In particular, when a job was failing due to retries, it wasn't incrementing the
'failed' or 'failures' stats for the associated queue. This also led to a bug when
trying to cancel jobs due to failed retries.
This is also related to a bug where the 'failures' count could become negative --
probably the longest-standing and until now un-replicated bug.
@myronmarston This should fix the bug you mentioned about getting failures when accessing boolean. It also fixes a long-standing bug about keeping track of the number of failed jobs.
A fix for: https://github.com/seomoz/qless-core/commit/670263e5380da502985079fb80101dbf9cf7b097#commitcomment-3927002
In particular, when a job was failing due to retries, it wasn't incrementing the 'failed' or 'failures' stats for the associated queue. This also led to a bug when trying to cancel jobs due to failed retries.
This is also related to a bug where the 'failures' count could become negative -- probably the longest-standing and until now un-replicated bug.