[QGIS-Developer] Stricter policy for LTR backport ? (was Re: Call for emergency 3.22 release)

[Even Rouault]

Hi,

I guess one point of improvement is to really implement a policy that 
was discussed in the past, that is that after let's say half the total 
period of support of the LTR only "critical" bugfixes are allowed into 
it, with critical being for example defined as:

- fix a regression introduced in a previous patch release ("patch" as in 
"major.minor.patch" version numbering scheme)

- fix a crashing bug

- fix data corruption

And also taking into account the complexity of the fix, how common the 
scenario is (a crashing bug in some "obscure" functionality has not the 
same severity as one at program opening), the risk of its side effects, 
how well covered it is by unit tests, etc.

This will always remain subjective, and there will always be the risk 
that a few-liner fix causes unanticipated issues. That's the reality of 
software development.

Perhaps half the period is quite restrictive, but maybe the last 3 
planned patch releases in the lifetime of a LTR should really enforce 
such super strict requirements.

Looking quickly at https://github.com/qgis/QGIS/commits/release-3_22, I 
can see this is far from being the case (and I should start by following 
my own advice as I can see a few commits of mine that definitely don't 
match the above criteria for critical bug fixes)

Even

Le 02/02/2023 à 17:19, Julien Cabieces via QGIS-Developer a écrit :
> Hi,
>
> I'm sorry that my PR led to a severe regression and thank you to the
> people who take care of fixing/releasing.
>
> I take the decision to merge this one because:
> - It was one month old
> - It has already been reviewed by someone else
> - It was not a critical modification (I believed...)
>
> I have to admit I rarely thoroughly review backport from bot because
> they are exactly the same as the one already reviewed. And I'd probably
> miss this kind of errors when reviewing too.
>
> If you have any advice on how to behave in the backport/review process,
> feel free to share.
>
> Regards,
> Julien
>
>
>> On Thu, 2 Feb 2023, 6:12 pm Alessandro Pasotti, <apasotti at gmail.com> wrote:
>>
>>   +1 for immediate release from me.
>>
>>   On a side note (because I sympathize thinking about what I would feel
>>   if it was a commit of mine) : Julien is doing an amazing job, a
>>   mistake can happen to anyone.
>>
>> 100% agree - it could easily have been any of us leading to this.
>>
>> I think the main lesson for us is "just" that we need to be extra careful with backports to ltr, and carefully assess whether every backport is really warranted or
>> not.
>>
>> Nyall
>>
>>   On Thu, Feb 2, 2023 at 7:38 AM Andreas Neumann via QGIS-Developer
>>   <qgis-developer at lists.osgeo.org> wrote:
>>   >
>>   > Hi Nyall,
>>   >
>>   > I agree, this is a pretty nasty one. Jürgen - can we release another release soon?
>>   >
>>   > Thank you Even for your reflections on the code patterns involved. I trust that the core devs will discuss and suggest improvements in our
>>   documentation/guidelines in order to avoid similar cases in the future.
>>   >
>>   > Thank you for your response.
>>   >
>>   > Andreas
>>   >
>>   > On 2023-02-02 02:19, Nyall Dawson via QGIS-Developer wrote:
>>   >
>>   > Hi PSC/list,
>>   >
>>   > I came across this horrible regression during bug hunting today:
>>   > https://github.com/qgis/QGIS/pull/51703
>>   >
>>   > If you load a project with any broken layers and then hover over the
>>   > layer in the layer tree, QGIS will instantly crash. It's a regression
>>   > caused by https://github.com/qgis/QGIS/pull/50256, and unfortunately
>>   > that PR was backported to 3.22 and accordingly the crash present in
>>   > the final release of 3.22.
>>   >
>>   > Given the extreme severity of this crash I believe we should be
>>   > pushing out another unplanned 3.22 patch release, as we cannot leave
>>   > the final 3.22 LTR with such a nasty regression in place.
>>   >
>>   > Nyall
>>   > _______________________________________________
>>   > QGIS-Developer mailing list
>>   > QGIS-Developer at lists.osgeo.org
>>   > List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
>>   > Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
>>   >
>>   >
>>   > _______________________________________________
>>   > QGIS-Developer mailing list
>>   > QGIS-Developer at lists.osgeo.org
>>   > List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
>>   > Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
>>
>>   --
>>   Alessandro Pasotti
>>   QCooperative:  www.qcooperative.net
>>   ItOpen:   www.itopen.it
>>
>> _______________________________________________
>> QGIS-Developer mailing list
>> QGIS-Developer at lists.osgeo.org
>> List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
>> Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer
>
-- 
http://www.spatialys.com
My software is free, but my time generally not.