Fork me on GitHub
#cljs-dev
<
2020-04-14
>
dnolen00:04:25

@niwinz it's hard for me understand that issue - the comments seem to say suggest there is a workaround?

niwinz06:04:26

That issue was reported a long time ago, i try to have time to check it aggain on the current cljs compiler. In any case, for complete reproducible steps, look at the referenced issue: https://clojure.atlassian.net/browse/CLJS-1682 , the CLJS-2399 is a regression.

dnolen12:04:40

@niwinz let's not use threads thanks - what I mean is that issue just has irrelevant information, and too much information - so I can't really understand it nor want to spend a lot of time to understand it if it can be made clearer first

dnolen12:04:06

first, let's not talk about other versions in that ticket - only master

dnolen12:04:29

second, we have test cases for ES6 modules - what's the difference between this ticket and the fact that the ES6 test has always passed

dnolen12:04:48

then when I look at I can have some idea of what's going on

niwinz12:04:05

as i have said, this issue is too old, on the past I have had many problems using ES6 modules declared on deps.cljs (take care that the same code but using :foreign-libs on the compiler options directly was working correctly)

niwinz12:04:25

I try to take some time and check it again with the current master

niwinz12:04:56

and i ping you with any result 😄

dnolen12:04:56

yes I understand, the ES6 modules stuff doesn't get that much usage so I'm not surprised there are issues

dnolen12:04:20

but it would be nice to separate these issus into distinct simple cases that we can go through one by one

dnolen12:04:53

and again the ticket needs to be reconciled w/ the fact that the tests work - if you don't have time for that - at least capture that bit in the issue

niwinz12:04:54

i think the ticked is pretty clear with a very simple reproducible case (as said, was reproducible in the time of creating the ticket)

dnolen12:04:53

just because a ticket has reproducible case doesn't mean I'm going to look at

dnolen12:04:12

people report all kinds of tickets with reproducible cases that there just isn't enough information to consider

dnolen12:04:47

the repro is necessary but it isn't sufficient for a ticket that's straightforward to assess

dnolen12:04:31

anyways more work has to be spent on that ticket before I'm personally going to spend time on it - also anybody else can attempt to add more information or cleanup that ticket

dnolen12:04:39

@niwinz I'm not suggesting it has to be you

dnolen12:04:46

also @niwinz note the comment - Bobby Calderwood appeared to have a solution - does that or does that not work for you?

dnolen12:04:00

if that does work for you then I can know that the fix is easy

dnolen12:04:12

if doesn't work for you then we know we need more information

dnolen12:04:17

I removed all the unnecessary information from the ticket

dnolen12:04:58

and called explicitly in the ticket is a reported regression

niwinz12:04:25

i understand, but i don't have time to look inside the compiler. My main use case for this is being able to write some javascript and incluide it together with cljs in a library. I tried several times use the es6 and commonjs module transformations and it din't worked reliably; so I stick with the plain closure files (that are imporable directly in cljs like any other cljs namespace) or using the shadow-cljs facilities that works just out of the box

dnolen12:04:23

I understand - please try installing the missing NPM dep and see if that fixes your issue, if not say so in the ticket

dnolen12:04:48

then we don't lose this information when we actually have time to look at it closely