Re: GStreamer version for 2.14
- From: Luis Villa <luis villa gmail com>
- To: Federico Mena Quintero <federico ximian com>
- Cc: Vincent Untz <vuntz gnome org>, desktop-devel-list gnome org
- Subject: Re: GStreamer version for 2.14
- Date: Wed, 18 Jan 2006 14:06:32 -0500
On 1/18/06, Federico Mena Quintero <federico ximian com> wrote:
> On Mon, 2006-01-16 at 16:14 -0500, Luis Villa wrote:
>
> > JWZ is an ass, and has been for a long time. If forced to take the
> > admittedly unpleasant choice between overwhelming maintainers so that
> > they never look at bugzilla at all, or incorrectly closing bugs which
> > might be reopened later, we should *always* choose the second option.
>
> In defense of asses, JWZ's bug reports tend to be detailed and he'll
> happily follow up on them if you ask nicely. I know because I fixed
> some of them.
Oh, his reports are great, and I wish more people were like him in
that respect. That doesn't mean he isn't an ass :)
> His tantrum was on our extremely irresponsible transition from 1.x to
> 2.0, where no one bothered to see if there were regressions, we didn't
> provide a migration path for user's settings, we didn't write migration
> documents for all the APIs that got replaced, and we just shoved
> everything under the rug.
No, that's not what his tantrum was about:
http://www.jwz.org/doc/cadt.html
Luis
> The bugsquad (did it exist then?) was probably not as good as it is now,
> etc. etc., but we cannot say "uh, I guess it may be fixed now" every
> time we switch versions. This includes minor and micro versions.
>
> If a bug has good info on how to reproduce a crash, and the code has
> changed so much that the provided stack trace is not relevant anymore,
> it is up to the maintainer to attempt to reproduce the crash with the
> original instructions, and close the bug if it is irreproducible.
>
> If a bug has only a stack trace and no info on how to reproduce it, feel
> free to mark it NEEDINFO immediately, or even INCOMPLETE/INVALID.
>
> Speaking of GStreamer...
>
> We should probably start to become stricter on API/ABI stability for the
> desktop suite, not only the platform suite. Users only see our desktop,
> never the platform; if the desktop breaks because a desktop-only library
> broke things, users will think that GNOME sucks as a whole.
>
> I'm thinking that we need to change our release model a bit, so that we
> have a user branch and a developer/experimental branch at all times, and
> it is ONLY the ongoing user branch that we ship to distros and thus
> users.
>
> Federico
>
>
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]