Re: DefaultObjectWrapper

It is fine to submit patches using ConcurrentHashMap for 2.4 (as it is  
declared to require Java 5 or higher). Best would be to open a feature  
or enhancement request in the SourceForge tracker and submit them there.

As for 2.4 timeline, I can't tell. Whenever Jonathan and Daniel  
jointly feel comfortable with the state of the code; I'm myself  
admittedly not doing much in it at the moment. If you can afford it,  
you can try testing the current SVN trunk tip; it's in a good enough  
shape that it shouldn't be risky to use it.

Attila.

On 2008.05.12., at 5:35, Bruce Ritchie wrote:

> Attila,
>
> On Sat, May 10, 2008 at 2:59 AM, Attila Szegedi <[email protected]>  
> wrote:
>> Completely threadsafe. You should have one per Configuration, and you
>> normally have one Configuration per your application. Struts should
>> really reference a single instance somehow, but instead of a static
>> singleton, the relevant object structure that does the tag parsing
>> would better have access to a Configuration object somewhere.
>
> Thanks, I implemented a quick fix and the performance improvement is
> noticeable in our automated tests.
>
> It seems to me from looking over many of the classes used by
> BeansWrapper that there is a lot of synchronization code that could be
> improved with the use of 1.5 classes such as ConcurrentHashMap. For
> example, the classCache in BeansWrapper or the cache variable in
> ClassBasedModeFactory could easily be changed to concurrent hash maps
> allowing for the unblocking of read operations for the typical case.
>
> Would there be any objection if I was to submit some patches for this
> against trunk?
>
> Also, is there a timeline for 2.4?
>
>
> Regards,
>
> Bruce Ritchie




-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft 
Defy all challenges. Microsoft(R) Visual Studio 2008. 
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/