This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2019-01-11
Channels
- # aleph (38)
- # announcements (6)
- # aws (1)
- # beginners (47)
- # calva (21)
- # cider (47)
- # cljs-dev (18)
- # clojure (40)
- # clojure-europe (7)
- # clojure-india (2)
- # clojure-italy (9)
- # clojure-nl (11)
- # clojure-norway (2)
- # clojure-sanfrancisco (1)
- # clojure-spec (17)
- # clojure-sweden (2)
- # clojure-uk (73)
- # clojurescript (10)
- # cursive (6)
- # datascript (12)
- # datavis (2)
- # defnpodcast (1)
- # duct (5)
- # emacs (36)
- # figwheel (2)
- # figwheel-main (10)
- # juxt (12)
- # leiningen (1)
- # midje (1)
- # nrepl (9)
- # off-topic (25)
- # pedestal (3)
- # portkey (3)
- # quil (2)
- # re-frame (45)
- # reagent (1)
- # ring (3)
- # ring-swagger (36)
- # rum (1)
- # shadow-cljs (48)
- # spacemacs (1)
- # speculative (50)
- # testing (2)
- # tools-deps (27)
- # yada (4)
https://github.com/thheller/shadow-cljs/commit/fe33a76fbd7454f4ede149bb123b7a6cf8741a2c#diff-5823118fcf420ad359fbb92728de1339R96 — would this line be a reasonable place to filter warnings via namespace? 🙂
@U0KHZ4HNG it is only supposed to warn once for library code?
the :from-jar
is supposed to filter everything where the source file is in a jar, ie. library
hmm, good point … our build system is a bit bespoke (we use BUCK) so at times our jars aren’t jars … but in this case I think they are in fact jars. its stuff like om.next
warning about *logger*
, etc. we also execute everything from the devtools api, not the command line, if that matters, e.g. (server/start!)
, (shadow/watch :foo)
I’m very much on board with the warnings in our namespaces — they’ve found actual bad stuff so far 😜
I just built a custom target type for nubank workspaces for shadow to auto-scan…was pretty easy. Thanks @thheller for making hooks to make that possible!
Damn that is way simpler than I would have expected
@tony.kay what kind of auto-scan? I have some built-in workspaces support but didn't quite figure out how to configure things like preloads and other dev stuff. would be interested to see what you built
Is there a way to get Cursive to resolve npm requires? I'm having an issue where doing something like (ns my.ns (:require ["firebase/auth" :as fb-auth]))
is showing up marked as unresolved. If not, is there a way to trick it?
@thosmos unfortunately not. I don't think string requires are undestood at all at this point. let alone npm resolve.
Is there a way I could pre-package it somehow so that I could do (:require [firebase])
and have cursive resolve it. Maybe with externs? I guess this is what cljsjs does?
So if I wanted it to resolve all the fns I'd need to create stubs for each namespace including each fn signature? Would that theoretically work? I imagine that an externs file could be used to auto-generate this.
ok thanks. I still much appreciate being able to code directly against the required symbol rather than js/whatever
I know that it sometimes asks to generate stubs for some deps, but I have no idea what that means
@thheller @tony.kay just tried using the new target, I have an issue because my old main had some requires that were used by all the other things, but none of then require it (specs file and workspace extension things), is there a way to indicate that those should be loaded before anything else? I tried just adding a new file that requires those, but the topological order is getting wrong and the code is trying to use the specs before they are loaded
I know I could just require it everywhere, but the point is trying to avoid it
can I just add those to :entries
?
you can but the target doesn't use that so you need to adjust the target code to use it
(assoc-in [::modules/config :main :entries]
(-> '[]
(into (get config :entries []))
(into dynamically-resolved-namespaces)
(conj 'nubank.workspaces.shadow-cljs.mount))))
Are you aware of any IDE that provides editor introspection based on js externs, even if it's only available in javascript? This might be easy to leverage for CLJS.
@thheller that worked great 🙂
sometimes in my CI the tests are failing with ../deploy/bin/shadow-karma-test: line 9: echo: write error: Resource temporarily unavailable
, had anybody here had this problem? ideas what could be about?
it only fails sometimes, I would say about 30% of the runs
I wonder if some IO resource is taking too long to respond and causing it