This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2016-08-23
Channels
- # admin-announcements (1)
- # alda (1)
- # bangalore-clj (5)
- # beginners (17)
- # boot (392)
- # capetown (4)
- # cider (16)
- # cljs-dev (24)
- # cljsrn (33)
- # clojure (106)
- # clojure-berlin (1)
- # clojure-nl (1)
- # clojure-russia (168)
- # clojure-spec (85)
- # clojure-uk (137)
- # clojurescript (83)
- # clojutre (4)
- # component (10)
- # cursive (6)
- # datavis (9)
- # datomic (11)
- # defnpodcast (15)
- # dirac (4)
- # docker (1)
- # ethereum (1)
- # hoplon (27)
- # jobs (5)
- # jobs-rus (1)
- # lein-figwheel (2)
- # luminus (5)
- # off-topic (5)
- # om (13)
- # onyx (60)
- # parinfer (2)
- # planck (12)
- # proton (2)
- # re-frame (45)
- # rethinkdb (5)
- # ring-swagger (9)
- # spacemacs (9)
- # specter (49)
- # test-check (1)
- # untangled (104)
- # yada (10)
@dnolen probably but will need some tweaking because of this: https://github.com/clojure/clojurescript/blob/master/src/main/clojure/cljs/closure.clj#L79-L92
the way it is currently is we’re saving is-old
and is-new
with 20160713, it’ll become incoherent: is-old = true, is-new = true
doesn’t make sense
I think the easiest would be to wait for 20160822 which has a substantial refactor of ES6ModuleLoader
to ModuleLoader
and just bump to that?
@anmonteiro I’m fine with waiting
@mfikes I wrote a patch to transit-java that I am curious if it resolves your parallel-build problems. If it doesn't it would disprove my hypothesis about separate containsKey()+get() being a problem. https://github.com/favila/transit-java/commit/3bb53013793b584496d61786c40a34bd0696065d
@favila It ultimately failed in what I suspect is the same way. Here is a gist containing a transcript: https://gist.github.com/mfikes/97d9ab512e428a792f3932102a572dc2
@favila FWIW, I was very careful to ensure I was testing with your patch, so I think the test above is valid.
@mfikes I guess it has to be something in WriteHandlerMap's mutation of the default handler that somehow doesn't get sychronized. Did you ever get a chance to try patch on http://dev.clojure.org/jira/browse/CLJS-1761 ? If it's still broken even with that, then I'm stumped.
@favila Are you referring to the change you describe in this comment? http://dev.clojure.org/jira/browse/CLJS-1761?focusedCommentId=43703&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-43703
@favila No. I’d have to actually comprehend what you wrote and apply that, which is more than I can afford to do at the moment. 😞 Perhaps later.
@mfikes Transit allows you to pre-create the handler map, when you do there is no handler cache involved. That is all the patch does: instead of read/write handlers only for custom types, creates handler map for all transit types ahead of time.
@favila Yeah, I’m sure it is a simple patch. I’m just unable to context switch right now.
FWIW the Closure Compiler 20160822 release is now out on Maven http://search.maven.org/#artifactdetails%7Ccom.google.javascript%7Cclosure-compiler-unshaded%7Cv20160822%7Cjar