[mythtv] issues with Windows NUV decode

[Lawrence Rust]

On Fri, 2011-04-15 at 23:07 +0100, jonathanjstevens at gmail.com wrote:
> Thanks Laurence - yes, I guess it was a bit of an assumption...
> 
> 
> I think the pertinent errors are:
> 
> 
> 2011-04-15 22:50:53.017 TV: StartPlayer(0, WatchingPreRecorded, main)
> -- begin
> 2011-04-15 22:50:53.017 TV: Elapsed time since TV constructor was
> called: 268 ms
> 2011-04-15 22:50:53.812 NuppelDecoder: lzo_init() failed, aborting

This error message only occurs if the function lzo_init() fails to
verify the compatibility of the caller's and callee's understanding of
various basic data types.  Could be several reasons:

1. mythfrontend is picking up an old installation of the libraries.  If
so track down and remove all previous myth installations on this Windows
system and then re-install.

2. Maybe a 64/32 bit architecture mismatch.  Was MythTV built on a
64-bit system or with the 64-bit mingw compiler?  If so it might be best
to re-compile on a 32-bit host or with the 32-bit mingw compiler.

> 2011-04-15 22:50:53.812 Player(0), Error: Could not initialize A/V
> decoder.
> 2011-04-15 22:50:53.812 Unable to open video file.
> 2011-04-15 22:51:13.814 playCtx, Error: StartPlaying() Failed to start
> player
> 2011-04-15 22:51:13.814 Player(0): StopPlaying - begin
> 2011-04-15 22:51:13.814 Player(0): Exited decoder loop.
> 2011-04-15 22:51:13.815 Player(0): StopPlaying - end
> 2011-04-15 22:51:13.815 TV: StartPlayer(0, WatchingPreRecorded, main)
> -- end error
> 
> 
> I obviously read NuppelDecoder as NUV - and (maybe because I've been
> upgrading for years now with much the same database?) the transcode
> jobs running on my backend produce NUV files. However, on testing, it
> would appear the same error occurs for both transcoded nuv files and
> pretranscoded mpgs. 

I suspect that this also is a version mismatch.

> 
> Anyway, here is the log...
> 
> 
[snip]

> There was also a string of the below errors, which escaped my ">"
> logging. Not sure if they are related.

Theses are not related, but they should be fixed - they're runtime
warnings from Qt.  If you can catch these in context with a
"mythfrontend --verbose most" log output then perhaps we can submit a
bug report and patch.

> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer:  invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> QEventDispatcherWin32::unregisterTimer: invalid argument
> _______________________________________________

-- 
Lawrence