Re: [Qemu-devel] [PATCH] MAINTAINERS: fix pattern for PC

[Laszlo Ersek <lersek@redhat.com>]

On 06/29/15 12:42, Peter Maydell wrote:
> On 29 June 2015 at 11:23, Laszlo Ersek <lersek@redhat.com> wrote:
>> On 06/29/15 12:00, Peter Maydell wrote:
>>> On 29 June 2015 at 10:54, Michael S. Tsirkin <mst@redhat.com> wrote:
>>>> Looks like hw/i386/ does not work, hw/i386/*
>>>> seems to work better.
>>>
>>> This is kind of vague... The documentation at the top
>>> of MAINTAINERS says the difference is that "hw/i386/"
>>> means "all files in and below hw/i386/", whereas
>>> "hw/i386/*" means "all files in hw/i386, but not below"
>>> (so won't match anything in hw/i386/kvm/ or hw/i386/xen/).
>>> Is this the effect you're trying to achieve? It would
>>> be nice to mention the symptoms of the problem this patch
>>> is fixing in the commit message...
>>
>> Assume I format the patch series:
>>
>>   http://lists.nongnu.org/archive/html/qemu-devel/2015-06/msg06677.html
>>
>> into a single file, with
>>
>>   git format-patch --notes --cover-letter --numbered --stdout
>>
>> and then run scripts/get_maintainer.pl on the resultant patch series file.
>>
>> Before this patch, I get:
>>
>>   Paolo Bonzini <pbonzini@redhat.com> (maintainer:X86)
>>   Richard Henderson <rth@twiddle.net> (maintainer:X86)
>>   Eduardo Habkost <ehabkost@redhat.com> (maintainer:X86)
>>   "Michael S. Tsirkin" <mst@redhat.com> (supporter:PC)
>>
>> Since Michael was listed at the bottom of that list, I didn't CC him. (I
>> wanted to give him a breather after my many PXB iterations.)
>>
>> Turns out that wasn't a good choice. With this patch for MAINTAINERS in
>> place, the script reports Michael at the top:
>>
>>   "Michael S. Tsirkin" <mst@redhat.com> (supporter:PC)
>>   Paolo Bonzini <pbonzini@redhat.com> (maintainer:X86)
>>   Richard Henderson <rth@twiddle.net> (maintainer:X86)
>>   Eduardo Habkost <ehabkost@redhat.com> (maintainer:X86)
>>
>> Maybe I should have considered something else than just the ordering of
>> the names in the list, not sure...
>>
>> Hm, yes, this is at least partly (if not fully) my fault. MAINTAINERS says
>>
>>         S: Status, one of the following:
>>            Supported:   Someone is actually paid to look after this.
>>            Maintained:  Someone actually looks after it.
>>
>> I didn't realize this distinction, and I also didn't realize that
>> "supporter" and "maintainer" were derived directly from "Supported" and
>> "Maintained".
>>
>> It would be helpful if developers with more jurisdiction (according to
>> the Supported / Maintained / Odd Fixes classification) were listed
>> higher in the output.
> 
> Mmm. If this is what we're trying to fix then messing with
> our MAINTAINERS file seems like the wrong thing.
> 
> Personally I think that the best approach would be just to
> cc everybody that get_maintainers.pl says is a maintainer
> or supporter; they're in the file because they *want* this
> email, after all...

Are you sure? See commit c6561586.

... Hm, although, that's specifically for the case when the MAINTAINERS
file has failed to help. So you can indeed say that whenever MAINTAINERS
matches, those people *all* want to be Cc'd.

> If your patchset touches two areas then
> that doesn't mean it's OK to drop a 'maintainer' email for
> area 2 just because area 1 happens to have two or three
> 'supporter' emails listed.

I didn't omit some Cc's because of area mismatches. I just wanted to
avoid spamming people, and took only the top of the list (Paolo). In the
future I'll include everyone. (Hopefully that won't lead to all people
on the Cc list thinking that someone else on that long Cc list should
review the patch...)

Thanks
Laszlo