Summary:ASTERISK-01909: [patch] Possible fix for empty option
Reporter:Tilghman Lesher (tilghman)Labels:
Date Opened:2004-06-27 23:19:14Date Closed:2008-01-15 15:01:03.000-0600
Versions:Frequency of
Environment:Attachments:( 0) 20040628__meetme.diff.txt
Description:Some time back, somebody reported on the list that empty wasn't working for them.  I'm posting this patch here to see if anybody who had that problem would care to test.
Comments:By: Mark Spencer (markster) 2004-06-28 10:05:56

Can you clarify what was supposedly broken about meetme?

By: Mark Spencer (markster) 2004-06-28 10:07:13

I'm also thinking this idea of posting bugs as a proxy for someone on the mailing list is getting to be an irritating one because the bug contact is not the person who has experienced the bug.

I think it should be recommended on the list that the people who have bugs place them in the bug tracker, not simply having an arbitrary third party place the bugs on their behalf.

By: Tilghman Lesher (tilghman) 2004-06-28 15:59:18

I have reproduced the problem and am now working on it.

By: Tilghman Lesher (tilghman) 2004-06-28 16:18:22

There's a lot of indentation changes in this patch, but the problem is fixed with the current patch.

The exact problem was there was a busy loop that wasn't getting incremented correctly (infinite loop).  There are also a number of other problems that I found that hadn't bothered anybody yet that are also fixed.

By: Rob Gagnon (rgagnon) 2004-06-28 16:43:05

I'm with Mark.

If the Description, summary, and Additional information fields are not verbose enough, how are people to know what was wrong, and how it was fixed?

By: Tilghman Lesher (tilghman) 2004-06-28 16:52:26

rgagnon:  that's a great theoretical question that might be answered elsewhere, as I've already noted in a previous bugnote that I'd duplicated the problem.

By: Ryan Courtnage (rcourtna) 2004-06-28 16:57:37

The patch fixes the problem:

exten => 80,1,MeetMe(|E)

Works great when a single extension dials into 80.
However, if a second extension dials 80 (while the 1st extension is
still in a conference), * will pin the CPU and require an * restart.

The fix appears to be very stable.

edited on: 06-28-04 17:41

By: Mark Spencer (markster) 2004-06-28 18:04:19

Just for reference:
# cat apps/app_meetme.c | perl -ane 's/^\s+//; print;' > m1
# patch -p0 < /tmp/20040628__meetme.diff.txt
# cat apps/app_meetme.c | perl -ane 's/^\s+//; print;' > m2
# diff -uNr m1 m2

That gives you the real diff without respect to whitespace changes.

Looking at the changes now, but nothing here suggests it would cause the symptoms you're suggesting, rcoutna, are you sure it wasn't some sort of fluke?

By: Ryan Courtnage (rcourtna) 2004-06-28 18:09:57

markster - absolutely positive.  There was a problem, and this patch fixes it.

If you'd like more info - I can easily revert back to my non-patched app_meetme.c and duplicate the problem again.

By: Tilghman Lesher (tilghman) 2004-06-28 18:24:34

rcourtna:  I think Markster is referring to the 1 second delay, not the CPU being pinned.

By: Mark Spencer (markster) 2004-06-28 18:24:49

I'm confused I thought there was some explanation that this patch caused an echo, that's what i meant didn't seem to be affected.

That comment seems to be gone now though, so I'm putting the patch in as-is.

By: Mark Spencer (markster) 2004-06-28 18:25:02

Fixed in CVS

By: Digium Subversion (svnbot) 2008-01-15 15:01:03.000-0600

Repository: asterisk
Revision: 3334

U   trunk/apps/app_meetme.c

r3334 | markster | 2008-01-15 15:01:02 -0600 (Tue, 15 Jan 2008) | 2 lines

Merge Tilghman's meetme updates (bug ASTERISK-1909)