VDR developer version 1.7.33 is now available at
ftp://ftp.tvdr.de/vdr/Developer/vdr-1.7.33.tar.bz2
A 'diff' against the previous version is available at
ftp://ftp.tvdr.de/vdr/Developer/vdr-1.7.32-1.7.33.diff
MD5 checksums:
7c21451360ac7959d0d95e533d34451c vdr-1.7.33.tar.bz2 c79257198f8569bc02f43dc470ee3076 vdr-1.7.32-1.7.33.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.
IMPORTANT: ==========
This version of VDR no longer sets LC_NUMERIC to "C" in order to make sure any floating point numbers written to configuration files use a proper decimal point. It rather explicitly converts such numbers using the new functions atod() and dtoa(). IF YOU USE PLUGINS THAT STORE FLOATING POINT NUMBERS IN THEIR OWN CONFIGURATION FILES, YOU SHOULD SET export LC_NUMERIC=C BEFORE RUNNING VDR, UNTIL THESE PLUGINS HAVE BEEN PROPERLY UPDATED.
The changes since version 1.7.32:
- In order to be able to play TS recordings from other sources, in which there is more than one PMT PID in the PAT, 'int cPatPmtParser::PatPmt(void)' has been changed to 'bool cPatPmtParser::IsPatPmt(int Pid)'. - Fixed learning remote control keys with the LCARS skin. - Updated the Macedonian OSD texts (thanks to Dimitar Petrovski). - Fixed getting only non-video packets in cCuttingThread::GetPendingPackets() (reported by Sören Moch). - Changed all occurrences of MPEG4 to H264 (pointed out by Sören Moch). - Fixed getting the number of editing sequences in case the last sequence has no actual end mark. - The cutter now only increments the TS continuity counter for packets that have a payload (pointed out by Sören Moch). - Fixed adjusting the DTS values in the cutter, to compensate for dropped B-frames (pointed out by Sören Moch). - Fixed a typo in skins.h (thanks to Lars Hanisch). - Simplified calculating the PTS offset in cPtsFixer::Fix() and fixed the overflow handling of PCR values (thanks to Sören Moch). - Fixed calling iconv_close() only with a valid iconv_t value (thanks to Juergen Lock). - Fixed faulty opening of the Recordings menu when pressing the Play key during normal live viewing mode in case there is a "last viewed" recording. - Fixed some #include statements in plugins to use <vdr/...> instead of "vdr/..." (thanks to Lars Hanisch). - Fixed some spellings in osd.h and svdrp.c (thanks to Ville Skyttä). - Fixed handling lowercase polarization characters in channel definitions if no DiSEqC is used (reported by Mike Hay, actual bug pointed out by Stefan Huelswitt). - Synchronizing system time to the transponder time is now done using adjtime() in order to avoid discontinuities (suggested by Manuel Reimer). If the time difference is more than 10 seconds, stime() is still used to do the initial sync. - The '7' and '9' keys now jump to the very beginning or end, respectively, of the recording, even if there is no mark set at that point (following a request from Andre Weidemann). - Now always setting the TDT EIT filter, because otherwise when turning on using the transponder time in the Setup menu, it would only be used after the next restart of VDR (thanks to Sundararaj Reel). - The new functions cDevice::CanScaleVideo() and cDevice::ScaleVideo() can be used by derived output devices to implement scaling the video to a given size and location (based on a suggestion by Lucian Muresan). - The SVDRP command HITK now discards any keys if the remote control is currently turned off (thanks to Alexander Hans). - The new remote control key "Play/Pause" can be used with remote controls that don't have separate keys for "Play" and "Pause", but rather have a single key for both functions (thanks to Stefan Hofmann for suggesting to implement support for such remote controls). - The new option "Setup/Replay/Pause on mark set" can be used to activate automatically going into Pause mode if an editing mark is set during replay (suggested by Andre Weidemann). - When regenerating the index of a recording, the frame rate stored in the info file is now automatically fixed if it differs from the value detected by the frame detector. - Fixed creating the edited version directory if a relative file name is given in the call to 'vdr --edit' (the '/video' part was stripped from the given file name even if it wasn't there). - The new option "Setup/Replay/Progress display time" can be used to activate automatically displaying the progress display whenever replay of a recording is started (suggested by Stefan Blochberger). - Changed reading and writing of floating point numbers into configuration files to make it independent of the decimal point used in the current locale. All calls to atof() have been replaced with the new function atod(), which makes sure the string representation of a floating point number using a '.' as decimal point will be handled correctly, even if the locale in use expects a ',' as the decimal point. Plugins that read floating point numbers from their own configuration files will also need to use atod() for this, or use a method of their own (this is not necessary if values are stored in VDR's setup.conf file, because VDR takes care of this). The reason for these changes is that floating point numbers presented to the user shall be displayed in the way defined by the current locale (suggested by Stefan Blochberger). If you use plugins that store floating point values in configuration files of their own and have not yet been adapted to this change, you should set export LC_NUMERIC=C before running VDR. Otherwise your plugin's configuration data may not be read or written correctly. - The new functions SetItemEvent(), SetItemTimer(), SetItemChannel() and SetItemRecording() of the cSkinDisplayMenu class can be reimplemented by skin plugins to display these items in a more elaborate way than just a simple line of text.
