6
I Use This!
Inactive

Commits : Listings

Analyzed 1 day ago. based on code collected 2 days ago.
Feb 04, 2025 — Feb 04, 2026
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Also fuzz acct_csv and catch_dtmf modules. More... almost 3 years ago
Add "longest_ipi" (largest inter-packet-interval), as well as rtpa_jlast, rtpa_jmax and rtpa_javg (last, max and average value of jitter respectively) into supported Q-command counters. Update RTPP_QUERY_NSTATS to match reality. More... almost 3 years ago
Properly consider getaddrinfo() results. This functionality has been somewhat lost during refactoring some 10 years ago. More... almost 3 years ago
Travis no more. More... almost 3 years ago
Make parent and child use fixed 16-bytes buffer to exchange PID, right now the child can send variable length so parent have to guess and rely on certain ordering of things. Factor in child's pid size while reading it from the rope socket, so that we don't accidentally read out part of the "startup completed" cookie. More... almost 3 years ago
When building debug version add wrappers for the pthread_mutex_lock() and pthread_mutex_unlock(). That should catch any errors trying to unlock object that we have not locked or lock object that has not been initialized. More... almost 3 years ago
Normalize deallocation of the sigterm. More... almost 3 years ago
Remove rtpp_memdeb_setbaseln() call now that we have most of the stuff deallocated already, add few missing pieces. More... almost 3 years ago
Re-gen. More... almost 3 years ago
Make sure the time we provide is current. The issue was masked before by the fact that the thread was awaken every 100ms, thus giving only slight discrepancy. More... almost 3 years ago
Update counters. More... almost 3 years ago
If wakeup fails, use pthread_kill() to end the worker thread. This should prevent glitching from causing test to hang. More... almost 3 years ago
Don't unlock mutex prematurely. More... almost 3 years ago
No need to run wakeup here anymore. More... almost 3 years ago
If we cannot wakeup command processing thread, reject connection otherwise we might deadlock on pfds mutex. More... almost 3 years ago
Re-gen. More... almost 3 years ago
Update counters based on the latest code. More... almost 3 years ago
Move TSTATE_XXX into its own header file. More... almost 3 years ago
Set psp->pfds earlier, so in case rtpp_cmd_connection_ctor() fails we are not leaving it dangling. More... almost 3 years ago
Emit forward declaration of the public object struct before fin() method. More... almost 3 years ago
Add a shadow copy for the changes log to swap it with the current one in one shot, reset the counter and unlock history mutex. This is so that we are not blocking command processing thread. +10% to max command throughput. More... almost 3 years ago
Split single rtpp_sessinfo lock into RTP and RTCP parts. This reduces contention and allows both thread synchronize in parallel. We still have a problem with synchronization in the RT[C]P threads locking respective history for the duration of the procedure, preventing command thread from making progress. We might be able to solve this with some kind of "shadow" history. There are still 10% of command processing speed to be had. More... almost 3 years ago
Re-work waking up of the rtpp_command_queue: instead of doing periodic check-ins, create notification socketpair and pull it along with other sockets. More... almost 3 years ago
Replace haszero() with a macro that works for full range of input data. More... almost 3 years ago
Use rtpp_strsplit() in place of rtpp_strsep() to do initial tokenization of the command string. This is supposed to be much faster. More... almost 3 years ago
Provide more info when child fails to report the status. More... almost 3 years ago
Add rtpp_strsplit() API to match all possible separators, put \0 in place of them and output map with '\0' representing where separators were found and 0xff elsewhere. All in one pass and using 64-bit operations for efficiency. More... almost 3 years ago
Disallow inputs > MAX_RPKT_LEN. More... almost 3 years ago
Initialize malloc/free in the acct_rtcp_hep interface correctly. More... almost 3 years ago
Busy-wait to give streams some time to finish. More... almost 3 years ago