[mythtv-users] everyday i have to make sure myth worked the day before

[David Engel]

On Fri, Apr 01, 2011 at 12:39:53PM -0400, Devin Heitmueller wrote:
> The second issue though, is the need for users to "poll" MythTV in
> general to find out if it did what it's supposed to.  I owned a Tivo
> in 2001, and when I turned on the TV it was nice enough to inform me
> when there were conflicts, or if something went wrong.  In 2011 we
> still don't have any comparable functionality in Myth (either in the
> main GUI or via some sort of email or other alerts).
> 
> Ask yourselves:  how many times have you sat down with some popcorn,
> turned on your MythTV box, hit play on a recording, and found out that
> there is a zero length file there?  Wouldn't it have been nice if Myth
> had proactively told you that much earlier?
> 
> Having a solution for the second problem makes the first class of
> problems much more bearable.

There is some very rudimentary support in MythTV for detecting missed
recordings, but it is sorely lacking.  The logical approach to
improving things is to extend the use of the oldrecorded table, which
essentially maintains past and current state for the scheduler, to
also maintain future state.  I've looked at doing this a few times,
but have never followed through because of two main problems.

First, unlike TiVo, MythTV reschedules everything whenever anything
changes.  This is nice because MythTV dynamically adapts to changes
much more nicely than TiVo.  The problem is that a lot of future state
would rewritten frequently and I'm afraid of how that would affect the
already long scheduling times that some people see.

Second, I've been afraid of the added complexity.  Because the
oldrecorded table doesn't contain scheduled recording end times, it's
not possible to directly tell which programs were missed.  That means
the scheduler has use some other way to determine when to ignore
future state (the normal case) and when to convert it to past state.

Anyway, I took another stab at things this weekend and the attached
patch appears to work in extremely limited testing.  If anyone's
feeling adventuresome, I'd like to hear how it works in more busy and
complicated configurations.  I'm limited to one tuner in my current
development setup so I can't test anything too elaborate right now.

Note: in order to try this patch, you will need to run the following
commands in mysql to handle the database changes.  If/when this gets
committed to trunk, these changes will be handled automatically.

ALTER TABLE oldrecorded ADD future TINYINT(1) NOT NULL DEFAULT 0;
UPDATE oldrecorded SET future=0;

David
-- 
David Engel
david at istwok.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: rsmissed1.patch
Type: text/x-diff
Size: 13357 bytes
Desc: not available
Url : http://www.mythtv.org/pipermail/mythtv-users/attachments/20110404/72ec8854/attachment.bin