[mythtv-users] Need sanity check re: repairing seektable and commflagging

[Craig Huff]

On Wed, May 13, 2009 at 5:43 PM, Michael T. Dean
<mtdean at thirdcontact.com> wrote:
> On 05/13/2009 06:01 PM, Craig Huff wrote:
>> On Wed, May 13, 2009 at 11:22 AM, Michael T. Dean wrote:
>>> On 05/13/2009 11:32 AM, Craig Huff wrote:
>>>
>>>>
>>>> snip...
>>>>
>>>> After reading wiki pages and e-mail archives it appears that my next
>>>> steps
>>>> are:
>>>>
>>>> $ for file in "list of last night recording filenames" ;
>>>> do
>>>>  mythtranscode --mpeg2 --buildindex --allkeys --showprogress --infile
>>>> $file
>>>>  mythcommflag --file $file
>>>> done
>>>>
>>>> Or, alternately take the mythcommflag command out of the do-loop and
>>>> use this after the do-loop:
>>>> $ mythcommflag --allstart 20090512130000  #for everything recorded
>>>> after 1pm yesterday
>>>
>>> Make sure you've run optimize_mythdb.pl first.  Then, build seektables
>>> for any new recordings.  The mythcommflag approach won't work for
>>> MPEG-2 (unless you're running trunk), so you'll probably need to do
>>> the mythtranscode approach.
>>>
>> My apologies for not being more clear.  I agree that I first need to
>> use mythtranscode since the MPEG-2 files are ones produced from
>> PVR-500 captures.  The questions were A) whether the mythtranscode
>> command had all the right options, and B) that my guess was correct in
>> that I afterwards need to run mythcommflag (one flavor or the other)
>> to get all the commercials flagged for skipping during viewing.
>>
>> It sounds like we are in agreement.
>
> Yep.  That's right.
>
> I'm sorry--I didn't even look at the options on the mythcommflag to notice
> it wasn't a --rebuild .
>
> Best info on rebuilding the seektable (and args to use) is:
> http://www.mythtv.org/wiki/index.php/Repairing_the_Seektable
>
Oops.  Got busy enjoying MythTV and forgot to report back on the results. %^)

Just to close this thread, it turned out that I only needed to run the
mythtranscode command on the files.  That ran faster than I expected
(less than 30 seconds / hour of recording, maybe even < 10 seconds /
hr), so I got a little paranoid ;-).  Opened a frontend session and
got into one of the recordings and opened up the edit to try jumping
around and see that the recording length made sense.  To my surprise
the  commercial flagging data was there, so I didn't need to run
commflag on any of the recordings. I checked another for good measure
and it was fine, too.

I am constantly amazed at how something so complex to set up and ...
erm ... "challenging" to configure works so well once you get it
going.  Just love using it.  Cheers for all the development effort
that has gone into this program!

Craig.