Someone here wanted cursive to remove unused imports, but I'm having this problem when not using type hints which is always:

Only when I import .getTime can be resolved, but then the import is unused.

@yonatanel the code uses now reflection. You can add (set! *warn-on-reflection* true) to get warnings of these. Add ^Date hint to the latter and it’s all good (into the let after a-date.

Sorry for the clojure question, but can I do (.getTime ^Date a-date)?

what about

(let [^Date a-date (:date m)]
  (.getTime a-date))

No problem. I'm curious, technically, is (.getTime ^Date (:date m)) possible?

(.getTime ^java.util.Date (java.util.Date.))
=> 1479372774370

yes 🙂

In (some-> m :date (.getTime)) getTime is not resolved but in (some-> m (:date) (.getTime)) it is. :thinking_face:

From http://clojure.org/guides/threading_macros: A bare symbol or keyword without parentheses is interpreted as a simple function invocation with a single argument.

that looks like a cursive bug, I don't think either of those is resolvable

Why not? It works

@yonatanel It might get guesses by Cursive, but the clojure compiler will still do reflection

so would your original snippet, but it would be slow

There is no chance for clojure to know that the map field :date is a Date

Enable *warn-on-reflection* and you'll see

@rauh You are right about reflection, but it should be resolved.

"not resolvable" warnings in cursive don't indicate "this won't work", they indicate "I don't know if this will work or not but it will be slow"

I see. So both should not be resolved...

well, I don't know if cursive aims to guarantee that it will always tell you something will be slow, but it would be useful if it does

anyone else having performance problems with cursive + intellij community these days?

something in my setup seems to be leaking memory and the longer I have a repl running the slower the whole ide becomes and eventually it runs out of java memory

the only recent change to my workflow was that I started using datomic

which could also be the culprit

@imre I actually saw that the other day. What it seemed to be was something holding the head of a seq in the nREPL client code.

I have some heap dumps but haven’t had time to investigate properly.

@bfabry Cursive does aim to do that, but has some bugs currently. In particular, it doesn’t do the right thing in a total absence of type hints (should warn always, but that perhaps should be a configurable strict mode).

It’s very close to being totally accurate, though.

@cfleming let me know if I can send you anything that could help tracking the issue down

Is there any way to replace a common keyword namespace on all occurrences, for example when rearranging clojure specs?

@cfleming greetings! is this possible in IDEA? https://twitter.com/mraleph/status/799284624115568641

@yonatanel You can rename a namespace alias, is that what you mean? So you can rename str/trim to string/trim or ::str/keyword` to ::string/keyword

Currently you can only rename the last segment of a namespace name, because renaming other segments involves moving things around which is more complicated, and not implemented yet.

@misha I don’t think so, no. You can use proportional fonts, but I believe the vertical space of every line has to be equal.

That would drive me nuts, personally - nothing would ever align.

also getting periodic repl slowdowns