[QGIS-Developer] Re-discussing the QGIS release schedule - in combination with the quarantine rule for LTR versions

[Karsten Tebling]

Hello,

I just wanted to share my thoughts from the users-perspective:

Whenever there is a LTR-Version of a software I use that for production 
environments. For QGIS however, I always have both the LTR and the 
current version installed. The reason for this is my personal experience 
with bug reporting in QGIS. In the past I submitted a few bug reports, 
one of them was patched within a couple of days, but it took two months 
until it was released in the LTR, while it was released within the 
current branch within one month. Afaik changes have to be approved for 
the LTR-release, so it may take more time. Another bug I reported three 
weeks ago was also fixed within days, but the backport release went 
stale and the backport was closed by the qgis-bot. I don't know if 
things will improve if qgis-devs who don't seem to have time to review 
backport-changes, need to spent even more time they don't have on manual 
testing.

Right now I use LTR for daily work and the current version if I 
encounter errors, if LTR is being released only every few months I'm 
gonna start using the current version and only switch to the LTR version 
if I encounter errors, because it most certainly has less bugs in the 
functions that exists in both versions and for production environments I 
prefer less bugs.

Cheers

Am 28.02.2023 um 08:55 schrieb Andreas Neumann via QGIS-Developer:
> Dear QGIS devs,
>
> The QGIS release schedule for LTR versions was recently "thinned out", 
> as part of a decision to introduce "manual testing" prior to release - 
> see QEP 239: 
> https://github.com/qgis/QGIS-Enhancement-Proposals/issues/239 - see 
> also the release schedule at 
> https://www.qgis.org/en/site/getinvolved/development/roadmap.html
>
> Quote from the QEP 239: "LTR releases will no longer have monthly 
> patch releases, but instead a 4 months cycle releases, coincident with 
> the release of the stable version".
>
> I can understand the reasoning behind that decision, because "manual 
> testing" is a lot of work and it was also heard that many users don't 
> install patch releases too often.
>
> However, the situation is, that there is also a "quarantine rule" - 
> which is not mentioned in QEP 239 - but it helps to prevent untested 
> patches to end up in LTR versions, by delaying the backports until the 
> backport was first tested in the non-LTR stable release. There had 
> been a number of examples where this quarantine rule helped prevent 
> regressions in the LTR version introduced by backports in the past. 
> So, I think this quarantine rule is useful to have.
>
> However, it doesn't match well with the decision to "thin out" the 
> release schedule of the LTR version. There can be situations where a 
> user will have to wait 4-5 months, until a backport ends up in an LTR 
> release, which is a rather long time. We should bring this down to 2 
> months, like in the past.
>
> My proposal is to "revisit" the decision of the "thinned out" release 
> schedule and only "thin out", 6 months after a version became LTR.
>
> Any thoughts? Especially from the commercial support providers? How 
> would your customers react to the fewer patch releases?
>
> Thank you for the discussion,
> Andreas
>
> --
> Andreas Neumann
> QGIS.ORG <http://QGIS.ORG> board member (treasurer)
>
> _______________________________________________
> 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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-developer/attachments/20230301/727dd82d/attachment.htm>