to determine PRECISELY which features of F77 have been deleted by straightforware reading of the manuals. Jim Pool Numerical Algorithms Group ARPANET: pool@anl-mcs USENET: ihnp4!nag!pool ------------------------------ Date: Mon, 17 Mar 86 10:28:36 EST From: Bernard_Tiffany%UMich-MTS.Mailnet@MIT-MULTICS.ARPA Subject: Microsoft Fortran & F77 The latest Microsoft MS-DOS version (3.31) of FORTRAN implements the subset of FORTRAN 77 with extensions to the full set. The last time I've heard is that Microsoft is working on the full set as their Apple MacIntosh version of FORTRAN 77 does implement the full set. The following are FORTRAN 77 full set features not implemented in MS FORTRAN: (1) Comments cannot be followed by a continuation line. (2) The following statements are not yet available: CHARACTER*(*) statements (length specification is an asterisk) ENTRY statement SAVE statement without a list PRINT format, list READ format, list character statement functions (3) The forms UNIT= or FMT= cannot be used in READ or WRITE statements. These forms can be replaced by discarding "UNIT=" and "FMT=". A character array name may not be used as an unit identifier or a FORMAT parameter in these statements. (4) The edit descriptor Iw.m is not used in FORMAT statements. All other edit descriptors are implemented. (5) The OPEN statement cannot have STATUS='SCRATCH', STATUS= 'UNKNOWN', BLANK='ZERO', or BLANK='NULL'. To simulate the BLANK parameter, the format edit descriptor BZ or BN may be used. The UNIT parameter cannot be specified in the OPEN statement. (6) The CLOSE statement does not have UNIT and ERR parameters. (7) BACKSPACE, REWIND, and ENDFILE do not have UNIT, ERR, and IOSTAT parameters. (8) All lower bounds of an array must be equal to 1. (9) The DO variable cannot be REAL or DOUBLE PRECISION. (10) DATA statements cannot have implied DO's. (11) SUBROUTINE statements cannot have dummy arguments (expressed as *). (12) Expressions (other than constants or PARAMETER names) cannot be used in dimensions or lengths for CHARACTER statements. (13) Character concatenation operator // is not implemented but can be simulated using substrings. The character functions INDEX and LEN are not available. ------------------------------ Date: Mon, 17 Mar 86 15:56 EST From: Paul.Birkel@A.CS.CMU.EDU Subject: NBI to PC-DOS (and CP/M) Connections I've received a number of responses to my query last month, and made a number of phone calls. Here's what I found out: There are three and two-halves solutions to my problem. As you would expect, functionality is proportional to cost. The first solution, and the most cost-effective one, is a piece of software which NBI had Microsoft write about two years ago. It took a little while to figure this out, but it's called: OASys 4000SC MS-DOS Conversion Utility This assumes, of course, 5 1/4" disk drives. This package allows you to read and write such disks (DSDD - 48 tpi) plus it allows some mapping between WP tokens and ASCII characters. The mapping is limited but does take care of things like CR/LF, tabs, backspaces and form feeds so that your ASCII files will be in a readable "generic" form. There is no "format code/command mapping". It runs on all MS-DOS machines (NBI has their own clone, the 4100PC), and costs $200. One caveat on the AT is that you have a problem with the 96 tpi drive since the 4000SC is (typically, unless you have the "high capacity" drives) equipped with 48 tpi drives. The solution is to buy a second 48 tpi drive, about $100 additional plus your effort in installing it (easy). That totals $300 and buys you a simple, direct file transfer not dependent on physical proximity. For a CP/M'er use a similar transfer program (Media Master, Uniform, etc.) to go from CP/M to MS-DOS first. Unfortunately, I couldn't find any software supporting a direct CP/M to 4000SC route. Sigh. Solution one and one-half is a NBI card known as the M-400. It costs $700 and buys you a 68B00 (remember the Commodore PET?)-based co-processor running an early version of WP called "Docu-Writer". It also reads and writes NBI disks and resides on the PC bus. Although "Docu-Writer" is less powerful than WP, it offers the same interface as WP so that your word processing skills transfer easily. It runs in most (all?) PC/PC-XT clones. It DOES NOT run in the AT/AT-clones due to the higher system clock speed. It appears likely that it never will as NBI has other plans for AT owners. If you don't have an AT-type machine and you would like to leverage your WP skills then this card may be for you since you still get the MS-DOS conversion utility functionality discussed above. The other solution one and one-half, is a commercially available box for disk conversions: Keyword Office Technologies, Inc. 2816 11th Street NE Calgary, Alberta, T2E 7S7 (800) 227-1817 x825A Two US (West Coast) offices: 649 Mission at New Montgomery San Francisco, CA 94105 (415) 543-5426 Contact: Deborah McAfee 2047 E. Hamilton Ave. San Jose, CA 95125 (408) 371-7770 The list price for the Keyword 7000 is $7996 (US). Solution number two is based on a serial link between the NBI and the target machine. This requires the NBI ASYNC/BISYNC Advanced Communications package which will run you $2500, plus the cost of an external modem and installation. All told you're touching $3,000. What you get is the ability to do terminal emulation-like things and serial file transfers. You also get a software package called TAILORS that lets you do very reasonable (but apparently not complete) mappings between the file formats on each end. You use your own word processor on the PC-DOS/MS-DOS or CP/M or what-have-you end and one of the commercially available communications packages. Of course you may need either two modems, or proximity in order to run a cable, but you get added flexibility. Like to take your dBase or 1-2-3 file and ship it over to the NBI to put into a document? Well, TAILORS is supposed to make that possible. But it's not cheap. Solution number three is the Rolls-Royce of systems; the System 64, or OASys. It'll run you upwards of $15,000 and more depending on configuration. What it buys you in this context is: 1) A star-organized LAN with a PC-DOS 8086 central processor, hard disk, and 6809-based channels. 2) Transparent file access from any "device" where "device" may be an NBI machine, or a PC-class machine, including file format conversions. 3) The ability to run WP through your PC-class machine (actually on the 6809 <I think>). You can even run WP from home over a dial-up line. This is only a brief description, but you get the general impression. Slick, but a rather expensive for my needs. This investigation has been an education for me. Although I still think that technically the 4000SC is not impressive for the price (one can go buy a PC plus a major word processing package for less), what you do get is hand-holding and a clear upgrade path: a networking solution. Although I won't pay that much for the hand-holding, many office automation managers will (and perhaps reasonably so). My problem in initial NBI contacts was probably due to too much knowledge and too little understanding of the NBI philosophy/market-place on my part. I've had many contacts with NBI personnel, and all have been pleasant and helpful, but it was a user of NBI equipment on this net that actually pointed me to the disk conversion solution. While I'll probably never buy NBI equipment, if I were responsible for automating a large word processing environment I'd certainly consider NBI, especially since they appear to be offering a reasonable LAN solution while IBM is still (basically) dithering. So . . . back to the beginning: Can anyone tell what the actual format of a 4000SC 5 1/4" floppy is so that I can roll my own CP/M conversion utility? Paul A. Birkel Dept. of Computer Science Carnegie-Mellon University Pittsburgh, PA 15213 (412) 628-3074 ------------------------------ Date: Mon, 17 Mar 1986 17:53 EST From: Jim Ennis <JIM%UCF1VM.BITNET@WISCVM.WISC.EDU> Subject: Turbo Pascal Colors To: info-ibmpc <INFO-IBMPC@USC-ISIB.ARPA> Hello, Can anyone tell me how to set the colors for the editor/envrionment in Turbo Pascal. I don't like the brown on black and I would appreciate any help in changing the colors. Thanks, Jim Ennis University of Central Florida JIM@UCF1VM.BITNET@WISCVM.ARPA ------------------------------ Date: Mon, 17 Mar 86 12:35:06 PST From: Bruce_A._Cowan%SFU.Mailnet@MIT-MULTICS.ARPA Subject: XT/370 Speedup I put a little module in my XT370 that runs the 8088 at about 6.7MHz instead of the normal 4.77. The 8088 runs fine, but the 370 part fails with the fast clock. The module I plugged in is potted in epoxy so I don't know exactly what it does, other than feed the 8088 a faster clock (and I don't have a PC/XT Tech reference to look at the schematic), but I'd guess that the processor clock also ends up on the CLK line on the bus and that the 370 card uses that for its processor clock. Since the AT/370 uses the same 370 processor chips, and it runs with a 6MHz clock, I'd guess that it ought to be possible to speed up the XT/370 at least that far. However, the CLK signal on the AT's bus is 6MHz with a 50% duty cycle, whereas on the PC it is 4.77MHz with a 33% duty cycle. So, 6MHz with 33% duty cycle might be outside the capabilities of the 370 processor chips. I'm planning to try changing my XT/370 to 7.16MHz with a 50% duty cycle as soon as I can get an 8MHz NEC V20. I'll let people know if that works. ------------------------------ Date: Tue 18 Mar 86 08:40:32-PST From: William Pearson <PEARSON@SUMEX-AIM.ARPA> Subject: XENIX Updates There have been several questions lately on what IBM updates are available and how to use non-standard (non 20 Mbyte) hard disk drives with IBM PC/XENIX for the AT. Although I was able to get the disk installed without the IBM