Pulseaudio fails to run when system's language is in certain non-English locales

Απρ 17 08:28:12 hostname pulseaudio[952]: [pulseaudio] module-alsa-card.c: Failed to parse ignore_dB argument.
Απρ 17 08:28:12 hostname pulseaudio[952]: [pulseaudio] module.c: Failed to load module "module-alsa-card" (argument: "device_id="0" name="pci-0000_00_03.0" card_name="alsa_card.pci-0000_00_03.0" namereg_fail=false tsched=ναι fixed_latency_range=όχι ignore_dB=όχι deferred_volume=ναι use_ucm=ναι card_properties="module-udev-detect.discovered=1""): initialization failed.
Απρ 17 08:28:12 hostname pulseaudio[952]: [pulseaudio] module-alsa-card.c: Failed to parse ignore_dB argument.
Απρ 17 08:28:12 hostname pulseaudio[952]: [pulseaudio] module.c: Failed to load module "module-alsa-card" (argument: "device_id="1" name="pci-0000_00_1b.0" card_name="alsa_card.pci-0000_00_1b.0" namereg_fail=false tsched=ναι fixed_latency_range=όχι ignore_dB=όχι deferred_volume=ναι use_ucm=ναι card_properties="module-udev-detect.discovered=1""): initialization failed.
Απρ 17 08:28:12 hostname pulseaudio[1021]: [pulseaudio] pid.c: Daemon already running.

strange that those parameters are localised

Status changed to 'Confirmed' because the bug affects multiple users.

Actually it does not concern only realtek but my intel hd audio card too. Thanks for the workaround...

This .po file seems to work without problems.

Compile the file to pulseaudio.mo and move it to /usr/share/locale-langpack/el/LC_MESSAGES/

$ msgfmt -o pulseaudio.mo -v ~/Downloads/el.po
$ sudo mv /usr/share/locale-langpack/el/LC_MESSAGES/pulseaudio.mo ~/pulseaudio.mo.bad
$ sudo mv ~/pulseaudio.mo /usr/share/locale-langpack/el/LC_MESSAGES/
$ sudo reboot

Right click and save as el.po

Then proceed with the instructions above (comment #6).

It works for me with solution of NickTh (Kubuntu 15.04) thank you

Further testing I've made showed different results.
I've downloaded the source code of pulseaudio in Vivid

$ apt-get source pulseaudio

then I compiled the pulseaudio-6.0/po/el.po , WITHOUT change anything and I replaced the pulseaudio.mo with the new compiled pulseaudio.mo.

Guess what, audio works.

So, maybe something is going on with the Machine Object file and has nothing to do with translation ?

The relevant message in /var/log/syslog is:

Apr 20 14:03:45 user-laptop pulseaudio[1661]: [pulseaudio] module.c: Failed to load module "module-alsa-card" (argument: "device_id="0" name="pci-0000_00_1b.0" card_name="alsa_card.pci-0000_00_1b.0" namereg_fail=false tsched=ναι fixed_latency_range=όχι ignore_dB=όχι deferred_volume=ναι use_ucm=ναι card_properties="module-udev-detect.discovered=1""): initialization failed.

What is says here is that "insmod" (or similar) was called with parameters like "tsched=ναι", and failed.

"git blame" says:
d806b197 src/pulsecore/core-util.h (poljar (Damir Jelić) 2013-06-27 19:28:09 +0200 92) static inline const char *pa_yes_no(bool b) {
cd13fb36 src/pulsecore/core-util.h (Tanu Kaskinen 2014-03-24 09:17:53 +0200 93) return b ? _("yes") : _("no");

which means that about a year ago a change was made to make yes/no localisable.

A quick fix is to make pa_yes_no() non-localisable again.

Sounds like sound is broken for everyone when in certain locales? That'll be Importance: High then.

I've tested and uploaded a fix now, it's waiting for an archive admin to approve. Thanks to all of you who helped reporting / triaging / analysing this issue, in this case it was just a matter of packaging up Simos patch, test and upload.

We currently use pa_yes_no to write module arguments, so they can not be
localised. Instead add a new pa_yes_no_localised function and use it in pactl
(and thus, revert all other places to use the non-localised version).

