VDR developer version 1.3.43 is now available at
ftp://ftp.cadsoft.de/vdr/Developer/vdr-1.3.43.tar.bz2
A 'diff' against the previous version is available at
ftp://ftp.cadsoft.de/vdr/Developer/vdr-1.3.42-43.diff
The changes since version 1.3.42:
- Removed an unnecessary toFile->SetReadAhead() from cutter.c (thanks to Artur Skawina). - The "Back" key now restores the original string when pressed while editing a string item (suggested by Markus Hahn). - Now stopping scanning the video directory if there are too many levels of symbolic links, which might indicate a recursive link loop (based on a patch from Helmut Auer). - Improved OSD area handling in cDvbSpuDecoder (thanks to Marco Schlüßler). - Now logging the description (if present) in case a thread is canceled (suggested by Marco Schlüßler). - cMenuText now uses the given font (thanks to Rolf Ahrenberg). - The ST:TNG skin now uses the fixed font if requested when displaying texts. - Fixed some typos in the CONTRIBUTORS file (thanks to Frank Krömmelbein). - Changed offset and size handling in 'libsi' from 'unsigned' to 'signed', so that overflows can be better detected (thanks to Marcel Wiesweg). - Checking data size in CaDescriptor::Parse() and LinkageDescriptor::Parse() of 'libsi' to avoid crashes with invalid data (thanks to Marcel Wiesweg). - Made CharArray::DataOwnData::assign() in 'libsi' more robust against invalid data (suggested by Oliver Endriss). Also changed CharArray::DataOwnData::Delete() so that it sets 'size' and 'data' to 0. - Now resetting the channel number if the number entered through the numeric keys exceeds the maximum channel number (thanks to Rolf Ahrenberg). - The language code in the 'X' component records of EPG data can now consist of two codes, separated by '+'. - If a recording starts and there is no EPG data available for the recorded channel, the 'X' audio component records for the 'info.vdr' file are now generated from the channel's PID data. - Externally provided EPG data (with table ID 0x00) now gets its component descriptors set from the broadcast data, so that language codes and descriptions are available (suggested by Andreas Brugger). - When setting the audio track descriptions, the language codes are now also set in case this is a replay session (based on a patch from Rolf Ahrenberg). - If a recording starts and the channel's audio PID data has more language code information than the EPG's component data, the code from the channel is taken. - Fixed handling DPID when deciding whether to switch to 'Transfer Mode' (thanks to Marco Schlüßler). - Fixed replaying recordings of radio channels with many audio tracks (thanks to Reinhard Nissl). - Added a comment to tChannelID::nid, explaining that is is actually the "original" network id.
Have fun!
Klaus
Klaus Schmidinger wrote:
VDR developer version 1.3.43 is now available at
ftp://ftp.cadsoft.de/vdr/Developer/vdr-1.3.43.tar.bz2
A 'diff' against the previous version is available at
ftp://ftp.cadsoft.de/vdr/Developer/vdr-1.3.42-43.diff
...
The original files contained 'overlay.tar.gz', which is a patch that was submitted to me, but I haven't decided yet whether to adopt it. I have now replaced the files with a version that doesn't contain this.
Klaus
Klaus Schmidinger wrote:
VDR developer version 1.3.43 is now available at ...
- Made CharArray::DataOwnData::assign() in 'libsi' more robust against invalid data (suggested by Oliver Endriss).
An error should be logged if assign() is called with invalid data. Otherwise you might never notice that a problem exists...
Oliver
Oliver Endriss wrote:
Klaus Schmidinger wrote:
VDR developer version 1.3.43 is now available at ...
- Made CharArray::DataOwnData::assign() in 'libsi' more robust against invalid data (suggested by Oliver Endriss).
An error should be logged if assign() is called with invalid data. Otherwise you might never notice that a problem exists...
Oliver
So far 'libsi' doesn't log anything, and I don't see how it could reasonably do this, because the incorrect data comes again every few seconds, so the log file would probably be flooded with such messages.
Klaus
Klaus Schmidinger wrote:
Oliver Endriss wrote:
Klaus Schmidinger wrote:
VDR developer version 1.3.43 is now available at ...
- Made CharArray::DataOwnData::assign() in 'libsi' more robust against invalid data (suggested by Oliver Endriss).
An error should be logged if assign() is called with invalid data. Otherwise you might never notice that a problem exists...
So far 'libsi' doesn't log anything, and I don't see how it could reasonably do this, because the incorrect data comes again every few seconds, so the log file would probably be flooded with such messages.
Well, if you don't like repeated messages you could increment an error counter which is displayed when vdr exits (like epg stats does).
Right now 'assign' is called with size==0 every few seconds. I don't know whether this points to a problem or not.
Oliver
Well, if you don't like repeated messages you could increment an error counter which is displayed when vdr exits (like epg stats does).
Right now 'assign' is called with size==0 every few seconds. I don't know whether this points to a problem or not.
That is all right. The private data in some descriptors has variable length, including 0.
Marcel
Oliver
Klaus Schmidinger wrote:
- cMenuText now uses the given font (thanks to Rolf Ahrenberg).
As a side effect, the commands menu and the recording commands now use fontFix instead of (in case of STTNG) the hard-coded fontSml before. Using a fixed font here is a good idea, however the default fixed font is imho awfully big. I'm now only getting 16 lines of 36 chars on osd, while fontSml got 19 lines of ~64 chars. Especially the 36 chars width is almost unusable for typical console command output.
I have no good idea how to avoid this. Maybe the fixed font size could be made a little bit smaller?
Cheers,
Udo