[mythtv] Some bugs

Bruce Markey bjm at lvcm.com
Fri Jan 30 16:23:48 EST 2004


Isaac Richards wrote:
> On Thursday 29 January 2004 04:00 pm, Bruce Markey wrote:
...
> Fixed these two.

Great...

> Are the same thing -- nuppeldecoder.cpp (and ivtvdecoder.cpp, really) needs to 
> be updated to match Doug Larrick's positionmap changes to avformatdecoder.  
> That should fix these two (and similar) issues.  I asked for easy bugs, not 
> time consuming ones =)

I was hoping to sucker you in with the first two =). Seriously,
I thought these might be easy.

> Not sure how to fix these two..  Would you think you'd want it to discard all 
> queued keys for new channel changes in the same direction after a channel 
> change has finished?  I could see that not being desirable, though..  Hmm.

I first discovered the channel change by holding the key until
I reached a station many channels away. It changed channels
for the next two minutes. You can't judge how many repeats
are queued to know if you will land on a certain station so
maybe it may be best to discard and move one channel at a time.
On the other hand if you wanted to go up three channels you
should be able to tap Up three times. Donno.

The skip forward is more noticeable because skipping forward
behaves differently for in progress than for recorded. However,
this might be indirectly related to how positionmapping is
handled.

>>- Manual Schedule fails if the starttime matches the starttime
>>and chanid of a show already in the listings.
> 
> 
> Fixed this one with a replace into instead of insert, I think.

Seems to work. Down side is that the user loses the original
program data which she must not have wanted anyway. Up side is
that you could now use manual record to re-enter a title and
description for an existing timeslot when needed.

Spoofing program entries may be a bad idea in general as they
are blown away by mythfilldatabase if you set them up more
than a day ahead of time. In the future there should probably
be a manual record type, of better yet, a manual record bool
in the record table that will tell the scheduler to forgo
checking against the program table and just accept the chanid,
starttime for the record rule.


David fixed a mis-behavior in the search popup and I've sent
Robert some filter doc updates. With that, I think I'm done
tinkering with 0.14 (unless I find a show-stopper =).

--  bjm


More information about the mythtv-dev mailing list