This includes other fixes like using methodResponse instead of
other crap.. Thing brings it more inline w/ the UPnP spec...
We now include the namespace bit, which is apparently what the PS3
needs...
note that the PS3 works in the readme w/ restrictions...
This removes the dependance on SOAPpy, and alsp fpconst...
[git-p4: depot-paths = "//depot/": change = 1105]
through the programs till we find the one w/ the PMT we want.. This
means an incorrect PAT doesn't slip through.. This now works since
we don't overload pid and attempt to use it's original value later...
[git-p4: depot-paths = "//depot/": change = 1097]
magic to simply replace the pat w/ only the program we want,
but keeps all the data there... it wastes a bit of bandwidth,
but makes the file seekable, which is good for the media player..
We need to extend this so that we generate the PAT's first, and
then start reading/writing the file so we don't pass a PAT unchanged..
[git-p4: depot-paths = "//depot/": change = 1096]
sometime between .5 and 1 seconds, and the second is between 1 and
5 seconds later...
Also, schedule a resend of the notify for sometime between a third
and a half of the max-age of it all...
send out two byebye notifications just incase...
Yes, this increases traffic, but it's required per the spec.. and
w/ 100Mbit it's not too big of a deal..
[git-p4: depot-paths = "//depot/": change = 1088]
This should prevent the DSM-520 from saying we've disappeared...
7 minutes means 4 sends before the 30 minutes max-age hits and is
a pretty good compromise, though we do send out five notifies each
time...
[git-p4: depot-paths = "//depot/": change = 1087]
mpegts.py, and makes tssel.py work by properly inserting
program_number into the PAT that matches the one in the PMT...
[git-p4: depot-paths = "//depot/": change = 1068]
two empty containers would compare equal causing the removal of
the first one if a later one was really removed... The problems
of inheriting from list.. :(
It was previously masked because I would expand everything by
default and now that I don't, more items are empty exposing the
bug...
[git-p4: depot-paths = "//depot/": change = 884]
support doing the entire title as a resource on the container now
that didllite stuff supports the pass through... just chain the
chapter iters together for this..
[git-p4: depot-paths = "//depot/": change = 872]
the most work (or are most "destructive") are last, and the short
and simple ones first... maybe we should do a sort on an attribute
the implies how many bytes are read or requires http access or
something..
[git-p4: depot-paths = "//depot/": change = 868]
the first function called in the chain, this mean no more anoying
'no mime-type for: media/.DS_Store' messages... should we just
ignore dot files?
[git-p4: depot-paths = "//depot/": change = 865]
maps to object 0's id, but we need to fix that...
when an object doesn't have an updateID, use the system's ID... this
means music presets work w/ the DSM-520 now...
[git-p4: depot-paths = "//depot/": change = 863]
waiting clients when an error occures to be passed back up... we
also clear the fetchingurls flag so that if an error occures, then
next query will try again...
[git-p4: depot-paths = "//depot/": change = 861]
are available.. the PID demuxer currently isn't fast enough (it's
written in python), but once fixed should be interesting...
[git-p4: depot-paths = "//depot/": change = 858]
don't repr the return value, the dict containing it will do that for
me...
don't go into subchildren and update them to pull in there children
just to find out the childCount... DIDLLite now doesn't include it
if there aren't any... (This may need to be tested w/ the Intel
software, but works great w/ the D-Link DSM-520)... It makes
browsing shoutcast genres doable now..
[git-p4: depot-paths = "//depot/": change = 852]
