Fork me on GitHub
#cljdoc
<
2018-10-15
>
justalanm06:10:50

@martinklepsch looks like it is possible to retrieve these libs https://dev.clojure.org/jira/browse/TDEPS-98?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&amp;focusedCommentId=50308#comment-50308 unfortunately I'm behind a proxy at the moment so I can't test it reliably

martinklepsch10:10:25

Yeah I saw Alexโ€™s updates and will take another look today. I was surprised though since we both seemed to reproduce this fairly reliably IIRC

justalanm10:10:00

The same for me ๐Ÿ™‚

martinklepsch15:10:07

Issue is within cljdoc ๐Ÿ˜“

justalanm15:10:04

No worries, the important thing is that we were able to track it down

dawran616:10:23

Hi everyone, I'd like to work on https://github.com/cljdoc/cljdoc/issues/32 . What I learned so far is the server code is using pedestal to build the urls. Is it recommended to get familiar with the library before jumping into the issue?

martinklepsch17:10:34

Hey Daw-Ran, happy to see you'd like to contribute ๐Ÿ™‚ For this issue the routing part actually isn't that important I would say but you should have a basic understanding of how interceptors work with pedestal: http://pedestal.io/reference/interceptors

martinklepsch17:10:31

It can also help to take a look at Pedestal's default not-found interceptor: https://github.com/pedestal/pedestal/blob/09dd88c4ce7f89c7fbb7a398077eb970b3785d2d/service/src/io/pedestal/http.clj#L93 and how it is used in Pedestal itself

martinklepsch17:10:02

You might notice that you can also pass a custom not found interceptor to Pedestal's default-interceptors function (it's in the same namespace).

martinklepsch17:10:30

We use this function in cljdoc so a good start might be to get up and running and make use of the option to supply a custom not found interceptor.

martinklepsch17:10:23

Would be happy to work through it together if you get stuck. Just let me know โœŒ๏ธ

martinklepsch17:10:46

oh and thanks again for the documentation fix ๐Ÿ˜Š

dawran618:10:26

Awesome! Appreciate your guidance. I'll start by understanding the interceptors. ๐Ÿ™‚

wilkerlucio19:10:41

@martinklepsch hello, I just had an idea, do you think would be feasable to add specs information to cljdoc? for example in Pathom I have a good number of functions that have specs for it (using s/fdef), I think could be useful to present these spec information with the docs for the fns, not sure how hard to get that in

martinklepsch20:10:07

This would be excellent of course. There are still some issues with the current Clojure.spec implementation that make this nontrivial unfortunately https://github.com/cljdoc/cljdoc/issues/67

martinklepsch20:10:07

I think we could fork clojure.spec and somehow enable this but itโ€™s obviously not a very nice solution

wilkerlucio20:10:18

yeah, I actually have done some inspection work in the past, but its not super reliable, I put some in a lib: https://github.com/wilkerlucio/spec-inspec it might be broken at this point

wilkerlucio20:10:42

but it did some helpful things, like proper follow back the attribute definition chain (when you make a spec pointing to another one)

martinklepsch20:10:19

The issue is that we need to be able to tell which artifact a spec belongs to in order to store it in a database conflict free

wilkerlucio20:10:03

yeah, and I can imagine a few complication cases, for example, spec lets (by design) you add specs to things that you don't own

martinklepsch20:10:06

Yeah thatโ€™s useful although I guess we would have to implement those things separately anyways since our data wouldnโ€™t be in a runtime ๐Ÿ™‚

wilkerlucio20:10:15

so, imagine that I get a lib and add a spec for a fn of a third part library in my library

wilkerlucio20:10:40

then the library adds its own spec for it, then how to know which one to keep?

wilkerlucio20:10:30

actually, keep from the one that owns it, hehe, if you can easely detect it ๐Ÿ˜›

martinklepsch20:10:44

Since we only depend on the artifact under analysis you could (optimistically) assume that none of its dependencies modify this kind of Stuff

martinklepsch20:10:26

I need to update this issue though, i feel like itโ€™s not properly describing the issue very well

martinklepsch20:10:50

But yeah, specs would be huge and we could do all kinds of cool stuff with them

wilkerlucio20:10:31

yeah, it's exciting to think about the possibilities ๐Ÿ™‚