Martin Desruisseaux wrote:
> No problem, I will not touch :). I will try to log to IRC tomorrow if it can
> be
> useful and I'm still at the laboratory at that time.
Hi Martin,
I believe i've fixed the failing test case.
We have been moving things around in referencing land, and also did some
Jody Garnett a écrit :
> get uses the read lock
> test uses the write lock
This is the little details that I missed :)
> Please wait until tomorrow - I am hooking the beast up right now (being
> very careful to leave existing classes alone). Indeed we could chat on
> IRC tomorrow and just do a
Yeah feedback! What you are looking at is the initial implementation
(so no weak references yet). I will try and answer your question ... but
i will also point you to the code examples in javadocs, and on the wiki
page.
-
http://docs.codehaus.org/display/GEOTDOC/09+Collection+Classes#09Collect
Just a few note that popup from my mind:
* Shouldn't DefaultObjectCache and ObjectCacheEntry be package-privates?
* ObjectCacheEntry.lock doesn't need to be volatile, since the field
is never modified (it can be final). Unless I'm missing something,
"volatile" just add overhead in this cont
