[Date Prev][Date Next] [Thread Prev][Thread Next]
Re: Xorg 6.9 Update for FC4?
- From: "Mike A. Harris" <mharris mharris ca>
- To: Development discussions related to Fedora Core <fedora-devel-list redhat com>
- Subject: Re: Xorg 6.9 Update for FC4?
- Date: Sat, 04 Feb 2006 05:06:19 -0500
Peter Gordon wrote:
With the recent release of the modular X.org X11R7, will Fedora Core 4 receive
an update to its monolithic counterpart X11R6.9?
Fedora Core 4 will most likely remain at Xorg 6.8.2, however we have
at least considered what you're suggesting. Engineering has been
focusing entirely on FC5 and RHEL for the most part for a while now,
which will probably continue through the remainder of the FC5 devel
cycle. There will likely be some FC4 X updates yet to come, but
barring critical security issues, bugfixes and driver updates for FC4
wont happen likely until FC5 is released, mainly due to resource
Having said that though, there are a growing number of issues in FC4's
X, which I would like to address in the next non-security triggered
update for FC4, including the new backport of the 'nv' driver from
X11R7. I'd also like to sync all RHEL4 bugfixes into the next FC4
X update too, among other things.
A 6.9 update however, is not on the radar for FC4, at least with our
current schedules. I may suggest it in our team meeting again in
the future though, but no promises.
The primary reason I inquire about this is for the updated drivers. Every so
often, OpenGL-intensive apps such as BZFlag and others will inevitably hardlock
my system, forcing a reboot to fix it. Some games such as gl-117 refuse to even
play without hardlocking. However, I upgraded a now-dead Gentoo install to one
of the later X.org release candidates and found that, unlike 6.8 (both on Gentoo
and Fedora), X11R7 seems to fix every issue that I have with the R200 driver.
(I use a Radeon 9200.)
I can neither agree or disagree with that, as I really done similar
comparisons between the two, however at the current point in time at
least, X11R7 is quite a bit less stable than 6.8.2 is overall. R7
needs a fair amount more stabilization yet.
While I'm aware that the shift to a modular, autotool build and distribution
system is a very fundamental change to the architecture of X.org, it did put out
a 6.9 release which is of the same codebase of 7.0, and thusly would (hopefully)
also fix this and other small issues I have with the radeon driver. It would
also provide better support for newer video cards (such as the i945 Intel GMA,
R300-series Radeon cards, et al.)
Releasing modular X for FC4 is definitely out of the question for sure,
as it inflicts a large number of changes into the system when done in
the context of an "official" OS update. It would require releasing
a great many updates to other packages to keep the system clean. Far
too many to really be considered in "official" context. Having said
that however, I run modular X from FC5 rebuild on my FC4 system locally
and it works fine. I dont mind the fact there are some dependency
conflicts, and that half the FC4 packages wont compile cleanly on top
of the modular X packaging without changes though. So it works fine
in "unofficial" context, but is not polished enough distro-wide for
consideration for official packages released to the public.
X11R6.9 is the same source code of course, and while that is
theoretically possible, a number of factors would need to be
considered very carefully. I'm not opposed to it personally,
but having had to make decisions like this for 5+ years now
across about 12 OS releases, I have a good gut feeling for
when we need to be very careful about updating certain things. ;o)
For now, I wouldn't say "yes", but I wouldn't say "no" either. It's
currently more of a "lets wait and see how X11R7 stabilizes first"
type of thing. Once that happens, if it appears we have engineering
resources available that could be allocated to doing it, I may
bring it up with the team again and see what team concensus seems
With FC5 coming out very very soon however, another factor is wether
it is even worth spending the time and effort on doing the work for
FC4. With fixed engineering manpower, 1 hour of work spent on FC4,
is an hour of work directly taken from fixing bugs on FC5. That's
very important to consider also.
Also, I'm not aware of any difficulties that we would face in doing this. It
would only be a matter of version-bumping the various xorg-x11 packages and
building a new release, right?
A 6.9 release would be a lot more work than that, but much much
less work than doing a 7.0 release. I already have 6.9 beta rpms
which I could probably update cleanly to 6.9 in a day or two, which
would then need a few weeks of cleaning up and bugfixing, etc.
I may do that at some point anyway, wether or not we release it for
FC4 officially. It'd be nice to have it as unofficial packages for
people who are interested, and it would help get the code more tested,
however everyone has to realize that 6.9 is very _dead_ code now. The
monolithic Xorg CVS is frozen now.
Alternatively, would it be better for me to grab a 6.9 source RPM and rebuild it
for my FC4 system? If this is the preferred method by the Fedora X.org team,
would someone please point me to instructions to do this? (Would it be a simple
"rpm --rebuild xorg-x11*.src.rpm"?)
There are a few people who have produced Xorg 6.8.99.x and possibly 6.9
rpms out there. The ones I have seen were initially based upon the FC4
packages, or upon 6.8.99.x packages that I created a while back, however
since the people who made them were making them for themselves first to
scratch a personal itch, I found they included too many experimental
patches and changes to the packaging. They'd probably be ok to use on
noncritical systems, knowing of course that they may contain highly
If there is strong enough interest in 6.9 rpms, I might be able to
be convinced to spend some of my personal time updating my previous
6.8.99.x packages on the weekend sometime. Of course it has to be
a pleasant experience for me to expend the effort though. The last
time I did it, the feedback I got from people was mostly inflammatory
and negative, with no appreciation for the effort, so I stopped doing
it. For an example of what I mean by this, simply look back a few
messages in this list to discover how rude and thankless people can
Mike A. Harris * Open Source Advocate * http://mharris.ca
[Date Prev][Date Next] [Thread Prev][Thread Next]