Gpc July 1996

gpc@gnu.de

6 participants
20 discussions

Files
by berend＠beard.nest.nl 01 Feb '02

01 Feb '02

* Moved (from: gpc) by Berend de Boer using timEd 1.10.b3+. Peter Gerwinski wrote in a message to Berend de Boer: PG> I tried to use Extended Pascal "binding" to assign a file PG> variable to an external -- existing -- file. I just copied PG> the "Assign" pro- cedure by Berend as published in PG> borland2ep.doc. It does not always work: I often get a PG> runtime error saying that the file did not exist. If you can repeat it, compile with another Extended Pascal compiler, for example Prospero's one. If that fails too, there is really something wrong with the code. But I suspect gpc. You're using which OS? Groetjes, Berend (-: fido: 2:281/527.23 email: berend(a)beard.nest.nl

4 3

GPC autoconf-ed
by J.J. van der Heijden 29 Jul '96

29 Jul '96

I mentioned my rts-autoconf adventures on the mailinglist; I dumped a modified RTS (rts-autoconf.tar.gz) in my homedir on agnes.dida.... Usage is straightforward: cd to your $(gpcbin)/rts directory and run ../../$(gpcsrc)/rts/configure ; make Autoconf is the way to go for GNU software and I think GNU Pascal should use it too. This is not finished: I have to do GPC itself, plus the result of a number of tests is not yet used in the RTS makefile. But it deals with all the -DSYSV -DNO_CASECMP -DMY_VERY_WEIRD_SYSTEM flags. It can also do crosscompilation: run CC=i386-go32-gcc AR=i386-go32-ar RANLIB=i386-go32-ranlib \ ../../$(gpcsrc)/rts/configure --target=i386-go32 ; make to get a djgpp RTS (on Linux) I also added a definition in Makefile.in to build libgpc as a shared lib. A 'make libgpc.so.2.8 CFLAGS="-O2 -fPIC"' results in: lrwxrwxrwx 1 root root 13 Jul 29 02:06 /usr/lib/libgpc.so -> libgpc.so.2.8* -rwxr-xr-x 1 root root 48960 Jul 29 02:06 /usr/lib/libgpc.so.2.8* This reduces the size of executables: -rwxr-xr-x 1 janjaap users 2856 Jul 29 02:09 hello* -rw-r--r-- 1 janjaap users 63 Jul 29 02:07 hello.pas Statically linked, the minimum is ~33K. BTW: this will work only on ELF un*x systems and I have only tested it on Linux. "ldd hello" shows: libgpc.so.2.8 => /usr/lib/libgpc.so.2.8 libm.so.5 => /lib/libm.so.5.0.5 libc.so.5 => /lib/libc.so.5.2.18 I will add this as an "--enable-shared" option to "configure" As for your desire to release 2.7.2: I can have this finished by friday (I think ...), then it should be tested on a number of different platforms. I'm especially curious about BSD unix systems, because I developed this on Linux, which is essentially SysV. I created a "configure.bat" for the dos world where there's no bash. Peter, you have the same with OS/2 so maybe you can take a look at it and create something simular for the EMX thing. (I ripped of binutils and GCC to create "configure.bat") /janjaap --- Thus spake the master programmer: "After three days without programming, life becomes meaningless." [The Tao Of Programming]

