Re: [Gnash-dev] updated binary packages

[John Gilmore]

>   The lucid packages were built in a chroot, and
> gstreamer0.10-fluendo-mp3 isn't even installed. Digging into this,
> somebody (could have been me ages ago) added this to the debian/control
> file. I just removed this from the control file, and uploaded fresh
> Lucid-i386 debs to our repository.

Thanks -- it installed cleanly and works pretty well.

While running, gnash identifies itself as

  "Gnash 0.8.10~git.master20401
   (0.8.10~git.master20401-20110426-none)"

So it's not a straight 0.8.10 release?

BTW, I'm getting sparkles of garbage along the upper right-hand margin
of some youtube clips, using this version.  It's in a narrow vertical
line to the right of the playback.  Here's a sample:

  http://www.youtube.com/watch?v=itl44eSdebg
  http://www.toad.com/Temp-gnash-sparkle.png

The sparkle isn't in the FLV; I have "save streaming media" set, and
playing the resulting file back with Totem works fine.

Also, when I play youtube embedded in a webpage here:

  https://www.eff.org/pages/UGC-test-suite
  (try the "Nixon Peabody" clip or "Something About Commandments")
  http://www.toad.com/Temp-gnash-trunc.png

I'm only getting the upper left corner of the video; it's cut off at
the bottom and along the right, rather than scaling to the size of the
video playback area on the web page.  However, when I stop the playing
with the pop-up Movie Control menu, then it suddenly resizes to fit,
and works thereafter.  See the screenshot on www.toad.com.

Also, Gnash still burns a lot of CPU time (but not 100%!) after
showing a youtube video.  It's better...for example this page:

  http://www.villagetelco.org/2011/02/in-a-village-telco-minute/

seems to burn about 11-15% of one CPU when doing nothing (just waiting
for someone to click on the big centered Play button to start the
video coming in from youtube).  Strace shows lots of poll() calls with
timeouts of 0 ms or 9 ms, plus dozens of gettimeofday() calls, an
ioctl(,FIONREAD) and a read() returning EAGAIN, continuously.

One more bug: the plugin isn't wait()ing for child processes (nor
asking the browser to do so), so it accumulates zombies:

% ps ax --forest   (edited down)
 6474 ?        S      0:00 /bin/sh /usr/lib/firefox-3.6.10/firefox
 6479 ?        S      0:00  \_ /bin/sh /usr/lib/firefox-3.6.10/run-mozilla.sh /u
 6483 ?        Sl     2:33      \_ /usr/lib/firefox-3.6.10/firefox-bin
 6560 ?        Sl     1:29          \_ /usr/bin/gtk-gnash -u http://s.ytimg.com/
 6630 ?        Z      5:07          \_ [gtk-gnash] <defunct>
 6686 ?        Z      2:34          \_ [gtk-gnash] <defunct>

There's some well defined way to handle this, because other plugins
that fork players (e.g. playing an Ogg directly in the browser from
its URL) don't do this.

        John