This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Using object lists for libf2c overflows command line length on AIX.
- From: Alexandre Oliva <aoliva at redhat dot com>
- To: Toon Moene <toon at moene dot indiv dot nluug dot nl>, dje at watsom dot ibm dot com dot redhat dot com
- Cc: gcc-patches at gcc dot gnu dot org
- Date: 31 Dec 2001 04:45:15 -0200
- Subject: Re: Using object lists for libf2c overflows command line length on AIX.
- Organization: GCC Team, Red Hat
- References: <E16HrpM-00004p-00@laptop.moene.indiv.nluug.nl>
On Dec 22, 2001, Toon Moene <toon@moene.indiv.nluug.nl> wrote:
> The attached patch, which you applied on December 4th, at 09:53 UTC,
> has the unwanted effect of aborting the bootstrap on AIX due to
> command line overflow (it probably has a similar effect on other
> OS's, but I haven't received complaints yet).
What is the exact command that exceeds the command-line length. I'd
expect libtool to correctly split any excessively-long command lines,
but it might be missing this splitting in certain cases, and I don't
have easy access to systems with low command-line length limitations
to verify at the moment. I think I sent e-mail to David asking
whether it really broke AIX again, or whether it just caused linking
to become slower because of the multi-phase linking, but I don't
recall having got an answer. I'd rather keep the patch in, because it
causes the static version of $(LIBG2C) to be created with non-PIC
code, which makes it a bit faster. Using intermediate convenience
libraries causes the non-PIC files to be unused, even if they're
created.
I'd rather fix the bug in libtool, if there is one, than reverting the
patch. But I need to know what the exact problem is first.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist *Please* write to mailing lists, not to me