Richard Shaw wrote: Looks like they've broken the libavcodec encoder plugin entirely. The fix (when they come up with one) should also fix AAC. (And as I wrote in Bugzilla, the 2.6 branch has AAC enabled in the libavcodec encoder plugin, so when 2.6.x will be released, we won't have to maintain that AAC patch anymore.) Kevin Kofler

Okay, good news and bad news. The patch works as I could encode a small video with MP2 audio with the (lav) codec but when I tried encoding the same video with AAC audio it didn't give any errors but sat with the conversion window open and maxing out one core but doing nothing. Richard

Ping... Anyone have any suggestions? What's the best way to troubleshoot a hangup with CPU usage spike? gdb? strace? valgrind? Thanks, Richard

Ok, I whoops'd and didn't notice my replies were going straight to Hans instead of the list so here's the short version for anyone else following this thread: On Tue, Jun 21, 2011 at 1:53 AM, Hans de Goede <j.w.r.degoede(a)gmail.com&gt; wrote: Yup, I installed the avidemux-debuginfo but not all the dependencies. I hope I don't need them as it's an extensive list. I usually run the program from gdb, but you're saying there's a way to attach gdb to it after the fact? On Wed, Jun 22, 2011 at 3:45 AM, Hans de Goede <j.w.r.degoede(a)gmail.com&gt; wrote: [SNIP] Well, I don't know why this works but it does... I first tried: # gdb /usr/bin/avidemux2_qt4 set logging on set logging file avidemux.log run <HANG> <CTL-C> thread apply all bt full quit This saved the output of avidemux but not the backtrace. Then I found a gentoo page[1] and they did things a little differently: # gdb /usr/bin/avidemux2_qt4 run <HANG> <CTL-C> set logging on set logging file avidemux_bt.log thread apply all bt full set logging off quit http://pastebin.com/iqsBF8fj Thanks, Richard [1] http://www.gentoo.org/proj/en/qa/backtraces.xml I looked at some of the output but I didn't see anything that jumped out at me as wrong. There must be a way to redirect the gdb output to a file which may make it easier to browse through. Thanks, Richard Update... I don't know if you've looked at the bt yet as I've updated it twice now :) I installed SOME of the missing debuginfo packages (at least the ones that might actually be useful, like glibc). I didn't both with the QT ones since it's huge and the problem is limited to the AAC encoder so it's probably not relevant. If you look at line 360 there seems to be a typo: vfPlugins = 0x16aedf0 "lugins/videoEncoder/" It looks like the P is missing but I would think that if that's the problem then it would mess up all the encoders, not just AAC. ... Grep'd though the source and didn't see the problem, guess it's not related. Thanks, Richard

On Thu, Jun 23, 2011 at 3:14 AM, Hans de Goede <j.w.r.degoede(a)gmail.com&gt; wrote: In the decoder and not the encoder? Strange, if that's the case then you would expect it to hang on all encoders, not just AAC. Well, I grep'd through the source and the only references I could find to "lugins/videoEncoder/" was in some Cmake files so I then assumed the problem was a cmake bug since the references had the "p". I ran cmake through rpmbuild so all the config options would be identical to what would get built under mock and grep'd through that as well. Of course after config there was a lot more hits so I grep'd for both "lugins" and "plugins" and diffed the two files and they were identical so this looks to be just a gdb quirk. So I'm still lost as to what's going on and would really like to get 2.5.5 released... Thanks, Richard

On Thu, Jun 23, 2011 at 9:19 AM, Hans de Goede <j.w.r.degoede(a)gmail.com&gt; wrote: I'll report it upstream and see what they say. Thanks, Richard