Previous mail
Next mail
Unformatted/full headers
Overview 10 days
Subject
Date
Thread
Author
From: Frank Heckenbach
Subject: GPC alpha
Date: 22 Sep 2005, 16:29:05
Tom Schneider wrote:
> % cat t.p
> program t(output);
> begin
> writeln(output,'hello world');
> end.
> % /usr/local/gpc-20050331/bin/gpc -v
> Reading specs from /export/data/local/gpc-20050331/bin/../lib/gcc-lib/sparc-sun-solaris2.8/3.3.3/specs
> Configured with: /export/home/bryantd/gpc-build/../gcc-3.3.3/configure --enable-languages=pascal --prefix=/usr/local/gpc
> Thread model: posix
> gpc version 20050331, based on gcc-3.3.3
> % /usr/local/gpc-20050331/bin/gpc -g t.p
> % gdb t
> GNU gdb 5.0
> Copyright 2000 Free Software Foundation, Inc.
> GDB is free software, covered by the GNU General Public License, and you are
> welcome to change it and/or distribute copies of it under certain conditions.
> Type "show copying" to see the conditions.
> There is absolutely no warranty for GDB. Type "show warranty" for details.
> This GDB was configured as "sparc-sun-solaris2.9"...
> (gdb) run
> Starting program: /home/strawberry/toms/work/emptytest/t
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x189dc in _p_ReturnAddr2Hex ()
> (gdb) quit
> The program is running. Exit anyway? (y or n) y
>
> I see now that my sysadmin did this:
>
> > I had compiled gpc on sparky so that it would be "solaris-8-
> > compatible", and then installed it on strawberry. I guess I could try
> > compiling the whole thing on strawberry and see if it makes any
> > difference.
>
> but
>
> uname -aimnprsvX
>
> gives
>
> SunOS strawberry 5.9 Generic_117171-12 sun4u sparc SUNW,Sun-Blade-1000System = S
> unOS
>
> Could that have caused this effect? Sparky is Solaris 2.8 and
> Strawberry is 2.9. Why GDB said 2.9 I don't know ...
>
> ********************************************************************************
I'm not so familiar with Solaris version issues, but it would seem
correct to me.
You could try entering "disass" in gdb to get more information.
But actually I'd rather suspect the "moving things around". GCC is
rather sensitive to moving things around, especially if other
versions are installed on the system as well.
So to save us all some guessing, it might be easiest to restart the
installation from scratch:
- extract the original source archives,
- configure with the prefix it will finally be installed in,
- build,
- make install or make pascal.install-with-gcc (the latter won't
overwrite a previous gcc installation if the prefix or the gcc
version is different -- in your case, the prefix should be
different AFAIUI)
What you can move around/copy/link/rename then are the main
executables (those in .../bin), nothing else (in particular not any
of the files installed under .../lib or .../libexec).
> > > | strawberry 44% /usr/local/gpc/bin/gpc -v
> > > | Reading specs from /export/data/local/gpc/bin/../lib/gcc-lib/sparc-sun-solaris2.8/3.3.3/specs
> > > | Configured with: /export/home/bryantd/gpc-build/../gcc-3.3.3/configure --enable-languages=pascal --prefix=/usr/local/gpc
> > > | Thread model: posix
> > > | gpc version 20050331, based on gcc-3.3.3
> > > |
> > > | When I try to use /usr/local/bin/gpc-2.1 in my gpcc script
> > > | (/home/strawberry/toms/work/emptytest/gpcc2.1), I get:
> > > |
> > > | gpc1: Invalid option `-Wno-identifier-case-local'
> > > |
> > > | so I can't compile until this is fixed ...
> >
> > Uhm, fixed? gpc-2.1 is 3 years old, and we don't actually plan to
> > backport new features to it. If it's just this one (you don't
> > require other newer features), you can take out this option for 2.1
> > (which simply didn't exist back then).
>
> I'm confused. All I knew was that I got this 'Invalid option'
> message, which seemed unrelated and off the wall. But it could mean
> that `-Wno-identifier-case-local' is no longer a valid option?
>
> [...]
