On Tue, Dec 31, 2013 at 3:03 AM, Timothy Pratley <[email protected]>wrote:
> Hi David > > > >> A ticket exists for this. Patch welcome :) >> > > Is this the ticket? > http://dev.clojure.org/jira/browse/CLJS-615 > It speaks of analyzer.clj, but I don't see such a file. > I do see analyze as a function in compiler.clj which sounds promising. > analyzer.clj exists. > closure.clj: > (defn cljs-dependencies > ;... > (map #(let [f (ns->cp %)] (hash-map :relative-path f > :uri (io/resource f)))) > (remove #(nil? (:uri %))) > > Looks to be ignoring missing files. I suspect though that it merges with > js dependencies later, and that the trick will be recognizing requires that > are in neither... I'll dig into it tomorrow and see what I can discover. > This is not the place to put this assertion check. The best location is ns form parsing in analyzer.clj. Not sure what difference it makes, warnings are errors but without blowing >> up the compilation process so you can get many warnings at once. >> > > Some compilers report multiple errors, it may be reasonable to classify > the output however you like. My opinion is that an error is appropriate > when you know 'it' just wont work, and refuse to produce an output... > whereas a warning is appropriate as a hint that probably 'it' will do > something you might not expect. In the specific case of an undeclared var > will it always be a runtime error? I can see an argument to say some js > output is better than none even if certain paths will fail. > This is already supported to some degree. Again look at analyzer.clj. Some errors will halt compilation - source reading errors, errors in the ns form, and those that would result in a mangled file. We haven't been super disciplined how we classify these. Yet another place for someone to contribute. > I believe that the reason the above does not trigger an error is because >> 'typomina' might be in a Javascript file somewhere. In GClosure there is >> not an 1-to-1 correspondence between namespaces and files. There is no sane >> way to find a file given a namespace symbol and a file might even define >> several namespaces. >> > > Ah that makes sense and makes me think that the code I was looking at > can't be patched so easily then. > It seems I need to read up on exactly what 'require' means in GClosure. I > was initially mistaken that JavaScript code would be imported to > distinguish from ClojureScript being required, but that is not the case. > Verifying that random JS libs exist on disk will require more work, but there's a lot of value in starting with .cljs and .clj macro files first. David -- Note that posts from new members are moderated - please be patient with your first post. --- You received this message because you are subscribed to the Google Groups "ClojureScript" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/clojurescript.
