|
From: Colin S. <col...@ex...> - 2005-09-21 15:35:51
|
That was the main reason. Most people don't want to bother (or often even know about) chaining view resolvers. So since they couldn't rely on the [Xml|ResourceBundle]ViewResolver being in use and taking care of this, there was a decent amount of controller code out there that just did manual instantionations of RedirectView. I always found that pretty nasty and unclean. A controller just shouldn't know about that kind of thing. Colin Darren Davison wrote: >Hi all, > >The "redirect:" prefix that was added as a feature last year enables >controllers to remain ignorant of the redirect since the String value >containing the prefix can be injected into the controller bean. That much I >understand. > >What I'm wondering is whether there was another benefit, or problem solved, by >taking this route? It seems to be a natural limitation of the choice of >ViewResolver that the problem actually solves.. if the app is using a >UrlBasedViewResolver or subclass, the redirect: prefix is needed. > >However, if the app defined an [Xml|ResourceBundle]ViewResolver and ordered it >first, then that could be used to resolve redirect views, falling back to the >UrlBasedViewResolver for everything else. It looks like this would mitigate >the need for the special handling of view name prefixes. > >I can't see anything in the discussions or JIRA issues from around the time >this feature was added eluding to this other than a comment I made myself [1], >so am I missing something here? > >[1] http://opensource2.atlassian.com/projects/spring/browse/SPR-387 > >Regards, > > |