>
> I would never ask you to 'fix' an older version! That's wasting your
> time! Isn't this 20050331 a newer version?
I think there's some confusion here. Yes, 20050331 is new, but 2.1
is older (2002-05-10). So the option works with the new version, but
not with the old one, and so it's no surprise that you get an
"invalid option" message with the old version (2.1, which I
understand you tried to use above). As the old version will never be
"fixed", you can either take out the option when you compile with
the old version, or just not old the old version anymore ...
> > : # -Widentifier-case-local
>
> Please accept my apologies for that. It was material that went into
> the script when the issue first came up. I have removed it and tried
> to clarify the issue in the documentation.
Now it says: "My programs use text modules that come from various
places, so keeping the case consistent between all modules is not
possible." -- Well, `-Widentifier-case-local' compares the case only
within modules/units, so it might still work for you. Or do mean
include files instead of modules?
> > BTW, your quick jumping to conclusions that "[GPC] is too alpha"
> > when in fact it may just be an installation problem on your (or your
> > admin's) part, or claiming that an older version not supporting
> > newer options is something that needs to be "fixed"
>
> I am reporting a problem to you as best as I can. I'm not an expert,
> I do not know the guts of GPC. I did not know that an incorrect
> installation (if that's what happened) could cause such an effect as I
> have never seen that happen before. All I meant was that "this
> particular installation of a GPC alpha didn't work". GPC itself is
> clearly quite robust (thanks to you!), which is why I use it. The
> alpha GPC that we have produced a binary that crashed. I don't know
> why that may be. Alphas in general can crash - this is true for
> Mozilla.
Some quotes from two articles also referenced in the GPC manual
("Reporting Bugs"):
http://www.catb.org/~esr/faqs/smart-questions.html
: Good and Bad Questions
:
: Finally, I'm going to illustrate how to ask questions in a smart way
: by example; pairs of questions about the same problem, one asked in a
: stupid way and one in a smart way.
:
: [...]
:
: Stupid: I can't get the code from project foo to compile. Why is it
: broken?
: He assumes that somebody else screwed up. Arrogant of him.
:
: Smart: The code from project foo doesn't compile under Nulix version
: 6.2. I've read the FAQ, but it doesn't have anything in it
: about Nulix-related problems. Here's a transcript of my
: compilation attempt; is it something I did?
: He's specified the environment, he's read the FAQ, he's showing
: the error, and and he's not assuming his problems are someone
: else's fault. This guy might be worth some attention.
http://www.chiark.greenend.org.uk/~sgtatham/bugs.html
: "It doesn't work."
:
: Give the programmer some credit for basic intelligence: if the program
: really didn't work at all, they would probably have noticed. Since
: they haven't noticed, it must be working for them. Therefore, either
: you are doing something differently from them, or your environment is
: different from theirs. They need information; providing this
: information is the purpose of a bug report. More information is almost
: always better than less.
So the point is not that alphas can crash, or that you don't like
that particular option -- it's the way you phrase it. To quote the
latter paper, if gpc-20050331 really couldn't compile a hello-world
program, don't you think we'd have noticed by now? So claiming that
"it is too alpha" on these grounds is rather silly.
I can only recommend, to you and anyone else who feels they should,
to read those two articles [again or for the first time]. Various
parts of them seem revelant to many bug reports on this list ...
Frank
Previous mail
Next mail
Unformatted/full headers
Overview 10 days
Subject
Date
Thread
Author
| Author | Subject | Date |
|---|---|---|
| Tom Schneider | GPC alpha | 19 Sep 2005, 20:29:55 |
Note: This page contains information that does not originate from the owner of this web site, but from the authors of the mails archived. The owner of this web site is not responsible for the content of such information. Any use of that infomation requires the consent of the respective author.
Where WWW addresses (URLs) in the mails archived are marked as hyperlinks, this is only for the comfort of the reader. The content of the web pages linked to like this does not necessarily reflect the opinion of the owner of this web site or of the authors of the mails archived. The owner of this web site is not responsible for the content of such web pages. Those pages are explicitly not to be considered as part of the content of this page, but merely as references.
This page was created by Crystal 0.999 (Linux 2.4.27/i686).