Peter, We asked us the same question some weeks ago, on clojuredev.
We took the path to follow how eclipse launches a java application when the user requires it to test it. So we created a customized "launch configuration" (sorry, eclipse jargon), that is just a classical eclipse java launcher with some options predefined. To enable the user work against the clojure version he wants, we require him to have clojure library as a standard library dependency on the classpath of his eclipse project. I think you could do quite the same thing for IntelliJ. We did that because of several reasons : - we wanted the user to be able use his own version of clojure, not an imposed one - we wanted the user to be able to have several clojures running at the same time - we absolutely wanted to prevent the user run his code in the eclipse VM ! - we wanted regular eclipse users to feel at home with clojuredev, having in mind that having a smooth migration path from java to clojure, first by just incorporating bits of clojure in a (n already existing) java project codebase, could be a good thing. So we took the road you described by quoting Stuart. We call a "bridge function" that takes strings and returns strings (or core clojure datastructures : maps, vectors, strings, keys, java.lang basic types). This bridge function runs on the eclipse JVM, and calls a server we systematically install in the remote JVM when the user launches his project. The code for the client part is here : http://code.google.com/p/clojure-dev/source/browse/clojuredev/trunk/src/clojuredev/debug/clientrepl.clj The code for the server part is here : http://code.google.com/p/clojure-dev/source/browse/clojuredev/trunk/src/clojuredev/debug/serverrepl.clj (and yes, it's yet another variation on the repl over socket) HTH, -- Laurent On 22 jan, 22:43, Peter Wolf <opus...@gmail.com> wrote: > This is a rejuvenation of the old "calling Java from Clojure" thread > > I have been looking at the solutions from Mark > / > 1) From a Java application, read a text file containing Clojure code > and invoke specific functions it defines from Java code. > 2) Compile Clojure code to bytecode and use it from a Java application > just like any other Java code. > > An example of the first option is provided at > > http://markmail.org/message/tx23zaxf77b6widh#query:%22calling%20Cloju.... > > / > > and Stuart > / > Here's how I do it: > > import clojure.lang.RT; > import clojure.lang.Var; > ... > RT.loadResourceScript("source/file/on/classpath/here.clj"); > Var myfunction = RT.var("my-namespace", "my-function"); > myfunction.invoke("arg 1", "arg 2", ...); / > > Where is the documentation on this Java API? In particular, all these > functions seem to side-effect a single Clojure image. > > Can I create several disjoint images? Can I stop or destroy an image > that has run amok? > > I want to use this for the IntelliJ plugin, and I don't want bogus user > code to clobber the IDE. I also want the defined symbols for a file to > be a function just of the code in that file, not all the files that > happen to be loaded by the IDE at the moment. How does Swank handle this? > > Thanks > Peter --~--~---------~--~----~------------~-------~--~----~ 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 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 -~----------~----~----~----~------~----~------~--~---
