This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PR debug/42896] introduce value uids to stabilize value compares
- From: Richard Guenther <richard dot guenther at gmail dot com>
- To: Alexandre Oliva <aoliva at redhat dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Wed, 3 Feb 2010 15:55:58 +0100
- Subject: Re: [PR debug/42896] introduce value uids to stabilize value compares
- References: <oreil8vyxw.fsf@livre.localdomain> <84fc9c001001300229r6eb68905od81c78515e560f4@mail.gmail.com> <or636jw4hn.fsf@livre.localdomain> <84fc9c001001301542r24eff387yf31acc6103c00195@mail.gmail.com>
On Sun, Jan 31, 2010 at 12:42 AM, Richard Guenther
<richard.guenther@gmail.com> wrote:
> On Sun, Jan 31, 2010 at 12:35 AM, Alexandre Oliva <aoliva@redhat.com> wrote:
>> On Jan 30, 2010, Richard Guenther <richard.guenther@gmail.com> wrote:
>>
>>> Any reason to make uid signed (I suppose because decl uids are signed)?
>>
>> Yup.
>>
>>> I suppose you initialize the cselib
>>> uid with next_decl_uid to avoid clashes between them?
>>
>> Yup.
>>
>>> Is that necessary for this particular patch
>>
>> Nope. ?The only difference is the arbitrary initial value. ?It has to be
>> chosen (and reset) one way or another, so we might as well set it in a
>> way that enables the bitmaps IIRC you suggested.
>>
>>> If so please remove that bits from this patch.
>>
>> I'd rather not waste time retesting another version of this patch just
>> to change an arbitrary value. ?What would we gain from such a change?
>
> The change is not to export next_decl_uid from tree.c via that function.
I have adjusted the patch accordingly, re-bootstrapped and tested
it on x86_64-unknown-linux-gnu with verifying it fixes the
non-bootstrap-debug bootstrap failure and installed it.
Richard.