]> git.ipfire.org Git - thirdparty/tvheadend.git/log
thirdparty/tvheadend.git
11 years agomkbundle: follow symlinks 295/head
Stefan Saraev [Tue, 24 Sep 2013 16:28:34 +0000 (19:28 +0300)] 
mkbundle: follow symlinks

11 years agoepggrab: remove out of date otamux config files
Adam Sutton [Tue, 24 Sep 2013 14:33:30 +0000 (15:33 +0100)] 
epggrab: remove out of date otamux config files

11 years agolinuxdvb: fix adapter number passing to capmt
Adam Sutton [Tue, 24 Sep 2013 14:28:45 +0000 (15:28 +0100)] 
linuxdvb: fix adapter number passing to capmt

11 years agowebui: fix stupid javascript typo.
Adam Sutton [Tue, 24 Sep 2013 09:30:37 +0000 (10:30 +0100)] 
webui: fix stupid javascript typo.

11 years agotsfile: fix double insert of mmi into mi_mux_active. Created infinite loop.
Adam Sutton [Sun, 22 Sep 2013 21:09:01 +0000 (22:09 +0100)] 
tsfile: fix double insert of mmi into mi_mux_active. Created infinite loop.

11 years agowebui: add support for hidemode in grids. Also tidied a few things up.
Adam Sutton [Sun, 22 Sep 2013 13:31:16 +0000 (14:31 +0100)] 
webui: add support for hidemode in grids. Also tidied a few things up.

11 years agowebui: minor clean up of toolbar (in services panel)
Adam Sutton [Sun, 22 Sep 2013 12:30:36 +0000 (13:30 +0100)] 
webui: minor clean up of toolbar (in services panel)

11 years agowebui: minor improvement to service mapping button
Adam Sutton [Sun, 22 Sep 2013 12:28:49 +0000 (13:28 +0100)] 
webui: minor improvement to service mapping button

11 years agoapi: make it possible to hide disabled services in the mux/service grids
Adam Sutton [Sun, 22 Sep 2013 12:15:16 +0000 (13:15 +0100)] 
api: make it possible to hide disabled services in the mux/service grids

This can be done in 3 ways:

none: don't hide anything
all: hide all disabled elements (you can't re-enable in this state)
default: hide those elements whose parents are disabled

11 years agoepggrab: ensure opentv gets registered during init scan.
Adam Sutton [Sat, 21 Sep 2013 21:43:49 +0000 (22:43 +0100)] 
epggrab: ensure opentv gets registered during init scan.

11 years agowebui: ensure tabs are in the right place!
Adam Sutton [Sat, 21 Sep 2013 21:07:24 +0000 (22:07 +0100)] 
webui: ensure tabs are in the right place!

11 years agowebui: add default options to various structures for better UI
Adam Sutton [Sat, 21 Sep 2013 21:00:07 +0000 (22:00 +0100)] 
webui: add default options to various structures for better UI

11 years agowebui: add some support for the new property fields
Adam Sutton [Sat, 21 Sep 2013 20:43:03 +0000 (21:43 +0100)] 
webui: add some support for the new property fields

11 years agoprop: added some extra options including defaults and advanced.
Adam Sutton [Sat, 21 Sep 2013 19:00:19 +0000 (20:00 +0100)] 
prop: added some extra options including defaults and advanced.

11 years agoepggrab: ensure that grabs that timeout will unsubscribe
Adam Sutton [Sat, 21 Sep 2013 18:45:55 +0000 (19:45 +0100)] 
epggrab: ensure that grabs that timeout will unsubscribe

11 years agolinuxdvb: add satconf delete handler
Adam Sutton [Sat, 21 Sep 2013 10:55:34 +0000 (11:55 +0100)] 
linuxdvb: add satconf delete handler

11 years agocwc: fix logic error in multi card procesing at service startup
amet [Fri, 20 Sep 2013 21:54:30 +0000 (01:54 +0400)] 
cwc: fix logic error in multi card procesing at service startup

11 years agowebui: make it possible to map selected services
Adam Sutton [Fri, 20 Sep 2013 21:39:27 +0000 (22:39 +0100)] 
webui: make it possible to map selected services

I probably need to improve the text on the button.

