Clojurians Log v2

we pondered that long and hard and decided that it was too widespread right now to warn

clj-kondo has a linter for it so people are in many contexts now being warned

the next step may be to create a warning and then eventually an error or some controllable version of either

Right. It appears the clj-kondo linter only catches invocations of, say, enums defined in the Java core libraries, but not custom enums. It does indeed seem to be so prevalent, an approach like *warn-on-reflection*, where you opt into it, might be warranted

Cursive also warns about this now too, and will catch all cases (and offer to unwrap them). It does have a bug I need to fix where it will warn in cases where there is a field and also a static method name, but that should be fixed shortly.

I had a metric ton of cases in the Cursive codebase, so I can imagine it's pretty widespread.

I found it easy to find cases by putting a breakpoint in Compiler.java (and then Cursive makes it trivial to fix them)

I was surprised at how many instances of this were in our code at work, once clj-kondo started warning about it!

@seancorfield if you’ve tried to eliminate all of them might be interesting to put a breakpoint in Conpiler.java and see if you find even more

That's too much effort 😄

Between you and I, this is analogous to incorrect grammar everyone ignores.

> it was too widespread right now to warn In order not to break builds/workflows, I'm assuming? Otherwise I'd consider widespread misuse a rather strong reason to do warn.

It would have been a deterrent to adoption

understandable