On Thu, Oct 25, 2007 at 10:36:30AM -0700, Gabe Rubin wrote:
> On 10/25/07, Richi Plana <[EMAIL PROTECTED]> wrote:
> > it's my guess that ATRPMs libraw1394 package was compiled given
> > the above instruction. I would suggest, however, that until such
> > time as mythtv works with current kernels, that a package be made
> > (say "mythtv-libraw1394") where the library is stored in a
> > mythtv-specific subdirectory (say /usr/lib(64)/mythtv/) and that
> > applications which need and link specifically against
> > libraw1394.so.8 be somehow redirected to that link from their
> > startup scripts. Perhaps and LD_LIBRARY_PATH override might be
> > set. In that way, the two "versions" of the same version of the
> > library can co-exist.
> 
> I believe this is correct.  I know Axel specifically made new packages
> for users (like me) who are daunted at compiling there own. (see
> http://www.gossamer-threads.com/lists/mythtv/users/283084)

Upstream suggests (http://wiki.linux1394.org/JujuMigration):

  At the time of this writing (10/2007), there are still multiple
  problems with the new FireWire kernel driver stack (alias Juju)
  compared to the old stack: [...] <!> Regarding Linux 2.6.22 and
  2.6.23, the best advice to Linux distributors (kernel packagers) as
  well as to regular users is: Build only the old IEEE 1394 drivers.

Fedora is a bit too experimental here. The idea is to force the juju
stack into the wild and get more people involved with fixing it. That
speeds up development (people want to see their systems working
again), but in reality it is PITA for production systems.

So one needs to rectify the world back to the old system until the new
system matures. This means *both* kernel and userland bits need to be
fixed. E.g. you don't only need proper libraw/libdc 1394 userland
libs, but also the ieee394 kmdls: http://atrpms.net/name/ieee1394/

> Again, Axel, appreciate the work you did on that!

Thanks! :)

> By the way, as a new user to this list, is there a reason that when I
> reply, it defaults to reply not to the list, but to the individual who
> sent the email I am replying to?

Well, because most MUAs are able to "list-reply" and some users get
angry when they wanted to do a private reply and the list's config
forced them to accidentially go public against their intentions. But
it's not that hard a criterion, I've often contemplated switching the
default (after checking with the current netiquette).
-- 
Axel.Thimm at ATrpms.net

Attachment: pgpvVEqlIVYZU.pgp
Description: PGP signature

_______________________________________________
atrpms-users mailing list
[email protected]
http://lists.atrpms.net/mailman/listinfo/atrpms-users

Reply via email to