Summary:ASTERISK-03715: [patch] More complete fix to address digit timeout on E&M T1 then emdigitwait
Reporter:wlloyd (wlloyd)Labels:
Date Opened:2005-03-19 22:04:39.000-0600Date Closed:2011-06-07 14:10:20
Versions:Frequency of
Environment:Attachments:( 0) chan_zap-emfix.txt
( 1) zappatch
( 2) zappatch2
Description:Interfacing a Norstar PBX with TE405 card.  Line is configured as T1 E&M wink start.  Norstar connects user set to E&M trunk immediately when user dials 9 for external line.  With user dialing manually digits are lost unless emdigitwait is increased to several seconds.

In this case dialing is slowed down considerably.  Function my_getsigstr is looking for a termination character of ' '.  Increased emdigitwait to several seconds means dialing is always slowed down considerably when number is finished dialing since it must wait for timeout as ' ' will never arrive.

Better to rewrite with standard FXO/S timing rather than have another parameter to control dialing timeouts.

Patch eliminates emdigitwait totally, and speeds up dialing.
Comments:By: wlloyd (wlloyd) 2005-03-21 09:13:26.000-0600

I've exchanged email with Aidan Van Dyk who originally submitted the emdigitwait patch.  Unfortunately he no longer has the equipment to test this setup anymore, but he took a quick look at the patch and agrees this is a better approach to fix his problem as well.

By: Michael Jerris (mikej) 2005-05-02 08:03:52

Is this up to date and working with current head?  Has anyone else tested this?  Do you have a disclaimer on file?

By: wlloyd (wlloyd) 2005-05-02 13:21:12

I'll check it against CVS head in the next day or two.  Without this patch dialing on the E&M in my setup is too problematic to be used no matter what value of emdigitwait I used.  

I've got 80 000 calls on this patch so far so at least in my setup it fixes all my problems.  Should help anyone using a T1 configured as E&M.

Nobody else has contacted me about using the patch.

I didn;t do a disclaimer.  I just copied the lines from another part of the same file.  It's so minor.  Is it really needed?  Please let me know.

The dual license is a hassle...


By: twisted (twisted) 2005-05-25 21:41:25

for a patch that modifies the code in more than just a very small way, a disclaimer for changes is needed.  Please file a disclaimer if you would like to see this go in.

By: Michael Jerris (mikej) 2005-06-19 12:29:41

Bug susspended due to no disclaimer.

By: Michael Jerris (mikej) 2005-06-19 12:30:44

Mark-  I closed this out due to no disclaimer, but wanted you to be aware so you could decide if you wanted to do it yourself?

By: wlloyd (wlloyd) 2005-07-14 10:08:31

I faxed in my disclaimer a few weeks ago now.  Should be ok to include this now..


By: Michael Jerris (mikej) 2005-07-14 10:18:07

Can you please update this patch to current cvs and get some testers to comment on it.

By: wlloyd (wlloyd) 2005-07-14 15:00:17

latest patches for HEAD.

Patch was 100% the same except line #'s changed.

By: Michael Jerris (mikej) 2005-07-20 23:05:47

ok, we need tester comments on this before it can be included.

By: wlloyd (wlloyd) 2005-07-26 10:20:20

Feedback from matt*@*bird*-*technologies*.*com

Yes... It does appear to solve any problem with dialing in to my
Asterisk system when manually entered digits are dialed... I tried
dialing very slowly and it worked which is pretty cool :)

and then today


    On further investigation the latest CVS head appears to be the
cause of my polycom ringing on forever and ever.  The patch worked as
advertised and I am going to recompile it back in and work on the
ringing issues.


By: Mark Spencer (markster) 2005-08-03 00:52:29

This patch looks good pending code review and the formatting fixes.

By: wlloyd (wlloyd) 2005-08-05 12:11:32

Do I need to change formatting and resubmit the patch?


By: Michael Jerris (mikej) 2005-08-07 14:10:29

Yes please if you can.  There is good documentation in the coding guidelines in cvs head that should guide your way.

By: wlloyd (wlloyd) 2005-08-21 13:08:34

Ok, patched to today CVS.

I've formatted from the guidelines.  If it's not right I'm not sure what elese to change...


By: Matthew Fredrickson (mattf) 2005-09-01 14:56:08

This is pretty cool, I thought about doing it this way a couple of weeks ago but couldn't really put the time into it.  I looked at the code and it looks good.  I'll commit it.

By: Matthew Fredrickson (mattf) 2005-09-01 15:02:08

Committed, thanks ;-)

By: Digium Subversion (svnbot) 2008-01-15 15:46:38.000-0600

Repository: asterisk
Revision: 6483

U   trunk/channels/chan_zap.c
U   trunk/configs/zapata.conf.sample

r6483 | mattf | 2008-01-15 15:46:38 -0600 (Tue, 15 Jan 2008) | 2 lines

New, more robust way of doing e&m digit collection.  Bug ASTERISK-3715