Fork me on GitHub
#cljs-dev
<
2016-01-01
>
mfikes06:01:53

At first blush, it looks like realized? can be made to work for lazy seqs by simply checking if fn is nil. I may dig further to see if this really holds true or if I’m missing something subtle. (Patch: https://gist.github.com/mfikes/ef1b535ab41d1e706a5c)

mfikes13:01:58

I’ll try a simple patch for realized? for discussion. If it is no good, we can reject / decline it. http://dev.clojure.org/jira/browse/CLJS-1531

mfikes13:01:02

@anmonteiro: A guess: It makes things consistent with Clojure, which can have only 20 positional arguments.

anmonteiro13:01:12

@mfikes: that makes sense, but I was wondering if there's also any performance implication

anmonteiro13:01:35

another example: David defined several arities in Om's update-state! (https://github.com/omcljs/om/blob/master/src/main/om/next.cljs#L551-L565)

dnolen17:01:41

@anmonteiro: yes arities can be directly dispatched to

mfikes21:01:10

@dnolen A few core predicates like iterable? and cloneable? delegate directly to satisfies? and lack ^boolean, while others like counted? and indexed? are hinted. Is there an interest in patches for these, or would it be (premature optimization) noise you'd rather not deal with?

dnolen22:01:27

@mfikes: patches welcome!