There are a few people that have worked on this problem for Java and other JVM projects by basically NOT starting a new JVM or pre-starting the JVM. Some places to start:
- Drip - https://github.com/flatland/drip (targeting Clojure specifically) - Nailgun - http://www.martiansoftware.com/nailgun/ (used heavily in the JRuby world) Jave SE 6u10 includes the Quick Starter too (really intended for applets) preloads class data to make it warm. Not sure if that can be leveraged somehow or that anyone has ever tried to specific to Clojure code. My impression is that the slow part of starting a Clojure program is not the Java startup as much as loading Clojure itself and the Clojure code to run. AOT should help some with loading your own code as you won't need to compile it. On Sunday, October 7, 2012 11:50:52 AM UTC-5, Brian Craft wrote: > > The two second.... > > delay to... > > do anything... > > is making... > > me crazy. > > I should probably be asking this on a java forum. I'm evaluating clojure > for a project that needs some number of cli tools (as well as server and > browser code) to be delivered to customers. Are there any good solutions to > the start-up delay? I've seen tools that run java as a service, so it's > always up, but I'm not crazy about the idea of requiring this of our users. > IIRC emacs had similar problems in the 90's, and the solution was to store > the initial VM state so it didn't need to be recreated on every invocation. > I'm probably getting the details wrong. And I expect there's some good > reason this doesn't work with java. Is there? > -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
