On Thu, 2004-09-23 at 01:59, Ceki GÃlcà wrote:

> Some would say heavy-handed, others would say simple and robust. Given
> the history of this discussion (see bug report 24159), I have a strong
> bias against modifying the existing synchronization code in
> AppenderSkeleton.  The problem in 24159 is just too contrived to be
> taken seriously.

I feel a bit frustrated because I've been bitten in the ass a few times
encountering this "contrived" problem and I haven't been taken very
seriously pointing this out.  As I have tried to explain, this is
something that has happened before on servers at this company, and it's
something we'd like to have fixed.

If you one day turned on DEBUG and soon watched your server lock up,
you'd be shaking your fist at whoever designed log4j.  It is not just
some annoying bug, it's also something that can result in your server
hanging.  And I'd rather have the server reboot than do that...  I admit
my test case was "contrived", but in the real world people employ
multiple levels of indirection and the actual triggering mechanism can
be quite convoluted.  And it can lurk until one unlucky day somebody
unknowingly turns on DEBUG.

I thought in the quest for helping fix this bug we could improve
concurrency and multi-threading.  Even if you think this bug is dumb,
wouldn't you like to improve in this area?

I sent out some changes that will maintain compatibility and fix this
issue, have you seen them?



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to