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

Re: XFree86 4.4.0 release plans - maybe it's a time to start debian packaging ?



On Wed, Nov 05, 2003 at 06:47:51AM +0200, Mantas Kriau?i?nas wrote:
> From http://www.xfree86.org/releaseplans.html :
> 
> The next full release of XFree86 will be 4.4.0. The schedule for this
> release is:
> 
> *Feature freeze: 15 October 2003.
> *Code freeze: 28 November 2003.
> *Release date: 15 December 2003.

There have been a couple of posts to the list about this already; we are
well aware of it. Some of us (like myself, up until recently) also read
xfree86-devel on a regular basis.

> Debian currently has Xfree version 4.2.1 in unstable and testing and 4.3.0
> pre1v4 in experimental. It seems, that 4.3.0 won't be included into
> unstable before 4.4.0 code freeze and because of this I suggest to start
> packaging 4.4.0 prerelease packages in experimental instead of 4.3.0

I suggest not.

XFree86 releases are not just a 'drop-it-into-a-deb' release. Last time
I checked, we were applying 103,000 lines of patches - it doesn't even
come close to building on half the architectures, with just a stock
release. Also, monitoring for important regressions is a pretty big
task.

It's also notoriously difficult to beat certain architectures into
shape; it took me some months to get a member of each architecture
team to do the porting.

We also have the sarge freeze to think about (I would consider it a
major failing on the behalf of myself and the X Strike Force if 4.3.0
does not make it into sarge, largely due to the i845g problem, nVidia
hardware in PowerPC, and more), which is scheduled to happen around the
time of the code freeze. This means that 4.4 for sarge isn't at all
realistic, and that we should instead focus our efforts on a realistic
goal - 4.3.

> I see at least 3 benefits if we start packaging 4.4.0 now:
> 
> 1. At last Debian will contain up-to-date X-Window-system (and because of
> this more hardware will be supported)

... on i386, at the expense of the other architectures.

> 2. Xfree will be more stable and will contain less bugs, because more debian
> users can test prerelease and report bugs, that can be fixed in upstream
> before 4.4 final will be released.

Heh. The likelihood of this happening and making a major difference (and
of making a difference upstream if it does) is pretty minimal.

> 3. For Debian maintainers it would be easier to maintain up-to-date
> version, than old version (this includes critical bug, security fixes, etc)

Not particularly; we have done backports before, and will happily do
them again.

> I think if debian developers start package 4.4 Xfree today then Xfree will
> be more stable in debian too, because debian xfree maintainers will improve
> upstream xfree and xfree packages instead backporting lots of changes from
> xfree 4.4 release to old xfree release.

Not likely, I'm afraid.

This is not, of course, to discourage anyone from working on 4.4
packages. If you can come up with good-quality stuff we can use as a
base (sort of like Ishikawa's initial work, but preferably closer to the
4.3.0-0ds* series, only without the regressions), more power to you. I
don't claim to be Branden's spokesman, but I really strongly doubt he'll
be working on 4.4 at the expense of 4.3 anytime soon. I know where I
stand.

-- 
Daniel Stone                                              <daniel@fooishbar.org>
Iterant hacker on Debian XFree86 packages, on a study vacation

Attachment: pgphmwaTtlfmQ.pgp
Description: PGP signature


Reply to: