Prior to this commit, if anything at all went wrong in
switch_ivr_phrase_macro_event() we would generate a warning like this:
[WARNING] switch_ivr_play_say.c:348 Macro [macro_name]: 'pattern_name' did not match any patterns
This is clearly misleading. The natural thing to do on seeing that
message is to verify that the language files are there, and that the
pattern really does exist in that macro. But none of that was usually
the problem. The message would be generated if the language wasn't
found, or if the channel had gone away, for example.
With this commit, we verify that we actually tried looking for the
pattern before displaying the warning about the pattern not matching.
For years we've been generating spurious messages like:
[WARNING] switch_ivr_play_say.c:348 Macro [voicemail_ack]: 'saved' did not match any patterns
This would happen when the caller hangs up during the playback of
certain prompts in the voicemail system where we weren't checking the
return value of vm_macro_get(). Looking closely at the log, it's
clear we were calling down into switch_ivr_phrase_macro() long after
the channel was gone.
The message above is also misleading -- switch_ivr_phrase_macro()
would have been able to find that pattern just fine, but it never
actually looked because the channel was gone. We'll clean up that
message in a follow on commit.
This commit also reverts 2 previous attempts to fix this very rare race issue spanning back to 2009
62ce8538974f727778f1024d0ef9549e438704fe Patch from MOC
3a85348cdfd0fd7df63a2a150790722c2d294b36 FS-2302 mutex added around switch_xml_toxml()
The real problem was switch_xml_toxml_buf() was actually temporarily modifying the xml structure being searialized to make it appaer to be a root structure then serializing it and restoring the pointers. This caused a non-threadsafe operation when some other thread was scanning the same xml structure.
This patch removes the modification and instead passes a new arg to switch_xml_toxml_r indicating to treat the structure as if it were a root structure.
This bug has been present since the induction of xml into FS.
Conflicts:
src/switch_xml.c
Cisco 7925G seem to work only with the correct conference_id2 and
rtptimeout set, so add protocol 11 definition fields and set
conference_id2 correctly.
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Nathan Neulinger <nneul@neulinger.org>
Cisco 7925g send access status message with just 8 byte of payload data.
Since we don't interpret the unknown 3rd field anyway, remove it. This
will prevent the first register to fail.
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Nathan Neulinger <nneul@neulinger.org>
WiFi phones like the 7925g may take longer than just one second to
acknowledge the open receive message. Increase the timeout to 5 seconds.
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Nathan Neulinger <nneul@neulinger.org>
This was momentarily called force_send_silence_when_idle, but that was
non-obvious as you had to set that value to true to be able to not
send silence when idle. This name describes the purpose much better.
We were handling the "send silence but not comfort noise" case in both
silence_stream_file_read and switch_generate_sln_silence. This
changes the former to rely on the latter.
If set to true, this prevents us from overriding the value of
send_silence_when_idle. When that is unset or set to zero and SRTP is
engaged, we typically override the value because many devices can't
handle gaps in the SRTP stream.
This variable is mostly for testing whether particular devices can
handle this behavior. Use at your own risk.
In commit 55d01d3defed4bfdc74704dbea0da9548a97a979 we set
send_silence_when_idle to -1 rather than 400 when SRTP is engaged.
But this left no way to enable white noise silence when desired.
When SRTP is engaged we can't simply not send RTP because it breaks
too many devices. So we need to prevent send_silence_when_idle from
being unset or being set to zero. This change allows it to be set to
other values so as to feed white noise rather than all zeros into the
codec.
When the channel variable send_silence_when_idle was set to zero,
switch_ivr_sleep was calling SWITCH_IVR_VERIFY_SILENCE_DIVISOR on it
anyway, causing it to be set to 400. The only way to get the behavior
of not sending silence when idle was to unset the variable completely.
This corrects the behavior such that setting the value to zero has the
same effect as leaving it unset.
write(3) can write fewer bytes than was requested for any number of
reasons. The correct behavior is to retry unless there is an error.
If there is an error, try to unlink the file; no sense in leaving
corrupted data laying around.
Unlike fread(3), read(3) will return -1 on error. We were assigning
the result of read to a potentially unsigned variable, and passing the
result down to switch_xml_parse_str() where it would end up
determining how many bytes to malloc(3).