11 years agolinuxdvb: safety check for possible bad config
Adam Sutton [Fri, 20 Sep 2013 21:33:03 +0000 (22:33 +0100)] 
linuxdvb: safety check for possible bad config

11 years agoservice mapper: make it possible to select services to map
Adam Sutton [Fri, 20 Sep 2013 21:32:17 +0000 (22:32 +0100)] 
service mapper: make it possible to select services to map

Also fixed some bugs that caused the mapping to fail and/or do the wrong
thing.

11 years agowebui: add variable to control number of items per page in idnode grids
Adam Sutton [Fri, 20 Sep 2013 20:14:42 +0000 (21:14 +0100)] 
webui: add variable to control number of items per page in idnode grids

11 years agolinuxdvb: added pre-defined mux configuration file support
Adam Sutton [Thu, 19 Sep 2013 21:17:02 +0000 (22:17 +0100)] 
linuxdvb: added pre-defined mux configuration file support

This had been left out for too long and makes it awkward for novices
to setup the networks. The config is possibly not as nice as before where
it was presented as a tree. But this can be improved later.

11 years agoidnode: allow NOSAVE options to be "loaded".
Adam Sutton [Thu, 19 Sep 2013 21:17:59 +0000 (22:17 +0100)] 
idnode: allow NOSAVE options to be "loaded".

This is needed since this routine is used to also load config on creation.
So if a NOSAVE option is included there it will not load it. If it really
is NOSAVE it shouldn't be in the config files anyway (due to NOSAVE check
on actual save).

11 years agowebui: UI crashed for non store based enumerations
Adam Sutton [Thu, 19 Sep 2013 19:45:25 +0000 (20:45 +0100)] 
webui: UI crashed for non store based enumerations

11 years agolinuxdvb: fix dvb-s config params modulation and rolloff
Adam Sutton [Thu, 19 Sep 2013 10:10:25 +0000 (11:10 +0100)] 
linuxdvb: fix dvb-s config params modulation and rolloff

11 years agompegts: change some config strings to be more user friendly
Adam Sutton [Thu, 19 Sep 2013 10:07:31 +0000 (11:07 +0100)] 
mpegts: change some config strings to be more user friendly

11 years agompegts: remove the wait for NIT in SDT processing
Adam Sutton [Thu, 19 Sep 2013 10:02:52 +0000 (11:02 +0100)] 
mpegts: remove the wait for NIT in SDT processing

This causes initial scans to time out where the SDT is waiting on
mux information from. Given that we'll rescan these muxes anyway
there is little point.

The only downside is the single point of info muxes (like 11778V
in the UK) may not get the opportunity to provide all the info they
could, since we might miss service info until later.

11 years agotvhlog: correct subsys processing, stuff was being incorrectly logged
Adam Sutton [Thu, 19 Sep 2013 10:01:58 +0000 (11:01 +0100)] 
tvhlog: correct subsys processing, stuff was being incorrectly logged

11 years agolinuxdvb: honour the autodiscovery option
Adam Sutton [Wed, 18 Sep 2013 13:09:32 +0000 (14:09 +0100)] 
linuxdvb: honour the autodiscovery option

11 years agocwc: added missing free() to stop memory leak
Adam Sutton [Wed, 18 Sep 2013 12:58:45 +0000 (13:58 +0100)] 
cwc: added missing free() to stop memory leak

11 years agoMerge remote-tracking branch 'amet/feature/dvb-rewrite_mgclient1' into feature/dvb...
Adam Sutton [Wed, 18 Sep 2013 12:58:09 +0000 (13:58 +0100)] 
Merge remote-tracking branch 'amet/feature/dvb-rewrite_mgclient1' into feature/dvb-rewrite

