[mythtv-users] DVD iso Playback Fails After Upgrade to 0.28

[Richard Hulme]

On 04/10/17 03:16, Jay Foster wrote:
> On 10/03/2017 11:17 AM, Richard Hulme wrote:
>> On 03/10/17 19:49, Richard Hulme wrote:
>>> On 02/10/17 17:42, Jay Foster wrote:
>>>>>
>>>> file THE\ MATRIX.iso
>>>> THE MATRIX.iso: # UDF filesystem data (version 1.5) 
>>>> 'THE_MATRIX_16X9LB_N_AMERICA'
>>>>
>>>> isoinfo -p -i "THE MATRIX.iso"
>>>> Path table starts at block 257, size 3222
>>>>     1:    1 105
>>>>     2:    1 106 CABINETS
>>>>     3:    1 107 SETUP
>>>>     4:    1 108 VIDEO_TS
>>>>     5:    1 109 WHV
>>>
>>> OK, so working on the theory that maybe the UDF library has an issue, 
>>> could you try running the frontend directly from the command line?
>>>
>>> Unforuntately the library writes its error messages to stderr  so we 
>>> can't easily redirect them to our logs.  If you only look at the log 
>>> file, you won't see them.  If you run the frontend from the command 
>>> line, everything will be written to the console.  It's probably not 
>>> necessary to start the frontend with any additional logging 
>>> parameters as we've already determined there's nothing there to help us.
>>
>> Actually, start the frontend like this to enable more logging within 
>> the library:
>>
>> UDFREAD_TRACE=1 mythfrontend
>>
>> Richard
>>
>> .
>> _______________________________________________
>> mythtv-users mailing list
>> mythtv-users at mythtv.org
>> http://lists.mythtv.org/mailman/listinfo/mythtv-users
>> http://wiki.mythtv.org/Mailing_List_etiquette
>> MythTV Forums: https://forum.mythtv.org
>>
> Here is the stdout/stderr output from the UDF library:
> udfread TRACE: ECMA 167 Volume, BEA01
> udfread TRACE: ECMA 167 Volume, NSR02
> udfread LOG  : Found Anchor Volume Descriptor Pointer from lba 256
> udfread TRACE: reading Volume Descriptor Sequence at lba 32, len 32768 
> bytes
> udfread LOG  : Primary Volume Descriptor in lba 32
> udfread LOG  : Partition Descriptor in lba 34
> udfread LOG  :   partition 0 at lba 665, 4097670 blocks
> udfread LOG  : Logical volume descriptor in lba 35
> udfread LOG  : Found UDF 1.02 Logical Volume
> udfread LOG  : File Set Descriptor location: partition 0 lba 0 (len 4096)
> udfread LOG  : Volume Identifier: THE_MATRIX_16X9LB_N_AMERICA
> udfread LOG  : Partition map count: 1
> udfread TRACE: map 0: type 1
> udfread LOG  : partition map: 0: type 1 partition, ref 0
> udfread TRACE: reading root directory fsd from part 0 lba 0
> udfread TRACE: reading metadata from part 0 lba 0
> udfread LOG  : root directory in part 0 lba 2
> udfread TRACE: file entry size 2048 bytes
> udfread TRACE: reading metadata from part 0 lba 2
> udfread TRACE: reading metadata from part 0 lba 3
> udfread TRACE: directory size 536 bytes
> udfread ERROR: unexpected tag 0 in directory file
> udfread ERROR: error reading root directory

OK, this seems to be the issue.  The latest version of libudfread has a 
change around the 'unexpected tag' line to prevent an out-of-bounds 
read.  I'm guessing that's what's happening here.  Possibly it's due to 
the image containing a lot of files in the root directory and so using 
more than one LBA to store them.

I've updated the sources in master to the latest greatest.  Assuming 
there aren't any problems with the buildbots, I'll push it to the v29 
branch too.

0.28 might take a little longer as that is on an older version of the 
library than 29.  I need to check that just upgrading libudfread won't 
break anything.

Richard

.