updates, I finally received them and they would certainly have made the process easier. There are two updated disks available from IBM (that I know of), I requested them through the IBM dealer I bought XENIX from (he was not able to find them until I gave him exact names) and received them from IBM PC Assistance Ctr/Zip 4409 1000 NW 51st Street Boca Raton, FL 33431 I received two disks, one dated 6/85 and titled: IBM Personal Computer Xenix 1.0 Soft Dev Sys Maintenance Diskette and one dated 9/12/85 titled: IBM PC XENIX INSTALLATION DISKETTE 1.01 9/12/85 850919A I was curious what had changed with the new installation diskette, so I checked the dates on the files, and found most changes dated 6/85 or 9/12/85. I have included a list of the modified files below. fd Jun 26 1985 bin Jun 25 1985 boot Jun 26 1985 etc Sep 12 1985 tmp Jun 24 1985 xenix Jun 25 1985 xenix.fd fd /bin Jun 26 1985 adb fd /etc Jun 24 1985 badtrack Jun 24 1985 boot1 Jun 24 1985 fd96boot0 Sep 12 1985 fdisk Jun 24 1985 hdboot0 Sep 12 1985 hdinfo Sep 12 1985 hdinit Sep 12 1985 mnttab Sep 12 1985 utmp Sep 12 1985 xinstall fd /dev Sep 12 1985 console Sep 12 1985 fd0 Sep 12 1985 fd096 Sep 12 1985 fd096ds15 Sep 12 1985 hd00 Sep 12 1985 hd0a Sep 12 1985 null Sep 12 1985 rhd00 Aug 30 1985 rhd0a Aug 30 1985 rhd10 Sep 12 1985 root Sep 12 1985 swap There are several other updates from Microsoft that IBM has apparently not yet distributed, which means they are virtually unavailable (the receptionist at Microsoft was very rude when I asked about them) including one dated 12/85 which provides the huge compiler model. But the Installation Disk 1.01 is the one you need for non-standard drives. Bill Pearson ------- ------------------------------ Date: Sun, 16 Mar 86 23:13:01 PST From: kegel%Juliet.Caltech.Edu@Hamlet.Caltech.Edu Subject: MS-Windows Drivers About Microsoft Windows: Microsoft only supports the IBM CGA and EGA graphics boards as of now (although I hear rumors of something great in the wind). Do any graphics board manufacturers, or graphics software houses, have Windows drivers for their products? (What I'm secretly hoping for is a 640x480 16 color board with a real palette and most of the Windows routines implemented in silicon...) - Daniel Kegel [As has been mentioned in recent digests the Genius display has a MS Windows driver. I spoke with the manufacturer in our unsuccessful attempt to get a working BIOS from them. At the time they were quite excited by the MS Windows drivers. I gather Microsoft developed the drivers even though the Genius isn't one of the "official" Microsoft supported devices. MSWindows already takes up enough distribution disks. -wab] ------------------------------ Date: Mon, 17 Mar 86 08:24:17 est From: Scott Guthery <sguthery%slb-doll.csnet@CSNET-RELAY.ARPA> Subject: Pascal to C Translator Does anyone know of a good Pascal to C translator for the PC? The operative word is `good'. Thanks, Scott ------------------------------ Date: 18 Mar 86 10:37-EST From: R.Rasulis <ext715%BOSTONU.bitnet@WISCVM.arpa> Subject: Symphony & MSWindows Has anyone had any luck running Symphony w/ MS Windows? It doesn't work using MS WIndows off the shelf, so you must have had to "fix" it. Please share your patch w/ me. Please reply directly to me *eom :- R.Rasulis -: ------------------------------ Date: 18 Mar 86 12:18 EST From: Jim Moore <moore@ncsc> Subject: DisplayWrite 3 EGA Support To: info-ibmpc@usc-isib I've tried this before with no response, but here goes again: If you use DisplayWrite X.XX with and EGA, how do you get ON-SCREEN underlining and boldfacing??? Any help is appreciated. [In a later message Jim Moore added that published reports claimed bold and underline was possible on the EGA, but he is only able to get colors. Unfortunately an editor crash lost this paragraph of text. -wab] Jim Naval Coastal Systems Center ------------------------------ Date: Sun, 16 Mar 86 04:56:45 PST From: dgb%DEImos.Caltech.Edu@Hamlet.Caltech.Edu Subject: FORMAT Equivalent Within a Program Does anyone out there know exactly what is involved in doing the equivalent of a DOS FORMAT from within a running program? We have an application where is it advantageous to leave a program running for days at a time, acquiring radio spectra. The observers aren't too good about keeping a supply of formatted floppies around, and one of these days they're going to find themselves with a full memory of data, and no blank floppies to write it to. There is no other PC for many miles of the observatory, so this could potentially be a real problem. Since we need all the available RAM we've got, I'm rather loathe to set aside the 30K+ that I would need to spawn a new command interpreter to run FORMAT. I would much rather use a program overlay. Does anyone know exactly what is involved? 