Fork me on GitHub
#cljs-dev
<
2017-12-22
>
mfikes13:12:45

@dnolen If you end up doing ClojureScript work today, I'd vote for this ticket. It is a regression currently only on master: https://dev.clojure.org/jira/browse/CLJS-2449

dnolen15:12:35

Yep gonna go through tickets probably cut a release unless there’s some reason not to?

mfikes15:12:57

That's the only regression on master I'm aware of.

rauh15:12:47

Is there a file where I can see which externs were inferred?

rauh15:12:10

Nvm, found it at the obvious place 🙂

dnolen16:12:31

what other issues should I look at?

dnolen16:12:03

I’m going to do :stable-names now since that’s an easy one and might mess around a bit with caching of JAR deps

dnolen17:12:24

@mfikes I’m fine with the ns special case approach on CLJS-2413

juhoteperi17:12:15

@dnolen Works for November Closure but not December, advanced build prints some warnings, not sure if those affect output

dnolen17:12:47

ok I’m going to leave it be - I feel like there’s too much churn around the modules stuff in Closure Compiler right now?

dnolen17:12:20

https://dev.clojure.org/jira/browse/CLJS-2216, this an old one but it’s nice and just needs a rebase

dnolen17:12:43

3-way merge didn’t work, I can manually do this - but happy to give you credit via a new patch

dnolen17:12:00

I bumped your JIRA privs too

kommen17:12:37

you think it’s a good idea to have :webworker it as a separate :target?

kommen17:12:47

I’m happy to rebase it for you

dnolen17:12:56

I don’t see why not, people have been asking for this for years

kommen17:12:24

also re the :closure-defines build-affecting-options thing from a few days ago: for some reason I didn’t look close enough and the :output-to file was actually updated and I though it was not, so all good there now

dnolen18:12:27

another case of direct invokes causing issues

mfikes18:12:56

Interesting

mfikes18:12:53

@dnolen If you end up releasing, I've written a draft of some copy that could go with the core.specs.alpha feature: https://gist.github.com/mfikes/a48780bec910e8c53a5b32e49f0d12f0

dnolen18:12:15

I’m thinking we want a hook like warning handler for spec errors

dnolen18:12:22

so that people can interpret them however they see fit

mfikes18:12:49

If you end up having separate guides for the release, we could put that into one, or if you just make overall release notes, feel free to incorporate that copy.

mfikes18:12:29

Yeah, a hook might be nice. I wonder if Clojure has figured out a "standard" way to do this that ClojureScript can follow.

Alex Miller (Clojure team)18:12:21

We have the explain-out dyn var

mfikes18:12:41

Ahh cool. So does ClojureScript.

mfikes18:12:45

cljs.user=> (source cljs.spec.alpha/*explain-out*)
(def ^:dynamic *explain-out* explain-printer)

dnolen18:12:18

but we probably want to have a compiler option to let people bind that through tooling

dnolen18:12:27

:explain-out

mfikes18:12:23

Thanks @alexmiller for the core.specs.alpha lib. I know it's your job and all, but... you know 🙂

dnolen18:12:49

sometimes porting Clojure code to ClojureScript is really mind bending - think I sorted this arity issue out

mfikes18:12:32

I get a downstream warning when building Planck; looking to see if it is innocuous

mfikes18:12:46

WARNING: Use of undeclared Var cljs.core$macros/m at line 756 out/cljs/core.cljc
related to https://github.com/clojure/clojurescript/commit/9ec796d791b1b2bd613af2f62cdecfd25caa6482

mfikes18:12:09

My initial hunch is that this should be core/fn

mfikes18:12:05

It is trivial to patch that... harder to test it

dnolen18:12:19

pushed that fix

mfikes18:12:15

Confirmed. Where previously

$ script/test-self-parity
WARNING: Use of undeclared Var cljs.core$macros/m at line 756 /Users/mfikes/Projects/clojurescript/src/main/clojure/cljs/core.cljc
WARNING: Use of undeclared Var cljs.core$macros/m at line 757 /Users/mfikes/Projects/clojurescript/src/main/clojure/cljs/core.cljc
WARNING: Use of undeclared Var cljs.core$macros/m at line 758 /Users/mfikes/Projects/clojurescript/src/main/clojure/cljs/core.cljc
Testing with Node
now cleanly
$ script/test-self-parity
Testing with Node

mfikes19:12:32

@dnolen Draft of some copy for the iterate, repeat, cycle change: https://gist.github.com/mfikes/f775e56c2566e8ac6c7f159badb2c95d

dnolen19:12:02

@mfikes cool if you want to put these together into a PR for the site that would be awesome

mfikes19:12:36

@dnolen I'm going to be AFK for the next several hours. If needed, feel free to pull this PR and post-edit.

dnolen20:12:03

@juhoteperi this looks fine to me but the patch has at typo - path isn’t there

rauh21:12:20

@dnolen Regarding the compiler perf ticket: It's a modest speedup of ~20% only for recompiles. It's slightly noticable in figwheel when I re-compile "master namespaces" which a lot of other namespaces depend on. It seems to have little impact on initial compiles. Quote: "I'll give it a try with something else". Does that mean I should come up with something else or you have some ideas?

kommen21:12:37

not sure though if it has to be adapted still to reflect the changes from https://github.com/clojure/clojurescript/commit/345041b87f8004102aa300b50829e70fa0d646c4 ?

kommen21:12:32

going to be afk for today

dnolen21:12:47

@rauh no I will try it with something else