[ATrpms-devel] The state of lirc...

[Paulo Cavalcanti]

On Fri, May 9, 2008 at 5:15 PM, Jarod Wilson <jarod at wilsonet.com> wrote:

> On Fri, 2008-05-09 at 12:25 -0700, David Rees wrote:
> > On Fri, May 9, 2008 at 12:10 PM, Jarod Wilson <jarod at wilsonet.com>
> wrote:
> > > On Fri, 2008-05-09 at 12:00 -0700, David Rees wrote:
> > >> Last time I tried using lirc strictly from Fedora, I had problems with
> > >> the lirc_serial driver. IIRC, at some point in time it wasn't included
> > >> with the kernel. I see that it is now on my Fedora 8 machine, so I'll
> > >> try removing the ATrpms versions of lirc and see if the upstream
> > >> versions work.
> >
> > Just switched over to the RPMs from koji and got everything migrated
> > and working. Here's what I had installed before:
> >
> > lirc-0.8.3-73_cvs20071109.fc8
> > lirc-lib-0.8.3-74_cvs20080314.fc8
> > lirc-kmdl-2.6.24.5-85.fc8-0.8.3-74_cvs20080314.fc8
> >
> > I removed the kmdl and then tried to upgrade using the new lirc and
> > lirc-libs, but lirc complained that it was older than the ATrpms
> > version. So then I tried --oldpackage, but then lirc-libs conflicted
> > with lirc-lib from ATrpms.
> >
> > Finally just removed lirc and lirc-lib with --nodeps then reinstalled
> > lirc and lirc-libs.
>
> Yeah, part of my original complaint was that the atrpms packages were
> rpm-newer than the fedora 0.8.3 final packages, despite being a ~2mo old
> pre-0.8.3 cvs snap. :)
>
>
> > ATrpms uses /etc/sysconfig/lircd, Fedora uses /etc/sysconfig/lirc so I
> > had to transfer my lircd options over.
>
> And similarly, service lircd start vs. service lirc start.
>
> > Unloaded the old lirc modules I had loaded, loaded the ones from the
> > kernel and stopped/started lircd. Restarted mythwelcome and
> > mythfrontend and everything seems to still work. So from my point of
> > view, looks good! :-)
>
> Excellent. I've been using the in-kernel stuff for quite a while myself,
> with lirc_i2c, lirc_mceusb and lirc_mceusb2 (I have something like 6
> different IR receivers now...)
>
> > > Yep, lirc patches have been in the Fedora kernels since sometime in the
> > > 2.6.23 timeframe, including the kernel F8 originally shipped with,
> iirc.
> >
> > I swear I tried using Fedora lirc when I upgraded this machine to F8,
> > but I could be wrong... Maybe it was some other issue that kept it
> > from working.
>
> Hrm, beats me, I think I've been using them since the day I patched them
> into the kernel.
>
> > Is Fedora still shipping lirc patches with recent kernels?
>
> Yes. Although I need to refresh them -- there's a
> just-fixed-right-before-0.8.3 segfault or oops with lirc_i2c and 2.6.25
> that I need to get taken care of.
>
> > Are you working on getting them upstream into Linus' kernel?
>
> Yes. Although working with upstream is... slow. Its high-latency, but
> upstream has already taken a ton of patches I've thrown at them to get
> lirc much closer to at least patching checkpatch.pl muster and replace
> some deprecated interfaces.
>
> > > I did talk with Ryan Pisani on this topic a bit, sounds like there may
> > > be a few drivers in the ATrpms package that aren't in the kernel for
> > > whatever reason. Assuming this is the only compelling reason for
> > > out-of-tree builds anymore, I'd either suggest the atrpms packages only
> > > carry those modules, or we figure out what's required to build them in
> > > with the kernel. User-space though, I can't think of any reason for
> > > duplication.
> >
> > Neither can I. And I also think that keeping as much as possible
> > upstream simplifies things for the end users as well.
>
> Axel, anything to add here?
>

Hi,

There are some minor problems, such as the use of lirc-lib-devel instead
of lirc-devel in ATrpms. This will require adjusting some spec files
accordingly.

At least kradio will have to be rebuilt (my bad, for using Requires:
lirc-lib),
and I have already uploaded an improved .src.rpm for it.

But I also tested your lirc packages, and they are working fine for me too.

These naming incompatibilities have always give me problems when using mock.

I really would like to see this issue solved once for all ....


-- 
Paulo Roma Cavalcanti
LCG - UFRJ
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.atrpms.net/pipermail/atrpms-devel/attachments/20080509/80eb2e3c/attachment-0001.html