Re: git: a9d9d3a42723 - main - lang/python: Revert "add bytecode trigger"

On Tue, Feb 28, 2023 at 08:32:22PM +0100, Matthias Andree wrote:
> Am 28.02.23 um 08:54 schrieb Baptiste Daroussin:
> > On Mon, Feb 27, 2023 at 09:05:09PM +0000, Matthias Andree wrote:
> > > The branch main has been updated by mandree:
> > > 
> > > URL: https://cgit.FreeBSD.org/ports/commit/?id=a9d9d3a4272303bf226b9deb55c42303e4fcebdc
> > > 
> > > commit a9d9d3a4272303bf226b9deb55c42303e4fcebdc
> > > Author:     Matthias Andree <mandree@FreeBSD.org>
> > > AuthorDate: 2023-02-27 21:01:32 +0000
> > > Commit:     Matthias Andree <mandree@FreeBSD.org>
> > > CommitDate: 2023-02-27 21:04:52 +0000
> > > 
> > >      lang/python: Revert "add bytecode trigger"
> > >      This reverts commit c17ddfbf66e2801ec620d49979aca3d7077d7002.
> > >      This causes breakage on several ports, and the next iteration
> > >      requires a full exp-run. See:
> > >      Differential Revision: https://reviews.freebsd.org/D34739
> > 
> > This is not how this is supposed to be done! while you have a point about the
> > issue and the full exp-run, the issue is easily fixable, you should have let
> > time to the committer to revert by himself, by asking him to revert instead of
> > abruply reverting!
> > 
> > Best regards,
> > Bapt
> 
> 
> Am 28.02.23 um 06:44 schrieb Tobias C. Berner:
> > Moin moin
> >
> > So you reverted the first major change to the python framework in ages
> > due to a leaf-port breakin non-fatally?
> > This looks a little bit like overstepping the bounds of good taste a bit
> to me.
> 
> Baptiste, Tobias,
> 
> First, let's get the timeline straight:
> 
> Feb 15, Charlie commits.
> Feb 18, Antoine claims the bug via mailing list.
> Feb 18, Mathieu asks why this was committed without prior -exp run,
>         via reviews.freebsd.org
> - silence -
> Feb 27, Matthias (yours truly) reverts that breaking change
> Feb 27 - 28, people pick at Matthias with emotional and non-technical
> reasons which are all besides the factual point.
> 
> There was NO acknowledgment of the bug, no hint that anyone was working, and
> we had broken sphinx and breathe ports, so we needed to fix things.
> 
> Then, half of the messages on IRC, mailing list, and the reviews were also
> non-technical, attacking me instead of the bug in Charlie's stuff and
> revealing who was getting touchy.
> 
> All in all, can we now please stop confusing cause and effect, and can those
> who have been touchy now please grow up?

I am perfectly aware of the timeline, and the reaction from Antoine and Mathieu.
They raised the issue and yes we would have expected an ack (at the very least)
from Charlie! However, both seemed to consider this was not worth a revert
otherwise they would have asked for the revert.

"Hey, can you quickly fix the following issue please, or revert until you can
come with a proper fix"

This is it should be asked.

On Charlie's side, yes he should have a been reactive, yes he should have
acknowledged the issue, his interractions should have been way better and he
needs to learn about. I didn't react on it because I felt this was being dealt
with somehow by others.

On the other side I saw your revert without prior benevolent message (like the
example above), such action is and as far as a I know has always been considered
a toxic behaviour not acceptable in the project. (yes it did happen in the past
and iirc always with the same reaction from portmgr, either public or private
depending on the cases).

It can only happen in rare cases in particular breakage that deeply impact the
package building for example.

This is why between a technical issue which clearly deservers lot of improvement
in the community interraction from Charlie and an aggressive revert, I chose to
react on the aggressive revert.

Best regards
Bapt