Fork me on GitHub
#cider
<
2019-06-07
>
borkdude14:06:21

I consider this the channel also for the Clojure style guide (because of a lack of one and because @bozhidar is here :)) So I’m going to link this here: https://github.com/borkdude/clj-kondo/issues/241 There have been discussions in several issues in different projects. I think the style guide should maybe take a stance on this

bozhidar15:06:20

@borkdude I prefer option 2. I totally agree we should document this in the style guide. I’ve many times deleted a namespace like this, because joker told me it was unused. 🙂

lread15:06:55

joker was just joking about that I guess simple_smile

borkdude15:06:55

@bozhidar I like that idea, except that it breaks alignment

borkdude15:06:33

[^:keep foo] would also be possible, but some tools might not see the metadata because you can’t actually add metadata onto a symbol 😉

bozhidar15:06:11

Good points.

bozhidar15:06:06

Well, I guess option is the most balanced then, but I assume some linters might consider an obsolete refer. 😄

bozhidar15:06:15

Tricky situation…

borkdude15:06:41

clojure-mode could be adapted to align vectors vertically in namespace declarations

borkdude15:06:08

about vertical alignment, it seems clojure-sort-ns already respects metadata, kind of:

(ns foo
  (:require
   ^:keep [bar.specs]
   ^:keep [foo.specs]
   [another.namespace.a]
   [another.namespace.b]))

borkdude15:06:27

so maybe this wouldn’t be a problem, since ‘kept’ libspecs go on top anyways

borkdude15:06:19

another idea:

(ns foo (:require (foo.specs)))

dpsutton15:06:01

i like that idea

bozhidar16:06:50

@borkdude Please, file a ticket on the style guide repo, so others can see/join the conversation and we’ll decide on the approach to adopt in the next few days.

borkdude16:06:22

Is it ok to redirect the conversation from there to the issue already open, since it already has a few comments?

bozhidar16:06:45

Yeah, that’s fine. I just want everyone watching the repo to see we’re discussing a new rule.

borkdude16:06:14

I’ll make an issue at joker and maybe cursive as well

julio-slack23:06:46

noticing a weird behavior in cider cljs repls: the content of the cider-classpatch doesn't match the lein deps: tree, specifically the cider-classpath reports older version of artifacts (transitive dependency), which is causing compilation errors. any ideas?

julio-slack23:06:24

Specifically: according to lein: [metosin/reitit "0.3.7"] [metosin/reitit-core "0.3.7"] [meta-merge "1.0.0"] according to cider-class path:

julio-slack23:06:45

.m2/repository/meta-merge/meta-merge/0.1.1/meta-merge-0.1.1.jar