1 0

Re: release?
by J.J. van der Heijden 29 Jul '96

29 Jul '96

On Sat, 27 Jul 1996 23:00:20 +0200 (MET DST, Peter Gerwinski <peter(a)agnes.dida.physik.uni-essen.de> wrote: >Hello again, > >there was a bug in my GPI stuff: File types did not >survive transport through a GPI file. I am sure this >was not the last bug, but now a very useful physics >program works, so I am optimistic. > >gpc-pg-1.2-2.7.2.tar.gz has been updated. > >Release? > I know of two good reasons not to release 2.7.2 (yet): 1) Remaining bug(s). Last time I looked, that "functions returning a string are called three times" bug was still there. That's a serious problem for anything bigger than "Hello world". Also (I mentioned before): gpc fails two of the "paranoia" Pascal compiler float tests because it loses significant digits. I'd be carefull with that physics program, Peter. 2) Almost all messages / email I read about GPC are not about "compiler bug X keeps me from using it" or "feature Y is missing" but "HOW DO I INSTALL THE THING ON MY SYSTEM ?" "Where's the docs" is a popular one too. That's why I'm busy rewriting configuration scripts using GNU autoconf and have created .bat files for the dos version. Then, building it is reduced to "configure ; make" I have it done for the RTS, am working on GPC itself now. I think it should be able to locate gcc sources and objects by itself (if they are in a reasonable location). It works great for the RTS already. Arguments in favour of release: 1) New features (These have to be mentioned in a "NEWS" file. I think we have (a) better BPascal Objects (b) Improved module support, but there must be other user-visible changes as well. 2) Releasing 2.7.2 because gcc-2.6.3 is obsolete is an argument, but I don't think we'll have to wait much longer for gcc 2.8 and from what I've seen so far, it's not radically different from gcc-2.7.x (internally). So porting gpc to 2.8 should take too long. If this thing is to be released in one or two weeks I think we should make a "release candidate" so everybody can check wether his patches are in and binaries for all platforms can be built. /janjaap --- Thus spake the master programmer: "After three days without programming, life becomes meaningless." [The Tao Of Programming]

1 0

GNU Pascal WWW home page
by Peter Gerwinski 26 Jul '96

26 Jul '96

Hello folks, I have constructed a GNU Pascal WWW home page: http://agnes.dida.physik.uni-essen.de/~gnu-pascal Comments, suggestions, bug reports welcome. Peter <peter.gerwinski(a)uni-essen.de>

1 0

forwarded Usenet article
by operator＠civpczaal.utwente.nl 26 Jul '96

26 Jul '96

[article forwarded from Student (operator(a)civpczaal.utwente.nl)] Path: news.nic.utwente.nl!surfnet.nl!howland.reston.ans.net!newsfeed.internetmci.com!btnet!news.enterprise.net!usenet From: eurgain(a)enterprise.net (Alistair Hamilton) Newsgroups: comp.os.msdos.djgpp Subject: Re: GPC112B.ZIP & GPT112B.ZIP Date: Tue, 23 Jul 1996 23:30:15 GMT Organization: Ty Eurgain Lines: 45 Message-ID: <4t3nq5$ih8(a)news.enterprise.net> References: <960718165325_74617.365_EHH119-1(a)CompuServe.COM> NNTP-Posting-Host: max02-063.enterprise.net X-Newsreader: Forte Free Agent 1.0.82 "Troy D. Van Horn" <74617.365(a)CompuServe.COM> wrote: >Hi all, > I just last night discovered these files for a PASCAL compiler last night >on a simtelnet mirror, and I was wondering where I could get some documentation >on how to use either compiler? The source file is really large (>8Meg) so I >would really prefer not to download it. Where should I look for documentation >on GPC and the differences between GPC and GPT version? (I already know >PASCAL). >Thanks, Troy Van Horn 74617.365(a)compuserve.com >Please E-mail all replys direct to my address, Thank you. PROGRAM GetSomeHelp (Input, Output|) CONST PascalExists = TRUE; VAR CIsBetter : BOOLEAN; BEGIN WRITELN('I would really like this info too. OK, as an ex Unix/C/X nerd, I cannot dispute that C cannot be beaten (except , perhaps, by C++). However, age has revealed that Pascal really is useful sometimes! Typing B E G I N and E N D sure as hell is! (I have programmed for accountancy apps ZZZzzzz in Pascal - but better P than C in that case).'); CIsBetter := TRUE; WRITELN ('I also have the GNUPascal binaries, and am held back from using them by lack of documentation. So come on, folks! At least you might tell us to b****r off and consult another really tesious group specially for Pascal bores, but if you do, it least you might tell us waht the group is called.'); WRITELN ('Regards'); WRITELN ('Alistair') END. :-) :-) :-)

