[mythtv-users] Re: libmyth-0.18.1-114.rhfc4.at VLD fails an my Via ...

[Jules Gosnell]

Axel Thimm wrote:

>On Mon, Nov 21, 2005 at 09:00:42AM +0000, Jules Gosnell wrote:
>  
>
>>Axel Thimm wrote:
>>
>>    
>>
>>>On Sun, Nov 20, 2005 at 12:55:04PM -0500, Isaac Richards wrote:
>>>
>>>
>>>      
>>>
>>>>On Sunday 20 November 2005 06:53 am, Jules Gosnell wrote:
>>>>  
>>>>
>>>>        
>>>>
>>>>>last night I found 113 and 114 SRPMs on an archive site somewhere,
>>>>>unpacked them, applied patches and then diffed them looking for any
>>>>>mention of XvMC - sure enough, I found one :
>>>>>
>>>>>diff -r 113/mythtv-0.18.1/libs/libavcodec/parser.c
>>>>>114/mythtv-0.18.1/libs/libavcodec/parser.c
>>>>>814c814,818
>>>>><     { CODEC_ID_MPEG1VIDEO, CODEC_ID_MPEG2VIDEO },
>>>>>
>>>>>          
>>>>>
>>>>>>    { CODEC_ID_MPEG1VIDEO,
>>>>>>      CODEC_ID_MPEG2VIDEO,
>>>>>>      CODEC_ID_MPEG2VIDEO_XVMC,
>>>>>>      CODEC_ID_MPEG2VIDEO_XVMC_VLD,
>>>>>>    },
>>>>>>            
>>>>>>
>>>>>I unpacked a 115, hacked its patch to undo this change, rebuilt
>>>>>overnight, installed and tested this morning - result - SUCCESS :-) !
>>>>>    
>>>>>
>>>>>          
>>>>>
>>>Does this imply that building w/o patching resulted in a failure, or
>>>did you just try that one build. Perhaps it is unrelated, and only a
>>>build w/o a patch will tell.
>>>
>>>
>>>      
>>>
>>I don't follow...
>>
>>113-5 all apply patches as part of the build. Vanilla 113 works for me, 
>>Vanilla 114/115 don't. All I did was to find a XVMC related difference 
>>between 113/114 and then remove that difference (from the patch that 
>>created it) in 115, creating a variant 115 binary which worked for me.
>>
>>I know that the vanilla 115 distro does not work for me.
>>I know that my hacked 115 distro (with a couple of lines removed from 
>>the svn patch) does work for me.
>>
>>I'm assuming that rebuilding a vanilla 115 SRPM would result in a 
>>vanilla 115 RPM so did not bother with that build.
>>    
>>
>
>The difference is not only in the patches, but also in the
>environment. I found out that FC's updates overwrote my libviaxvmc
>lib, so the packages had no VLD support indeed.
>
>I assume that if you simply rebuild the 115 sources you'll have
>success, too.
>  
>
Well, much to my chagrin, Axel, you are right :-)

I rebuilt the 115 SRPM last night and it works fine.

I think it probably comes down to the fact that the xorg that I am 
running includes a unichrome backend along with the relevant C header 
files, which eclipse those in the xvmc package. I guess these were 
similar enough for 113 to work without a rebuild, but by 114/115 libmyth 
needs rebuilding in my environment to work properly...

Anyway, a successful resolution - apologies for all the wasted bandwidth ...

Jules

>  
>
>>Is that clearer ?
>>
>>Jules
>>
>>    
>>
>>>      
>>>
>>>>>I have probably not so much fixed a bug as disabled some code that was
>>>>>enabled in 114, but it works for me - I now have h/w accelerated MPEG2
>>>>>on libmyth-*115* so I am happy.
>>>>>
>>>>>Thanks for the help, Axel.
>>>>>
>>>>>Should I let someone know about this somehow, or will you feedback to 
>>>>>them
>>>>>?
>>>>>    
>>>>>
>>>>>          
>>>>>
>>>>Looks to be an atrpms problem - this code is how it should be in the 
>>>>release-18-fixes branch.
>>>>  
>>>>
>>>>        
>>>>
>>>I think so, too. I'm looking into this, but I have no VLD hardware to
>>>test with, so I'll be fixing things blindly. Anyway, better than it
>>>not working at all, and I can verify that non-VLD stuff is still
>>>working.
>>>
>>>
>>>
>>>_______________________________________________
>>>mythtv-users mailing list
>>>mythtv-users at mythtv.org
>>>http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
>>>
>>>
>>>      
>>>
>>    
>>
>
>  
>
>------------------------------------------------------------------------
>
>_______________________________________________
>mythtv-users mailing list
>mythtv-users at mythtv.org
>http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
>  
>


-- 
"Open Source is a self-assembling organism. You dangle a piece of
string into a super-saturated solution and a whole operating-system
crystallises out around it."

/**********************************
 * Jules Gosnell
 * Partner
 * Core Developers Network (Europe)
 *
 *    www.coredevelopers.net
 *
 * Open Source Training & Support.
 **********************************/