Re: [Fedora-haskell-list] Mass closure of submissions?
by Miles Sabin
On Wed, Jun 18, 2008 at 1:11 AM, Bryan O'Sullivan <bos(a)serpentine.com> wrote:
> Most certainly. Someone showed up from nowhere and used a Perl script
> to generate spec files for 40 packages, which he then submitted en
> masse. I can certainly believe that the submissions were in good
> faith, but the way the submitter went about things was not okay.
>
> Volunteering is a two-way thing: it is important to treat newcomers
> with respect and patience as they learn the ropes, provided they
> display a basic modicum of clue in the first place. I was quite
> annoyed to have my time wasted by a deluge of autogenerated rubbish.
Auto-generation isn't a bad thing per-se, surely? Generating RPMs from
cabalized packages ought to be relatively straightforward as your own
cabal-rpm shows. If there was something concretely wrong with the
generated packages that would be one thing, which could presumably be
fixed relatively easily by improvements to the script used to create
them.
I hadn't realized there were as many as 40 subitted tho'. That makes a
difference, certainly ... I agree with your comment that it would been
better if he'd submitted one and started a discussion about his script
here.
Cheers,
Miles
15 years, 10 months
[Fedora-haskell-list] Mass closure of submissions?
by Miles Sabin
Hi folks,
As a potential submitter of Haskell packages I'd welcome some
clarification on what just happened here.
>From the outside it looks like,
* New submitter created BZ's for packages in good faith but without
being aware of the (currently non-existent?) Haskell packaging
guidelines.
* Various borderline rude comments were made.
* The BZ's were closed with extrordinary haste without giving the
submitter a great deal of time to respond (I mean, maybe he's been
away on holiday for the last few days?).
Now, perhaps that's not how it really was, but I don't feel
particularly motivated to do work on any packages myself unless I'm
reasonably confident that I won't be on the receiving end of similar
(apparently) high-handed treatment.
To that end it would be very helpful if someone could outline,
* What in particular that guy did wrong.
* What are the most important things for a package submitter to do to
ensure a successful outcome.
Cheers,
Miles
15 years, 10 months
[Fedora-haskell-list] Thinking about changes to ghc packaging
by Bryan O'Sullivan
One of the things that Yaakov has suggested, in his draft Haskell
package management guidelines, is a simplification of the layout for
Haskell library packages. I think that some time ago, there was a
desire to support multiple versions of GHC installed concurrently. In
all the time that we've had the infrastructure to support that more or
less in place, the packaging energy to actually do anything about it
has never materialised, and I think it never will.
I'd like to take Yaakov's suggestion for packages and apply it to GHC
itself. Instead of splitting GHC into ghc and ghc682 packages, we'd
have a single ghc package that would represent the current version.
I need to perform some surgery on the GHC spec file over the coming
few days in any case, because it violates a few of the packaging
guidelines (e.g. binaries in %{_libdir} instead of %{_libexecdir}) and
has some bugs besides (it shouldn't be messing with SELinux labels in
the %post script).
Thoughts?
15 years, 10 months