This is the mail archive of the
mailing list for the GCC project.
Re: Question on -Werror usage in Makefiles...
- From: Eric Botcazou <ebotcazou at libertysurf dot fr>
- To: Andreas Jaeger <aj at suse dot de>
- Cc: "Zack Weinberg" <zack at codesourcery dot com>,gcc-patches at gcc dot gnu dot org,"Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>
- Date: Mon, 3 Nov 2003 14:22:45 +0100
- Subject: Re: Question on -Werror usage in Makefiles...
- References: <firstname.lastname@example.org> <email@example.com> <firstname.lastname@example.org>
> Then please send a patch for e.g. adding it to config/sparc/t-sol2.
Sorry, I tested on my locally modified files...
It turns out that, with an unmodified tree, adding
# mprotect not declared (see config/sol2.h)
_trampoline.o-warn = -Wno-error
# pthreads initializer problem
unwind-dw2.o-warn = -Wno-error
to config/sparc/t-sol2 doesn't work, supposedly because libgcc.a is built by
a separate script (mklibgcc).
> If we figure out that we cannot get some platforms to bootstrap with
> this change, then let's revert it. But let's give it a try, I see the
> benefits more than the problems.
I certainly don't question the long-term benefits. I simply don't understand
why a patch that had been on hold for a while (covering stage1/stage2)
because of known potential issues was merged during stage3.
Here's the only fix I can come up with (I don't know the makefile machinery
very well), but it is ugly because we now have '-Werror -Wno-error' on the
2003-11-03 Eric Botcazou <email@example.com>
* config/sparc/t-sol2: Add -Wno-error to TARGET_LIBGCC2_CFLAGS.
RCS file: /cvs/gcc/gcc/gcc/config/sparc/t-sol2,v
retrieving revision 1.11
diff -u -r1.11 t-sol2
--- config/sparc/t-sol2 16 Dec 2002 18:21:59 -0000 1.11
+++ config/sparc/t-sol2 3 Nov 2003 13:15:19 -0000
@@ -21,4 +21,4 @@
# routines in crtstuff.c.
CRTSTUFF_T_CFLAGS = -fPIC
-TARGET_LIBGCC2_CFLAGS = -fPIC
+TARGET_LIBGCC2_CFLAGS = -fPIC -Wno-error