I was wondering if it is possible to convert (:a x) to (.-a x) if we know x is a record and a is in its basis. An experimental patch for this is in https://dev.clojure.org/jira/browse/CLJS-2950 Would be interested in feedback on whether this might cause any unintended consequences.

I just finished going through the few obvious spec issues I saw - this is the last one https://dev.clojure.org/jira/browse/CLJS-2728

needs to be rebased I think? @jbiserkov ^

I don't recall ever seeing anywhere the minimum version of Clojure required by the ClojureScript compiler being documented on https://clojurescript.org project.clj has 1.10.0-alpha in it script/bootstrap has 1.9.0 in it and script/build uses 1.8.0 The last seems to be the deciding factor, as the code being used in the compiler gets AOT compiled with 1.8.0.

On a different subject, I ran master trough Canary. A couple of @darwin’s projects might be affected by master (perhaps they just need some tweaks to their tests)

well only what’s declared in the pom.template.xml matters

all the other stuff is conveniences

@mfikes for cljs-devtools confirmed, they just need to update some expected test results, I will do that shortly after clojurescript release, for cljs-oops, it looks like previously generated google closure optimized code was incorrect in some cases, newly it emits what I would expect a) before: https://github.com/binaryage/cljs-oops/blob/1c878ff336feff7b3591e8f598c6e7eb169d86a1/test/transcripts/expected/oget_static_goog.js#L65 b) after: https://travis-ci.org/binaryage/cljs-oops/builds/448515341#L850-L853

Cool! (ClojureScript is currently sitting at zero release-blockers; just wanted to ensure these Canary test failures were OK.)

well, not really incorrect, but confusing (probably)

would have to look how goog.object.get works with only one arg, if it resturns it as-is, then it works technically

@mfikes I actually think we need to get the spec macro disable thing in

that’s the only outstanding thing I saw

I just pulled in all the big spec issues I found, that was the only patch that didn’t apply

👍

Minimum Clojure version stuff above captured for a potential future tweak to the site: https://github.com/clojure/clojurescript-site/issues/274

Hah, I had CLJS-2728 applied in a branch and, rebaselining it automatically retained Jordan Biserkov's attribution, so I attached that rebaselined patch to the ticket..