This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Avoid unneeded TREE_CONSTANT_POOL_ADDRESS_P test
- From: Zack Weinberg <zack at codesourcery dot com>
- To: Roger Sayle <roger at eyesopen dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Sun, 01 Jun 2003 15:39:56 -0700
- Subject: Re: [PATCH] Avoid unneeded TREE_CONSTANT_POOL_ADDRESS_P test
- References: <Pine.LNX.4.44.0306011538220.15234-100000@www.eyesopen.com>
Roger Sayle <roger@eyesopen.com> writes:
> On Sun, 1 Jun 2003, Zack Weinberg wrote:
>> This is the right general idea, but...
>> this check is wrong; the TREE_CONSTANT_POOL_ADDRESS_P flag is
>> private to varasm.c. Anyhow, you don't need to look at it.
>
> Hi Zack,
>
> Many thanks. I wasn't sure whether the test was strictly needed
> or not, and the use of TREE_CONSTANT_POOL_ADDRESS_P in integrate.c
> seemed to hint that it wasn't completely private to varasm.c.
The appearance in integrate.c is a layering violation, but it's
unavoidable right now and I'm not particularly motivated to
restructure the RTL inliner, since it's going to die when tree-ssa
is merged.
zw