Bill Barker wrote: > Firstly, let me add my +1 to the proposal. tomcat-dev could use more > warm-and-fuzzies. ;-)
Thanks :) >>I have been following the discussion regarding the Tomcat 5 proposal. >> >>I have some general comments. >> >>Improve Performance Goal: >> >>I agree with Chritopher that in order to make improved performance >>a goal you need to have metrics by which you can measure whether you >>have met that goal, not anecdotal evidence. Catalina consists of >>a number of different components; mod_jk, http, JSP, SSI, CGI, etc. >>Some of these components can have different performance characteristics >>based on how they are being used. For example a simple HelloWorld.jsp >>versus Complex500Tags.jsp. This begs for a performance testing suite >>which can provide consistent reproducable results. If there is enough >>interest in this perhaps a separate repository could be created called >>jakarta-tomcat-bench. The benchmarks could be used to measure performance >>of different tomcat versions and other containers as well. > > > +0 > I'd probably use it if somebody else did the work, but I agree with Remy > that any single benchmark suite isn't going to tell you how your particular > web-app will perform. +1. It is not Tomcat buisness to define benchmarks or a workload. Others spent years doing that. If you're interested to do it, maybe you can start a project in commons, but this is off-topic here. Or, you can pick up your favorite load, run it once in a while, and post the results for us to enjoy. >>Proposal in General: >> >>The proposal is pretty vague on details. I have seen a number of >>replies stating "That's an implementation detail". I for one would >>like to see the proposal broken out into much more detail before >>work starts. Perhaps we should take a step back and start asking >>questions first so that there is more information and consensus for >>a formal proposal. Questions like: >> >> 1. What code in Tomcat really smells bad? > > > This is an evolution. Anything that smells bad can be fixed in 4.1.x (and > will be picked up by 5.0.x). If you want a revolution, that is another > proposal. *NO* code will be removed, except o.a.c.connector.*, which is indeed a bad implementation. That particular item is explicitely stated in the proposal. Other than that, other small refactorings will probably happen in both 4.1 and 5.0, but they are small refactorings which we will consider on a case by case basis. I am strongly in favor of maintaining compatibility (source and binary) with the 4.1 modules. We've lost a lot of time rewriting them, and I don't want to do that again. >>My fear is that work on Tomcat 5 will turn into a CVS version of >>the wild wild west if the proposal isn't detailed enough. > > > Again, this is an evolution. Currently in 4.1.x the only supported > connectors (with the exception of Warp, which Remy wants to bring in) use > Coyote. The proposal is little more than to expose a little bit more of > Coyote to the servlet-container to allow for some additional optimizations. > In addition, it removes the (currently deprecated) o.a.c.connectors.**, and > o.a.ajp.**. Think of it as spring cleaning. +1. Good summary. Remy -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
