This patch does the following:
* only starts MOH if no other file (sync or async) is currently playing
* adds a variable "conference_permanent_wait_mod_moh" that controls the
behavior of how the enter and exit sounds interact with the MOH when
wait_mod is set. When the variable is set, the MOH keeps playing and
the enter and exit sounds are mixed with the MOH. When the variable
is unset, then any playing MOH is first stopped, then the enter or
exit sound is played and the MOH is started again.
This functionality is useful in case the enter and exit sounds are
used to announce the name of the caller, who is joining or leaving a
conference.
FS-5159 #resolve
If the member do an attended transfer this loop stops and
the agent is set to Available state again, when in fact he still
busy with other channel.
This was happening, for example, when a member calls support queue, then
the support operator do an attended transfer to sales queue, while the sales
operator is talking the member that was transferred, the sales queue will
send calls to the operator, which is not expected.
This patch allows conference flags to be set dynamically from the
dial plan by either passing them to the conference application in
the +flags{ } string or by setting the "conference_flags" dial plan
variable.
The +flags{ } string is currently used to set *user* flags only.
This patch changes this by allowing the +flags{ } string to contain
conference related flags as well (for example wait_mod). It shouldn't
be a problem to pass both types of flags via +flags{ } as long as
the user and conference flag names are kept unique.
FS-5099 #resolve
when video floor is locked by a member, changing audio floor on del_member
will cause the video floor lock cleared unexpectedly, this commit fixes that.
When we specifically release all limits on a channel we destroy the
hash table stored in the "limit_hash" private channel data but we
don't destroy the private data as it will be reclaimed as part of the
session. If limit increment is called after the limit release we can
reuse that channel private, but we need to check whether the hash
table is null first. Fortunately this makes the code look better
anyway.
FS-6775 #resolve
FS-6783 #resolve
* Added BERT stats channel variables
* Check if the channel is going down when out of sync to avoid flagging it
as out of sync, if the channel is going down it is expected to have some errors
* Calculate sampling error rate over a time window
* Create the following variables to control bert testing behavior:
bert_timeout_ms - Milliseconds to wait for bert sync before timing out
bert_window_ms - Window of time to calculate sampling error rate percentage
bert_max_err - Max percentage of error tolerated before going out of sync
bert_hangup_on_error - Hangup when error occurs (Timeout or going out of sync)
What if the hash were destroyed by one thread holding the lock while
another thread was waiting for the lock? The waiting thread would
have already checked that the hash was non-null and would proceed on
that assumption after acquiring the lock.
With this commit we check only after acquiring the lock.
ref: FS-6783
ref: FS-6775
