| Summary: | ASTERISK-07696: [patch] Jabber timeout enhancements | ||
| Reporter: | jmls (jmls) | Labels: | |
| Date Opened: | 2006-09-08 07:02:39 | Date Closed: | 2006-10-04 10:38:53 |
| Priority: | Minor | Regression? | No |
| Status: | Closed/Complete | Components: | Channels/General |
| Versions: | Frequency of Occurrence | ||
| Related Issues: | |||
| Environment: | Attachments: | ( 0) jabber.patch2 | |
| Description: | This patch A) fixes a little buglet - Although client->timeout was set to 20s, the counter was decremented before the test, so it timedout after 19s B) client->timeout can now be set in jabber.conf ****** ADDITIONAL INFORMATION ****** However, I think that there still is a bug of sorts lurking, as the original logic (which I haven't changed) decrements the timeout regardless of what happens in the loop. This makes it possible for the timeout counter to decrement *past* zero, (ie become negative) in which case the test for "timeout == 0" would never be met. | ||
| Comments: | By: jmls (jmls) 2006-09-08 07:10:56 always decrementing the timeout counter would also have the effect of having a "timeout" even if events ocurred. Should not the timeout only be decremented if nothing has happened ? By: jmls (jmls) 2006-09-08 11:29:20 sorry, just noticed that vi mangled my original patch (_timeout instead of timeout). New patch uploaded. Please delete original patch. By: Anthony LaMantia (alamantia) 2006-09-08 17:30:02 Looks good, we are going to review this patch. By: Matt O'Gorman (mogorman) 2006-10-04 10:38:53 fixed in a different way, code is much better now, jmls wouldnt you agree? | ||