11 years agodescrambler: configure with --enable-dvbcsa flag
amet [Tue, 17 Sep 2013 09:04:46 +0000 (13:04 +0400)] 
descrambler: configure with --enable-dvbcsa flag

11 years agomove detect_card_type to commonly accessible place
amet [Tue, 17 Sep 2013 17:06:31 +0000 (21:06 +0400)] 
move detect_card_type to commonly accessible place

11 years agoextend cwc to act as Mgclient, one port many caids
amet [Sun, 15 Sep 2013 18:49:13 +0000 (22:49 +0400)] 
extend cwc to act as Mgclient, one port many caids

11 years agowebui: started updating the idnode functionality to be cleaner
Adam Sutton [Tue, 17 Sep 2013 08:23:35 +0000 (09:23 +0100)] 
webui: started updating the idnode functionality to be cleaner

It's still a mess, but it mostly does the job

11 years agoui: I have slightly hacked the multi combo to return a list for the form value
Adam Sutton [Tue, 17 Sep 2013 08:23:01 +0000 (09:23 +0100)] 
ui: I have slightly hacked the multi combo to return a list for the form value

11 years agoapi: updated idnode objects to support new rend() method in prop_t
Adam Sutton [Tue, 17 Sep 2013 08:21:50 +0000 (09:21 +0100)] 
api: updated idnode objects to support new rend() method in prop_t

This allows enumerated (ish) types to now be fully sorted and filtered in
the way you'd expect from the UI.

11 years agoidnode: just a few bits of tidying up
Adam Sutton [Tue, 17 Sep 2013 08:21:02 +0000 (09:21 +0100)] 
idnode: just a few bits of tidying up

11 years agoservice_mapper: added extra (probably redundant) list check
Adam Sutton [Mon, 16 Sep 2013 20:46:51 +0000 (21:46 +0100)] 
service_mapper: added extra (probably redundant) list check

11 years agoidnode prop: make it possible to properly sort/filter enum/list types
Adam Sutton [Mon, 16 Sep 2013 20:41:21 +0000 (21:41 +0100)] 
idnode prop: make it possible to properly sort/filter enum/list types

This does require a certain level of extra input from prop_t users where
the type is enumerated. A new routine rend() must be implemented to provide
the rendered value (for lists that are not enum, this is done transparently).

11 years agohtsmsg: add routine to convert a list to a CSV string
Adam Sutton [Mon, 16 Sep 2013 20:39:12 +0000 (21:39 +0100)] 
htsmsg: add routine to convert a list to a CSV string

Note: this will not work with all lists, only lists of basic types.

11 years agosupport: update config migration script for new idnode lists
Adam Sutton [Mon, 16 Sep 2013 14:04:44 +0000 (15:04 +0100)] 
support: update config migration script for new idnode lists

11 years agoservice: update to new idnode API for lists
Adam Sutton [Mon, 16 Sep 2013 14:03:40 +0000 (15:03 +0100)] 
service: update to new idnode API for lists

11 years agochannel: updated to the new idnode/prop API for fields that represent lists
Adam Sutton [Sat, 14 Sep 2013 22:26:36 +0000 (23:26 +0100)] 
channel: updated to the new idnode/prop API for fields that represent lists

11 years agoapi: added channeltag list API call
Adam Sutton [Mon, 16 Sep 2013 13:17:11 +0000 (14:17 +0100)] 
api: added channeltag list API call

11 years agoidnode prop: new API for specifying lists
Adam Sutton [Sat, 14 Sep 2013 22:24:51 +0000 (23:24 +0100)] 
idnode prop: new API for specifying lists

This is to avoid the need for string processing within the C code to combine
and split lists from CSV strings.

