revert 44bfcf1d this breaks the build for everyone who does not have python-dev package

FS-3172 this also fixes the incorrect usage of L16 on payload 10 which may or may not break interop with other sip devices if we do it right.  also added rtp_disable_byteswap variable that can be set to false to disable byteswap when a device is encountered that is incompat (inluding all precious version of FS up till now)

freetdm: add support for setting the channel tx/rx queue size from config file (wanpipe.conf) and from CLI

FS-3170 this was specific to the user channel which is not a real channel in every sense of the word as it has no running thread or any usable state changes so this new line of code in 233d3164 to wait for the state machine to stabalize before returning from originate caused an issue with user/ channels

FS-3168 this will remove the reported symptom but does not change the fact that 1khz resolution is ideal for proper performance

OPENZAP-146 - this restores the correct behavior regression from Commit:03dc3b7b

add string params to ladspa so you can connect files to audio ports (string params don't count towards number params)

switch_xml: Remove commented out mmap.  With the changes in the past 2 year, mmap can't really be put back in it current state.

switch_xml: reloadxml will(should) never lock again.  It will load the XML structure into a new XML structure, and just replace the currently available ROOT XML.  It then the job of the last user of the switch_xml structure to free it.

Merge branch 'openmethods-merged-unimrcp-asr-dtmf-bugfix-0' of git://scm.dashjr.org/var/scmroot/git/freeswitch

mod_callcenter: FS-3158 Change the default and sequentially-by-agent-order strategy to include the longest-idle-agent.  This should offer a default consistant way to go through all the agent within the same tier/position. (Before, it was left to the DB to return the order of the result)

Add mod_ladspa (Audio plugin framework for linux) http://http://www.ladspa.org/ try it with autotalent http://web.mit.edu/tbaran/www/autotalent.html

http://bugzilla.maptools.org/show_bug.cgi?id=2297
in libtiff. A really screwed up 2D T.4 image, or a maliciously
constructed T.4 2D or T.6 image should potential run off the end
of an image decoder buffer.