Have fun!
Klaus
On Saturday 08 December 2012 12:38:30 Klaus Schmidinger wrote:
- In order to be able to play TS recordings from other sources, in which there is more than one PMT PID in the PAT, 'int cPatPmtParser::PatPmt(void)' has been changed to 'bool cPatPmtParser::IsPatPmt(int Pid)'.
there is one more change you did not mention:
int PmtPid(void) const { return pmtPid; }
has been removed.
this breaks xineliboutput.
On 08.12.2012 13:18, Wolfgang Rohdewald wrote:
On Saturday 08 December 2012 12:38:30 Klaus Schmidinger wrote:
- In order to be able to play TS recordings from other sources, in which there is more than one PMT PID in the PAT, 'int cPatPmtParser::PatPmt(void)' has been changed to 'bool cPatPmtParser::IsPatPmt(int Pid)'.
there is one more change you did not mention:
int PmtPid(void) const { return pmtPid; }
has been removed.
this breaks xineliboutput.
Sorry, that was a typing mistake. It should have been
- In order to be able to play TS recordings from other sources, in which there is more than one PMT PID in the PAT, 'int cPatPmtParser::PmtPid(void)' has been changed to 'bool cPatPmtParser::IsPmtPid(int Pid)'.
Klaus
8.12.2012 17:28, Klaus Schmidinger kirjutas:
On 08.12.2012 13:18, Wolfgang Rohdewald wrote:
On Saturday 08 December 2012 12:38:30 Klaus Schmidinger wrote:
- In order to be able to play TS recordings from other sources, in
which there is more than one PMT PID in the PAT, 'int cPatPmtParser::PatPmt(void)' has been changed to 'bool cPatPmtParser::IsPatPmt(int Pid)'.
there is one more change you did not mention:
int PmtPid(void) const { return pmtPid; }
has been removed.
this breaks xineliboutput.
Sorry, that was a typing mistake. It should have been
- In order to be able to play TS recordings from other sources, in which
there is more than one PMT PID in the PAT, 'int cPatPmtParser::PmtPid(void)' has been changed to 'bool cPatPmtParser::IsPmtPid(int Pid)'.
Klaus
Xine-plugin is broken too. Is it possible to provide patch for fixing plugin compiling?
br, A
On Sat, Dec 8, 2012 at 9:03 AM, Arthur Konovalov artlov@gmail.com wrote:
- In order to be able to play TS recordings from other sources, in
which there is more than one PMT PID in the PAT, 'int cPatPmtParser::PatPmt(void)' has been changed to 'bool cPatPmtParser::IsPatPmt(int Pid)'.
there is one more change you did not mention:
int PmtPid(void) const { return pmtPid; }
has been removed.
this breaks xineliboutput.
Sorry, that was a typing mistake. It should have been
- In order to be able to play TS recordings from other sources, in which
there is more than one PMT PID in the PAT, 'int cPatPmtParser::PmtPid(void)' has been changed to 'bool cPatPmtParser::IsPmtPid(int Pid)'.
Klaus
Xine-plugin is broken too. Is it possible to provide patch for fixing plugin compiling?
That's really a job for the plugin maintainers. Have you tried contacting them to report the plugins need to be updated?
On 08.12.2012 18:03, Arthur Konovalov wrote:
8.12.2012 17:28, Klaus Schmidinger kirjutas:
On 08.12.2012 13:18, Wolfgang Rohdewald wrote:
On Saturday 08 December 2012 12:38:30 Klaus Schmidinger wrote:
- In order to be able to play TS recordings from other sources, in
which there is more than one PMT PID in the PAT, 'int cPatPmtParser::PatPmt(void)' has been changed to 'bool cPatPmtParser::IsPatPmt(int Pid)'.
there is one more change you did not mention:
int PmtPid(void) const { return pmtPid; }
has been removed.
this breaks xineliboutput.
Sorry, that was a typing mistake. It should have been
- In order to be able to play TS recordings from other sources, in which
there is more than one PMT PID in the PAT, 'int cPatPmtParser::PmtPid(void)' has been changed to 'bool cPatPmtParser::IsPmtPid(int Pid)'.
Klaus
Xine-plugin is broken too. Is it possible to provide patch for fixing plugin compiling?
The plugins I encountered (burn is affected too) had something like if (Pid == PatPmtParser->PmtPid() ) in them. I changed those into something like if (PatPmtParser->IsPmtPid(Pid) )
Hope that helps for fixing it yourself.
/Malte
br, A