May 1999 - Gpc

Re: Restricting posting to the GPC list? (was: SNN - Star Wars - WCLT)
by Frank Heckenbach 26 May '99

26 May '99

Peter Gerwinski wrote: > David James wrote: > > I take it from this that the GPC list doesn't restrict posting to > > members only. Since the SPAM/UCE merchants seem to be everywhere > > would it be a good idea to do this - I believe Majordomo has such a > > capability. > > This would have the drawback that the GPC mailing list could > no more serve as a place to call for help with GPC without > subscribing in the first place. > > Majordomo has some more capabilities that allow to check for > keywords first, so suscipious mails get sorted out and must be > approved by a moderator before further submission. Would it be possible to forward mails that are not from subscribers (and perhaps don't contain clear keywords like `GPC') to a moderator? Since most of the mails are from subscribers, this should eliminate all spam, without causing very much work for a moderator... Frank -- Frank Heckenbach, frank(a)fjf.gnu.de, http://fjf.gnu.de/ GPC To-Do list, latest features, fixed bugs: http://agnes.dida.physik.uni-essen.de/~gnu-pascal/todo.html

3 3

GPC Html Docs for ver 19990409
by GPC on Ken Linder's Linux Box 23 May '99

23 May '99

Hi All, I have just successfuly brought up my own little WWW server & have placed the html docs to the latest ver (19990409) on it. Also, I have a rather strange itch to scratch by adding information to the docs and sort of 'fleshing' them out. I will post the newest html docs on my server, along with diff -u files of the originating .texi files, a change log, and soon... tar'd & gzipped html documentation files. Take a look at the server at http://kc7rad.lvcm.com/gpc/index.html If you have any questions or comments, send them to pas(a)kc7rad.lvcm.com or ken(a)kc7rad.lvcm.com Please note that this server is under construction and you may hit a link that points to nowhere. These will be resolved shortly. Thanks! Ken Linder

1 0

Re: Problem exporting enumerated type from Extended Pascal module
by david＠tcs01.demon.co.uk 22 May '99

22 May '99

In message <19990522190631.C292(a)esmeralda.gerwinski.de> Peter Gerwinski writes: > > I can reproduce this bug. This is one more instance of the > "two-interface problem" (see http://fjf.gnu.de/gpc-todo_2.html: > baby.pas et al.). > > Thanks for the detailed report. (-: You're welcome ... > > Since I am currently working on related parts of the compiler > there is some hope to have it fixed in the next snapshot. It would be wonderful if it was - this is a showstopper for me in using Extended Pascal modules; I could probably move to units, but that's quite a lot of extra work as the existing code is in modules. Will the next snapshot be in the alpha subdirectory? Do you have any estimate of when it might be? -- David James mailto:david@tcs01.demon.co.uk

2 1

Problem exporting enumerated type from Extended Pascal module
by david＠tcs01.demon.co.uk 22 May '99

22 May '99

I think I'm seeing a problem in exporting an enumerated type from an Extended Pascal module. My system is RedHat Linux release 5.2 (Apollo) running kernel 2.0.36 on an i586. gpc -v reports as follows: ---------------------------------------------------------------------- Reading specs from /usr/lib/gcc-lib/i586-pc-linux-gnu/2.8.1/specs gpc version 19990118, based on gcc-2.8.1 -------------------------------------------------------------------- Below is a small module (dajmod2.p) and a program (daj2.p) which imports it. These have been cut down from a much bigger piece of code. When compiling daj2.p the compiler reports: -------------------------------------------------------------------- daj2.p: In function `program_Daj2': daj2.p:7: incompatible types in assignment daj2.p:8: invalid operands to binary = -------------------------------------------------------------------- 8<---- daj2.p ------------------------------------------------------ program daj2; import progtypes in 'dajmod2.p'; globalvars in 'dajmod2.p'; begin pgmtype:=prog_press; if (pgmtype=prog_press) then writeln('OK') else writeln('Error'); end. 8<--- dajmod2.p ----------------------------------------------------- module dajmod2 interface; export progtypes =(tprogtype, prog_none..prog_last); globalvars =(pgmtype); TYPE TPROGTYPE=(prog_none, prog_press, prog_last); VAR pgmtype:tprogtype; end. {of module header} module dajmod2 implementation; end. {of module implentation} -------------------------------------------------------------------- If I embed just the TYPE and VAR declarations from dajmod2 within daj2, the program compiles and runs correctly. -- David James mailto:david@tcs01.demon.co.uk

