I think if you do comment and close, it doesn't trigger a close event the same was just clicking a close button does. We have a cron job running hourly for each project to sync up differences (one at :15 after and one at :45 after so as to not process everything all at the same time).

Those GitHub hooks are kinda strange beasts....
Actually the issue is displayed correctly (as closed by @eddieajau). So yes, there is a "lag".
If this becomes an issue we might increase the frequency the cron job is running.

Well, it was supposed to update but didn't because: [2014-04-12 17:38:28] JTracker.ERROR: Error updating GitHub issue joomla/joomla-cms #1569 in the tracker: A build is required.

I could have sworn we did something about that...

Umm.... in which of the log files did that message end up? (seems that they have been ""rotated"" since )

Looking at IssuesTable->check() there are several other restrictions that would not apply to GitHub issues. Time to re-think-the-thing™...

It was in the issue log.

In the hooks, we added code to set a build when inserting issues, but didn't handle an update condition. So, records which have been in the database for a while wouldn't pass current validation if we repeated those activities.

We might need to add a source to at least the IssuesTable object so that it can act accordingly if a record is being inserted/updated via hook, user input, or a CLI script.

We might need to add a source to at least the IssuesTable object so that it can act accordingly if a record is being inserted/updated via hook, user input, or a CLI script.

Sounds like a plan ;)

We might need to add a source to at least the IssuesTable object so that it can act accordingly if a record is being inserted/updated via hook, user input, or a CLI script.

Sounds like a plan ;)

A good one ;)

I believe this has been "fixed" since.