I guess I don't understand the meaning of 'regression' actually.

Anyway, I see some of my help files updates were considering 0.53, like the unification of sliders, so yeah, even though there's not much new, there it is :) 

Em dom., 16 de out. de 2022 às 13:47, Alexandre Torres Porres <porres@gmail.com> escreveu:
What are the regressions? I could only see bug fixes. 

Em sáb., 15 de out. de 2022 às 17:01, IOhannes m zmölnig <zmoelnig@iem.at> escreveu:
Am 14. Oktober 2022 18:13:00 MESZ schrieb Alexandre Torres Porres
<porres@gmail.com>:
>Em sex., 14 de out. de 2022 às 12:12, IOhannes m zmölnig <zmoelnig@iem.at>
>escreveu:
>
>> Am 14. Oktober 2022 15:32:42 MESZ schrieb Antoine Rousseau <
>> antoine@metalu.net>:
>> >has been made obsolete by the
>> >recent rework of the interface to GUI
>>
>> Speaking of which: I think a bugfix release 0.52.3 should probably *not*
>> contain the "rework of the interface to GUI", which is a rather big
>> change...
>>
>
>Why not? We're talking about things that had already been merged before,

So what?

i'm not really sure i understand what you mean exactly with "have been
merged before".
the fact that the PRs have already been merged into master?
how is that related to "bugfix release"?


>right? It seems to me most changes are "under the hood" and not proper new
>features.


And all of them include regressions.


imho, a "bugfix release" is a promise to the user that says "nothing has
changed with respect to the previous release, except we fixed these
annoying bugs".
so i worry not about "features" but about "anything not a bugfix".

and with 0.52-3, plenty of things have changed (and a few annoying bugs
have been fixed as well).
~70 C-files have been modified, and 3 header files have been modified
(suggesting additions to the API)

so i just fail to understand why we can't just call it "0.53-0", as this
clearly is beyond a "bugfix" release.


i'm probably highly influenced by semantic versioning [1], which is
targetting versioning for libraries (API/ABI compatibility) rather than
applications, but i think the model is so great that it can be applied
to any software (that is not released very frequently).

in any case: i'm not the one who is deciding on the versioning.


gfmdsa
IOhannes


[1] https://semver.org/
_______________________________________________
Pd-dev mailing list
Pd-dev@lists.iem.at
https://lists.puredata.info/listinfo/pd-dev