[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Frustrated package submitters?

>>>>> "MS" == Michael Schwendt <bugs michael gmx net> writes:

MS> I would prefer if those unnamed people contacted FESCO or
MS> fedora-extras-list directly. In case you picked yourself as the
MS> spokesman of that group, provide some details, please.

Well, I've been on both ends of the process and I think it generally
works pretty well, but it could work a little better.  Here are a few
observations after having submitted a few packages and having reviewed
several others.

First, the sponsor process hurts a bit; I think we could relax the
requirement that the initial sponsor and reviewer be the same person.
I just went through a big review for multitail but I think that there
was no notation that a sponsor was needed.  Under the current rules I
shouldn't have done the review, but what does it hurt?  And in any
case we probably could use more sponsors.

One thing that's bothered me is the nonspecificity of the guidelines.
Someone noted a blocker on one of my packages, but when asked where
the prohibition was in the guidelines the commenter didn't respond.
Some people say that my reviews haven't been picky enough, and some
have said they're too picky.  There's not a lot of consistency.  In my
own reviews I'm trying to strike a balance.

Doing good reviews takes a lot of time and besides lowering standards
the only thing we that can be done to help is to provide better and
more complete templates and then require adherence to them.

 - J<

MS> perspective (and I admit I've been doing a *lot* less reviews
MS> myself compared with fedora.us era), many reviews are still quite
MS> difficult and time-consuming. The usual packaging mistakes range
MS> from "simply fails to build" to "does not work at run-time" and
MS> "does not erase without errors". There are packages which
MS> NEEDSWORK, not seldomly due to severe packaging mistakes. There
MS> are packages, where the reviewers becomes an instructor (same
MS> thing applies to some upstream projects). And there are packages,
MS> where the packagers seem to spend less time on packaging and
MS> testing than the reviewer(s) do. The recent repo breakage caused
MS> by invalid "Provides" plus some bugs in new packages and updates
MS> are reason enough not to "lower the hurdle" by altering the review
MS> process for new packages.

MS> In general, reviews and approvals can be sped up by bringing
MS> packages in shape, doing test-builds and reviews and run-time
MS> tests and only then declaring a package as "ready". Better check
MS> your own package more carefully instead of relying on
MS> reviewers. Everybody can help with that in bugzilla. New packagers
MS> can demonstrate good packaging practice and that they are aware of
MS> things discussed in the packaging guidelines/policies.

MS> -- fedora-extras-list mailing list fedora-extras-list redhat com
MS> https://www.redhat.com/mailman/listinfo/fedora-extras-list

  Jason L Tibbitts III - tibbs math uh edu - 713/743-3486 - 660PGH - 94 PC800
       System Manager:  University of Houston Department of Mathematics 
And with death The knowledge comes It was the life all along We'd been afraid of

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]