[mythtv-users] MythArchive import failing if archived chanid isn't defined on local machine

[John Pilkington]

On 02/02/14 03:32, HP-mini wrote:
> On Wed, 2014-01-29 at 14:13 +0000, John Pilkington wrote:
>> failed
>> partway through the mythutil --copyfile process.
>>
> I've seen that in recent past 0.26-0.27..
> Except for seektable info, mytharchive native export/import used to work
> very reliably.
>
>> The UI shows suggested local values for these fields and for Channel
>> Name and Callsign.  At present attempts to change the offered chanid
>> 3104 suggest 1104 but won't accept it and just hang.  I can probably
>> rescan my Video Sources in a different order to get a match, or perhaps
>> just define a matching dummy channel, but this looks like a bug.
>>
>> ..or am I doing something wrong?
>>
> I've successfully(**) mytharchive export/imported 2 recordings (<1GB)
> into a isolated BE/FE system (running master) that did not have matching
> channel numbers/names/callsign.
>
> I found it strange that it appears to offer options to change these
> details but refuses to accept any...
>
> **- no seektable or markup import & TFW buffer warnings..files appear to
> be complete. Source USB HDD 2.5" & small 800MB files.
>
> There are recent changes in mytharchivehelper/main.cpp &
> D/th-readed-file-writer code.
> Changes in mytharchivehelper were meant to fix the seektable/markup
> export/import..
>
> When fed with almost unlimited input (e.g. dummy import recorder) the
> TFW used to eat all RAM as buffer & corrupt output file & not release
> memory.
>
> I would manually copy over the recording after using mytharchive import
> to be sure the file is complete & not corrupted.
>

I did, IIRC, get import to report completion after changing the order of 
definition of video sources to give a matching chanid.   Title, 
subtitle, recording date were ok but no seektable.  By then I had made 
several attempts and the DB might well have been confused.  I've gone 
back now to an SD only main box so that I can play with HD recordings on 
the testbed.  And, in passing, my/Panasonic's upnp discrimination 
between the same mpeg2-ps file as a recording or a video still puzzles 
me.  I've been trying mythbackend --setverbose upnp (on a running 
backend) but tend towards info overload.

The simplest way I have found of getting a recording from one box to the 
other is to copy the file on top of a short new recording and 
mythcommflag --rebuild.  But some people want metadata...  Parts of that 
are editable, but the rebuild doesn't update the format icons or, I 
would guess, the artwork.