Closed GoogleCodeExporter closed 8 years ago
I'm gonna need to see a gnome-mplayer -v log of it doing this. gnome-mplayer
should
only access the cdda device when you select, file -> open disc -> open audio cd
Can you also post your $HOME/.mplayer/config file?
Original comment by kdeko...@gmail.com
on 5 May 2010 at 12:58
There isn't much in it:
liviu@debian-liv:~$ cat .mplayer/config
# Write your default config options here!
[gnome-mplayer]
vo=xv
msglevel=all=5
ao=alsa
vf=eq2
As for the -v log, whenever this occurs I kill the app asap, otherwise I risk a
system freeze and a reboot. Is it possible to redirect the log output to a file
that
will survive the app kill?
Original comment by landroni...@gmail.com
on 5 May 2010 at 1:24
The logging is not enabled unless you add -v to the command line so something
like
this should work.
gnome-mplayer -v > ~/gnome-mplayer.log
So you are saying it doesn't happen all the time?
Also, is there a /etc/mplayer/config or /etc/mplayer/mplayer.conf file?
Also what version of mplayer do you have?
Original comment by kdeko...@gmail.com
on 5 May 2010 at 1:28
No, it happens randomly, usually couple of times in a week. This log command
works
fine, so I'll send a debug as soon as I encounter the behaviour again.
See /etc/mplayer/mplayer.conf attached.
Original comment by landroni...@gmail.com
on 5 May 2010 at 2:00
Attachments:
What version of mplayer do you have?
Also, I run gnome-mplayer all the time, and have never seen this behavior. Is it
possible you have the default playlist enabled? Edit->Preferences->Interface
(4th
checkbox down) and it has cdda items in it?
Original comment by kdeko...@gmail.com
on 5 May 2010 at 3:06
[deleted comment]
Synaptic says it's 1.0.rc2svn20100313-0.0 (on Debian testing)
liviu@debian-liv:~$ mplayer -v
MPlayer SVN-r30656 (C) 2000-2010 MPlayer Team
CPU vendor name: AuthenticAMD max cpuid level: 1
CPU: AMD Turion(tm) X2 Dual-Core Mobile RM-72 (Family: 17, Model: 3, Stepping:
1)
extended cpuid-level: 26
extended cache-info: 33587520
Detected cache-line size is 64 bytes
CPUflags: MMX: 1 MMX2: 1 3DNow: 1 3DNowExt: 1 SSE: 1 SSE2: 1 SSSE3: 0
Compiled with runtime CPU detection.
get_path('codecs.conf') -> '/home/liviu/.mplayer/codecs.conf'
Reading /home/liviu/.mplayer/codecs.conf: Can't open '/home/liviu/.mplayer/
codecs.conf': No such file or directory
Reading /etc/mplayer/codecs.conf: Can't open '/etc/mplayer/codecs.conf': No
such
file or directory
Using built-in default codecs.conf.
"Use default playlist" is unchecked, and the playlist contains no cdda items.
Original comment by landroni...@gmail.com
on 5 May 2010 at 4:56
if you create a new account on your machine and don't customize the desktop can
you
still cause this to occur? I'm wondering if you have a shortcut somewhere that
is
sending a dbus message to gnome-mplayer to open the cd?
Original comment by kdeko...@gmail.com
on 6 May 2010 at 10:25
I have trouble replicating this consistently even in my user account. As a
rule, the
behaviour always happened when I selected "Open with Gnome-mplayer" from
Thunar. I
don't know if any dbus message is being sent when doing so. Now I'm always
opening
video files in debug mode, so I hope it's a matter of time to get a pertinent
log.
Original comment by landroni...@gmail.com
on 7 May 2010 at 7:51
A small update. Since I began starting gnome-mplayer in debug I never
encountered
the offending behaviour. But using the same method as before (double-click in
Thunar), today I got the beahviour 4 times during five minutes. I'm not sure
what's
happening, but I have this theory:
- in the debug session, the file is played using a `gnome-mplayer -v /path/to
file
with spaces' (1st screenshot)
- in the normal file open, Thunar calls `gnome-mplayer
file:///path/to%20file%20with
%20spaces' (2nd screenshot)
I don't know whether it's the "file://" or "%20" causing the issues, but it
might be
a lead.
Original comment by landroni...@gmail.com
on 11 May 2010 at 8:09
Attachments:
Is gnome-mplayer compiled with gio support? It will tell you when you start it
with
the -v option. If not, recompile with that support enabled.
Original comment by kdeko...@gmail.com
on 11 May 2010 at 12:53
It seems that GIO is enabled.
liviu@debian-liv:~/Videos$ gnome-mplayer -v file:///tmp/copy%20of%20Flashf39k0f
GNOME MPlayer v0.9.9.2
vo = xv ao = alsa
Running with GIO support
[..]
I finally managed to get a debug, by opening the file as shown above. The
player
started playing all cdda:// tracks and towards track 21 (according to
libnotify) I
managed to xkill it. CPU was full-throttle. For comparison, I attach a normal
session (where the video is opened and played as expected) on the same video,
same
command, a couple of moments later.
Is there anything wrong in there?
Original comment by landroni...@gmail.com
on 11 May 2010 at 2:30
Attachments:
[deleted comment]
Minor update: the behaviour is not caused by the "%20" since I get the same
behaviour on
liviu@debian-liv:~/Videos$ gnome-mplayer -v file:///tmp/Flashf39k0f
Sofar I can only think of "file://" causing it.