11 years agoapi: started to migrate EPG functions to the common API
Adam Sutton [Fri, 13 Sep 2013 21:25:58 +0000 (22:25 +0100)] 
api: started to migrate EPG functions to the common API

11 years agowebui: fix issue that caused UUIDs (not strings) to be displayed on update
Adam Sutton [Fri, 13 Sep 2013 15:52:22 +0000 (16:52 +0100)] 
webui: fix issue that caused UUIDs (not strings) to be displayed on update

11 years agoinput: include missing information about stream info
Adam Sutton [Fri, 13 Sep 2013 15:30:25 +0000 (16:30 +0100)] 
input: include missing information about stream info

11 years agoinput: fix error in BER reporting
Adam Sutton [Fri, 13 Sep 2013 15:19:29 +0000 (16:19 +0100)] 
input: fix error in BER reporting

11 years agolinuxdvb: fix mistake in diseqc setup, missed LNB config entirely!
Adam Sutton [Fri, 13 Sep 2013 14:50:31 +0000 (15:50 +0100)] 
linuxdvb: fix mistake in diseqc setup, missed LNB config entirely!

11 years agompegts: ensure that mux descriptor callback is only called with valid mux
Adam Sutton [Fri, 13 Sep 2013 11:56:02 +0000 (12:56 +0100)] 
mpegts: ensure that mux descriptor callback is only called with valid mux

11 years agolinuxdvb: fix diseqc switch control (thanks to amet for fixing/testing).
Adam Sutton [Fri, 13 Sep 2013 10:02:58 +0000 (11:02 +0100)] 
linuxdvb: fix diseqc switch control (thanks to amet for fixing/testing).

The problem was that I played fast and loose with the diseqc spec. I set the
line volts much later than you're supposed to, for various reasons.

I've now sorted that, though because of the complexity if a rotor exists I
run a full re-program of the switch (only if a rotor is present). And I double
set the line volts, both in the switch (to ensure it happens early) and then
in the LNB tuning (in case no switch exists).

11 years agosubscriptions: reporting of subs output caused passthru streams to crash
Adam Sutton [Thu, 12 Sep 2013 21:48:22 +0000 (22:48 +0100)] 
subscriptions: reporting of subs output caused passthru streams to crash

I had forgotten that MPEGTS passes raw pktbuf not pkt and so was attempting
to incorrectly extract the length and causing it to crash.

11 years agosubscription: fix double list insertion on detection of bad service
Adam Sutton [Wed, 11 Sep 2013 21:12:40 +0000 (22:12 +0100)] 
subscription: fix double list insertion on detection of bad service

If a service was marked as bad (no input) it was not unlinking the
subscription from the service before re-applying, this resulted in
a double entry into the service subs list and ultimatey a livelock
or possibly worse.

11 years agotvhlog: added build define to remove logging thread.
Adam Sutton [Wed, 11 Sep 2013 21:12:05 +0000 (22:12 +0100)] 
tvhlog: added build define to remove logging thread.

I think eventually I'll make this is cmdline switch so I can turn it on if
I need it.

11 years agocapmt: fix problems with capmt not being properly enabled
Adam Sutton [Wed, 11 Sep 2013 20:04:30 +0000 (21:04 +0100)] 
capmt: fix problems with capmt not being properly enabled

This was due to failure to properly reported which adapters have
been enabled. There are other problems with the capmt setup code
that really need addressing, but I don't have the time right now!

11 years agoidnode: make tree's default sorted by title
Adam Sutton [Wed, 11 Sep 2013 09:00:22 +0000 (10:00 +0100)] 
idnode: make tree's default sorted by title

11 years agodebug: just some tidying up of debug output
Adam Sutton [Wed, 11 Sep 2013 08:07:21 +0000 (09:07 +0100)] 
debug: just some tidying up of debug output

11 years agompegts: ensure that potentially valid muxes are not ignored
Adam Sutton [Wed, 11 Sep 2013 08:03:31 +0000 (09:03 +0100)] 
mpegts: ensure that potentially valid muxes are not ignored

It's possible a mux is not tunable, yet that doesn't imply the presence
of a free tuner means its invalid. It could just be that its an S2 mux
with only an S tuner available (just one example).

