Gpc

gpc@gnu.de

5400 discussions

DJGPP installation problems
by Peter Gerwinski 16 Oct '96

16 Oct '96

Newsgroups: comp.os.msdos.djgpp References: <01bbbab6$1e0cf340$LocalHost@default> Organization: Universitaet Essen, Germany Reply-To: peter.gerwinski(a)uni-essen.de Distribution: Niels Heirbaut (heirbaut(a)pi.net) wrote: > I don't know if this is the right group to post to (I haven't found an > alternative), so please forgive me if this post is off topic. Since this is a DJGPP-specific problem, I think this NewsGroup is appropriate. However, there is a GNU Pascal mailing list, gpc(a)hut.fi. To subscribe, write to gpc-request(a)hut.fi. (This followup is also posted to the GPC mailing list.) > I downloaded the binary files for GPC (version 2.7.2) and extracted them to > my DJGPP directory. But when I tried to compile the program hello.pas (see > below) with: C:\>gpc hello.pas -o hello.exe > > I got the next message: > C:\DJGPP\BIN/ld.exe: cannot open crt0.o: No such file or directory (ENOENT) > > I checked if crt0.o existed (it did). I also tried to put the directory > (C:\DJGPP\LIB) in my PATH-variable, but still the same response. In the > info file I read something about setting the variable GPC_EXEC_PREFIX. I > tried this (in DJGPP.ENV and AUTOEXEC.BAT) but it still didn't work. > Here are my program, my AUTOEXEC.BAT and my DJGPP.ENV files: > > [...] > > DJGPP.ENV: > > [...] > > [gcc] > COMPILER_PATH=%/>;COMPILER_PATH%%DJDIR%/bin > LIBRARY_PATe=%/>;LIBRARY_PATH%%DJDIR%/lib;%DJDIR%/contrib/grx20/lib > > [...] > > [gpc] > GPC_EXEC_PREFIX=%/>;GPC_EXEC_PREFIX%%DJDIR%/lib >From the GPC installation instructions (you might have an older version which does not yet contain this): 8< ------------ Finally, add these lines to your \djgpp\djgpp.env file: --------- cut here ------- cut here --------- [gpc-cpp] C_INCLUDE_PATH=%/>;C_INCLUDE_PATH%%DJDIR%/include [gpc] COMPILER_PATH=%/>;COMPILER_PATH%%DJDIR%/bin LIBRARY_PATH=%/>;LIBRARY_PATH%%DJDIR%/lib 8< ------------ Hope this helps, Peter e-mail: peter.gerwinski(a)uni-essen.de home address: D\"usseldorfer Str. 35, 45145 Essen, Germany WWW: http://agnes.dida.physik.uni-essen.de/~peter/

1 0

Word and LongWord
by Peter Gerwinski 16 Oct '96

16 Oct '96

According to Andreas Bauer: > JJvdH> type > JJvdH> word = __unsigned__ integer; > You should call it longword. > I had some trouble, because I declared a word like this and i wondered > that the entries after flags had some errors :) > Perhaps longword... A matter of taste and self-discipline while counting bits. ;-) I prefer "word" to stand for the "natural" unsigned integer - which has 32 bits with GNU Pascal. (But I also had trouble with this being a Borland Pascal programmer for many years ... ;-) Peter e-mail: peter.gerwinski(a)uni-essen.de home address: D\"usseldorfer Str. 35, 45145 Essen, Germany WWW: http://agnes.dida.physik.uni-essen.de/~peter/

1 0

GPC Bugs ....
by Phil Nelson 16 Oct '96

16 Oct '96

Hello GPC fans, Recently a gnats database was set up for GNU Pascal. Gnats is the GNU bugtracking system. This will allow the developers of GNU Pascal follow bugs and bug fixes much better. It will also allow users of GNU Pascal to see if a bug has been reported before and/or fixed. You many both search the bug database and submit new bug reports at the URL http://fawn.cs.wwu.edu/~phil/gpc. A link to this site will be added to the GNU Pascal home page soon. (The GNU Pascal home page is http://agnes.dida.physik.uni-essen.de/~gnu-pascal.) Currently, a report of a bug to the GNU Pascal gnats site will send a message to gpc(a)hut.fi so we can all see which bugs have been reported. -- Phil Nelson NetBSD: http://www.netbsd.org e-mail: phil(a)cs.wwu.edu LPF: http://www.lpf.org http://www.cs.wwu.edu/~phil !gifs: http://www.lpf.org/Patent/Gif/Gif.html

1 0

Re: Call for installation patches and hacks.
by Peter Gerwinski 16 Oct '96

16 Oct '96

According to Kevin A. Foss: > I did have to make a couple changes to the OS/2(EMX) makefile in order to get > everything working. :-( That's strange because I made *a lot* of tests with the EMX version on FAT partitions as well as on HPFS partitions to ensure that you just type [I:\emx\gnu\gpc-272] configure (* configure.cmd *) and [I:\emx\gnu\gpc-272] make (* make.cmd - calls dmake *) and GPC compiles through (assumed that you have compiled GCC before). Are you sure that you aren't using the 2.6.3 EMX patch? (Of course. ;-) Please send me your modified Makefile. > Specifically, even though gcc_bcmp.c is copied into the > RTS directory, it is never referenced in the makefile. (MakeHPFS.RTS) You are right! =:-O > It was > only after I tried to compile some code that apparantly needed _gcc_bcmp() > that I noticed the problem. So I had to go back and edit the Makefile. > > Or did I do something wrong? I couldn't find any reference to gcc_bcmp.c/.o > in the MakeFAT.RTS file either. Again: Please send me your modified Makefile. > I also experienced confusion over which version of Make to use. INSTALL says > to use nmake, but README.EMX says to use dmake. `dmake' is correct. > (Sorry if this has been addressed before, I just recently joined the list..) No, it's a new bug. :-)-: Thank you for the report. Peter e-mail: peter.gerwinski(a)uni-essen.de home address: D\"usseldorfer Str. 35, 45145 Essen, Germany WWW: http://agnes.dida.physik.uni-essen.de/~peter/