Original comment by landroni...@gmail.com
on 11 May 2010 at 2:34
I think the playlist parser is opening the flash file and is thinking it is a
audio
cd. I would need the flash file to debug this issue. Are you sure these are FLV
files
and not SWF files. SWF files are not supported by gnome-mplayer.
Original comment by kdeko...@gmail.com
on 13 May 2010 at 5:27
It happens with just about any MegaVideo file. For example, try this [1]. I
would
think they're flash since mplayer plays them and I get video and audio.
[1] http://www.megavideo.com/?v=E2BFED4D
Item: /tmp/FlasheiLaGF
Type: Macromedia Flash Video
Mime: video/x-flv; charset=binary
Original comment by landroni...@gmail.com
on 13 May 2010 at 7:36
Sorry, but I do not want to sign up for that site.
Original comment by kdeko...@gmail.com
on 13 May 2010 at 7:58
Hmm, you wouldn't need to sign up for the site to access a stream. Would you
prefer
for me to send you/post a 5sec sample that produces the issue on my system?
Original comment by landroni...@gmail.com
on 13 May 2010 at 8:25
Yes, a sample would be sufficient, make it 10-15 seconds so something can be
seen in
mplayer. And attach it to this ticket if possible.
Original comment by kdeko...@gmail.com
on 13 May 2010 at 8:39
OK, here's a ~15sec sample. After several tries (here, between 5 and 20) it
should
produce the offensive behaviour.
Original comment by landroni...@gmail.com
on 13 May 2010 at 8:47
Attachments:
Ran about 100 tests of it with the SVN code... can't make it do it here.
Perhaps you
should try SVN.
Original comment by kdeko...@gmail.com
on 13 May 2010 at 8:57
Yesterday I tried SVN r1699, and I still get the offensive behaviour (less than
20
tries is usually sufficient). To be sure, I also confirmed this with a default
~/.config/gnome-mplayer. Did you make sure to open it as shown below?
liviu@debian-liv:~/Videos$ gnome-mplayer file:///tmp/Flashf39k0f
Otherwise, perhaps you could add some temporary debug messages that would help
isolate the issue?
Original comment by landroni...@gmail.com
on 14 May 2010 at 1:44
Still can't duplicate it.. I used this as my command line (same command line I
used
yesterday)
gnome-mplayer file:///home/kdekorte/Downloads/copy\ of\ Flash00TtgC
I also tried
gnome-mplayer file:///home/kdekorte/Downloads/copy%20of%20Flash00TtgC
Perhaps the --reallyverbose option will give a hint to the issue. Please run
with
that option and upload the output when it fails. I'm guessing when you have
problems
it will be around "opening playlist", but still unsure at this point.
Original comment by kdeko...@gmail.com
on 14 May 2010 at 2:01
See attached.
Original comment by landroni...@gmail.com
on 14 May 2010 at 3:13
Attachments:
Please apply the attached patch and reproduce the error (with --reallyverbose).
This
will help locate the problem area.
patch -p0 < debug.patch
in the gnome-mplayer directory
Original comment by kdeko...@gmail.com
on 14 May 2010 at 3:24
Attachments:
Here's the second --reallyverbose dbg using the patch.
Original comment by landroni...@gmail.com
on 14 May 2010 at 4:04
Attachments:
ok that is odd... none of my debug code was hit.
Can you add this line to line 160 in dbus-interface.c
printf("Adding %s via dbus\n", s);
should go right before: g_idle_add(add_to_playlist_and_play, g_strdup(s));
also add
printf("parse cdda with %s\n",filename);
at line 600 in support.c
should go right before: if (g_ascii_strncasecmp(filename, "cdda://", 7) != 0) {
And retest..
Original comment by kdeko...@gmail.com
on 14 May 2010 at 4:12
With the changse, it seems that the "parse cdda" message caught up, while "via
dbus"
didn't.
Original comment by landroni...@gmail.com
on 14 May 2010 at 7:46
Attachments:
I'm wondering if it something with the desktop file. Perhaps you should put
single
quotes around the value from Thunar.
Exec=gnome-mplayer %U
Maybe it should be
Exec=gnome-mplayer '%U'
Thunar should be passing files with spaces as %20 for the space. Since uri's
should
not have spaces in them.
Original comment by kdeko...@gmail.com
on 14 May 2010 at 7:52
Also, I just noticed in the log that the file is being reported by g_stat as a
block
device. This is obviously wrong in this case. So that is what is triggering the
cdda:// loading code. That code is really early in the loading, and that is why
it is
happening. So something is causing the filename to be picked up incorrectly.
I noticed something here that could be done a little better around here, so
please
give SVN r1706 a try.
Original comment by kdeko...@gmail.com
on 14 May 2010 at 8:30
I don't think this is particularly linked to Thunar, since I could replicate by
opening the file from the command line. In any case, I'm currently debugging
via
Thunar's custom actions "gnome-mplayer -v --reallyverbose file://%f >
/tmp/gnome-
mplayer.log"
I've just tried r1706 and so far so good. To be honest, today I'm probably
tired of
performing an extensive check. But I'll continue opening the files as
previously and
should the offensive behaviour pop back, I'll report here. Thanks for looking
into
this!
Original comment by landroni...@gmail.com
on 14 May 2010 at 9:15
Hello
I think it's pretty safe to close this bug. After several days I haven't bumped
into
the same issue, so it looks like the fix worked.
Original comment by landroni...@gmail.com
on 19 May 2010 at 9:54
Fixed
Original comment by kdeko...@gmail.com
on 19 May 2010 at 9:56
Original issue reported on code.google.com by
landroni...@gmail.com
on 5 May 2010 at 12:09