Wrapping up 7.4 (finally)
daniel at fooishbar.org
Thu Jun 12 01:24:16 PDT 2008
On Wed, Jun 11, 2008 at 04:48:28PM -0700, Rob Currey wrote:
> >We also need to start planning what we want in 7.5.
> Hint: try building the git on Redhat 5, Solaris10, HPUX 11.11+ ... (and god
> forbid AIX 5.3)
> Couldn't xnest be a baseline xserver target for these "non-linux" platforms
> ??? (this *is* X.org and not Xfree86 correct?)
If the build is broken, please report bugs. I've just committed (but
not pushed as of this moment) a fix to master for Xvfb and Xnest missing
We have tinderboxes running regularly, and someone kicking us whenever
they break. I don't have any HPUX hardware, and I don't know anyone who
does, but anyone who does is welcome (likewise AIX). I know someone
tried to build xproto on HPUX a while ago, which I fixed with the aid of
some manpages I found on Google from which I guessed how to get
fds_bits, but I've not heard of anyone using HPUX trying to build the
Maybe it'll just work. Probably not. Want to find out?
> <Adam Jackson ... from a RH5 virgin system, what rpm list is necessary?
> (hint: python 2.5?)>
Why would python 2.5 be necessary?
> <Alan Coopersmith ... what's it take for Solaris10? got a pkglist?>
> <is there a core developer from HP any more? would be nice to see the depots
> <AIX ???? (ya, I know who cares)>
FWIW, for Debian, the answer is pretty much covered by sudo apt-get
> Besides portability to non-linuxen ... I'm sure many have seen
> Some of the rants are noise, but there are many people who simply give up
> because ...
> "git sucks"
> "builds just don't work"
> "'modular'?, still seems pretty 'monolithic'!"
> Would be nice if 7.5 could fix some of *those* issues ...
I did see Slashdot comments, and while I'd like to say I ignored them, I
read them. Don't get me wrong, there are a hell of a lot of things
wrong with X, but I didn't see any insightful comments on there that
were actually pointing out things wrong with X. (The insightful ones
were all defending X, which actually kind of makes me feel bad.)
Is there a _specific_ problem with the build you are encountering?
Assuming that we're not switching away from git, that every build
failure is a bug (either configure.ac fails and tells you what to
install, or it's a bug) and should be reported as such, and that we are
actually ridiculously modular?
The modularity often requires changes in a couple of other modules, due
to the rapid pace of development. But again, see above: if it doesn't
file in ./configure, file a bug. Uncoupling the X server from Mesa was
one of the last big changes here.
> One last thing, it's probably time to cut loose much of the "cruft"
> 1) app/* has bits than can be cut loose to fend for themselves (e.g. xeyes,
> xlogo, ...) ... let sourceforge or redhat or someone pick up the pieces (or
> else they simply die) ... IOW, not X.org's problem anymore!
What impact does this have?
> 2) drop xlib ... embrace xcb and be done with it. When venders (e.g. QT,
> gtk, etc) get wind of this, they'll also transition to xcb in a hurry.
No, when vendors get wind of this, they'll fork Xlib or just ignore it.
Qt are definitely moving in the XCB direction, and GTK+ possibly are;
for the latter, I get the impression it's as much lack of developer time
(which I'm very sympathetic to) as anything else.
> 3) if a parts are dead, cut 'em off (e.g. mfb/cfb, XTrap, LBX, XPrint?, etc)
mfb/cfb have been deleted. XTrap is absolutely not dead. LBX was
deleted years ago. XPrint was deleted relatively recently. Most major
server releases, as far as I'm aware, deleted more code than they added,
and I believe that both ajax and myself (the #3 and #2 committers to the
server, respectively) have actually had a net _negative_ impact on lines
of code in the server.
> I'm *wanting* to move past R6.9 and into R7.x ... but it is *so* painful.
> <putting on my flame suit>
I'd love to say that this mail was insightful and pointed out many
actual problems, but instead it was just a metaflame. 'Look, people on
Slashdot are saying that things suck! You should do things which are
already done! Other assertions without reasoning!'. Please, for my sake
(this is hardly motivating me to get out of bed) and yours (it's your
name on it), and X.Org's (having our problems coherently documented is
hardly a bad thing, but this wasn't it), try to do better next time.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 197 bytes
Desc: Digital signature
More information about the xorg