Fork me on GitHub
#cljdoc
<
2019-02-28
>
martinklepsch18:02:36

@jayzawrotny whoops, just saw this again, sorry for the delay 😅

martinklepsch18:02:36

@jayzawrotny it looks like you're building docs for a leiningen template, which is not really supported very well

jaide19:02:31

Ah ok, thought that may be the case and as I thought about it the less sense it made to generate docs at all. The README is probably just fine, thanks for looking into it.

dominicm19:02:57

I suppose a list of files or directories to ignore would solve this?

dominicm19:02:01

Oh actually

dominicm19:02:12

I might have a better fix

dominicm19:02:44

hmm, you're not using tools.namespace directly?

dominicm19:02:08

Sorry, that was more directed at martin & cljdoc

jaide21:02:25

oh! My mistake sorry

martinklepsch19:02:12

at dinner now will reply later, we use codox under the hood which may use tools namespace, not 100% sure

dominicm19:02:44

:cljdoc.api/namespaces could one use this to perform the appropriate filtering perhaps? currently it's restricted to just the hardcoded things though 🙂

dominicm19:02:50

I can't see how namespaces are found normally

dominicm19:02:57

maybe analysis is the wrong place

dominicm19:02:17

now I know where to look though 🙂

dominicm19:02:52

https://github.com/weavejester/codox/blob/59cafeffcadc37d0eae0973011977e3916ddadbb/codox/src/codox/reader/clojure.clj#L7 looks like it does use it yeah, okay. So you're basically running into TNS-45. There is a JUXT fork of tns which applies the patches there, but it requires a bit of annoying work (globally excluding tools.namespace). Doing that would fix the problem though. You could probably monkey patch tns from cljdoc if you wanted.

dominicm19:02:02

https://dev.clojure.org/jira/browse/TNS-45 tns-45 is the bane of my existence 🙂

dominicm19:02:41

I have considered going on a crusade like nrepl did with a forked version of tns which fixes this. I think tns has more users though, so might be harder, and it doesn't seem to bother that many people?