On Thu, Sep 30, 2021 at 12:13 PM Fabio Valentini <decathorpe(a)gmail.com> wrote:
On Thu, Sep 30, 2021 at 11:17 AM Mikolaj Izdebski <mizdebsk(a)redhat.com> wrote:
>
> On Wed, Sep 29, 2021 at 1:51 PM Miro Hrončok <mhroncok(a)redhat.com> wrote:
> > Since the @java-maint-sig group is esentially non-responsive, I suggest we do
> > the following:
>
> Thanks for making this distinction - @java-maint-sig group is not the
> same as Java SIG.
> Some of the most active members of Java SIG (like me or OpenJDK
> maintainers) are not
> (and never were) members of @java-maint-sig.
That's true. The "old" Java SIG was never properly set up as a FAS
group. Its Wiki page has contained obsolete information for *years*,
and the list of members there hasn't been accurate for years, either.
Java SIG is an informal group. By definition, SIGs are informal groups
within Fedora Project. Therefore there is no formal membership of Java
SIG. Anyone interested in contributing to make Java in Fedora better
can be considered a member of Java SIG.
On the other hand, you were explicitly welcome to join the newly set
up group, which you never bothered to do, since you were pretending
@java-maint-sig evolved from Stewardship SIG, which was formed with a
goal of preventing unmaintained packages from being retired, which I
don't appreciate. My opinion is that retirement of unmaintained
packages is desired. I want Fedora Linux to be a high quality
distribution and I believe it's better to have fewer, but better
quality packages. This is the primary reason for me not joining
@java-maint-sig.
Besides that I'm not a big fan of collaborative package maintenance
groups such as @java-maint-sig. I prefer a model with a single primary
maintainer who owns the package, like a product owner - has the
authority to make technical decisions about the package. Similarly to
cathedral vs bazaar, both of which are valid software development
models.
that Modularity will solve all of humanity's problems, and
obviously
preferred to work alone, never offering help or feedback, unless
*maybe* when explicitly asked. Now that modules won't solve Java
I never claimed that modularity was perfect, nor that it was better
than traditional package maintenance. Modularity solves some important
problems, but introduces others.
From my PoV, the most important feature of modularity that I wanted to
take advantage of - building packages in a controlled, isolated
environment - is now implemented as Koji dynamic sidetags (BTW, I was
the original author of sidetag-koji-plugin). Another important feature
- private dependencies - is now solved by allowing bundling much more
freely and by exempting compat packages from the package review
process. Therefore I no longer see modularity as a good approach to
maintain default versions Maven and Ant - it could still be used for
alternative versions.
packaging either, you're back, and bad-mouthing all the work the
SIG
did while it was active to keep the shit from hitting the fan while
you were AWOL, which I find a bit rich.
I am back to maintaining non-modular Java packages because I want to
keep contributing to Fedora as a Java package maintainer and Fedora
engineering leadership decided that non-modular packages are strongly
preferred over modules. I don't disagree with that decision and I
respectfully obey it.
--
Mikolaj Izdebski