cljdoc

https://cljdoc.org/ & https://github.com/cljdoc/cljdoc
martinklepsch 2020-10-14T09:03:28.069200Z

@miikka good morning, nice to see the sourcehut thing in production 🙂 what do you think about stripping the leading ~ from the username in the SCM link text? Since it’s not something users choose I feel it’s non-essential and almost like a detail related to sourcehut that isn’t very relevant

martinklepsch 2020-10-15T11:35:50.092600Z

That differentiation doesnt exist in the same consistent way across the different scm providers, which makes me feel it shouldn’t “leak” into cljdoc?

martinklepsch 2020-10-15T11:41:00.093500Z

But idk, at the same time sourcehut repos aren’t going to be many for the foreseeable future so if sourcehut users consider the extra character useful I’m not gonna get in the way :)

dominicm 2020-10-15T11:43:19.093700Z

Drew considers it significant. +cljdoc and ~cljdoc are completely separate.

1đź‘Ť
miikka 2020-10-18T07:48:18.094900Z

Ah, I didn't know about this group syntax plan, but in that case I'd say let's keep the extra character in

1đź‘Ť
lread 2020-10-18T12:29:40.096500Z

For the unaware, like me before googling, Drew is Drew DeVault, the person who created sourcehut.

martinklepsch 2020-10-19T10:42:12.097200Z

@dominicm @miikka let’s keep it in then :)

1đź‘Ť
miikka 2020-10-14T09:33:37.069500Z

:man-shrugging: I thought about it and decided to keep it since sourcehut keeps it everywhere, but if you feel like removing it is better, that's fine by me

timo 2020-10-14T11:48:18.070100Z

Hi there

lread 2020-10-14T11:48:43.070500Z

Hi @timo!

timo 2020-10-14T11:49:20.071100Z

Is it possible to configure cljdoc? I need to avoid cljs analysis on cljc files in my project

timo 2020-10-14T11:49:36.071400Z

it throws errors and I need to postpone that

lread 2020-10-14T11:51:03.071900Z

We don’t currently have a way to tell cljdoc to ignore cljs files.

lread 2020-10-14T11:52:49.072900Z

Are the throws due to something cljdoc might be doing wrong?

timo 2020-10-14T11:53:36.073600Z

I am not certain, but the error might be on my side and it errors still if I put ^:no-doc on top

martinklepsch 2020-10-14T11:54:18.074800Z

We can hardcode an exception if it’s not something that can be fixed on the side of the lib

martinklepsch 2020-10-14T11:55:06.076100Z

Does the file work with the Clojure script Compiler otherwise?

timo 2020-10-14T11:57:08.076800Z

we are working on cljs-compat right now so it might be broken for cljs currently, but that's beyond me

timo 2020-10-14T11:57:21.077100Z

it's datahike

timo 2020-10-14T11:57:50.077700Z

I would like to add only the api-ns

timo 2020-10-14T11:58:22.078300Z

so I add no-doc everywhere and released a snapshot

lread 2020-10-14T11:59:23.079600Z

no-doc just controls what parts of your api get exposed. Cljdoc will still load your entire project.

martinklepsch 2020-10-14T12:00:31.080900Z

Yeah, and if a namespace required by the api ns can’t be loaded then that will also cause the api ns to fail to load

lread 2020-10-14T12:01:52.082200Z

I wonder… have you thought about maybe keeping cljs work in a branch until it is ready for release?

lread 2020-10-14T12:02:18.082400Z

Or does that not work?

lread 2020-10-14T12:03:20.082900Z

Yeah I guess not. If you have cljc files. You are clj ready but not cljs ready… hmm…

timo 2020-10-14T12:06:33.084400Z

yeah right

lread 2020-10-14T12:06:41.084800Z

Personally, I think I’d just push through and get cljs side working before release. But that’s me. And maybe not what works for you?

timo 2020-10-14T12:06:46.085Z

as it comes from datascript mostly it was working once

timo 2020-10-14T12:07:09.085400Z

that's actually not what I am working on

timo 2020-10-14T12:07:18.085700Z

and I think it is a lot of work

timo 2020-10-14T12:07:39.086Z

so I just want to add the api-ns as docs

timo 2020-10-14T12:07:45.086200Z

but I can find another way

timo 2020-10-14T12:09:26.086400Z

thanks anyway

lread 2020-10-14T12:15:23.086500Z

I think both arguments are valid. (How helpful of me! :simple_smile:). We are linking to sourcehut and that’s the syntax they use for a repo. We are linking from cljdoc and its not the syntax it has used for all other repos.

miikka 2020-10-14T12:17:55.086700Z

I admit it does look a bit weird

lread 2020-10-14T12:19:24.086900Z

On coin toss things like this, I usually just go with what @martinklepsch prefers. :simple_smile:

martinklepsch 2020-10-14T14:13:57.088500Z

We can totally hardcode it if it makes your life easier, there’s a file that explicitly supports this kind of hard coding so it wouldn’t have any extra maintenance overhead

timo 2020-10-14T15:46:18.089200Z

ok perfect. that would be cool for now. it will take a few weeks until clojurescript is restored I guess.

timo 2020-10-15T15:47:07.094400Z

cool! thanks a lot. that version needs a few more no-docs though. but never mind

martinklepsch 2020-10-14T15:50:42.089500Z

the hardcoding works based on group/artifact ID — which ones should cljs get disabled for?

timo 2020-10-14T15:55:10.089700Z

io.replikativ/datahike

1đź‘Ť
martinklepsch 2020-10-14T16:20:19.089900Z

Is there an issue to track cljs compatibility?

dominicm 2020-10-14T19:46:34.090600Z

Fwiw, in the future there will be a syntax for groups which differs

dominicm 2020-10-14T19:48:11.090800Z

~ for users and something else like + for groups. So it's useful in a sense for distinguishing.