| Age | Commit message (Collapse) | Author |
|---|
|
- Fixed detecting transponder lock in cDvbTuner (based on a patch from Stefan
Meyknecht).
- What was previously marked with WAIT_FOR_LOCK_AFTER_TUNING is now permanently
active and uses a cCondVar to signal when a transponder is locked.
- Added some missing 'const' to cChannel.
- Added a sample setup for 'DisiCon-4 Single Cable Network' to 'diseqc.conf'
(thanks to Oliver Endriss).
- Fixed attaching a cPlayer to a cDevice, so that 'Operation not permitted'
errors don't occur any more (thanks to Marco Schlüßler).
- Fixed a case where the resultBuffer in cRemux ran full before getting a sync.
- Removed the usleep() call from cDvbPlayer::Action() to make VDR run on NPTL
systems (thanks to Alfred Zastrow). The NPTL check at startup has also been
removed.
- Taking the complete size of available data into account when deciding whether
to clear the transfer buffer to avoid overflows (thanks to Reinhard Nissl).
- Updated Romanian language texts and the iso8859-2 fonts (thanks to Lucian Muresan).
- Now actually using the iso8859-15 fonts (thanks to Lucian Muresan).
- Some minor code cleanups (thanks to Prakash K. Cheemplavam).
- Fixed missing cleanup at program exit in case there is a problem with a plugin
(thanks to Mattias Grönlund for pointing this out).
- Increased the required free buffer space in the resultBuffer of cRemux to
2 * IPACKS to avoid a buffer overflow in case a cTS2PES writes one complete
packet and then (within processing the same TS packet) wants to write another
small packet.
- Removed the signal handler and WakeUp() call from cThread (it is no longer
needed).
- Added some checks when canceling a thread and removed the usleep() in
cThread::Start() (suggested by Ludwig Nussel). Also removed 'running' from
cThread and using only childTid to indicate whether a thread is actually
running.
- Added cCondWait::Sleep() and using it to replace all usleep() calls (based
on a suggestion by Werner Fink).
- Only assigning events to timers if the related schedule has actually been
modified.
- When searching for the present event, the running status is now only taken
into account if the event has been "seen" within the past 30 seconds.
This avoids shortly seeing the wrong events in the channel display when
switching to a channel that hasn't been tuned to in a while.
|
|
- Added new entries to 'ca.conf'.
- Fixed closing unused PID handles (thanks to Stefan Schluenss for reporting this
one).
- Added more examples to 'diseqc.conf' (thanks to Oliver Endriss).
- Fixed disabling multiple recordings on a single DVB card (comment out the definition
of the macros DO_REC_AND_PLAY_ON_PRIMARY_DEVICE and DO_MULTIPLE_RECORDINGS in
dvbdevice.c).
- Plugins can now have their very own OSD setup in the object they return from
a call to cPlugin::MainMenuAction(). In order to implement this, the return type
of cPlugin::MainMenuAction() had to be changed from (cOsdMenu *) to (cOsdObject *).
So in case you are compiling an existing plugin with this version of VDR and you
get an error message, simply change cOsdMenu to cOsdObject in the plugin's source
for the MainMenuAction() function.
Plugin authors who have so far (ab)used the cControl mechanism to implement their
own raw OSD should take a look at the new demo plugin 'osddemo'. It implements
a very primitive game that shows how a plugin can have its own raw OSD. Especially
look into cLineGame and see how it implements the Show() function. See also
the chapter on "User interaction" in PLUGINS.html.
- Added three new fields to the lines in 'channels.conf': NID, TID and RID. NID and
TID are the Network and Transport Stream IDs, respectively. RID is an additional ID
that can be used to tell apart channels that would otherwise be indistinguishable.
This is typically the case with radio channels, which may have the same NID, TID
and SID, but different "radio IDs". This new field is therefore called RID ("radio
ID"). Currently NID and TID are not yet used by VDR and should always be 0. The
RID is actually used when building the "unique channel ID", so if you have channels
in your 'channels.conf' file that cause error messages when loading, you can set
the RIDs of these channels to different values.
When reading an old 'channels.conf' these new fields will be automatically
initialized to 0 and once the file is written back to disk they will be appended
to the channel definitions.
Thanks to Régis Bossut for pointing out that with some providers the channels can
only be distinguished through the RID.
- The "unique channel ID" now contains an optional 5th part (the RID). See man vdr(5).
- Updated 'channels.conf.cable' and made some channels unique using the new RID
(thanks to Andreas Kool for pointing out the problems).
- Made some channels unique in 'channels.conf.terr' using the new RID.
- Extended the '-l' option to allow logging to LOG_LOCALn (n=0..7) by writing, for
instance, '-l 3.7' (suggested by Jürgen Schmidt).
- Now deleting stale lock files if they have a time stamp that is outside the window
'now +/- LOCKFILESTALETIME'. This improves things in cases where the system time
makes far jumps, so that a lock file might end up with a time stamp that lies
in the distant future (thanks to Oliver Endriss).
|
|
- Fixed a missing Flush() call in the remote control learning procedure (thanks
to Oliver Endriss).
- Modified channel handling to cover all parameters necessary for DVB-C and DVB-T
(see man vdr(5) for the meaning of the additional parameters stored in the field
previously named 'polarisation'). Thanks to Uwe Scheffler and Andy Carter for testing.
If you have a system with different kinds of DVB cards, like DVB-T and DVB-C,
for instance, there is no more need to distinguish the channels through the
'Ca' parameter in order to assign them to the various DVB cards. This is now
taken care of by the "source" parameter. So a channel marked as "terrestrial",
for example, will only be received on DVB-T cards.
Note that the cChannel class has been moved into a separate file (channels.[ch]),
and that all data members have been made private and are now only accessible
through member functions. You may have to change any plugin code that accesses
cChannel data accordingly.
- The new configuration file 'sources.conf' contains the various signal sources
(satellites, cable and terrestrial) which are used in 'channels.conf' and
'diseqc.conf' (thanks to Reinhard Walter Buchner for adding some satellites to
'sources.conf' and Oliver Endriss and Lauri Tischler for testing and debugging).
- The 'diseqc' parameter in the channel definitions has been redefined to hold the
"source" of the given channel (which can be either a satellite, cable or terrestrial).
For compatibility with channels.conf files from older versions, numeric values in
this parameter will be tolerated, but they have no meaning. If you want to use
DiSEqC you will need to replace these old values with the proper source identifiers
defined in the new configuration file 'sources.conf'. See how this is done in the
'channels.conf' file that comes with the VDR package.
- The new configuration file 'diseqc.conf' can be used to set up the individual
diseqc configuration (see man vdr(5) for a description of the file format).
- The "Edit channel" menu has a new entry "Source:" in which the source of this
channel can be selected (either a satellite, cable or terrestrial). The set of
parameters at the end of this menu will change according to the type of source.
- The "Use DiSEqC" parameter in the "Setup/LNB" menu has been moved to the beginning
of the list and disables the rest of the parameters when set to "yes", since these
are now only meaningful if DiSEqC is _not_ used.
- Removed some unnecessary #includes from eit.c and changed cMenuRecordings::Del()
to cMenuRecordings::Delete() to avoid warnings in gcc-3.2 (thanks to Andreas
Schultz for pointing this out).
- Improved skipping channels that are (currently) not available (thanks to Stefan
Huelswitt).
- Updated channels.conf.terr and channels.conf.cable (thanks to Uwe Scheffler).
- Fixed a bug when pressing the "Blue" button in the main menu without having
displayed it (thanks to Oliver Endriss for reporting this one).
|
