Digging into the related code I see that if my new test has no listener interface then everything is OK. So I reverted the revert and it is as in 6.x again.
For the next tickets I may need to move it again. Martin Grigorov Wicket Training and Consulting On Fri, Jan 10, 2014 at 12:07 PM, Martin Grigorov <[email protected]>wrote: > Hi, > > > On Thu, Jan 9, 2014 at 3:38 PM, Sven Meier <[email protected]> wrote: > >> Hi Martin, >> >> it seemed reasonable to me to let the mapper decide to drop the >> parameters. It already has all information at hand, why delay the decision >> and put the burden on the PageProvider? >> >> >> >My reasoning to null-ify in PageProvider is: if there is parsed pageId >> and the page cannot loaded from the >> >stores then it is expired, so the page parameters are ignored and new >> page instance is created. How do >> >you decide to drop the parameters in the mapper ? >> >> If there's a pageId, the request parameters are uninteresting for the >> page provider, aren't they? >> > > Maybe not: https://issues.apache.org/jira/browse/WICKET-4441 > I need to debug some more to see what params come with the following > requests (request caused by click on a link for example). > > PackageMapperTest#decodeNamedParameters() is a new test introduced for > WICKET-4686 and it also needs the parameters to be able to read the named > parameters from the url. > > >> >> https://git-wip-us.apache.org/repos/asf/wicket/repo?p= >> wicket.git;a=commitdiff;h=5e1bf8d8169a8d01f041a1d2bf41a8b8fe170dbd >> >> But feel free to revert this, if null-ifying the parameters in >> PageProvider helps you in solving the puzzle of these settings. WICKET-5068 >> is really important IMHO. > > > For now I'll revert it and keep an eye on it when dealing with the other > tickets in this area. > > Thanks! > > >> >> >> Sven >> >> >> On 01/09/2014 02:12 PM, Martin Grigorov wrote: >> >>> Hi Sven, >>> >>> What was the reason to move the null-yfing of the parsed PageParameters >>> from PageProvider to AbstractBookmarkableMapper in >>> https://git-wip-us.apache.org/repos/asf/wicket/repo?p= >>> wicket.git;a=commitdiff;h=b3982a4beff352cd5e61521490a397a926c13eed >>> >>> and >>> https://git-wip-us.apache.org/repos/asf/wicket/repo?p= >>> wicket.git;a=commitdiff;h=5e1bf8d8169a8d01f041a1d2bf41a8b8fe170dbd >>> >>> I am working on https://issues.apache.org/jira/browse/WICKET-4686 and I >>> had >>> an issue >>> with org.apache.wicket.core.request.mapper.PackageMapperTest# >>> decodeNamedParameters >>> because of this change. So I reverted it in branch sandbox/WICKET-4686. >>> >>> My reasoning to null-ify in PageProvider is: if there is parsed pageId >>> and >>> the page cannot loaded from the stores then it is expired, so the page >>> parameters are ignored and new page instance is created. >>> How do you decide to drop the parameters in the mapper ? >>> >>> Other related tickets that I'd like to address for Wicket 7: >>> https://issues.apache.org/jira/browse/WICKET-5068 >>> https://issues.apache.org/jira/browse/WICKET-5070 >>> >>> Martin Grigorov >>> Wicket Training and Consulting >>> >>> >> >
