that's odd, sounds like some kind of corrupted cache, could you try rm -rf .lsp/.cache and try again?

Now, it worked! Thank you for your help.😀

Good, I need to debug further later why sometimes it stuck like that

What do you mean by test documentation?

Well i need to document the tests i have in my project. As to gain an overview of what business logic is being tested. Instead of having to update a document everytime a test is written, i thought it might be a start to generate this documentation using some framework. Maybe adding some metadata to every test and thereby keeping the documentation close to the code and easy to update

Similar to Javadoc, I imagine?

Yes exactly

Maybe one of the tools discussed here could help you: http://clojure-doc.org/articles/ecosystem/generating_documentation.html

yes, although why not do them both in the same require?

just for organization, as I have many namespaces that are only for ns keywords, not code namespaces

I can have all in one require, this was a nice to have if it was possible to organize the requires that are for code from the ones that are for ns-kw

You may find some linters object to multiple :require clauses.

you can always just put a (require …) call outside fo the ns

thanks all for the answers, clj-kondo doesn’t seem to complain, good to know it works

call a protocol method on it

Feel free to copy https://github.com/nedap/speced.def/blob/70fc9c7afce9d93677eb5bf639d66b8e7233f969/src/nedap/speced/def/impl/satisfies.cljc

in general, I think many uses of satisfies? are a smell that you are putting conditionality outside the protocol instead of inside a default protocol implementation

I agree, generally I only use it for specs/assertions or such, not for dispatch logic

can a protocol specify a default implementation?

i know multimethods can but a protocol?

Yes, indirectly. From https://clojure.org/reference/protocols: > To define a default implementation of protocol (for other than nil) just use Object

or alternately by wrapping protocol method calls in functions and handling these default in some other way

@U2FRKM4TW that Object impl will take precedence before any metadata based ones so that will effectively disable extension by meta

Huh, indeed, I didn't know about that.

Ah, wait, I tested incorrectly. What you wrote seems to be wrong:

=> (defprotocol P :extend-via-metadata true (a [_]))
P
=> (extend-protocol P Object (a [_] (println "a from Object")))
nil
=> (a (with-meta {} {`a (fn [_] (println "a from metadata"))}))
a from metadata
nil

Per https://clojure.org/reference/protocols, "Protocol implementations are checked first for direct definitions (defrecord, deftype, reify), then metadata definitions, then external extensions (extend, extend-type, extend-protocol)."

extensions to Object are the last category (external)

Hm it seems I read that wrong

strings are mapped OK. Clojure uses long for integer literals. in the library there is a type cast to the sql type and it fails because long is bigger than Integer

using Apache Calcite. I wrote a small library to interface with clojure. the above example is to demonstrate how to use it on q clojure datastructure.

I can write sql to filter data from that table

will make an announcement this week