Fork me on GitHub
#cljs-dev
<
2015-12-01
>
martinklepsch15:12:30

@spinningtopsofdoom: let me know if it works or not, should def. print minimal case in the failure case you pasted above

martinklepsch15:12:23

@spinningtopsofdoom: oh and are you aware that my fork uses boot so you need to install using boot? boot build-jar

martinklepsch15:12:45

maybe lein install still works as well but I haven’t explicitly checked that

spinningtopsofdoom16:12:47

@martinklepsch: I cleaned the maven cache and reran the installs and I'm still missing the actions leading to failure. Here's my branch for this https://github.com/bendyworks/lean-map/tree/collection-test

martinklepsch16:12:18

Tests failed:

    seed = 1448988800788
 actions = (-> coll)
@spinningtopsofdoom: when tests are failing I get something like the above above all the FAIL messages

spinningtopsofdoom17:12:31

Hmmmm I'm running this in repl via Cursive, I'll try to get a minimal case for you.

nberger20:12:58

@spinningtopsofdoom: I couldn't reproduce the "no test output" issue (well, it's not exactly "no output" what you were talking about, more precisely some output was missing). I set up lein-doo and ran the tests on phantom and on node 4.2.2, I always get something like:

Tests failed:
 
    seed = 1448999945948 
 actions = (-> coll (assoc (cljs.core/with-meta [0] {:foo 0}) (cljs.core/with-meta [0] {:foo 0})))

ERROR in (assert-lean-map-core-map-like) (TypeError:NaN:NaN)
map-like
expected: (= (map (fn* [p1__11530#] (a p1__11530#)) ks) (map (fn* [p1__11531#] (b p1__11531#)) ks))
  actual: #object[TypeError TypeError: a.call is not a function]
[...]
(there are 4 errors like that)

nberger20:12:17

I pushed the lein-doo setup to my doo branch which is based on your collection-test branch: https://github.com/nberger/lean-map/tree/doo

nberger20:12:10

So @martinklepsch I couldn't reproduce the issue that I was thinking was being caused by test.chuck. Sorry for the noise

martinklepsch20:12:17

Cool thanks for confirming :)

spinningtopsofdoom20:12:26

@nberger: thanks for the double check I thought it was only on my end, I justs didn't know enohgh about the test internals to be sure

r0man21:12:30

@dnolen I think I found the issue why I couldn't upgrade one of my CLojureScript apps that uses modules from 0.0-3308. We were talking about this some time ago on IRC. I hope I fixed it: http://dev.clojure.org/jira/browse/CLJS-1500

dnolen21:12:47

@r0man: nitpick but what’s the point of (map str) in module-entries?

r0man21:12:38

@dnolen: I think the items in :entries are symbols and Closure wants strings

r0man21:12:43

let me check

dnolen21:12:58

@r0man: pretty sure that’s not the case here follow the code path

dnolen21:12:04

@r0man: other thing that needs fixing

dnolen21:12:09

do not break signatures

dnolen21:12:21

add-dependency-sources

dnolen21:12:26

needs to retain the old signature

r0man21:12:27

ok, I thought it was the only use

dnolen21:12:36

at this point we never change signatures

dnolen21:12:51

unless we absolutely have to

dnolen21:12:05

so add an arity but do not remove the old one

dnolen21:12:23

@r0man: otherwise looks good

dnolen21:12:26

thanks for sorting this out

r0man21:12:09

np, can't wait to use all the new goodness simple_smile

dnolen21:12:40

also thanks for adding the test case, super helpful

r0man22:12:00

@dnolen: Ok, I hope I addressed your comments. No need for (map str), entries are symbols and work as you said. And I added back the one arity case. Also forgot to require set, had a stale REPL.

dnolen22:12:44

looks great

dnolen22:12:17

@r0man: applied to master

r0man22:12:55

Cool thanks, any pre/release on the radar? 😉

dnolen22:12:51

there are some other big things I’d like to address before the next release, but I’m OK doing a pre-release with this fix in it

r0man22:12:19

Thanks, that would be nice.

dnolen22:12:50

1.7.189 pre-release cut