11 years agoClose a few file handles that have been left open unecessarily.
Adam Sutton [Wed, 11 Sep 2013 07:47:31 +0000 (08:47 +0100)] 
Close a few file handles that have been left open unecessarily.

11 years agolinuxdvb: add option to stop closing FEs and attempt to stop on retune
Adam Sutton [Tue, 10 Sep 2013 23:18:44 +0000 (00:18 +0100)] 
linuxdvb: add option to stop closing FEs and attempt to stop on retune

Rather than adding an explicit flag (to stop mux calls) on retune I
have moved the frontend close operation to be done in the background
monitor timer. This isn't foolproof but will cover most things
and I can probably add a delay to cause that to not execute immediately
should the timer be about to expire.

11 years agosubscriptions: add extra debug
Adam Sutton [Tue, 10 Sep 2013 22:56:38 +0000 (23:56 +0100)] 
subscriptions: add extra debug

11 years agompegts: stop network scanning from failing due to bad mux
Adam Sutton [Tue, 10 Sep 2013 22:27:12 +0000 (23:27 +0100)] 
mpegts: stop network scanning from failing due to bad mux

11 years agolinuxdvb: add NULL ptr check for frontend on satconf
Adam Sutton [Tue, 10 Sep 2013 22:26:36 +0000 (23:26 +0100)] 
linuxdvb: add NULL ptr check for frontend on satconf

11 years agosubscription: for stupid mistake causing double free()'s
Adam Sutton [Tue, 10 Sep 2013 20:15:02 +0000 (21:15 +0100)] 
subscription: for stupid mistake causing double free()'s

11 years agotsfile: fix debug output format error
Adam Sutton [Tue, 10 Sep 2013 16:25:34 +0000 (17:25 +0100)] 
tsfile: fix debug output format error

11 years agotranscoder: check for language config is NULL
Adam Sutton [Tue, 10 Sep 2013 16:23:08 +0000 (17:23 +0100)] 
transcoder: check for language config is NULL

11 years agodebug: output used non word sized format strings
Adam Sutton [Tue, 10 Sep 2013 16:07:15 +0000 (17:07 +0100)] 
debug: output used non word sized format strings

11 years agowebui: some attempt to make enum lists more usable.
Adam Sutton [Tue, 10 Sep 2013 15:27:22 +0000 (16:27 +0100)] 
webui: some attempt to make enum lists more usable.

11 years agosubscription: add in/out bandwidth indicator
Adam Sutton [Tue, 10 Sep 2013 15:18:30 +0000 (16:18 +0100)] 
subscription: add in/out bandwidth indicator

Now have the concept of both input and output bandwidth usage of a sub.

The reason for this is that while data may be flowing into the subscription,
i.e. a signal is being received. Things like timeshift, or even potentially
flow control, may mean the actual outgoing rate is quite different.

Input rate is added by the subscription code as data enters the sub, however
the output rate must be added by the subscription handler as close to the
client as possible.

11 years agosubscription: reworked the subscription handling to avoid re-entrancy
Adam Sutton [Tue, 10 Sep 2013 12:50:04 +0000 (13:50 +0100)] 
subscription: reworked the subscription handling to avoid re-entrancy

This is particularly important for null susbcriptions that receive no
stream data and therefore handle the SMT_STOP inline (actually was
direct before). This causes list corruption, therefore these subs
are now stored to a secondary list that are cleaned up within the
standard rescheduler.

11 years agohtsp: add NULL ptr check on channel name
Adam Sutton [Sat, 7 Sep 2013 19:36:22 +0000 (20:36 +0100)] 
htsp: add NULL ptr check on channel name

11 years agoinput: forgot to commit generic input processing.
Adam Sutton [Sat, 7 Sep 2013 10:36:49 +0000 (11:36 +0100)] 
input: forgot to commit generic input processing.

