Fatal errors not respected (tasks still re-queued)

[nrusch]

I’ve got a callback method bound to the CheckExitCodeCallback of my plugin, and in it, I’m remapping process exit codes to Deadline abort levels. I then call self.AbortRender('Some message', abortLevel), but even when abortLevel is Fatal, tasks are still getting re-queued until they hit the repository’s threshold for failing tasks.

This is in 6.2 RC1 on Linux (Fedora 19).

[rrussell]

Thanks for reporting this! We’ve fixed the issue, so this will be fixed in RC3.

Cheers,
Ryan