On Fri, Nov 28, 2014 at 11:29:19AM -0500, Lennart Sorensen wrote:
> On Fri, Nov 28, 2014 at 05:24:23PM +0100, Gilles Chanteperdrix wrote:
> > the integration of the xenomai package into the Debian repositories
> > is largely a failed experiment. I think the problem is that the
> > debian xenomai package does not have enough users for them to look
> > hard enough and report bugs to reach a good package quality. And
> > there will never be enough users for that, because real-time is a
> > niche, whether people want it or not. So, our goal should be to
> > not cut users from the project, and try and keep them reporting
> > back to us instead of downstream distributions, and to work for
> > these users, not to try and work with intermediates.
>
> Getting a kernel chosen in Debian for which there is a working
> ipipe/xenomai patch is certainly a problem.
>
> Still having the runtime be packaged is quite handy.
It would be handy if the runtime package was built correctly, but
since you were the first one to use the word "shitty", if there one
thing shitty it is Debian's debian/rules. This file has essentially
been unmaintained for years, just modified enough to compile the
next version, without taking the release notes into account, and
removing options or compilation flags that were no longer useful.
The method to provide patches for the kernel has disappeared from
make-kpkg around Lenny or something. But since no user was paying
attention, nobody noticed.
No, really, stay away from Debian's xenomai packages.
--
Gilles.
_______________________________________________
Xenomai mailing list
[email protected]
http://www.xenomai.org/mailman/listinfo/xenomai