1 0

extend bug
by Jim Brander 16 Oct '96

16 Oct '96

Running gpc 2.6.3 on ELF format Linux 1.2.13 _p_extend() in rts/rts-file.c has following problem. Program is writing out 3000 line file, extending file for each line (don't grumble, it works this way). File extends OK for several hundred lines (a line or 2 each time), then the contents of the file buffer is spilled, say 20 lines written twice but mucked up on line endings, then extends OK for several hundred more lines, then does it again. The file buffer changes to a different address when spill occurs - that is, while it happens to be the same buffer address used each time, no problem. Doesn't display this behaviour on gpc 2.6.3 on Linux 1.1.59 with a.out format. Is fixed by commenting out line in rts/rts-file.c in _p_extend() int ch = getc(m_FILNUM(File)); File pointer is already pointing at end of file without this line. Regards Jim

1 0

Re: Call for installation patches and hacks.
by Kevin A. Foss 15 Oct '96

15 Oct '96

On Tue, 15 Oct 96 15:33:43 CST, J.J. van der Heijden wrote: >We are finishing up preparations for the "final" 1.2 GPC release, and now I >would like to know, did any of you have to apply special hacks to get GPC >up and running? Need to edit a config file, Makefile etc., add special >CFLAGS? I did have to make a couple changes to the OS/2(EMX) makefile in order to get everything working. Specifically, even though gcc_bcmp.c is copied into the RTS directory, it is never referenced in the makefile. (MakeHPFS.RTS) It was only after I tried to compile some code that apparantly needed _gcc_bcmp() that I noticed the problem. So I had to go back and edit the Makefile. Or did I do something wrong? I couldn't find any reference to gcc_bcmp.c/.o in the MakeFAT.RTS file either. I also experienced confusion over which version of Make to use. INSTALL says to use nmake, but README.EMX says to use dmake. (Sorry if this has been addressed before, I just recently joined the list..) -Kevin --- Kevin A. Foss -- kfoss(a)atlsysnet.com ---

1 0

STABS format -- solaris users around?
by J.J. van der Heijden 15 Oct '96

15 Oct '96

Hello, I have started a project to add a "Pascal" mode to GDB. This means GDB can demangle Pascal symbols, and evaluate Pascal expressions. The debug format for most platforms is stabs, and exploit all possibilities, GPC and GDB have to support a number of Pascal specific stabs. These are rumoured to be ducumented by SUN for their SUN pascal compiler, and this document might be on the Solaris SDK CD-ROM. So, if any of you have this CD-ROM around, and could pass me this document, maybe I can follow an existing standard, instead of reinventing the wheel. I don't care about the format, I can handle most formats between raw text and PostScript. Sorry for my vagueness, JanJaap --- "Nothing shocks me, I'm a scientist", Indiana Jones

1 0

Call for installation patches and hacks.
by J.J. van der Heijden 15 Oct '96

15 Oct '96

Hello, We are finishing up preparations for the "final" 1.2 GPC release, and now I would like to know, did any of you have to apply special hacks to get GPC up and running? Need to edit a config file, Makefile etc., add special CFLAGS? We know about the problems with stor-layout and friends (undefined symbols when linking gpc1) and BSD/SYSV confusion with signals in rts/rts-rt0.c We would like to hear about any other hacks you had to do, and this is your chance to get it fixed before the release. In the ideal world, `configure ; make' should work. I have heard people failing to build GPC on Solaris. Is this still true, and if yes, can you give me a detailed description what goes wrong? Greetings, JanJaap --- "Nothing shocks me, I'm a scientist", Indiana Jones

1 0

getenv() and system()
by Peter Gerwinski 15 Oct '96

15 Oct '96

According to John Michael Black: > IMHO, I would think that the function getenv() and the procedure system() > would be two widely used routines for a lot of people's programs. Instead > of expecting everyone to "create" them each time, what possibility is > there of including them in a standard library? My BO5 project is intended to become a standard library for GNU Pascal. Its Unit "Tools" already contains a Borland- compatible version of getenv(). > I know they're not > standard or extended pascal, GetEnv is in Borland's DOS Unit, so it's de-facto-standard on DOS-like platforms. Berend already ported some features of DOS to an Extended Pascal module, other things are in BO5, so it should be straightforward to write (at least parts of) a true Borland-compatible DOS Unit. > but like I said it seems like a lot of people > would need to use them. If there isn't an include for operating system > functions, perhaps this could start one. Includes are not part of ISO-Pascal, modules are. With Borland or UCSD Pascal, you can include a file via (*$I foo *) (supported by GPC, too). However in my opinion, a Module or Unit would be preferable. If somebody would like to write such a thing (or already has written), please let me know. Greetings, Peter e-mail: peter.gerwinski(a)uni-essen.de home address: D\"usseldorfer Str. 35, 45145 Essen, Germany WWW: http://agnes.dida.physik.uni-essen.de/~peter/

1 0

Bug report :-)
by Jukka Virtanen 15 Oct '96

15 Oct '96

Reported by: Juki <jtv(a)hut.fi> GPC version: 1.2-2.7.2.1 Platform : Digital Alpha, DEC unix 4.0 Description: ------------ The attached program does not compile. narya gpc-1.2-2.7.2.1 58 % gpc fail.p fail.p: In function `zap': fail.p:5: parse error before `Absolute' Attached program: ----------------- program zap; var enumvar : (absolute); begin end.

1 0

Jump to page:

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