Date 11.5.2013 00:32, Jordi Mallach wrote:

Hi!

I've been wanting to put some order in the collection of patches we've been building up in Debian due to me neglecting forwarding anything for far too long.

Let's try to correct this.

I have a number of patches, most of which should be trivial and no-brainers, all updated to 1.0.27.

I'm going to outline all the individual patches below. The patch tracker has the patches just as I would send them upstream, but if you prefer git formatted patches, I can go ahead and do that too, if it makes your life easier.

Thanks!

ALSA-LIB: http://patch-tracker.debian.org/package/alsa-lib/1.0.27-3

http://patch-tracker.debian.org/patch/series/view/alsa-lib/1.0.27-3/add-main... add-maintainer-mode.patch is something I'd like to see in all ALSA configure scripts: have maintainer mode enabled, but add the possibility to disable it via --disable-maintainer-mode. See http://blogs.gnome.org/desrt/2011/09/08/am_maintainer_mode-is-not-cool/ for reference.

http://patch-tracker.debian.org/patch/series/view/alsa-lib/1.0.27-3/multiarc... The patch header is self explanatory: it fixes the plugin lookup path for Debian/Ubuntu's multiarch paths.

http://patch-tracker.debian.org/patch/series/view/alsa-lib/1.0.27-3/spelling... Little to explain here. ;)

The alsa-lib patches looks good. Post them to this list in the patch format (or 'git format-patch').

ALSA-UTILS: http://patch-tracker.debian.org/package/alsa-utils/1.0.27-2

http://patch-tracker.debian.org/patch/series/view/alsa-utils/1.0.27-2/spelli... Again, trivial.

http://patch-tracker.debian.org/patch/series/view/alsa-utils/1.0.27-2/fix_mi... Even more trivial.

http://patch-tracker.debian.org/patch/series/view/alsa-utils/1.0.27-2/utils_...

Fixes a build failure when using -Werror=format-security, which is default in Debian and Uubntu's toolchain.

http://patch-tracker.debian.org/patch/series/view/alsa-utils/1.0.27-2/alsa-r...

This one I want to discuss. We think the After condition is wrong, as it's depending on an optional service (alsa-state might not happen if the config file is not present). Instead, it should be checking for sysinit, to ensure local-fs.

The condition is not wrong and it is expected that both services alsa-state/alsa-restore will be in the system, even if one is disabled via the ConditionPathExists . The systemd just skips services with false conditions, but the dependency is retained.

The main reason to serialize these two services via dependency is to have only one service for the sound state restore (alsa-restore) for other systemd units. Fedora uses this in canberra-system-bootup.service for an example. Perhaps this dependency may be solved using a new target, but it seems more complicated than use 2 services at the moment for me.

http://patch-tracker.debian.org/patch/series/view/alsa-utils/1.0.27-2/manpag...

Extensive formatting fixes for all alsa-utils manpages, plus an English improvement (allows to → allows one to).

Other patches are ok..

ALSA-PLUGINS: http://patch-tracker.debian.org/package/alsa-plugins/1.0.27-2 ===========================================================================

http://patch-tracker.debian.org/patch/series/view/alsa-plugins/1.0.27-2/arca... Fix linking to avoid undefined symbols in arcam-av.

http://patch-tracker.debian.org/patch/series/view/alsa-plugins/1.0.27-2/use_... The libav changes in 1.0.27 were insufficient: a52 uses av_free(), which is part of libavutil. The patch needs to be completed to look for the correct header files, though, but it's enough to get the package compiling in our setup.

Looks good.

ALSA-TOOLS: http://patch-tracker.debian.org/package/alsa-tools/1.0.27-2

http://patch-tracker.debian.org/patch/series/view/alsa-tools/1.0.27-2/init_l... Fix trivial bashisms.

http://patch-tracker.debian.org/patch/series/view/alsa-tools/1.0.27-2/manpag... Spelling fixes.

http://patch-tracker.debian.org/patch/series/view/alsa-tools/1.0.27-2/ld10k1... More -Werror=format-security issues.

http://patch-tracker.debian.org/patch/series/view/alsa-tools/1.0.27-2/hda-ve... Extend check for sys/io.h for other platforms in Debian that don't support it. Alternatively, the check could be reversed to chef for the architectures currently know to have the header, as stated in the comment. This fixes the build in a number of Debian architectures.

http://patch-tracker.debian.org/patch/series/view/alsa-tools/1.0.27-2/envy24... This one is to be discussed. envy24control currently creates a visible directory in $HOME, which is quite unacceptable. This patch just moves the directory to a dotdir, but an upgrade path might need to be coded. I would just move the thing away.

As Mark suggested, the envy24control should look for the old config files, too.

http://patch-tracker.debian.org/patch/series/view/alsa-tools/1.0.27-2/as10k1... For convenience to our build systems, add an alias for clean.

http://patch-tracker.debian.org/patch/series/view/alsa-tools/1.0.27-2/usx2yl... While not ready for submission, this one probably reveals a hardcoded path that should be made configurable, if not all distros store firmware in this dir. Not sure about Fedora, at least. /etc/hotplug looks plain old, anyway.

Other patches look good, please, post.

ALSA-OSS: http://patch-tracker.debian.org/package/alsa-oss/1.0.25-1

http://patch-tracker.debian.org/patch/series/view/alsa-oss/1.0.25-1/aoss_wra...

Make it obvious aoss is not doing anything useful if ALSA is not present.

http://patch-tracker.debian.org/patch/series/view/alsa-oss/1.0.25-1/aoss_lib... Fix path to libaoss on the multiarch case.

No objections, please, post.

Jaroslav