This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: patch rereview requested for PRs 6860, 10467 and 11741
- From: law at redhat dot com
- To: Gabriel Dos Reis <gdr at integrable-solutions dot net>
- Cc: Roger Sayle <roger at eyesopen dot com>, Geoff Keating <geoffk at geoffk dot org>, gcc-patches at gcc dot gnu dot org
- Date: Sat, 20 Sep 2003 23:09:24 -0600
- Subject: Re: patch rereview requested for PRs 6860, 10467 and 11741
- Reply-to: law at redhat dot com
In message <m3u176rdux.fsf@uniton.integrable-solutions.net>, Gabriel Dos Reis w
rites:
>Roger Sayle <roger@eyesopen.com> writes:
>
>| Patches must be assumed innocent until proven guilty!
>
>I disagree.
>
>Software development is not about lawyering. If we go that way, then
>there are high chances that reviewers would have an unreasonble duty
>of proving that every patch is correct. That is wrong. Instead, a
>submit should explain why he believes the patch is correct. Argument
>that if the reviewer can't prove it is incorrect, then the patch is OK
>is not acceptable.
I agree with Gaby. Particularly so when I believe that there is no sensible
way to handle PARALLELs as independent hunks in LCM.
Fundamentally, I do not believe there is a placement method that will
work if one uses one expression of a PARALLEL to eliminate a later
redundancy which is outside a PARALLEL. I can't stress this enough.
If someone added this kind of capability, then I believe that person
made a huge mistake and the original patch probably needs to be reverted.
A PARALLEL with multiple expressions can/should only be used to eliminate
other identical PARALLELs.
Any patch which does anything different for LCM is wrong.
Note that handling PARALLELs in const/copy propagation is safe because
we do not have to do any code motions.
Jeff