This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Combine four insns

> I suppose Paolo could comment as to what he intended.  It seems pretty
> reasonable to me, though, to say that if you can approve dataflow
> changes, you can approve insertion of a call to the dataflow machinery.

That's arguably questionable, given that Paolo himself said that he actually 
could not approve the two lines.

> I don't want to see it get too procedural to get something checked in.
> I'd rather have a culture where it's not too hard to get things in, but
> where we are responsive to problems raised after the patch is checked
> in.  I'm less concerned about something going in than about people being
> unwilling to take something back out, or to fix something that's broken.
>  You have my assurance that if Bernd breaks something, and refuses to
> fix it, his CodeSourcery management will poke him with a sharp stick.

Thanks, but that wasn't really necessary given Bernd's responsiveness.

> Eric, did you object after Bernd checked in his patch?  He posted a
> commit message in:
> and the mail archives don't show up a follow-up from you.  That could be
> that it flowed in to the next month and I can't find it, though.  In any
> case, I think that if you object to an interpretation of an approval
> message it's appropriate for you to speak up and we can resolve it at
> that point.

No, I didn't, I only replied to the first message in the thread:
and stated again my disagreement with Bernd's approach.  But I'm not a fanatic 
either, the end result was reasonable so I stopped there.

> In any case, are you objecting to the change now?

No, it's in, let's keep it in.

> If so, what's your concern?

I was replying to Bernd's accusation of "attempting once again to block one of 
[his] patches".  Of the 3 problematic patches, I stepped down for the first 
one and didn't say anything when the second one was installed without 
(indisputable) approval.  Great attempts at blocking something.  The third 
one, yes, I tried to block it in its original form for the reasons already 

I think it's just responsible maintainership.  I'm often at the other end of 
the review table, I sometimes try to argue with the maintainer and in some 
cases was frustrated because I didn't understand at all why one of my patches 
was rejected.  I never called his position absurd or declared that the review 
process had failed.

I reviewed several patches from Bernd over the past year but, after the last 
two, I think I'm done with that for the few years to come. :-)  Therefore I'd 
suggest, if he agrees of course, that the SC promotes Paolo to whatever 
position it deems appropriate for him to be able to review patches touching 
all the RTL optimizers.

Eric Botcazou

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]