This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Fix for "libgloss uses STARTUP directives consistently"
- From: cgd at broadcom dot com
- To: fnf at intrinsity dot com
- Cc: fnf at public dot ninemoons dot com,gcc-patches at gcc dot gnu dot org,echristo at redhat dot com
- Date: 14 Dec 2003 19:11:21 -0800
- Subject: Re: Fix for "libgloss uses STARTUP directives consistently"
- References: <20031214212639.7B1A02F2A26@beeville.vert.intrinsity.com>
At Sun, 14 Dec 2003 15:26:39 -0600 (CST), fnf@intrinsity.com wrote:
> I'd like to revisit this old patch since it is still required to build
> gcc correctly when you do separate builds and installs of binutils,
> gcc, gdb, and newlib (I.E. NOT an integrated tree where libgloss
> sources can be found).
It's been a while since i thought about this, and i don't really have
much time to think about it now, but wasn't the problem that some
files, e.g. idt32.ld, did use STARTUP directives, and some,
e.g. idt.ld did not? IIRC, the check was to see if idt.ld did, since
at the time it was made to do so, all the rest were made to do so
also, i.e., if does, all do.
I think if you want to make this more robust, you should probably
check for STARTUP's presence in all of the linker scripts, i.e., don't
just assume that if one has it, all do.
> * configure.in (gcc_cv_mips_libgloss_startup): When checking the
> installed mips linker scripts, match all installed copies, not just
> "idt.ld", which might not even be installed. For example,
> mipsisa32-elf newlib installs only idt32.ld, idt64.ld, and cfe.ld.
> * configure: Regenerate.
You've committed the sin of making a ChangeLog information contain
useful information^W^Wthe rationale for the change. You're not
supposed to do that! 8-)
chris