PING^2: resubmitted IRA improvement patches

[Vladimir Makarov]

  On 10/04/2010 11:41 AM, Richard Guenther wrote:
> On Thu, Sep 30, 2010 at 10:51 PM, Vladimir Makarov<vmakarov@redhat.com>  wrote:
>>
>> I can not get an approval of the following patches since middle of August.
>>
>> Jeff Law started a review of the 1st version of these patches in July
>> (before their rework to resolve conflicts with big Bernd's patch) but he can
>> not continue this work for some personal reasons.
>>
>> Although it would be more helpful for me to get a maintainer status for my
>> code to avoid such situations in the future.
> The patches are ok for trunk unless you hear otherwise til this Friday
> (that's basically a RM ok, as a middle-end maintainer I simply trust
> that you understand your own code).
>
> I do find the current situation
> less-than optimal as well and suggest that we either find more
> register allocation reviewers (I see a bunch of people there, but none
> of them is very active on this list) or consider promoting you to
> full maintainership.  CCed Mark to eventually raise this issue on
> the steering committee list.
>
> I hoped that eventually Bernd or Matz would have chimed in for
> a review.  That said, I'd rather have the patches in early to have
> time to deal with possible fallout than to delay them even longer.
Thanks, Richard.

I've just had a talk with Kenny.  He was kind to review the patches (3 
of them because the 1st one is actually approved).

I think it is a good idea to commit the patches on Friday (earlier than 
later).  Although I don't expect major disruption after committing I can 
not guarantee it because it is the most machine-dependent part of GCC 
and having it earlier on the trunk could save us some time.

I'll commit it on Friday unless I have major no from Kenny before 
Friday.  I am going to promptly fix comments from Kenny's review after that.

Probably, I should have CCed the patches to other reviewers.  I'd 
definitely do this if I knew what situation I came into (no final 
reveiwer of the patches after some big and active discussion with Jeff 
since June).

I did not expect that other reviewers did not know about my work because 
I did it on branch.  I myself always follow and check the branches which 
in my area of expertise.

I think it would be nice to have some policy to avoid such situations in 
the future.