Hi,
a few month ago streamdev's CVS got lost when www.vdr-developer.org moved to a new server. In the meantime I moved the project to projects.vdr-developer.org. Unfortunately only the latest releases plus a few snapshots could be saved from the old CVS tree and imported into git.
Streamdev's old address http://streamdev.vdr-developer.org now forwards to the new home http://projects.vdr-developer.org/projects/plg-streamdev.
I just released the new version 0.5.1, mostly a bugfix release. Some of the fixes have also been applied to the 0.4 branch (for VDR 1.4.x). In case there's still interest in this old releases, please contact me and I'll publish a final 0.4.1.
Changelog of 0.5.1: - updated copy of GetClippedNumProvidedSystems to the version used since VDR 1.7.15 (reported by carel@vdrportal) - fixed the code deciding if a device is in use for live TV or not. It did not work as expected for FF cards (reported by wtor@vdrportal) - increased client side timeout for TUNE command - more dsyslog messages to help troubleshouting channel switch issues - improved the channel switch code trying to move live TV to different card - make sure that a client doesn't interrupt replaying on server's FF card (reported by wtor@vdrportal) - switching away live TV failed even when "always suspended" (reported by Michal Novotny) - fixed regression: no receiver created for ES/PS/PES (reported by Gavin Hamill) - VTP no longer uses a static priority value for its server-side receivers. The server stores channel and priority requested with the PROV command and re-uses these values in a subsequent TUNE for the same channel. The new PRIO command is used to update the receiver's priority if necessary. - added parameter HEIGHT to externremux.sh - fixed syslog messages reporting local instead of remote IP and port - fixed regression of the GetDevice(...) change. Filter streaming to clients with a recent VDR version no longer worked. - log an error if externremux.sh is missing or not executable - since VDR 1.5.0 cDevice::GetDevice(...) is no longer a query only method. It detaches all receivers of the device it returns. So it is no longer suitable for testing the availability of a device. Added a copy of VDR's cDevice::GetDevice(...) without the detach receivers part as a workaround until a better solution is available - added dsyslog messages to help troubleshouting channel switch issues - VTP command SUSP didn't attach the player to the primary device - fixed incompatibilities with older make versions - replacing a connections receiver is now an atomic operation. Solves stuttering audio/video due to lost TS packets when adding/removing PIDs - disabled attribute warn_unused_result in libdvbmpeg - slightly increased thread priorities of cStreamdevWriter/Streamer (suggested by Rolf Ahrenberg) - fixed missing support for invisible channel groups (groups without name) in HTTP menu (reported by Timothy D. Lenz) - don't quote actual program call in externremux.sh, so you can run the program through e.g. nice or taskset just by extending the variable which holds the program name - in externremux.sh each mencoder audio and video codec has a dedicated variable for a default option string now. Still you can override each default option with an URL parameter - externremux.sh mencoder now uses scale parameter with negative height instead of -xy for scaling (suggested by vel_tins@vdrportal) - added FPS (frames per second) parameter to externremux.sh (suggested by vel_tins@vdrportal) - don't use std::map.at(). It's not available in older libstdc++ version (reported by Matthias Prill) - fixed extremux x264 using value of ABR for VBR (thanks to vel_tins@vdrportal)
Have fun, Frank