1 0

Re: Programmers wanted: CRT, Graph, etc. for GNU Pascal
by Peter Gerwinski 25 Jul '96

25 Jul '96

Berend de Boer wrote: > The point of having an extended pascal compiler > is being portable, at least that's the goal. I agree to the goal of being portable. But I don't think that we can reach this goal just by obeying the Extended Pascal standard. (More about this in a separate mail.) > Assembler and Dos are not the way to go. The point of having a > System, Crt, Printer of Graph unit is entirely uninteresting if it is not > portable except if you want to port Dos Borland Pascal programs to Extended > Pascal. I want to port DOS Borland programs to *GNU Pascal*, and I am sure that many people intend the same. System, Crt, Printer and Graph for GNU Pascal would be *highly interesting* for those programmers who are accustomed to Borland Pascal and think about changing to GNU Pascal (instead of changing to Delphi or C++) in the future. To have those Units on the DOS platform only would be a starting point -- portable versions would be preferred, of course. > Why should anyone want to do this? I assume everyone is programming for > Windows 95/Windows NT or Unix. Are programmers really running Dos today?? Yes, I am for example. More concrete, I am working in an OS/2 DOS box and in the Linux DOSemu (with Novell DOS 7). I do not intend to use Windows 95 or Windows NT, and it will take a long time to get my clients using UNIX (Linux). Up to that day, I will prefer writing good DOS programs to writing poor Windows programs (with a lot of special effects, but slow, uncomfortable, and unstable). In the long term, I intend to become completely platform independent with my BO5 library, i.e. my programs will run under Windows 95/NT, too, just be recompiling. But for the moment, I am stuck with DOS. > Quite a few calls in System or Crt do simply not apply in non-Dos > environments. There are really few of them, but since the source exists, I suggest that somebody ports *everything* to the DOS version of GPC, and we can make the "potentially portable" parts portable in a second stage. > A portable Printer library?? What's the problem? It would just be a Unit (Module) which "knows" the file name of the printer device for the actual operating system and opens a text file "lst", so the application program doesn't need to care about how the printer is named on the operating system. > The calls from System that are portable can for 90% be written in plain > Extended Pascal. However, it would be nice to have a compatibility Unit (Module) which simplifies the port of Borland Pascal programs to GNU Pascal. > We should do better by following existing standards and provide EP > interfaces for that. Borland's Units *are* an existing standard. There may be other standards, but I think these Units are very popular. Furthermore, we *have* the source for them, so I ask again: ******************************************************* We have free sources of Borland Units (CRT, Graph, ...) which cry for being ported to GNU Pascal. Anybody wants to do the job? ******************************************************* Greetings, Peter

2 2

Re: The purpose of GNU Pascal
by Berend de Boer 24 Jul '96

24 Jul '96

> By the way: Where are those many different Extended Pascal compilers? You have Prospero's compilers for Dos/Windows and OS/2, DEC has a Pascal compiler on VAX which implements almost all of the Extended Pascal standard. I'm not sure of the Macintosh, there is a Pascal compiler on the Mac from Warriar, but if that one implements Extended Pascal?? I'm going to research that today. > You don't lose portability by simply including bit-shift operators, objects, etc. into > the Extended Pascal standard. But they didn't include them. No idea, why. Objects are in another standard. bit-shift operators have to depend upon memory layout and byte ordering. But I agree with you that they should be available. > And what about, say, user-defined operators which are almost a *must* for scientific calculations? I completely agree with you that user-defined operators are a must in a field where those operators defined. In most other areas they are not. We can probably atttach a dozen explanations to the following construct: House := PersonA + PersonB; The example you gave had exactly one meaning clear to everyone. > But I don't intend to spend my time with trying to satisfy standard specifications I don't realize their > sense. I can quite agree with you. But don't forget: there are millions, literally millions of code written in some form of Standard Pascal out there in the world. For such people an alternative in the form of GNU Pascal could really be interesting. > How would you define "real Pascal spirit"? Let the compiler do as much as possible for you: basically that comes down to no untyped pointers. I *have* to use much more untyped pointers in Borland Pascal (in obtaining addresses from procedures, strings, etc.) than would be necessary. > Extended Pascal may be good, but GPC will be even better. It is time to set up the next standard! Completely agreed. I can accept the modifications to GPC.GUIDE you presented. Of course a compiler which accepts Borland Pascal syntax is a major advantage. Makes porting much easier. And if you really have good alternatives for Borland's standard libraries... Groetjes, Berend.

