OK, that got me a bit further but now I have this failure https://circleci.com/gh/cljdoc/builder/36329 -- paredit
isn't mentioned in my code so I'm not sure what's going on here. Is there a way to tell cljdoc to only do a Clojure parse and ignore the cljs aspect?
Hiya @seancorfield, cljdoc analyzer will analyze for cljs when it finds cljs or cljc files in a library jar. Cljdoc doesn’t currently let library authors specify platforms to analyze, but we do have a few hardcoded overrides for specific projects in cljdoc itself. Without digging in at all that fail on paredit sure seems odd. I’d love to help more but am a somewhat out of commission at the moment.
NP. I’ll create an issue on the cljdoc repo about it so it doesn’t get “lost”. It’s unfortunate that I can’t build the full docs for Expectations at all because of that tho’… 😞
@seancorfield does expectations target cljs for browser and node? it does seem a bit strange to be that it relies on planck in this way
Self-hosted testing via doo
(using planck
) is the intended usage.
It’s not what I would have chosen, but it’s what the person who contributed all the cljs work needed.
That’s why I think it would be ideal to be able to tell cljdoc — only do Clojure analysis for this project.
ok, I can do that later today
@lee actually, do you remember if hardcoded-projects.edn
translates to the new cljdoc-analyzer
setup?
Think it happens here now https://github.com/cljdoc/cljdoc-analyzer/blob/master/resources/config.edn
I’m taking a look at this build https://app.circleci.com/pipelines/github/cljdoc/builder/16909/workflows/64cbb502-fe58-4465-a28d-778c179f96a5/jobs/33284, and in looking at the error thats thrown, I can’t explain why tools.analyzer is throwing a can’t recur error here: https://github.com/protojure/lib/blob/master/src/protojure/internal/grpc/codec/io.clj#L99. I’d very much like to get these docs publishing again, but am not sure how to fix this issue. Any tips?
(would annotating this ns with https://github.com/cljdoc/cljdoc/blob/master/doc/userguide/for-library-authors.adoc#hiding-namespaces--vars-in-documentation cause the analyzer run to bypass the ns? I expect not, but this was the easiest potential work around I could see)
and even the setup I need to run this locally, — i imagine theres some deps.edn file present on the build allowing the clojure -Sdeps "${CLJDOC_ANALYZER_DEP:-$DEP_FALLBACK}" -m cljdoc-analyzer.cljdoc-main "${CLJDOC_ANALYZER_ARGS:-$FALLBACK}"
to run in the container?