Re: [Exist-open] problem masking in ReentrantReadWriteLock

[Dannes W. <di...@gm...>]

Hi,

On Jan 25, 2008 3:48 AM, Alexei Guevara <ale...@gm...> wrote:
> This code intrigued me because if the calling thread is not the one owning
> the luck that means that there is a bug somewhere else. I took a look at the
> same class in the 1.2 branch, thinking that maybe this was an oversight that
> could have been fixed in 1.2.  I found pretty much the same code there,
> except that now the whole stack trace is logged as well as the error
> message, something that will most certainly help diagnose the problem.

Last summer wolf has spend many many weeks is solving concurrency
issues in eXist. He actually redesigned significant parts of the code
and was very successful here.

I don't think that looking at just one class (ReentrantReadWriteLock)
will give a good picture.

We are absolutely interested in your experiences with eXist 1.2 under
a high load. The problems found in our install base (we have heavy
users) have disappeared since last summer....

regards

Dannes


-- 
# Dannes Wessels # The Netherlands #