11 years agowebui: some cosmetics
Adam Sutton [Sat, 7 Sep 2013 00:17:54 +0000 (01:17 +0100)] 
webui: some cosmetics

11 years agochannel: add ability to delete channels
Adam Sutton [Sat, 7 Sep 2013 00:12:25 +0000 (01:12 +0100)] 
channel: add ability to delete channels

11 years agowebui: add channel create button and service mapping
Adam Sutton [Sat, 7 Sep 2013 00:07:15 +0000 (01:07 +0100)] 
webui: add channel create button and service mapping

The service mapping button appears to have gone missing during the
integration with master.

11 years agoapi: add channel create method
Adam Sutton [Sat, 7 Sep 2013 00:06:55 +0000 (01:06 +0100)] 
api: add channel create method

11 years agostatus: added back in basic status info
Adam Sutton [Fri, 6 Sep 2013 21:49:22 +0000 (22:49 +0100)] 
status: added back in basic status info

It will now show subscriptions and streams (tuners/muxes combined).

11 years agompegts: ensure that services on already tuned muxes will start
Adam Sutton [Fri, 6 Sep 2013 21:54:08 +0000 (22:54 +0100)] 
mpegts: ensure that services on already tuned muxes will start

Previously it was setting the weight for already tuned muxes which
made it appear as if the service couldn't be started!

11 years agosubscription: ensure that direct service subs work properly
Adam Sutton [Thu, 5 Sep 2013 23:15:52 +0000 (00:15 +0100)] 
subscription: ensure that direct service subs work properly

11 years agodebug: hide thread start messages
Adam Sutton [Thu, 5 Sep 2013 23:15:02 +0000 (00:15 +0100)] 
debug: hide thread start messages

we get a lot of spam due to HTTP server setup so this needs to be debug

11 years agompegts: ensure muxes are properly stopped when unsubscribing svcs
Adam Sutton [Thu, 5 Sep 2013 22:48:02 +0000 (23:48 +0100)] 
mpegts: ensure muxes are properly stopped when unsubscribing svcs

11 years agosubscriptions: stop re-entrant call to subs_reschedule()
Adam Sutton [Thu, 5 Sep 2013 22:46:43 +0000 (23:46 +0100)] 
subscriptions: stop re-entrant call to subs_reschedule()

THe problem is that mux subs have been done such that unsubscribe
(which in turn calls subs_reschedule()) can itelf be called in line
from subs_reschedule().

Really this is a bit of a mess but this simple hack fixes things.

11 years agodebug: add wrapper for pthread_create to add a meaningful name
Adam Sutton [Thu, 5 Sep 2013 21:19:25 +0000 (22:19 +0100)] 
debug: add wrapper for pthread_create to add a meaningful name

This is very useful for debugging. Though tbh I need to do this code
properly.

11 years agoepggrab: further fixes to stop chicken/egg scenarios
Adam Sutton [Thu, 5 Sep 2013 14:21:49 +0000 (15:21 +0100)] 
epggrab: further fixes to stop chicken/egg scenarios

too many A calls B calls A depending on state etc... but can't be
bothered at this stage to re-engineer the lot!

11 years agoopentv: ensure that we can't double delete table
Adam Sutton [Thu, 5 Sep 2013 14:18:07 +0000 (15:18 +0100)] 
opentv: ensure that we can't double delete table

THis might happen if registering that the scan is complete removes
the table.

