Okay I poked at the system timer problem a bit. Putting this in modprobe.conf is
supposed to get snd-rtctimer loaded automatically:

options snd-seq seq_default_timer_device=1
alias snd-timer-1 snd-rtctimer

But I can't seem to convince rosegarden to use it. Setting it in the preferences
doesn't seem to do anything. And if you go back to the preferences, reverts back
to auto. In fact, no matter what I set it to, it seems to not take effect and it
stays stuck on auto. Bleh. I haven't been able to google up anyone else with
this problem, everyone seems to just compile their kernels with a 1000hz timer
instead...

Created attachment 128983 [details]
Patch for x86-64 ladspa

This patch tells rosegarden4 to look for plugins in lib64 instead of lib.
Just conditionally apply it in the spec file like so...

# Conditionally apply patch to read ladspa plugins from lib64 dir.
%ifarch x86_64
%patch1 -p0
%endif

Admittedly, this is a bit of a hack.  I'll follow up with upstream to make sure
there's a cleaner solution in the next release.

Okay, so I tried building rosegarden against jack, and now the timer setting
will stick, and /proc/asound/seq confirms it is actually using the RTC timer,
however it STILL complains about the system timer resolution on startup. Bleh.

I'm going to hold off on updates until its various (supposedly) optional
dependencies are accepted. Looks like liblo was accepted, and I'm working on
reviewing jack.

Why do you not use %find_lang macro?

You may uncomment BR: jack-aufio-connection-kit due to it is built for devel,
fc4,5 repos.

Actually the updated package I'm sitting on uses %find_lang. Today I plan to
finish up my dssi review, which is the last remaining soft-dependency for this
package.

Alright, so I've run into a problem on the timing front. On my x86_64 system,
using snd-rtctimer will cause a total system lockup within seconds, when the
ALSA sequencer is used. So snd-rtctimer does in fact still not get along with
SMP, and it appears simply *running* an SMP kernel, even on a UP system, will
cause lockups with snd-rtctimer. (x86_64 always runs with an SMP kernel, as it
apparently doesn't suffer the speed hit i386 does when running an SMP kernel on
UP systems.)

This happens with any ALSA MIDI app, not just Rosegarden. However, I'm not sure
if its something I did, but somehow Rosegarden is causing snd-rtctimer to be
loaded when it starts up, even if I take it out of my modprobe.conf. And it
seems to insist on using it even if I tell it not to. I had to blacklist
snd-rtctimer in modprobe.conf to keep my system from locking up.

Since I approved dssi, I'll get an updated rosegarden package up tomorrow
hopefully. (Summer session just started and I'm taking a few classes. 3 months
worth of class mashed into one month...)

Alright, shiny new package update:

http://www.haxxed.com/rpms/rosegarden4-1.2.3-2.src.rpm
http://www.haxxed.com/rpms/rosegarden4.spec

- Removed which from Buildrequires, mock needing it is confirmed to be a bug.
- Use find_lang macro.
- Look for ladspa plugins in the correct place on x86_64.
- Build against liblo, jack, dssi, lirc.
- Salvage the upstream desktop file, rather than using our own.

(In reply to comment #7)
> Alright, shiny new package update:
> 
> http://www.haxxed.com/rpms/rosegarden4-1.2.3-2.src.rpm
> http://www.haxxed.com/rpms/rosegarden4.spec

Do you know what all this is about?

# rpmlint /usr/src/redhat/RPMS/i386/rosegarden4-1.2.3-2.i386.rpm
E: rosegarden4 zero-length
/usr/share/apps/rosegarden/fonts/feta/6/custodes-medicaea.xpm
E: rosegarden4 zero-length
/usr/share/apps/rosegarden/fonts/feta/4/custodes-rvaticana.xpm
E: rosegarden4 zero-length
/usr/share/apps/rosegarden/fonts/feta/8/custodes-mensural.xpm
E: rosegarden4 zero-length
/usr/share/apps/rosegarden/fonts/feta/12/custodes-rmedicaea.xpm
E: rosegarden4 zero-length
/usr/share/apps/rosegarden/fonts/feta/12/accordion-accSB.xpm
E: rosegarden4 zero-length
/usr/share/apps/rosegarden/fonts/feta/12/custodes-rvaticana.xpm
E: rosegarden4 zero-length
/usr/share/apps/rosegarden/fonts/feta/4/accordion-accDot.xpm
etc 
etc
etc

That's part of the pre-rendered music notation font its using. Dunno why there's
a bunch of null files in there. I figure we can ignore it for now, and I'll
report it upstream along with all the patches once the review is done.

I don't have the "fedorabugs" membership yet, so this I can't approve yet, but
here's my formal review anyways...

* package meets naming and packaging guidelines.
* specfile is properly named, is cleanly written and uses macros consistently.
* dist tag is present.
* build root is correct.
* license field matches the actual license.
* license is open source-compatible.  License text included in package.
* source files match upstream:
md5sum is e7fb7ebcb21ac6841ac5cfd6683f5fb2
* latest version is being packaged.
* BuildRequires are proper.
* package builds in mock ( ).
* rpmlint is silent apart from empty file complaints which we've discussed in
bugzilla.  Those will be fixed upstream.
* final provides and requires are sane:
    rosegarden4-1.2.3-2.x86_64.rpm
    rosegarden4 = 1.2.3-2
=
    /bin/sh
    /usr/bin/perl
    desktop-file-utils
    libDCOP.so.4()(64bit)
    libX11.so.6()(64bit)
    libXft.so.2()(64bit)
    libXrender.so.1()(64bit)
    libasound.so.2()(64bit)
    libasound.so.2(ALSA_0.9)(64bit)
    libfontconfig.so.1()(64bit)
    libfreetype.so.6()(64bit)
    libjack.so.0()(64bit)
    libkdecore.so.4()(64bit)
    libkdeprint.so.4()(64bit)
    libkdeui.so.4()(64bit)
    libkio.so.4()(64bit)
    liblirc_client.so.0()(64bit)
    liblo.so.0()(64bit)
    liblrdf.so.2()(64bit)
    libqt-mt.so.3()(64bit)
    libstdc++.so.6()(64bit)
    libstdc++.so.6(CXXABI_1.3)(64bit)
    libstdc++.so.6(CXXABI_1.3.1)(64bit)
    libstdc++.so.6(GLIBCXX_3.4)(64bit)
    libz.so.1()(64bit)
    perl(File::Basename)
    perl(File::Copy)
    perl(Getopt::Long)
    perl(XML::Twig)
    perl(strict)
* no shared libraries are present.
* package is not relocatable.
* owns the directories it creates.
* doesn't own any directories it shouldn't.
* no duplicates in %files.
* file permissions are appropriate.
* %clean is present.
* code, not content.
* documentation is small, so no -docs subpackage is necessary.
* %docs are not necessary for the proper functioning of the package.
* no headers.
* no pkgconfig files.
* no libtool .la droppings.
* not a web app.

I would approve this if I had "fedorabugs" membership (which I've requested).

Ok, I'm in fedorabugs.   This package is ACCEPTED!

Created attachment 132518 [details]
Patch to look for DSSI plugins in lib64 directories

Please apply this patch when you check rosegarden4 in.	It tells rosegarden
where to find DSSI plugins on x86-64 systems.

Patched and built on devel and FC5! FC4 is missing liblrdf, liblo and dssi though.

(In reply to comment #13)
> Patched and built on devel and FC5! FC4 is missing liblrdf, liblo and dssi though.

Great news! 

I wasn't planning on putting those packages in FC4, as I don't have an FC4
machine to test with.  Is this still something you would really like?