On 3/5/06, Remy Maucherat <[EMAIL PROTECTED]> wrote: > > Costin Manolache wrote: > > Using less memory and supporting users who don't run tomcat on huge > servers > > is not > > ridiculous. And 'this is not my use case' is not a valid reason to veto > IMO. > > We're talking about a very small amount of memory. Webapps which have > lots of large JSPs are not small webapps, obviously: they are going to > require lots of resources anyway. Using straight classes is quite > efficient besides the small initial memory cost.
I have a feeling you only look at one use case ( high perf, heavy loaded server ) and can't accept there are other reasonable use cases. Well - probably I do the same :-), as I care more about my use case ( tomcat using low resources ) than the big server. Maybe you have a lot of large files and just a bit of 'dynamic' behavior - and most pages won't get 100 requests per second, maybe just few requests per minute. There are many web servers like this. Not all JSP uses are complex 'build a pet store with as many taglibs as possible'. Having a lot of content doesn't mean you need a lot of memory or CPU resources ( just a big hard drive ). The only thing that requires lots of resources - even if the load is very low - is the implementation of jasper, and even this patch can solve some of the worse effects. Anyway - there is not point to argue, at least until someone has a better patch. I agree that unloading JSPs is not the best solution. Costin
