Clojurians
#cljs-dev
<
2018-09-16
>

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

raymcdermott10:09:03

using latest master I have FAILs on the tests

mfikes11:09:32

Hmm. Master is currently OK in CI https://github.com/mfikes/clojurescript/commits/master Will try locally.

mfikes11:09:38

What version of NPM are you running @raymcdermott?

mfikes11:09:42

OK, I'm running 5.6.0 as well, and node v8.11.3

mfikes11:09:51

Hmm. I recall some problems with Node 9 even in another ClojureScript project I was running. Let me try that with ClojureScript's tests.

raymcdermott11:09:11

the versions of these (npm and node) are not mentioned anywhere in the pre-requisites

mfikes11:09:44

Right. The only problem I'm aware of is one that occurs with an ancient version of NPM / Node https://dev.clojure.org/jira/browse/CLJS-2878 And I had assumed my Node 9 problems may have been related to something half-baked in Node 9. (I think Node 8 and 10 worked for me IIRC.)

mfikes11:09:04

Hrm. I can't repro any failures locally

$ node --version
v9.7.1
$ npm -version
5.6.0
$ lein test :only cljs.module-processing-tests

lein test cljs.module-processing-tests

Ran 5 tests containing 14 assertions.
0 failures, 0 errors.

raymcdermott11:09:22

I get the same results with node v10.10.0

mfikes11:09:30

@raymcdermott You are experiencing https://dev.clojure.org/jira/browse/CLJS-2703 A workaround it to check out the compiler into a directory that doesn't have any hyphens in its name.

mfikes11:09:23

It is not exactly CLJS-2703, but close. New ticket: https://dev.clojure.org/jira/browse/CLJS-2914

raymcdermott12:09:57

ok, rebuilding / re-running

raymcdermott12:09:55

I hope that periods in the path are OK

raymcdermott12:09:11

with a period, 9 fails

raymcdermott12:09:17

so it’s going really well

mfikes12:09:48

Probably worth a JIRA

raymcdermott12:09:43

I’ll first check if simple paths work

raymcdermott12:09:55

ok, works with simple paths

raymcdermott12:09:19

I’ll redo with periods and open a ticket

raymcdermott13:09:57

and that’s blown my time budget for the first foray into CLJS development

raymcdermott13:09:08

at least we got a ticket out of it

raymcdermott16:09:19

do you want me to add a small note of caution to the Dev build instructions?

raymcdermott16:09:44

(for example *ensure the directory path has no special chars including ‘.’ and ‘-’)

mfikes16:09:47

My opinion: We can just fix the bug

raymcdermott16:09:02

that’s better

mfikes16:09:03

(It will probably be vanquished shortly)

mfikes16:09:55

(It probably affects more than just compiler devs... in that it could affect any ClojureScript project that has a dot in its name)

slipset19:09:09

@mfikes Looking through the cljs/newbie-friendly tasks. Quite a few of the older ones have patches that no longer apply.

mfikes19:09:08

If I find tickets like that, I usually leave a comment that the patch no longer applies.

slipset19:09:22

Yes, those are the ones I’m referring to :slightly_smiling_face:

slipset19:09:44

So I’m wondering is there a magic incantation to make the patch apply and still keep attribution?

alexmiller21:09:13

There’s no magic but if you can manually reapply the change and make a patch, you can just swap the author line with the one from the original patch to retain attribution

slipset19:09:04

Or is it up to the author of the patch to create a new one?

mfikes19:09:51

Those are open questions in my mind, especially in the case of non-trivial patches. But for newbie tickets may be easy to create new patches. Another concern is the case when the patch author hasn’t signed the CA.

slipset19:09:31

I guess for the last case, I understand one should do a clean room reimplementation of the patch, right?

mfikes19:09:35

Definitely. You don’t want to derive a patch from a patch authored by someone who hasn’t signed the CA.

mfikes19:09:32

(I suppose ideally you wouldn’t even review such patches.)

mfikes23:09:27

Many of my inchoate thoughts about the recent type inference stuff I've been brewing up were captured in a discussion with Daniel Compton https://therepl.net/episodes/3/

mfikes23:09:36

(I think the type inference part is the first 20 minutes or so IIRC.)

andy.fingerhut23:09:20

Sorry, nothing on topic here. Just had fun noticing that "inchoate" is a so-called "unpaired word", i.e. "choate" isn't a word that means the opposite of "inchoate". Humorous article "How I Met My Wife" is an exercise in using many of these all in one short story: https://www.newyorker.com/magazine/1994/07/25/how-i-met-my-wife