Tvheadend

dear colleagues,

this commit with comment

"add streaming error for active recordings"

is effectively making impossible (if used together with XBMC) to browse recording,
or literally to perform any action on them (if recording is active).

in idle state (xbmc doing anything but no PVR related action) all is fine. it refresh
schedules, recording and EPG changes in the defined intervals.

problem starts, as soon as XBMC reads recordings list. normally XBMC would start
action (read directory), switch GUI into "please wait" mode and finish sooner or later.

unfortunately this commit makes XBMC think, that something went wrong and as soon
as it finishes, it starts again with reading the list and notifying "please wait" - and this
loops again and again.

even in case we do not trigger "list recordings" view, if TVHEADEND is still recording
and XBMC finds this out (through normally scheduled recording list re-read - each 5 minutes),
in that case it schedules new scan in 30s. so loops one circle again.

if TVHEADEND is hosted on remote server, having ~1000 recordings and few XBMC clients
connected, this is quite a BURN OUT for complete infrastructure.

is this something you can reconsider / reimplement, or should be something
communicated to XBMC / Lars and change in code there ?

thanks