- Gpc - gnu.de

GPCC
by Peter Gerwinski 26 Jul '99

26 Jul '99

Hi, everybody, as you know, The African Chief wrote a program `GPCC' that adds Borland-style command syntax to GPC. It works as a wrapper, translates options and then calls `gpc'. Since we got no reports about problems with this utility, we can integrate it into GPC now. What about creating a new directory `utils' in the GPC source directory where GPCC and other development tools especially written for GPC can be stored? Its `Makefile' would (currently) only have to `gpc --automake' each utility and then install it in the `<prefix>/bin' directory. Any comments? Peter -- Peter Gerwinski, Essen, Germany, http://home.pages.de/~Peter.Gerwinski/ Maintainer GNU Pascal - http://home.pages.de/~GNU-Pascal/ - gpc-19990118 PGP key on request - 6C 94 45 BE 28 A4 96 - 0E CC E9 12 47 25 82 75 *** Vote against SPAM! ********* http://www.politik-digital.de/spam/ ***

3 2

Re: Drop C-numbers and multiple compilations per file?
by Nick Burrett 25 Jul '99

25 Jul '99

Peter Gerwinski <peter(a)gerwinski.de> writes: > * Multiple compilations in one file. > > One source file contains one program and several units, or > an interface and an implementation module, or whatever which > has more than one `end.' (please note the dot). > > Hammering this behaviour of GPC into the GCC/EGCS backend is > a nontrivial change which I do not expect the GCC steering > committee to accept. (They already rejected much less > crucial changes.) Thus, removing this feature would > probably be necessary for full integration into GCC/EGCS. Can the `program'/`end.', `unit'/`end.' tokens be treated like C #ifdef, #endif pairs such that there must be an equal number of each within a source file otherwise it is flagged as an error. Perhaps the pre-processor should combine a program and it's units in much the same way as you would #include a file in C rather than leaving it up to the compiler. Nick.

1 0

I want to contribute
by Prana 24 Jul '99

24 Jul '99

To whom it may concern, I want to contribute my SVGA graphic library for the development of GNU Pascal. Anyone interested ? It has been made since 1996. Limitation : it will only run on Intel processors, since some functions were made with assembly. You can download the files from http://cyest.hypermart.net, but the version there was outdated. But it's still good. It works much better than Borland BGI engine, without requiring any BGI drivers at all. The compiled executable size is also smaller than that of Borland's with its BGI. It support up to High Color 1024x768 Please e-mail me if you're interested. I'll distribute the FULL source code, with, of course, new license. Prana

2 1

Re: GPC BUG: Problems in subrange definition if exponentiation operator used
by Frank Heckenbach 24 Jul '99

24 Jul '99

pdietrich(a)synstar.de wrote: > The exponentiation operator cannot be used in expressions for lower/upper bound > of subrange definitions. > The compiler issues two different error messages: > If a unary minus is present, the error message is "parse error before **", > otherwise it is "subrange bounds are not of the same type". > There is also a lot "extra semicolon" messages for lines following the "parse > error". The result of the `**' operator is of type real, so it can't be used where an integer is expected. For integer exponentiation, you can use the `pow' operator. (That's like `/' and `div' for real/integer division.) > program test (output); > > type > utriple = 0..2**24-1; > triple = -2**23..2**23-1; > begin > end. > > F:\PROJECTS\ISIDATA\isidef>gpc -c test6.pas > test6.pas:4: subrange bounds are not of the same type > test6.pas:5: parse error before `**' > test6.pas:5: extra semicolon There is a real bug in GPC in that expressions can't be used as lower subrange bounds. We'd like to fix it, but it turns out to be very difficult with the parser we're using, so if anyone has experience with parsers and could help here... As a workaround, you can declare a constant, so the following works: program test (output); const foo = -2 pow 23; type utriple = 0..2 pow 24-1; triple = foo..2 pow 23-1; begin end. Frank -- Frank Heckenbach, frank(a)g-n-u.de, http://fjf.gnu.de/ GPC To-Do list, latest features, fixed bugs: http://agnes.dida.physik.uni-essen.de/~gnu-pascal/todo.html

1 0

GPC BUG: Problems in subrange definition if exponentiation operator used
by pdietrich＠synstar.de 24 Jul '99

24 Jul '99

Hello, I found the following bug(s) in the Beta version of GPC. Best regards, Peter Dietrich -------------------- GPC-Version (gpc -v): Reading specs from C:\usr\local\lib\gcc-lib\i386-mingw32\2.8.1\specs gpc version 19990118, based on gcc-2.8.1 System-Version: Windows NT 4.0 on i386 Problem description: The exponentiation operator cannot be used in expressions for lower/upper bound of subrange definitions. The compiler issues two different error messages: If a unary minus is present, the error message is "parse error before **", otherwise it is "subrange bounds are not of the same type". There is also a lot "extra semicolon" messages for lines following the "parse error". The compiler is happy, if I compute the result of the expression myself, and substitute the computed result. The syntax is ok for the Digital VMS Pascal compiler. program test (output); type utriple = 0..2**24-1; triple = -2**23..2**23-1; begin end. F:\PROJECTS\ISIDATA\isidef>gpc -c test6.pas test6.pas:4: subrange bounds are not of the same type test6.pas:5: parse error before `**' test6.pas:5: extra semicolon

1 0

Re: GPC on BeOS (was: Re: Cool :) GPC on BeOS :)...)
by Frank Heckenbach 23 Jul '99

