[mythtv] Datadirect 3 day limitations -- one thought on reducing impact

[David Shay]

In looking at the long thread over at zap2it labs bulletin board on the 
3 day limitation in data and doing some more investigation, I have come 
to the realization that there is some degree of further optimization 
that could be done on the mythtv end [beyond the support of the 
suggested download time that aloready made it into .19].

When you have multiple data sources, for example I have 3 (one for 
digital cable, one for regular cable, and 1 for HDTV OTA), there are 
multiple pulls of data from datadirect, including one for each day.  A 
normal pull of today, tomorrow, and +13 will therefore do 9 pulls of 
data, if my math and understanding of the code is correct. 

The existing code for the "dd-grab-all" option actually reduces this 
down to a single pull (as long as the datadirect userid for each 
videosource is the same).  This only does 1 pull, but of more data.  In 
real-world analysis that I have done, before datadirect limited the data 
to 3 days, this one pull of data was actually LESS bandwidth than the 9 
other pulls in total.  Although this did hog a datadirect app server for 
a longer period of time on the first one, again in total it should have 
been less usage than all 9 together.

With some minor modifications, this re-use of the downloaded XML file 
could be extended to at least reduce the number of datadirect pulls to 3 
(one for each day, and not multiplied by the number of videosources).

With some more work, we could reduce the number of pulls to 2 (one pull 
to account for today and tomorrow, and a second for +13), which should 
also reduce the strain on the datadirect servers, even for those folks 
without multiple data sources.

Thoughts?  I am willing to code a patch up, but wanted some feedback 
first.  Not sure how many people in the total mythtv user base have 
multiple videosources, but the second change would help out even those 
with only one.