BugLink: https://bugs.launchpad.net/ubuntu/+source/pulseaudio/+bug/1445358
Signed-off-by: David Henningsson <email address hidden>
---
 src/pulsecore/core-util.h | 4 ++++
 src/utils/pactl.c | 20 ++++++++++----------
 2 files changed, 14 insertions(+), 10 deletions(-)

diff --git a/src/pulsecore/core-util.h b/src/pulsecore/core-util.h
index dcdc40e..d5a2d39 100644
--- a/src/pulsecore/core-util.h
+++ b/src/pulsecore/core-util.h
@@ -90,6 +90,10 @@ int pa_parse_boolean(const char *s) PA_GCC_PURE;
 int pa_parse_volume(const char *s, pa_volume_t *volume);

 static inline const char *pa_yes_no(bool b) {
+ return b ? "yes" : "no";
+}
+
+static inline const char *pa_yes_no_localised(bool b) {
     return b ? _("yes") : _("no");
 }

diff --git a/src/utils/pactl.c b/src/utils/pactl.c
index a13c7fb..d6791d4 100644
--- a/src/utils/pactl.c
+++ b/src/utils/pactl.c
@@ -196,7 +196,7 @@ static void get_server_info_callback(pa_context *c, const pa_server_info *i, voi
              pa_context_get_server(c),
              pa_context_get_protocol_version(c),
              pa_context_get_server_protocol_version(c),
- pa_yes_no(pa_context_is_local(c)),
+ pa_yes_no_localised(pa_context_is_local(c)),
              pa_context_get_index(c),
              pa_context_get_tile_size(c, NULL));

@@ -302,7 +302,7 @@ static void get_sink_info_callback(pa_context *c, const pa_sink_info *i, int is_
            pa_sample_spec_snprint(s, sizeof(s), &i->sample_spec),
            pa_channel_map_snprint(cm, sizeof(cm), &i->channel_map),
            i->owner_module,
- pa_yes_no(i->mute),
+ pa_yes_no_localised(i->mute),
            pa_cvolume_snprint_verbose(cv, sizeof(cv), &i->volume, &i->channel_map, i->flags & PA_SINK_DECIBEL_VOLUME),
            pa_cvolume_get_balance(&i->volume, &i->channel_map),
            pa_volume_snprint_verbose(v, sizeof(v), i->base_volume, i->flags & PA_SINK_DECIBEL_VOLUME),
@@ -409,7 +409,7 @@ static void get_source_info_callback(pa_context *c, const pa_source_info *i, int
            pa_sample_spec_snprint(s, sizeof(s), &i->sample_spec),
            pa_channel_map_snprint(cm, sizeof(cm), &i->channel_map),
            i->owner_module,
- pa_yes_no(i->mute),
+ pa_yes_no_localised(i->mute),
            pa_cvolume_snprint_verbose(cv, sizeof(cv), &i->volume, &i->channel_map, i->flags & PA_SOURCE_DECIBEL_VOLUME),
            pa_cvolume_get_balance(&i->volume, &i->channel_map),
            pa_volume_snprint_verbose(v, sizeof(v), i->base_volume, i->flags & PA_SOURCE_DECIBEL_VOLUME),
@@ -578,8 +578,8 @@ static void get_card_info_callback(pa_context *c, const pa_card_info *i, int is_

         printf(_("\tProfiles:\n"));
         for (p = i->profiles2; *p; p++)
- printf("\t\t%s: %s (sinks: %u, sources: %u, priority: %u, available: %s)\n", (*p)->name,
- (*p)->description, (*p)->n_sinks, (*p)->n_sources, (*p)->priority, pa_yes_no((*p)-...

On Mon, 2015-04-20 at 17:41 +0200, David Henningsson wrote:
> We currently use pa_yes_no to write module arguments, so they can not be
> localised. Instead add a new pa_yes_no_localised function and use it in pactl
> (and thus, revert all other places to use the non-localised version).
>
> BugLink: https://bugs.launchpad.net/ubuntu/+source/pulseaudio/+bug/1445358
> Signed-off-by: David Henningsson <email address hidden>
> ---
> src/pulsecore/core-util.h | 4 ++++
> src/utils/pactl.c | 20 ++++++++++----------
> 2 files changed, 14 insertions(+), 10 deletions(-)

Looks good to me.

--
Tanu

The attachment "Make pa_yes_no() non-localisable." seems to be a patch. If it isn't, please remove the "patch" flag from the attachment, remove the "patch" tag, and if you are a member of the ~ubuntu-reviewers, unsubscribe the team.

[This is an automated message performed by a Launchpad user owned by ~brian-murray, for any issues please contact him.]

This bug was fixed in the package pulseaudio - 1:6.0-0ubuntu6

---------------
pulseaudio (1:6.0-0ubuntu6) vivid; urgency=medium

  * debian/patches/0099-pa-yes-no.patch:
    - Fix pulseaudio failure to load in some non-English
    locales (LP: #1445358)
 -- David Henningsson <email address hidden> Mon, 20 Apr 2015 16:19:10 +0200