cljdoc

https://cljdoc.org/ & https://github.com/cljdoc/cljdoc
seancorfield 2019-06-18T02:17:27.059300Z

@martinklepsch I happened to notice that one of my next.jdbc namespaces doesn't make it into the nav tree on the left (because it has no vars) https://cljdoc.org/d/seancorfield/next.jdbc/1.0.0/api/next.jdbc.connection -- I wondered if there was an option in cljdoc via metadata to force inclusion of a namespace?

martinklepsch 2019-06-18T09:08:16.060200Z

The decision of not showing "empty" namespaces in the navigation was made ad-hoc and could be revisited. Do you have any thoughts on that general decision? @seancorfield

borkdude 2019-06-18T11:17:25.061Z

@seancorfield you can preview cljdoc docs on your own machine (e.g. via Docker): https://github.com/cljdoc/cljdoc/blob/master/doc/running-cljdoc-locally.md#running-locally-with-docker this prevented surprises before I deployed my libs a few times

seancorfield 2019-06-18T14:46:49.062Z

@borkdude yeah, no surprises. I was just curious about the decision.

šŸ‘ 1
seancorfield 2019-06-18T14:48:42.064200Z

@martinklepsch it's a perfectly reasonable default. A metadata option on the ns to override it would be nice. Not a big deal. I definitely think "empty" namespaces are an edge case.

borkdude 2019-06-18T14:50:21.064700Z

Iā€™m also curious why you would want a namespace without vars to be documented. what is the reason for its existence?

seancorfield 2019-06-18T16:50:08.065500Z

@borkdude It has implementations of protocols so it's "important" from that point of view but the only docstring in it, I think, is on the ns itself.

seancorfield 2019-06-18T16:51:07.066200Z

(it also contains a bunch of private vars and functions that support those implementations)

borkdude 2019-06-18T17:25:54.066400Z

makes sense