2017-01-26 22:13 GMT+01:00 Mark Thomas <ma...@apache.org>: > On 26/01/2017 14:38, Rémy Maucherat wrote: > > 2017-01-26 15:29 GMT+01:00 Mark Thomas <ma...@apache.org>: > >> On 26 January 2017 13:53:22 GMT+00:00, Konstantin Kolinko < > >> knst.koli...@gmail.com> wrote: > > <snip/> > > >>> It needs some actual numbers - how big of a JSP can be compiled by > >>> Tomcat. > >> > >> It took around 250 tags with a scriplet on a single page to trigger the > >> problem. I need to test how much difference the try/finally fix made. > >> > > Working with Tomcat 6 (9 won't be that much different), before the > try/finally clean-up code was added a JSP page could handle 299 > instances of the foo tag (from the examples webapp) before the method > got too big. After adding the try/finally clean-up, that drops to 228. > > I've been looking at the generated code for a while. One obvious > optimisation had no effect (I'm guessing the compiler found it anyway). > > With a couple a small changes, I got the maximum tag count back up to 249. > > To take this further, I think the next step is to take the setters and, > if the tag processed in _jspService(), move the setters for that tag > into a separate method. > > I intend to explore this in 9.0.x with a view to back-porting if it works. > > If the try/finally fix is supposed to be important, it would be good to use this opportunity to resurrect my try/finally patch for simple tags now that the release cycle is done (otherwise, they have the same bug). With the addition of Violeta's idea though.
Rémy