I will this evening. This has been a major gripe of mine for years now. (Matthias probably also currently curses a lot on those issue)
It is in my opinion the biggest problem to get people into jsf and also the biggest problem to find people willing to put time into component development. Actually to sum it up, JSF has a not of nice areas, but the component api is simply a desaster. Adrian Mitev schrieb: > Hi, Werner! You could post that as comment to my issue at > javaserverfaces-spec > > 2007/3/16, Werner Punz < [EMAIL PROTECTED] > <mailto:[EMAIL PROTECTED]>>: > > Adrian Mitev schrieb: > > I've posted RFE for jsf 2.0 [1] about api for component development. > > > > > > https://javaserverfaces-spec-public.dev.java.net/issues/show_bug.cgi?id=246 > > <https://javaserverfaces-spec-public.dev.java.net/issues/show_bug.cgi?id=246> > > > > <https://javaserverfaces-spec-public.dev.java.net/issues/show_bug.cgi?id=246> > > > > Hi Adrian I think this does not go too far enough, a handful of common > utils classes does not really help with the problem > > what is needed is to tackle the problem on various ends > a) way too much configuration, try to get rid of it entirely, move the > entire configuration into annotations and probably also get rid of > the entire taglib binding classes that way. > > that way we also probably could get rid of the entire save and restore > code, and the properties simply would be setters and getters with the > save restore bindings woven around > > b) we need a renderer templating engine, but not like jsp more along the > lines of velocity which can be used optionally > > c) we need something high level (goes hand in hand with b) > eventually adding facelets to the mix, but with the possibility > to hook in custrom control classes in a rather seamless way (this would > be the same combination as in b, with facelets as templating for the > component) > > >