1) This routine would only need to operate on data floppies, so that I don't have to deal with getting a system set up. The drives in question are 'standard' D9 format. Ie. DSDD, 360K, 40 track, etc... 2) The BIOS can do the hard work of actually formatting the disk. The tricky part is going to be setting up all the DOS work areas such as the boot track, directory, FAT(s). 3) It will be running on a two floppy system, so I can cheat, and copy the boot track from the system disk. (I would prefer to avoid this, and might just store a copy of the boot track in memory, to avoid problems when running on the HD based development systems. Any better ideas?) 4) I have the usual refs for such things. DOS technical apendix, Tech Ref, Peter Norton's book, etc. I might be able to piece a workable routine from that, but I would be worried that some critical field wasn't getting initialized or something. I don't want to have data floppies floating around with 'bombs' in them from my homebrew format routine. I would really like to hear from someone who knows the subject! If it's too much of a hassle, I won't worry about it, and just renew my efforts to get our people to format the floppies when they buy them. But I just know that we're going to see trouble from this one day.... As usual, thanks in advance for any help. --Daniel Briggs dgb@DEImos.Caltech.Edu ------------------------------ Date: 16 Mar 1986 18:33:35 PST Subject: FORMAT Equivalent Within a Program From: Billy <BRACKENRIDGE@USC-ISIB.ARPA> To: dgb%DEImos.Caltech.Edu@HAMLET.CALTECH.EDU The Tall Tree products for RAM disk emulation and the JFORMAT program all come with source code. It would be a good place to start looking to write such a routine. Tall Tree Systems (415) 964-1980 Tech Info (415) 964-1397 ------- ------------------------------ Date: 16 Mar 86 11:25:02 GMT From: brad @ DCA-EUR.ARPA Subject: Zenith PCs at Govt Contract Prices When Zenith had their first contract with the AF/Navy a couple of years ago for Z-100s, Government Employees Association (GEA) worked out an arrangement with Zenith to sell Z-100s to individuals connected with the Govt at the Contract Price. Question: Does anyone know if an organization, such as GEA, is working out a similar arrangement with Zenith for the new contract with the Miliary?? Those of us working for the Military in Stuttgart Germany are very interested. Brad ------------------------------ Date: Mon, 17 Mar 1986 12:34 O From: Guy Sirton <MLSIRTON%WEIZMANN.BITNET@WISCVM.WISC.EDU> Subject: Xenix Device Driver I am having problems writing a device driver in XENIX (V1.0 running on a regular PC-AT). The problems are: 1) Accessing storage (for memory-mapped devices). I overcame this problem by defining a 'far' pointer but I am wondering if there is any built-in kernel routine to do it. 2) Creating the new kernel. When I type 'make' in /usr/sys/conf I get the following result: ld: Unresolved extrenals: __stkgro in file(s): /lib/Mlibc.a(~_chkstk.o) __Syspc in file(s): /lib/Mlibc.a(exit.o) __syscal in file(s): /lib/Mlibc.a(exit.o) *** Error code 1 Stop. 3) Poor documentation. The documentation at the lowest level is really poor some routines (such as seterror() ) are mentioned but not explained and I'm pretty sure (by 'string'ing lib_io and the kernel) that there are routines that aren't even mentioned. The index is also in a very poor state... Anyone who can help me in these problems and/or has a list of un-documented routines please mail me back. Guy ------------------------------ Date: 17 Mar 86 13:36 PST From: Ghenis.pasa@Xerox.COM Subject: Document Indexing Program What are the available programs for creating a table of contents and index for a WordStar document, both public domain and commercial? How capable and user-frienldy are they? Thanks in advance. ------------------------------ Date: Sun, 16 Mar 86 10:43:14 pst From: George Cross <cross%wsu.csnet@CSNET-RELAY.ARPA> Full-Name: George Cross To: info-ibmpc@usc-isib.arpa Subject: BRIEF vs EPSILON I am trying to decide between BRIEF and EPSILON. Is there someone out there who has used both? The way I read it, BRIEF and EPSILON are emacs-like, fully customizable in Mock-Lisp (respectively Mock-C), but EPSILON supports a kind of multi-tasking so you can continue editing while compiling and BRIEF does not. What about performance? (Especially assuming a floppy-only system) ---- George - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - George R. Cross cross@wsu.CSNET Computer Science Department cross%wsu@csnet-relay.ARPA Washington State University faccross@wsuvm1.BITNET Pullman, WA 99164-1210 Phone: 509-335-6319 or 509-335-6636 [We still keep our version 2 epsilons around for running on floppy disks. While the newer versions run just as fast on a hard disk system the older is still useful when floppy disk space is at a premium. Even with a floppy only system for another $400 you can get a 2MB JRAM card and edit any sized file with either version. -wab] ------------------------------ End of Info-IBMPC Digest ************************ -------