x.org is Hacker Trash
gmane at colin.guthr.ie
Thu Mar 29 12:25:36 PDT 2007
Joseph Parmelee wrote:
> The linux kernel project, arguably the gold standard of open source
> projects, puts up release candidates for testing which can be quite easily
> built by persons with only a modicum of experience. Why is that important?
> Because it permits widescale testing of all the various different hardware
> out there.
A fully configured kernel to support all your hardware etc. will still
take a while to compile. Say you get a new USB pen, do you just compile
that and "add it on" to your kernel or do you have to build the whole
damn thing again? Now I know you can keep your built tree so it will
compile fast but say you rm -fr'ed? No joy, build from scratch. And I
know there are dkms systems (which are great) but this wasn't your
point! You were defending the BIG project approach as a way to get more
people to test things! The kernel is a hell of a confusing thing to
configure correctly for a novice user. It's managed to get round this by
having a fairly nice GUI to set the various options.
Your argument that it's easy to get out there to test on a variety of
systems where as modular X does not, just doesn't hold true! I packaged
x11-server13 and x11-driver-video-intel13 for Mandriva to add the
xserver 1.3 and intel driver RCs to the mix. It was very easy to build
this on top of our existing xserver but just use a different prefix for
these parts. It works flawlessly and meant I didn't need to build (and
thus have users install and download) a whole heap of other
libs/utils/drivers too. By being able to build these packages so easily
it makes for a far better testing ground for new X features. It's just a
matter of finding out (e.g. by asking or reading) what the dependencies
for new feature X are and build them in order. It's not rocket science.
It's all very easy.
More information about the xorg