Re: [Qemu-devel] Maintainers, please git-am -m

[Markus Armbruster]

Philippe Mathieu-Daudé <address@hidden> writes:

> On 2/8/19 8:30 AM, Markus Armbruster wrote:
>> Short story: please add
>> 
>>     [am]
>>             messageid = true
>> 
>> to your .gitconfig.
>> 
>> Long story.  git-am can add a Message-id: tag.  Looks like this:
>> 
>>     commit 335dbb5de1e98c4dc73590349f17bb2a4d72596c
>>     Author: Thomas Huth <address@hidden>
>>     Date:   Mon Feb 4 09:25:43 2019 +0100
>>     Commit:     Peter Maydell <address@hidden>
>>     CommitDate: Mon Feb 4 15:25:21 2019 +0000
>> 
>>         tests/docker/test-mingw and docs: Remove --with-sdlabi=2.0
>> 
>>         Patchew currently reports failures with the mingw docker test - this
>>         is due to --with-sdlabi=2.0 configure flag which does not exist 
>> anymore.
>>         Remove this remainder from the docker test and the docs now.
>> 
>>         Signed-off-by: Thomas Huth <address@hidden>
>>         Reviewed-by: Daniel P. Berrangé <address@hidden>
>>         Reviewed-by: Philippe Mathieu-Daudé <address@hidden>
>>         Tested-by: Philippe Mathieu-Daudé <address@hidden>
>>         Acked-by: Alex Bennée <address@hidden>
>> --->    Message-id: address@hidden
>>         Signed-off-by: Peter Maydell <address@hidden>
>> 
>> The Message-id identifies the patch e-mail.  It makes finding the review
>> thread easier and more reliable.  It's also a valid key on Patchew[*].
>
> It is sometimes confusing however, see:
> https://lists.gnu.org/archive/html/qemu-devel/2019-02/msg01992.html
>
> We have:
>
> Signed-off-by: Paolo Bonzini <address@hidden>
> Message-Id: <address@hidden>
> Signed-off-by: Paolo Bonzini <address@hidden>
>
> This is probably a special case, and eventually a Patchew limitation,
> but I first wondered who wrote this patch? Since the Message-Id is from
> Yang Zhong, is the Signed-off-by tag incorrect?
> Then I went thru the previous version and understood the author is
> indeed Paolo, but his patch was first sent by Yang Zhong, then he
> applied his own patch sent by Yang.
>
> And yet this isn't a pull request but a simple patch, so the Message-Id
> might not be very relevant there.

git-am appears to add a Message-Id only if none is present.

Ideally, people don't submit patches with Message-Id, and maintainers do
use git-am -m.  Given how losely we enforce our process, we're bound to
deviate from the ideal, just like for other patch submission details.
All I'm asking for is a best effort.

>> Sadly, not all of our commits don't carry it.  Here's how top committers
>
> 'committers' -> 'maintainers'?
>
> (Actually committers don't know their patch Message-Id before sending
> the patch).

I used "committers" in the sense of "whoever git-log shows in the
"Commit:" line.  I agree these are maintainers in our work flow.
Calling them maintainers would've been clearer.

>> have been doing recently[**]:
>> 
>>    with without
>>     264      21 Peter Maydell <address@hidden>
>>      65       2 Gerd Hoffmann <address@hidden>
>>      64       0 Laurent Vivier <address@hidden>
>>      64       0 Eric Blake <address@hidden>
>>      62       1 Eduardo Habkost <address@hidden>
>>      56       0 Markus Armbruster <address@hidden>
>>      29      72 Richard Henderson <address@hidden>
>>      24      24 Paolo Bonzini <address@hidden>
>>      10      31 Marcel Apfelbaum <address@hidden>
>>       0     107 Kevin Wolf <address@hidden>
>>       0     106 David Gibson <address@hidden>
>>       0      93 Michael S. Tsirkin <address@hidden>
>>       0      81 Aleksandar Markovic <address@hidden>
>>       0      69 Samuel Thibault <address@hidden>
>>       0      54 Alex Bennée <address@hidden>
>>       0      50 Thomas Huth <address@hidden>
>>       0      29 Anthony PERARD <address@hidden>
>>       0      28 Marc-André Lureau <address@hidden>
>> 
>> Not bad, but there's room for improvement.
>> 
>> 
>> [*] Try 
>> https://patchew.org/search?q=id%3A1549268743-18502-1-git-send-email-thuth%40redhat.com
>> 
>> [**] git-log master --no-merges --pretty=fuller --since '3 months ago' | 
>> gawk '/^Commit: / { if (c) h[c][m]++; c=$0; m=0 } /^    Message-[Ii]d: / { 
>> m=1 } END { for (c in h) if (h[c][0] + h[c][1] > 25) printf "%7d %7d %s\n", 
>> h[c][1], h[c][0], substr(c,13) }' | sort -nr
>>