i get a dejavu feeling now... We had this discussion long before. Add detach to Behavior... Back then it was voted down i believe for 1.2.x Because somebody could implement IBehaviour directly instead of AbstractBehavior..
But i am +1 by cleaning it up, especially if it fixes things. I will also try to backport the AccessStackPAgeMAp leak fix and the open file problem. I guess both are not yet in? johan On 2/14/07, Matej Knopp <[EMAIL PROTECTED]> wrote:
Why wouldn't you be allowed to do that? Does anyone object? -Matej Igor Vaynberg wrote: > thats because we have ten thousand methods for attaching and detaching, and > behaviors dont have a nice clean detach() like they should! > > thus the whole nasty thing with clean() in abstractbehavior > > some of this i have cleaned up in 1.x but i dont think i am allowed to add > behavior.detach() and cleaning it up further > > -igor > > > On 2/14/07, Matej Knopp <[EMAIL PROTECTED]> wrote: >> >> +1 >> >> I hope it will work this time. We are fighting this since 1.2.1 or so. >> >> -Matej >> >> Igor Vaynberg wrote: >> > currently things are not properly detached in 1.2.x after an ajax >> request. >> > this patch [1] fixes it. >> > >> > the root of the problem is that only components that are rerendered get >> > behavior.cleanup() called. but if you have something like this: >> > >> > page >> > +- ajax link >> > +- panel >> > >> > and that ajax link repaints the panel the links's behavior.cleanup is >> not >> > called because the link itself is not rerendered and it is the end of >> > render >> > that calls behavior.cleanup(). >> > >> > [1] http://papernapkin.org/pastebin/view/4627 >> > >> > -igor >> > >> >> >