3 3

Re: Rhide and gpc
by Frank Heckenbach 22 May '99

22 May '99

Olivier PECHEUX wrote: > I have loaded the gpc (5/99). I use rhide and windoze. And I have > troubles: > When I compile, Rhide call > "gpc -o etape.exe --automake="-g" -g etape.pas -lm -lgpc" > > First, --automake="-g" aren't in the documentation and I have the > message > "ignoring AutoMake option -g"; This was changed in the last release of GPC. Instead of `--automake="-g" -g', you should now simply say `--automake -g' (i.e., no need to repeat the options for automake). > Is it possible to change the call? If you have the RHIDE source and can compile it, it should be easy to change the GPC invocation. Otherwise you'd have to ask Robert, the RHIDE author, to do it. > Second: with the -lgpc the program dosn't reconize the gpc librairie > (GPC_randomize is't reconize at all). But if I put "uses gpc" it does. > Why? > > Third: What does mean -lm? Is there an m librarie? There is a difference between libraries and units/modules. Libraries are originally a C concept -- they consist only of object code without any interface declarations (these are in header files in C), while Pascal units/modules consist of code and declarations. GNU Pascal can use libraries with `-l' switches, but this only provides the object code, not the interface declarations. E.g., the GPC Run Time System is in libgpc.a which is linked with the `-lgpc' switch. However, to make available the declarations, you have to use the unit GPC (which is in gpc.pas), as you noticed. (BTW, don't use GPC_Randomize in your programs. It's an internal routine. The compiler knows `Randomize' which will ultimately call GPC_Randomize, but is more portable. For `Randomize', you don't even have to use GPC, but for other declarations, you do.) Similarly, there's a math libarary (libm.a, -lm) which contains routines like sin and cos. For them, you don't need to use any unit, because the compiler knows about them. Actually, GPC knows already that it needs -lgpc and -lm, so one doesn't have to specify those two libraries explicitly (and this has been so for a few years at least). I don't know why RHIDE adds them, but that's unnecessary. To put it shortly, don't worry about libraries (i.e., `-lfoo' options). Just use the units you want, and you should be fine. This also holds for units like CRT which need additional libraries (curses), because they take care of the libraries automatically. > When I compile a librairie, I have an "undefinied reference to main" > message. For now I ignore it because it's work. What is your position? > What I have to do? I suppose you mean a unit. Indeed, it doesn't cause any problems except this error message, but you can give the `-c' option, and the message will disappear. (Perhaps in the future, you won't need `-c', but currently, you do.) Frank -- Frank Heckenbach, frank(a)fjf.gnu.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

info
by GPC Pascal on Linux User 22 May '99

22 May '99

info

1 0

Rhide and gpc
by Olivier PECHEUX 21 May '99

21 May '99

Hi, ----------------------- I have loaded the gpc (5/99). I use rhide and windoze. And I have troubles: When I compile, Rhide call "gpc -o etape.exe --automake="-g" -g etape.pas -lm -lgpc" First, --automake="-g" aren't in the documentation and I have the message "ignoring AutoMake option -g"; Is it possible to change the call? Second: with the -lgpc the program dosn't reconize the gpc librairie (GPC_randomize is't reconize at all). But if I put "uses gpc" it does. Why? Third: What does mean -lm? Is there an m librarie? ------------------- When I compile a librairie, I have an "undefinied reference to main" message. For now I ignore it because it's work. What is your position? What I have to do? ------------------- I inform also that there is a french maling list on the Pascal and on the gnu: http://www.onelist.com/subscribe/prog_pascal_gnu --------Publicite--personnelle--------------------------------------- ........ABC de la Programmation: ................http://www.multimania.com/opecheux Abonnez-vous aux listes francaises: - Assembleur: http://www.onelist.com/subscribe/prog_nasm - C : http://www.onelist.com/subscribe/prog_c_gnu - Pascal : http://www.onelist.com/subscribe/prog_pascal_gnu

