This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [rfa] fix c++/13693
- From: Mark Mitchell <mark at codesourcery dot com>
- To: Richard Henderson <rth at redhat dot com>
- Cc: gcc-patches at gcc dot gnu dot org, Jason Merrill <jason at redhat dot com>
- Date: Thu, 29 Jan 2004 23:35:44 -0800
- Subject: Re: [rfa] fix c++/13693
- Organization: CodeSourcery, LLC
- References: <20040130015324.GA17125@redhat.com>
Richard Henderson wrote:
While this shows up on tree-ssa at the moment and will *not* show
up on mainline, due to mips now implementing backend thunks, I
still think it's a bug.
The problem here is that we build a void target_expr in use_thunk.
This blows up later cause we can't copy VOIDmode values.
Does this seem the right fix?
We really just shouldn't call force_target_expr if the return value from
the thunk satisfies VOID_TYPE_P. I think you should just change
use_thunk not to call force_target_expr in that case and have
target_expr use my_friendly_assert to check that we're not trying to
make a VOID_TYPE_P TARGET_EXPR. If that works, that patch is certainly
OK for mainline.
Thanks,
--
Mark Mitchell
CodeSourcery, LLC
mark@codesourcery.com