Im GIT (git.tvdr.de) gibt es jetzt die VDR Developer Version 2.5.2.
ACHTUNG: Verwendung dieser Version für den Alltagsbetrieb auf eigene Gefahr!
Hier die Änderungen seit Version 2.5.1:
Code
- Fixed strreplace() to handle NULL strings (reported by Jürgen Schneider).
- Somewhere down the road the 'x' bit of Doxyfile.filter got lost, so the
Makefile now makes sure it is set before calling doxygen.
- Using strgetlast() in more places.
- Fixed a crash in the SVDRP command CLRE in case a non-existing channel number is
given (reported by Manuel Reimer).
- Fixed handling $(PKG_CONFIG) in newplugin (thanks to Winfried Köhler).
- Fixed initializing the cChannel::schedule pointer after reading EPG data with PUTE
(reported by Manuel Reimer).
- Fixed VPS handling when spawning a pattern timer, in case Setup.UseVps is false
(reported by Jürgen Schneider).
- Increased the number of possible modulation systems in cDevice::GetDevice()
(reported by Ulf Grüne).
- Now explicitly triggering respawning of pattern timers.
- If an event in the Schedules menu is marked with a 'T' or 'I' and the user presses the
Red button to edit the timer, spawned timers are now preferred over pattern timers
in case there is more than one timer that will match that event.
- Made the functions cRecordingInfo::SetData() and cRecordingInfo::SetAux() public
(thanks to Peter Bieringer).
- Fixed setting the 'title' of a recording's info to the recording's name if there
is no info file (the change in version 1.7.28 broke the fallback to the old 'summary.vdr').
- Now making sure a spawned timer only fully overlaps the given event (reported by
Jürgen Schneider).
- Added some missing user command calls for copying, renaming and moving recordings
(thanks to Peter Bieringer).
- Improved cSectionSyncer to make sure that no sections are missed, and to allow
handling partially used segments (as in the EIT) and processing sections in random
order. Segment syncing is now done with the two member functions Check() and
Processed(). The old functions Sync() and Repeat() are deprecated and may be
removed in a future version. See the comments in filter.h for a description on
how to use these new function.
- When checking whether a particular recording has already been made by a pattern
timer, the characters ' ' (blank), ':' and '-' are now ignored, making
"TITLE - EPISODE" and "TITLE: EPISODE" the same.
- Decreased the scrambling timeout for CAMs known to decrypt a certain channel, so
that it won't collide with MAXBROKENTIMEOUT in recorder.c.
- Fixed scaling subtitles with anti-aliasing (thanks to Peter Bieringer).
- Improved handling EPG data from the EIT tables:
+ Table 0x4F is now completely ignored.
+ Once a schedule has seen events from 0x5X, tables 0x6X are ignored for that
schedule.
+ When looking up an event in its schedule, the start time is used for tables 0x6X, and the
event id for tables 0x4E and 0x5X.
+ When hashing events by event id or start time, existing older entries in the hash
tables are now deleted before entering the new ones.
+ The function cSchedule::GetEvent() is now deprecated and may be removed in a future
version. Use GetEventById() and GetEventByTime() instead.
+ On channels that use proper event ids a change of the start time no longer
causes a new event to be created, but rather modifies the existing one. This
avoids possible interruptions in VPS recordings in case the event's start time
is changed while the recording is already going on.
- The margins for timer recordings are now always limited to the duration of the
previous and next event.
- Spawned timers that don't use VPS now automatically adjust their start/stop times
to changes in the respective event's times.
- Fixed handling the timer counter in cSchedule::DelEvent().
- Fixed MakePatternFileName() in case the event doesn't yet have a short text (reported
by Jürgen Schneider).
- No longer switching devices for pattern timers (thanks to Helmut Binder).
- cTimer::TriggerRespawn() now only acts on local timers.
Alles anzeigen
Das mit der automatischen Anpassung der gespawnten Timer an Verschiebungen der Event-Zeiten konnte ich nur theoretisch testen, da solche Fälle ja nicht so oft vorkommen.
Bitte ausgiebig testen!
Bug-Reports etc. bitte hier in diesem Thread posten.