Clojurians
#cljs-dev
<
2018-03-28
>

This page is not created by, affiliated with, or supported by Slack Technologies, Inc.

mfikes00:03:58

FWIW, I'm nearing getting the Windows unit test situation cleaned up. https://dev.clojure.org/jira/browse/CLJS-2711 (ready) https://dev.clojure.org/jira/browse/CLJS-2669 (ready) https://dev.clojure.org/jira/browse/CLJS-2712 (assigned to me) Once that last one is done, we'll be in good shape with not regressing on Windows.

mfikes03:03:32

^ Solved, now down to just two Windows test failures :slightly_smiling_face: https://dev.clojure.org/jira/browse/CLJS-2713 https://dev.clojure.org/jira/browse/CLJS-2714

mfikes12:03:02

Oh, it is Wed, so we can post the Shared AOT Cache news post if nobody sees anything wrong with it https://github.com/mfikes/clojurescript-site/blob/issue-183/content/news/2018-03-28-shared-aot-cache.adoc

mfikes12:03:33

I did some light c/e to ^, it LGTM at the moment.

john12:03:35

It sorta feels like it's almost big enough to break out into sections. Any smaller and it seems it should go into its parent doc context... Definitely reads cogent though

dnolen12:03:15

@mfikes looks good! go for it

mfikes13:03:02

I think we said we are cool with updating the original news post to point to this article. This PR would do that https://github.com/clojure/clojurescript-site/pull/213

dnolen14:03:11

@mfikes merge away!

dnolen15:03:47

didn’t GraphQL have the same problem?

dnolen15:03:10

if so let’s switch over to a less annoying dep

dnolen15:03:08

@juhoteperi so it’s same problem yes?

dnolen15:03:16

it’s what I recalled anyway

juhoteperi15:03:50

Problem with pg was that main points to a directory (with index file inside), not file

dnolen16:03:07

not totally related but I do think we should take the approach of Closure Compiler and support testing w/o going to the file system

dnolen16:03:12

it would really simplify testing

dnolen16:03:38

and probably speed it up too

dnolen16:03:59

we could test more without doing full builds

jannis17:03:05

It would indeed be good to find another npm package to replace pg/pg-native with in the test. There must be more packages that do the same (point their main entry to a directory).

mfikes17:03:22

goog-define is not often used directly in a REPL, but if :def-emits-var is set, it might be nice for the macro to return the var being defined.

mfikes17:03:55

(The var is actually created by the declare, just not being returned.)

juhoteperi17:03:11

I wonder if it will be easier to create dummy npm package with main entry folder than to search for one

mfikes17:03:05

That's what I wanted to do, but was afraid of missing a salient aspect.

juhoteperi17:03:06

I'll add patch to use this

mfikes20:03:54

I'll be out-of-town starting tomorrow, probably not returning online until sometime Monday.