We group our errors together so that we can prioritise and fix the most
frequently occurring ones, and it allows us to do additional filtering (as
we can pick individual errors to exclude, rather than have to pick a type
of exception)
The hash is computed from the stacktrace of the exception, meaning that
even if a different message is generated, the error hash will be the same.
(Although this has a known issue at the moment that if, for example, you
catch a sql error, the stacktrace will be the same whether the error is
transient or permanent).
I've attached a patch to this error log. Hopefully it's nice and
straightforwards. It is only a small change.
Please note I have only made the changes (so far) to the SQL and XML stores
(and in-memory is done by default)
Original issue reported on code.google.com by mat.stee...@gmail.com on 28 May 2009 at 5:48
Original issue reported on code.google.com by
mat.stee...@gmail.com
on 28 May 2009 at 5:48Attachments: