[
https://issues.apache.org/jira/browse/TS-3429?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15246147#comment-15246147
]
ASF GitHub Bot commented on TS-3429:
------------------------------------
Github user jpeach commented on a diff in the pull request:
https://github.com/apache/trafficserver/pull/576#discussion_r60103328
--- Diff: proxy/InkAPI.cc ---
@@ -991,7 +991,14 @@ INKContInternal::handle_event(int event, void *edata)
Warning("INKCont Deletable but not deleted %d", m_event_count);
}
} else {
- return m_event_func((TSCont)this, (TSEvent)event, edata);
+ int retval = m_event_func((TSCont)this, (TSEvent)event, edata);
+ if (event == EVENT_INTERVAL) {
+ // In the interval case, we must re-increment the m_event_count for
+ // the next go around. Otherwise, our event count will go negative.
+ if (ink_atomic_increment((int *)&this->m_event_count, 1) < 0)
--- End diff --
I think this would be better as
```C
ink_release_assert(ink_atomic_increment((int *)&this->m_event_count, 1) <
0);
```
I agree this is a legitimate assertion; it is more helpful to have the
actual assertion expression in the log.
> TSContScheduleEvery does not increment event count correctly
> ------------------------------------------------------------
>
> Key: TS-3429
> URL: https://issues.apache.org/jira/browse/TS-3429
> Project: Traffic Server
> Issue Type: Bug
> Components: Core
> Reporter: Bin
> Assignee: Susan Hinrichs
> Fix For: 6.2.0
>
>
> TSContScheduleEvery only increments the event count the first time it is
> scheduled. When the event handler gets invoked, it decrements the event
> count. So it triggers the assertion at InkAPI.cc:987.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