1 0

Re: Restricting posting to the GPC list? (was: SNN - Star Wars - WCLT)
by Nick Burrett 20 May '99

20 May '99

Frank Heckenbach <frank(a)tim.gerwinski.de> writes: > Peter Gerwinski wrote: > > > David James wrote: > > > I take it from this that the GPC list doesn't restrict posting to > > > members only. Since the SPAM/UCE merchants seem to be everywhere > > > would it be a good idea to do this - I believe Majordomo has such a > > > capability. > > > > This would have the drawback that the GPC mailing list could > > no more serve as a place to call for help with GPC without > > subscribing in the first place. > > > > Majordomo has some more capabilities that allow to check for > > keywords first, so suscipious mails get sorted out and must be > > approved by a moderator before further submission. > > Would it be possible to forward mails that are not from subscribers > (and perhaps don't contain clear keywords like `GPC') to a > moderator? Since most of the mails are from subscribers, this should > eliminate all spam, without causing very much work for a > moderator... One solution, that is used by the egcs.cygnus.com mail server, is to modify the mail delivery agent to check addresses against MAPS RBL and ORBS. This is quite easy if your MDA is sendmail 8.9. Nick.

1 0

Re: SNN - Star Wars - WCLT
by david＠tcs01.demon.co.uk 19 May '99

19 May '99

In message <359DFE77.4AC9(a)erols.com> oho87(a)jboy.com writes: > StockNetworking News > [etc., etc., etc] I take it from this that the GPC list doesn't restrict posting to members only. Since the SPAM/UCE merchants seem to be everywhere would it be a good idea to do this - I believe Majordomo has such a capability. -- David James mailto:david@tcs01.demon.co.uk Special Stage Rally results archive <URL:http://www.tcs01.demon.co.uk/>

2 1

Re: GPC for AIX 4.3.1
by Frank Heckenbach 19 May '99

19 May '99

Marina Markus wrote: > stage1/xgcc -Bstage1/ -c -Wall -Wpointer-arith -Wmissing-prototypes > -Wmissing-declarations -O2 -g -O2 -I. -I../.. -I/usr/local/src/egcs-1.1.2/ > gcc/p/rts/../.. -I/usr/local/src/egcs-1.1.2/gcc/p/rts/../../config /usr/loc > al/src/egcs-1.1.2/gcc/p/rts/pexecute.c > In file included from /usr/local/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/egcs-2. > 91.66/include/sys/wait.h:49, > from /usr/local/src/egcs-1.1.2/gcc/p/rts/pexecute.c:38: > /usr/local/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/egcs-2.91.66/include/sys/sign > al.h:121: warning: `SIGEMT' redefined > rts-config.h:182: warning: this is the location of the previous definition > /usr/local/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/egcs-2.91.66/include/sys/sign > al.h:126: warning: `SIGSYS' redefined > rts-config.h:178: warning: this is the location of the previous definition > /usr/local/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/egcs-2.91.66/include/sys/sign > al.h:130: warning: `SIGURG' redefined > rts-config.h:186: warning: this is the location of the previous definition > /usr/local/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/egcs-2.91.66/include/sys/sign > al.h:164: warning: `SIGIOT' redefined > rts-config.h:190: warning: this is the location of the previous definition To get rid of the warnings, you can insert #include <signal.h> into rts/pexecute.c, before the line #include "rts-config.h" > make: Cannot find a rule to create target heap.o from dependencies. > Stop. I think the problem is that the RTS makefile uses "pattern rules" which your version of make does not understand. So you can either use GNU make, or replace %.o:%.pas by .SUFFIXES: .pas .pas.o: This should fix the problem. Both things will be fixed in the next snapshot. Frank -- Frank Heckenbach, frank(a)fjf.gnu.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