11 years agoepggrab: change order of init to ensure things work properly
Adam Sutton [Thu, 5 Sep 2013 14:14:41 +0000 (15:14 +0100)] 
epggrab: change order of init to ensure things work properly

11 years agoepggrab: sort out configuration for epggrab ota muxes
Adam Sutton [Wed, 4 Sep 2013 21:15:10 +0000 (22:15 +0100)] 
epggrab: sort out configuration for epggrab ota muxes

11 years agoepggrab: corrections to the way the whole ota mux system works
Adam Sutton [Wed, 4 Sep 2013 00:07:27 +0000 (01:07 +0100)] 
epggrab: corrections to the way the whole ota mux system works

11 years agompegts: improvements to DVB SI processing
Adam Sutton [Wed, 4 Sep 2013 00:04:54 +0000 (01:04 +0100)] 
mpegts: improvements to DVB SI processing

I've made the table extraid 64 bits, as some entries require 32+X bits
so a 32-bit field wasn't enough.

Added callback for registering destruction (for dealloc'ing memory
alloc'd for opaque).

Added method of registering desciptor tag handlers, saves duplicating
BAT processing used for OpenTV, which uses 1 custom descriptor.

Improved the table completion detection and reduced level of debug
that could swamp tvhlog queue.

11 years agoopentv: updated to new APIs
Adam Sutton [Tue, 3 Sep 2013 12:14:10 +0000 (13:14 +0100)] 
opentv: updated to new APIs

This uses a significantly simpler setup, table is processed in a
linear fashion.

11 years agosubscriptions: corrections to the way ota muxes are unlinked
Adam Sutton [Wed, 4 Sep 2013 19:44:20 +0000 (20:44 +0100)] 
subscriptions: corrections to the way ota muxes are unlinked

11 years agotvhlog: ensure the logging Q is bounded in size, else it can become a leak
Adam Sutton [Wed, 4 Sep 2013 00:03:11 +0000 (01:03 +0100)] 
tvhlog: ensure the logging Q is bounded in size, else it can become a leak

If processing because so high that the log queue becomes swamped we can
end up with a sort of avalanche scenario where everything gradually slows
down and the size of the buffer continues to grow. The main culprits of
this have been eliminated at source, however this is still a useful
safe guard.

11 years agotvhlog: performance improvement when trace/debug is disabled.
Adam Sutton [Tue, 3 Sep 2013 22:14:31 +0000 (23:14 +0100)] 
tvhlog: performance improvement when trace/debug is disabled.

The processing performed in the hexdump routine, commonly used in
SI table trace output. Was having a significant impact on performance
as it was not pre-processing the config to see if trace was even
enabled before continuing to build up the (large and frequent) hex
strings.

11 years agolinuxdvb: fixed leaving dangling file handles and stop FE on stop
Adam Sutton [Mon, 2 Sep 2013 10:21:58 +0000 (11:21 +0100)] 
linuxdvb: fixed leaving dangling file handles and stop FE on stop

Note: really we don't want to close the FE if we're planning to
retune, but I need to change the API to support that operation.

11 years agosubscription: some minor corrections for mux sub
Adam Sutton [Mon, 2 Sep 2013 10:12:26 +0000 (11:12 +0100)] 
subscription: some minor corrections for mux sub

11 years agowebui: make it possible to subscribe to a mux
Adam Sutton [Mon, 2 Sep 2013 10:02:03 +0000 (11:02 +0100)] 
webui: make it possible to subscribe to a mux

I'm not entirely convinced this is fully working. I'm sure the data
rate was too low.

11 years agoepggrab: updated to the new mux sub API
Adam Sutton [Mon, 2 Sep 2013 10:01:18 +0000 (11:01 +0100)] 
epggrab: updated to the new mux sub API

11 years agosubscription: Created a proper mux subscription mechanism
Adam Sutton [Mon, 2 Sep 2013 09:58:22 +0000 (10:58 +0100)] 
subscription: Created a proper mux subscription mechanism

It's a bit ugly to link subs and mpegts directly, but its not the end
of the world (and probably unavoidable without lots of duplication).

I'm still not convinced its robust to mux deletions if subs exist on
that mux. Probably needs mmi to be ref counted.

There is also a special kind of sub that expects to receive no stream
data, i.e. all data will come from the SI tables only. This special
sub is automatically unsubscribed (but that might need changing).