Merry Christmas to all VDR users!
It's been a very busy year for me, in which I was unable to
spend as much time on VDR as I would have liked to. But now things
are settled again and I managed to prepare a new developer version
with the most important fixes and improvements. Please feel free
to tell me if I missed something important - some things may well
have slipped under my radar ;-).
So here's my Christmas gift for you!
VDR developer version 2.3.2 is now available at
ftp://ftp.tvdr.de/vdr/Developer/vdr-2.3.2.tar.bz2
A 'diff' against the previous version is available at
ftp://ftp.tvdr.de/vdr/Developer/vdr-2.3.1-2.3.2.diff
MD5 checksums:
6dbb208ea3d59658a18912b49af175b3 vdr-2.3.2.tar.bz2
68a0ed9f01048026333939d30e0a6474 vdr-2.3.1-2.3.2.diff
WARNING:
========
This is a *developer* version. Even though *I* use it in my productive
environment, I strongly recommend that you only use it under controlled
conditions and for testing and debugging.
The changes since version 2.3.1:
- Fixed a crash when deleting a recording (reported by Oliver Endriss).
- Fixed an overflow of PIDs in a receiver (thanks to Robert Hannebauer).
- Updated the Italian OSD texts (thanks to Diego Pierotto).
- Fixed initializing device specific parameters in cDvbTransponderParameters.
- The function SetCurrentChannel(const cChannel *Channel) is now deprecated and
may be removed in a future version. Use SetCurrentChannel(int ChannelNumber)
instead.
- The SVDRP command DELC now refuses to delete the very last channel in the list,
to avoid ending up with an empty channel list.
- The cRwLock class now allows nested read locks within a write lock from the
same thread. This fixes possible crashes when moving or deleting channels in
the menu or through SVDRP (as well as other operations that try to acquire a
read lock within a write lock).
- Fixed a crash when trying to delete a channel that is being used by a timer.
- Fixed setting the current item and counter values in the Recordings menu after
deleting the last recording in a subfolder.
- Fixed a crash when deleting a recording that is currently being replayed.
- Fixed a crash when moving a recording to a folder on a different volume.
The cRecordingsHandler now performs its actual operations in a separate thread,
thus avoiding locking problems and reducing the time between subsequent
operations.
- Added a note to the description of cFont::Size(), regarding possible differences
between it and cFont::Height() (suggested to Thomas Reufer).
- Made the cPlayer member functions FramesPerSecond, GetIndex and GetReplayMode
'const' (thanks to Thomas Reufer).
- Fixed resuming replay at a given position, which was off by one frame (thanks
to Thomas Reufer).
- Improved handling frame numbers to have a smoother progress display during
replay of recordings with B-frames (thanks to Thomas Reufer).
- Fixed replaying recordings to their very end, if they don't end with an I-frame
(thanks to Thomas Reufer).
- Implemented a frame parser for H.265 (HEVC) recordings (thanks to Thomas Reufer).
- Added cFont::Width(void) to get the default character width and allow stretched
font drawing in high level OSDs (thanks to Thomas Reufer).
- Fixed regenerating the index of audio recordings (thanks to Thomas Reufer).
- Fixed building VDR with systemd >= 230 (thanks to Ville Skyttä).
- Sorted sources.conf by continous azimuth (thanks to Lucian Muresan).
- Added 'S58.5E Kazsat 3' to sources.conf (thanks to Aitugan Sarbassov).
- Fixed truncated date/time strings in the skins on multi-byte UTF-8 systems
(reported by Sergey Chernyavskiy).
- Updated the Estonian OSD texts (thanks to Arthur Konovalov).
- Added a 'const' version of cTimers::GetTimer() (thanks to Lars Hanisch).
- Fixed a typo in the description of cTimers::GetTimersRead() (thanks to Lars
Hanisch).
- Fixed a possible buffer overflow in handling CA descriptors (suggested by
Lars Hanisch).
- Avoiding some duplicate code and unnecessary work in nit.c (thanks to Ville
Skyttä).
- Added support for the systemd watchdog (thanks to Marc Perrudin),
- Added a short sleep to cTSBuffer::Action() to avoid high CPU usage (thanks to
Sergey Chernyavskiy).
Have fun!
Klaus