that's a good point On Mon, Aug 29, 2011 at 9:39 PM, Igor Vaynberg <[email protected]> wrote: > yes, but we still iterate over *all* to build that list... > > -igor > > On Mon, Aug 29, 2011 at 12:37 PM, Pedro Santos <[email protected]> wrote: >> Helps a bit because it will be tested first in the list of MapperWithScore >> in CompoundRequestMapper >> >> 2011/8/29 Igor Vaynberg <[email protected]> >> >>> you can do this once in SystemMapper. if url starts with the namespace >>> give it a high value. >>> >>> however, this wont solve the problem where you have thousands of >>> mappers and we have to try them all. >>> >>> -igor >>> >>> On Mon, Aug 29, 2011 at 12:07 PM, Martin Grigorov <[email protected]> >>> wrote: >>> > Hi, >>> > >>> > Currently the compatibility score of PIM and BM is 0 so that users' >>> > mappers have priority. I think this is a bit wrong because mounting >>> > more pages in YourApp#init() will increase the time to get to PIM's >>> > mapRequest(). >>> > Most of the time stateful apps work with PIM because every callback >>> > url is processed by PIM (e.g. >>> > wicket/page?3-1.IBehaviorListener-form-button) >>> > I.e. there is no need to ask N MountedMappers before PIM when the >>> > chance that the request is for PIM is quite high. >>> > >>> > I suggest to make its #getCompatibilityScore() logic the same as >>> > BufferedResponseMapper, i.e. if the request url starts with >>> > 'wicket/page' then the score should be high (Int.MAX_VALUE). >>> > I see no problems with that for small apps but I see big gain for apps >>> > like Topicus' with 1000+ page (@Topicus devs: are they mounted pages?) >>> > >>> > The same is valid for BM. >>> > >>> > What do you think ? >>> > >>> > -- >>> > Martin Grigorov >>> > jWeekend >>> > Training, Consulting, Development >>> > http://jWeekend.com >>> > >>> >> >> >> >> -- >> Pedro Henrique Oliveira dos Santos >> >
-- Martin Grigorov jWeekend Training, Consulting, Development http://jWeekend.com
