GNU bug report logs - #64151
[PATCH] etc: Stop making sendemail behave strangely.

Previous Next

Full log

View this message in rfc822 format

From: Maxim Cournoyer <maxim.cournoyer <at> gmail.com>
To: Liliana Marie Prikler <liliana.prikler <at> gmail.com>
Cc: Ludovic Courtès <ludo <at> gnu.org>, Christopher Baines <mail <at> cbaines.net>, 64151 <at> debbugs.gnu.org
Subject: [bug#64151] [PATCH] etc: Stop making sendemail behave strangely.
Date: Sat, 01 Jul 2023 12:03:05 -0400

Hi Liliana,

Liliana Marie Prikler <liliana.prikler <at> gmail.com> writes:

> Am Freitag, dem 30.06.2023 um 23:03 -0400 schrieb Maxim Cournoyer:
>> > There's nothing wrong with automation per se, but you are confusing
>> > automating your own process knowingly with automating someone
>> > else's process without their knowledge or permission.  I'd also
>> > argue that your approach doesn't maximize etc/teams.scm, but rather
>> > makes it exhibit the weirdest behaviours imaginable by applying it
>> > blindly.
>> 
>> What is weird?  People opt to be in a team to be notified; the
>> default git configuration when submitting patches causes the
>> submission to notify them when appropriate.  I don't understand how
>> that qualifies as as the "weirdest behaviour imaginable" ?
> It's about the sending end rather than the receiving end, really.  As a
> sender, if you have a series that "invades" the territory of several
> teams, each will get CC'd exactly on the patches that overlap with
> them.  I argue, that this is the worst possible configuration.
>
> For a recent example, Christopher sent a series that just renames ruby
> everywhere.  I got 08/24 because gnome-team receives changes to webkit.
> As a member of gnome-team I said LGTM, but I really only had that one
> mail to go of.  If I don't investigate the full series, Christopher
> could have introduced a wrong ruby here; either by accident (typo) or
> maliciously, and no you wouldn't catch on from that single mail.

Maybe we just need to agree that each patch should have a LGTM unless
the reviewer explicitly mentioned e.g. "LGTM for this patch and the
previous ones".  I believe that's already the way most reviewers
understand it but documenting this explicitly wouldn't hurt.  Then even
if you review a single patch with no context your LGTM can still be
useful for that single patch and not taken erroneously as an overall
badge of approval.

> Thus, at the very least, the cover letter ought to go to all teams who
> have major stakes in any particular patch.  But here's the second
> thing: gnome-team doesn't have major stakes in a minor patch out of
> twenty-four.
>
> With the current automation in place, users unknowingly and without
> ever intending to alert team attention to where it wouldn't be needed
> in a manner that leaves major context clues to be found in the aether.
> I don't think this "maximizes" the potential of teams in any way
> whatsoever.

I think patches touching overlapping scopes will be the edge case rather
than the common one, so I don't see it as a serious issue, especially if
we explicit the LGTM as suggested above.

What do you think?

-- 
Thanks,
Maxim

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.