23 Jul '99

Orlando Llanes wrote: > On Thu, 22 Jul 1999, Frank Heckenbach wrote: > > (OT: Linux also has sound and since version 2.2 also graphics > > built-in if you select them...) > > But the only standard way of accessing sound in Linux costs $$$ :) > (OpenSound, open, cha! Right!). Yes, this naming is misleading, but at least for me, the default (free) drivers work fine. > > So, you could just get the GCC and GPC sources and try to build it. > > Perhaps it will just work, or there will be only minor issues that > > can easily be solved. And if you succeed, you can upload the binary > > to Agnes. :-) > > I'd love to :) I'll post as soon as I have any results. I wish I had a > Mac Performa 6360 (6460? I forget which one), it runs the PowerPC version > of BeOS (BeOS is on a hybrid CD), would be cool to have a PPC GNU Pascal > with assembler and everything :) Good luck! Frank -- Frank Heckenbach, frank(a)g-n-u.de, http://fjf.gnu.de/ GPC To-Do list, latest features, fixed bugs: http://agnes.dida.physik.uni-essen.de/~gnu-pascal/todo.html

1 0

Re: GPC on BeOS (was: Re: Cool :) GPC on BeOS :)...)
by Frank Heckenbach 23 Jul '99

23 Jul '99

Orlando Llanes wrote: > I have BeOS 3.x and 4.5, they are both POSIX compatible far as I > know. To me, BeOS looks like Linux with graphics and sound capabilities > built-in. (OT: Linux also has sound and since version 2.2 also graphics built-in if you select them...) > I don't know about sh (though it uses a "Terminal" program which is > prolly compatible), but it does have a version of GCC with it, so all > tools should be in place (4.5 is better than 3.x because GCC, and its > tools, is in the path). So, you could just get the GCC and GPC sources and try to build it. Perhaps it will just work, or there will be only minor issues that can easily be solved. And if you succeed, you can upload the binary to Agnes. :-) Frank -- Frank Heckenbach, frank(a)g-n-u.de, http://fjf.gnu.de/ GPC To-Do list, latest features, fixed bugs: http://agnes.dida.physik.uni-essen.de/~gnu-pascal/todo.html

2 1

Re: GPC on BeOS (was: Re: Cool :) GPC on BeOS :)...)
by Frank Heckenbach 22 Jul '99

22 Jul '99

Orlando Llanes wrote: > Maybe someone oughta make a "real" BeOS port of GPC :) I don't know if > it's possible on version 3.x of BeOS since it does not use the COFF format > (but Be is not going to support BeOS 3 for long anyway). But the weird > thing is that BeOS 3 has a GCC compiler. I think that the OS specific > calls use a lib anyway, so it *might* be a simple matter of recompiling > GPC. I don't know if BeOS' GCC is capable of creating GUI apps, if it is, > GPC can easily do so as well. All it takes is a port of the binutils (i.e., as and ld), a GCC backend (the harder part is the processor specific part, and this, of course, exists already for the other OS's), and some standard libraries, in particular a more-or-less POSIX compatible libc. I don't know if BeOS is (or aims to be) POSIX compliant, but even if not, it is possible to write a POSIX layer on top of the system, like e.g. DJGPP which provides a very faithful and extensive POSIX layer on top of Dos which is about as far from POSIX as possible. ;-) If GCC works on BeOS, then there probably are already binutils and the backend, and also the libraries unless the compiler is used only for non-standard programs, so indeed it might be a simple matter of recompiling. (Compiling natively on BeOS requires some more tools like sh, make, sed, etc., but if they don't exist and can't be compiled, it's also possible to cross-compile GPC for BeOS from, say, a Linux host.) Frank -- Frank Heckenbach, frank(a)g-n-u.de, http://fjf.gnu.de/ GPC To-Do list, latest features, fixed bugs: http://agnes.dida.physik.uni-essen.de/~gnu-pascal/todo.html

2 1

Pascal Bug
by Ricky W. Butler 22 Jul '99

22 Jul '99

The following program: program bug(input,output); var x : real; begin x := 0.5000000000000E+00; writeln(' x = ',x:10:3); end. incorrectly prints out x = 5.000 It should print out x = 0.500 This program was compiled with GNU Pascal version 19990118, based on gcc-2.8.1, compiled Jun 15 1999 09:46:03. gpc --verbose Reading specs from /usr/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/specs gpc version 19990118, based on gcc-2.8.1 ld -m elf_i386 -dynamic-linker /lib/ld-linux.so.2 /usr/lib/crt1.o /usr/lib/crti.o /usr/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/crtbegin.o -L/usr/lib/gcc-lib/i686-pc-linux-gnu/2.8.1 -lgpc -lm -lgcc -lc -lgcc /usr/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/crtend.o /usr/lib/crtn.o /usr/lib/crt1.o(.text+0x36): undefined reference to `main' Rick Butler

2 2

Re: Cool :) GPC on BeOS :)...
by Frank Heckenbach 21 Jul '99

21 Jul '99

Dr A A Olowofoyeku wrote: > > > they have a port of GPC 2.1 :) > > > > 8-) Cool! Since we do not have v2.1 ourselves yet, perhaps we > > should ask them for a copy of their source? ;-) > > They do seem to have a port of GPC v2.1 (or so they claim). Unfortunately, the latter... In src/Read me!, it says: GNU Pascal for BeOS gpc-19980830.tar.gz This is, as you know, of course, not 2.1, but a beta version more or less close to the upcoming 2.1... I suppose we should point out this difference to them, as well as the fact that distributing the binary without the source, as they do, is illegal according to the GPL. I know that Be has violated the GPL in another case, but perhaps some day they'll learn... Frank -- Frank Heckenbach, frank(a)g-n-u.de, http://fjf.gnu.de/ GPC To-Do list, latest features, fixed bugs: http://agnes.dida.physik.uni-essen.de/~gnu-pascal/todo.html

4 3