Hi Cyril, Thank you for the compliment. Logback code has still a lot of room for improvement. Your suggestions for improvement are welcome. I might be mistaken but given that the failures are due to an assertion error and not a timeout, a deadlock is probably not a root cause. I should also add that failures in lossyAppenderShouldOnlyLooseCertainEvents are more frequent than failures in workerShouldStopEvenIfInterruptExceptionConsumedWithinSubappender. Although both failures are rare and hard to reproduce, failures in the former might be a little easier to reproduce than failures in the latter. It would be quite wonderful if you could identify the reason for the failures and explain your reasoning on this list so that we can all learn something. -- Ceki On 4/22/2013 2:14 PM, cyril.lapinte@free.fr wrote:

Hello Ceki,

I have started to work on defect #298. I really appreciate it. Logback is a such a nice piece of code ;-)

As far as I know the defect looks like a deadlock. I will have to dive deeper to find where it is possible for the 3 (Main, Worker and Reader) threads to acquire more than one lock.

Cyril

------------------------------ Hi Cyril,

Thank you for your proposal. Are any topics you'd be particularly interested in? Would like to see a particular bug fixed or feature added? How familiar are you with the logback code base?

Looking forward for your response,

-- Ceki

On 4/20/2013 1:28 PM, cyril.lapinte@free.fr wrote:

...

Greeting,

I recently read a tweet from Ceki where he called for volunteers in the Qos Logback project. That's why I am reporting here.

In few words, I am a professional open source Java developer in Geneva. I already have the opportunity to use logback in some real production environments.

I will be to use my skills for whatever it is best for the project. From the volunteers page tasks list, I can work both on the 10min tutorial and/or fixing some bugs.

Best regards, Cyril Lapinte