2 1

Re: Extended Pascal
by Berend de Boer 24 Jul '96

24 Jul '96

>> Last time I really looked at gpc (beginning 1995) it missed simply too much. > What is missing? If you publish a list, perhaps somebody would like > to do the job of implementing it. OK, I hope to present a list of missing features to this list within a few days. Groetjes, Berend.

1 0

Re: Programmers wanted: CRT, Graph, etc. for GNU Pascal
by Berend de Boer 24 Jul '96

24 Jul '96

>> What use does it have when you already have a good Dos compiler, i.e. Borland >> Pascal? It fullfills all my needs for the Dos environment. > Not mine, and when you look into comp.lang.pascal.borland you will find Yes, I did see your problem, but I never encountered these things. My units are small, so I don't have codesegment faults, and for large arrays I have a few objects of my own which either use XMS/EMS or conv memory to present large arrays. But I see. I make a Posix interface for GPC first and present any difficulties I encounter to you so you can make the necessary changes to the compiler ok? > Note, I HAVE the source of the Graph unit!!! Ah, I now remember you saying this, but shouldn't this be: I have the source of a Graph unit? >> - qualifed imports available? > What's that? To reference a symbol from these modules you always need to add the module name as prefix. Maybe I don't need this now, I'll see. OK, I see if I can write something with gpc. I'll tackle the Posix interface and the Dos unit first. I start using djgpp's port. Which one should I take? Groetjes, Berend.

1 0

Re: The purpose of GNU Pascal
by Berend de Boer 24 Jul '96

24 Jul '96

> Furthermore, I could not write my real-world programs > just with Extended Pascal; I need, for example, bit shift operators, GetMem, > FreeMem, and Objects. You don't need GetMem and FreeMem as I did show you :-) But you need the other two. > and integrate them in one powerful compiler which will set up the next standard. I certainly don't agree with this if the compiler does not have switches to accept only Extended Pascal or only ISO Pascal or only ... Programs conforming to the Extended Pascal will probably not only work in many different environments but also with many different compilers. I don't want to loose that feature/possibility. Extending a compiler with more features is not a problem, but IMO, one should complete the standard parts first. > It is okay to support Extended Pascal (why not?), but in my opinion, > Extended Pascal is not clear and powerful enough to serve as "the new > Pascal standard", so programmers are not motivated to switch from > Borland Pascal to Extended Pascal. I think, this is the reason why > computer vendors do not provide Extended Pascal compilers. Hmm, I don't think that Extended Pascal is not powerful enough. It's true that it does not contain everything one needs, but that's mainly due to the goal of creating a portable language. And the new object standard also rectififies some omissions. I can create everything I want in Extended Pascal programs with only a few routines accessing non-standard parts, mainly concerning bit-operators (you can do this with Extended Pascal, but it's a bit more tedious), and system call access (assembly). A compiler which accepts every Pascal standard and does everything would be nice of course. But there is quite some work to do. So I would like to add that we first focus on the ISO Pascal and Extended Pascal part. Even the ISO Pascal standard, according to Scott Moore (have email address if anyone wants it), needs much looking after. Don't overlook the Pascal standards boys (no girls here I assume). They do much more than you think, but different than Borland Pascal which has quite a C feeling and lost too much of it's real Pascal spirits. Groetjes, Berend.

2 1

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

Gpc July 1996