[mythtv] Scheduling and repeats

[J. Donavan Stanley]

Bruce Markey wrote:

> J. Donavan Stanley wrote:
> ...
>
>>> Would this omission explain why a non-repeat show would not be 
>>> recorded when checking dups in current and previous recordings but 
>>> WOULD be recorded when only checking dups in current recordings? The 
>>> shows I think should be recorded are now showing with a 'P' 
>>> (previously recorded), which is an improvement over showing an 'r'. 
>>> However, the shows in question were not recorded previously because 
>>> they are brand new.
>>
>>
>>
>> The section of SQL I had left out would have done the proper lookups 
>> on the titles and descriptions so probably.  When I saw it my first 
>> thought that it would allow dups through but I guess it was in 
>> inverse.  I need to do more testing of this, but I can't do so until 
>> I have a chance to either get a capture card in my dev box or time to 
>> test with my MBE.
>>
>>>>>
>>>> That is correct.  New recordings only is currently implemented as a 
>>>> "dup in" not a "dup loc".  When set something to be recorded as new 
>>>> episodes only it looks for duplicates using subtitle and description.
>>>
>>>
>>>
>>> I also notice there is not way to specify what to check for (i.e. 
>>> description, subtitle) when you are recording new episodes only.
>>>
>>> I'm not sure what you mean by "dup in" and "dup loc". If shows 
>>> typically have no subtitle, would it make sense to have an option to 
>>> allow the new recordings to look for duplicates using subtitle 
>>> and/or description?
>>
>>
>>
>> I was smoking crack when I wrote that.  The new episodes is a "dup 
>> method" as in "match using subtitles and descriptions" not a dup 
>> location as in "look in current and old recordings".  If the show is 
>> not a repeat (kDupCheckNewEpi) then the scheduler should treat it as  
>> "match using subtitles and descriptions" (kDupCheckSubDesc) and look 
>> in the area specified by the recording rule.
>
>
> This is what I had mentioned last week but may have stated it in
> a confusing way (since you had enabled the dupin selections which
> was not the solution I was suggesting ;-).
>
> Dupmethod is the choice of fields to compare when two items are
> compared.
>
> Dupin is the choice of which items are elegible to be compared.
>
> New episodes only seems to be more a matter of which items are
> eligble to be compared and not a choice of the feilds that are
> compared.
>
> Take for example a grabber that doesn't have programids but does
> have original air dates or at least decent repeat markings. A series
> has unique subtitles, no descriptions and each episode has three
> showing per week. If kDupCheckNewEpi is a dupmethod, there is no
> way to set kDupCheckSub and so it will record all three showings.
>
> If instead it was kDupsInNewEpi ( = 0x13 with the 0x10 flag to
> check previouslyshown) to limit it to check both recorded and
> oldrecorded and mark all other previouslyshown as rsRepeat, then
> any dupmethod could be applied.

I reworked the first run stuff on my machine this morning before work so 
that it's now kDupsNewEpi (0x04), and then falls back on kDupsInBoth 
after that check passes.  I need to schedule some of the new shows to 
verify things a bit more.  I'm seeing rejects I wouldn't expect to see, 
but it may be due to subtitle reuse by the series.