Sponsorship Process (Was: Re: Packaging Help)
loic.martin3 at gmail.com
Mon Sep 28 19:43:49 BST 2009
Daniel Holbach wrote:
> Am Freitag, den 25.09.2009, 20:44 +0200 schrieb Loïc Martin:
>> There's no visibility at all. Basically, your patch/diff.gz can be
>> uploaded the same week, or go for month without being looked at. Nothing
>> in Launchpad tells prospective contributors (or upstreams) which
>> packages they have chances to see their work used, and which packages
>> they should ignore because no sponsor is interested in seeing them improve.
>> There's also no timeframe, for example no way to see if a backport
>> request that has been tested and reported working by 2 or 3 contributors
>> is going to be ACKed in a meaningful time. A backport or a fix to a
>> stable release looses much point if it only gets ACKed a few weeks
>> before the development release is delivered, because at this point most
>> of your users have moved to the newer version (excluding LTS). How do
>> you convince upstream it's better to work within Ubuntu (so a better
>> release is backported, or a fix done), if they can see that only a PPA
>> gives them the possibility to make Ubuntu better?
>> The sponsorship process doesn't tell you what to do when a patch is
>> roting in Launchpad. For contributors, it's just a black box. It's hard
>> to plan your work that way.
> Do you think the idea of IRC reviewers that can be pinged will help with
> that? If not, what do you think would help?
Don't IRC reviewers address a different issue? They're probably good for
when you need help in fixing a bug, and for teaching, but how do they
deal with the sponsoring issue? Even if they were also a place to ask
for it, that's just adding another layer to the sponsoring process, and
discriminates against new contributors, who don't use IRC (and asking
upstream to find someone to nag on IRC each time there's a patch, a
backport or a new release?) Where does it helps for people that have
been maintaining a package in Ubuntu for a few releases already (only
name in the changelog), and suddenly nobody has the time to upload the
new version anymore? How long till they give up?
Actually, nothing in the documentation tells you there's an added step
after hitting "Confirmed" and subscribing the relevant sponsors.
However, it's often necessary, and since being obnoxious on #ubuntu-motu
is counterproductive, you progressively learn to pick up the bugs you
think, and not to much, because that way you can leave a few days
between asking on the channel.
Just as an illutration of what I'm talking about, gnucash crashes for a
lot of users in Hardy/Intrepid/Jaunty (and Debian :P) because of a
faulty patch that got reverted upstream long ago (I only remembered that
one because a backport request is on atm). It's a trivial job to revert
the patch and ask for an SRU; the sponsoring part is where it hurts.
Fact is, only time I did an SRU took at least two emails to
#ubuntu-motu, and much nagging - there's no other word, sorry - on IRC,
for a debian/control one-word-change. For gnucash, I already had some
sponsor requests for other packages on (and some planned), and not
burning all the vessels on a problem where the users that could test
-proposed have probably given up already is something you have to learn
quickly if you want some of your work uploaded in Ubuntu. So no SRU for
gnucash, even though the actual work for submitting the request was
never the problem in the first place (work was done for Karmic anyway).
(Same for xvid - requesting the sync with pkg-multimedia would be
suicidal when nobody's been interested in an improved package, even with
the work done already.)
> The problem with "a long list of stuff that need to be done" is that
> people naturally will pick what interests them, what they work on
> generally anyway or what they have time for right now.
> Which other ideas do you have to fix it?
Having other ideas won't do much, eventually I'm not a sponsor and the
thread looks like a monologue already. Better if it were coming from the
ones that do the sponsoring itself, since ideas coming out of the blue
can't get much support. I don't sponsor, and for sure don't understand
enough of it.
My posts where only to give you an idea of what happens on the side of
the fence I'm in, if it could help. Please don't take it as a rant, if
it was one I'd have folded up and unsubscribed from the bugs already.
Contributors have to either give up, or learn to pick their battles,
which limits what little work we can do.
It would be nice if people on -devel stopped putting the blame on REVU
though, especially when, as a matter of fact, what they're asking for in
words isn't welcomed as much as they would paint it... XP
More information about the ubuntu-devel