2010/7/20 Meikel Brandmeyer <m...@kotka.de> > Hi, > > On Jul 19, 11:09 pm, Laurent PETIT <laurent.pe...@gmail.com> wrote: > > > You got me thinking that all the stuff I and you were talking about > > concerning namespace dependency graph, etc. for only reloading the > > appropriate namespaces should be possible to be run from the running > REPL, > > and thus provided as a regular tool to be available to all environments: > > REPL, any IDE / back-end server etc. > > > > In its "pure" form it could recompute the dependency graph of the > namespaces > > every time "on-the-fly". > > > > With this function, even from the REPL could you be sure that you have > all > > the functions depending on the macros of some modified namespace > up-to-date. > > It is even possible to solve the "functions should be removed if no more > in > > the namespace" issue, by adding temporary watchers to the namespace vars, > > and then removing all the vars whose watchers were not triggered and are > not > > defonce vars (this defonce issue being the last problem I currently see: > how > > to know that defonce vars have been removed ? The only solution I > > temporarily envision is to rebind the defonce macro during the reload > with a > > decorator which will store information about defonced vars). > > > > Of course, some corner cases will still exist: direct access to vars in > > other namespaces via other-ns/the-var while not explicitly > requiring/using > > other-ns (maybe "oddly" relying on other-ns to be loaded preliminary, by > > other means). > > I had hopes that this discussion > > http://groups.google.com/group/clojure-dev/browse_thread/thread/b9d15561e8a3b5aa/0541da4e70ad682d > triggers something, but Stephen obviously didn't have the time to work > further on this topic. >
Oh, I hadn't noticed that the discussion also solved the "reload the dependent parts" as well. It's indeed a shame that it did not trigger anything in the end. One question : if I would like to give a try to just the "create a function which reloads the namespace as well as its dependent parts", on which primitives should I rely ? Not theoretically, but in practice (maybe "require" is the theoretic answer, but if "load", although lower level, "just works" and require "works in most cases", I would go for "load") ? Of course, if Steve could resurrect his code and finish the work this would be /good/ ! :) -- 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
