Okay, I understand. Actually, I don't really want OGNL
CompoundPropertyModel to be replaced. I just want to have the
possibility of non-ognl compound property model in 1.1. So maybe It
could be added to wicket-extensions?
-Matej
Martijn Dashorst wrote:
I'd rather finalize 1.2 sooner,
Very very nice!
then what i did see in the profiler also is very real in the real world!
And there is also another factor that you will only really see if the load is high..
and that is that there is also the same factor how many memory is gc'ed!
johan
On 11/17/05, Matej Knopp [EMAIL PROTECTED]
How about the percentage it takes up in comparison of the total
computing for one request? Any other bottlenecks you can identify? Or
was - like Johan said - OGNL one of the biggest hurdles?
Eelco
On 11/17/05, Matej Knopp [EMAIL PROTECTED] wrote:
Okay, I took CompoundPropertyModel from 1.2 and
Well, the numbers I've sent are rought, I know, I didn't profile it.
It's just a filter that checks how much time does it take to process a
request. The only change between the two numbers was that I switch old
CompoundPropertyModel to new model (called FastCompoundPropertyModel to
are you really seeing big performance improvements when comparing the same app with 1.1 and 1.2?
Thats nice to hear :)
Backporting to 1.1.1 ... we could do it but it is a big change.. that
could affect running installs (you can't just swap 1.1 to 1.1.1 then if
you depend a bit more on ognl)
This
I'd rather finalize 1.2 sooner, cutting off new features, than
backport the OGNL replacement into 1.1. It is a breaking change, and
by its nature, a 1.1.x release should not break stuff.
Martijn
On 11/16/05, Johan Compagner [EMAIL PROTECTED] wrote:
are you really seeing big performance
