[mythtv] Mythtv 0.23 RC Segfault in Mythfrontend

[tim at electronghost.co.uk]

On Mon, March 29, 2010 3:01 pm, R. G. Newbury wrote:
> On 03/28/2010 07:18 PM, stuart wrote:
>>
>>
>> Steve Berlo wrote:
>>> Hello all,
>>>
>>> I just upgraded to Mythtv 0.23 and am getting a segfault when I
>>> attempt to watch any recordings or videos. Here are a few things I
>>> have tried in troubleshooting: Since there were audio changes in this
>>> release, I have tried setting audio to NULL, I have tried several of
>>> the different video output profiles and methods, and several different
>>> recordings (I have an HD-PVR, Air2PC, and PVR-250, and all of the
>>> recordings have the same results when I attempt to play them), I am
>>> running the current version of the NVidia drivers from their web site
>>> - I just tried the newest NVIDIA-Linux-x86_64-195.36.15-
>>> pkg2.run, but previously I was running 190.53 and had the same symptoms.
>>>
>>> The backtrace shows this segfault:
>>>
>>> Program received signal SIGSEGV, Segmentation fault.
>>> 0x0000003a7d86ca2c in QMutex::lock() () from /usr/lib64/libQtCore.so.4
>>>
>>> (I can email the whole backtrace file and mythfrontend logs if
>>> necessary - they are too large to post to the list)
>>>
>>> I am running on Fedora 12 x64 on AMD (I have another system that is
>>> running Fedora 12 x32 on Intel - same symptoms)
>>>
>>> Here is some system info that might be relevant:
>>>
>>> [root at server ~]# uname -r
>>> 2.6.31.12-174.2.22.fc12.x86_64
>>> [root at server ~]# rpm -qa|grep qt -i
>>> PackageKit-qt-0.5.7-1.fc12.x86_64
>>> avahi-qt3-0.6.25-5.fc12.x86_64
>>> PyQt4-4.7-1.fc12.x86_64
>>> dbus-qt-0.70-7.fc12.x86_64
>>> google-gadgets-qt-0.11.1-4.fc11.x86_64
>>> qt-sqlite-4.6.2-3.fc12.i686
>>> qt3-devel-3.3.8b-28.fc12.i686
>>> dbus-qt-0.70-7.fc12.i686
>>> qt-4.6.2-3.fc12.i686
>>> polkit-qt-0.95.1-3.fc12.x86_64
>>> poppler-qt4-0.12.4-2.fc12.x86_64
>>> qt-devel-4.6.2-3.fc12.x86_64
>>> qt-sqlite-4.6.2-3.fc12.x86_64
>>> qt-mysql-4.6.2-3.fc12.x86_64
>>> qt3-devel-3.3.8b-28.fc12.x86_64
>>> qt-x11-4.6.2-3.fc12.i686
>>> qt3-3.3.8b-28.fc12.x86_64
>>> qtscriptbindings-0.1.0-10.fc12.x86_64
>>> pinentry-qt-0.7.6-4.fc12.x86_64
>>> qt-x11-4.6.2-3.fc12.x86_64
>>> qt-4.6.2-3.fc12.x86_64
>>> qt3-3.3.8b-28.fc12.i686
>>> avahi-qt3-0.6.25-5.fc12.i686
>>>
>>> Thanks for all of your hard work! I appreciate any guidance you can
>>> give me that can help me solve this problem.
>>>
>>> Steve
>
>> I have seen this as well on my Debian systems. I tried to create a
>> trouble ticket:
>
>> Feel free to add your back trace to the above trouble ticket if you
>> think that might help.
>>
>> ---
>> I've been using mythtv for years - thanks everyone for contributing!
>
> User error.
>
> Your uname -r says you are running x86_64 but you seem to be running a
> bunch of plain  i686 packages, especially qt.
>
> Here's my output of the same test, ALL X86_64:
>
> [root at tor1 etc]# rpm -qa|grep qt -i
> qt-x11-4.6.2-3.fc12.x86_64
> qt3-3.3.8b-28.fc12.x86_64
> qt3-MySQL-3.3.8b-28.fc12.x86_64
> polkit-qt-0.95.1-3.fc12.x86_64
> qt-4.6.2-3.fc12.x86_64
> PyQt4-4.7-1.fc12.x86_64
> pinentry-qt-0.7.6-4.fc12.x86_64
> poppler-qt4-0.12.4-2.fc12.x86_64
> qt-mysql-4.6.2-3.fc12.x86_64
> qtscriptbindings-0.1.0-10.fc12.x86_64
> qt-devel-4.6.2-3.fc12.x86_64
> avahi-qt3-0.6.25-5.fc12.x86_64
> PackageKit-qt-0.5.7-1.fc12.x86_64
> qt-sqlite-4.6.2-3.fc12.x86_64
>
>
> Your repo's are messed up. You might have to browse to a repo mirror,
> and force an install the correct architecture.

Err, Fedora's multilib. You can install both 32 and 64-bit libraries. The
32-bit libraries go in /usr/lib and the 64-bit libraries in /usr/lib64. The
OP's list contains both sets. There's nothing wrong with that, or with his
repositories. If you tell yum to install a package without specifying the
architecture, it will normally install both 32 and 64-bit versions if both are
available, since it doesn't know what sort of binary you will run.

If there were anything wrong with that, myth wouldn't run at all, as the
runtime linker would fail to find the right libraries.

-- 
Tim Smith <tim at electronghost.co.uk>
There are spaces in the world which should not be filled.
This is one of them.