Fork me on GitHub
#clojure-dev
<
2015-10-09
>
bronsa15:10:08

@alexmiller: I just realized I didn't document the new :load-ns opt to deftype/defrecord in my patch for CLj-1208

bronsa15:10:15

should I open a new ticket or reopen that one?

alexmiller15:10:25

you mean docstring?

alexmiller15:10:44

new ticket - I will just pull it up into 1.8 as that seems important

alexmiller15:10:13

I'd also welcome a para for the http://clojure.org/datatypes page too

bronsa15:10:33

ok -- I'm not on my dev laptop rn so I can't make a patch, I'll just open a ticket and handle that tomorrow

fdserr16:10:10

Hi, there. Is it known/okay that & gets silently bound in this case:

(defprotocol Foo
  (foo [this & b]))
=> Foo
(deftype T []
  Foo
  (foo [o & b] [o & b]))
=> user.T
(foo (->T) 1 2)
=> [#object[user.T 0x7e01f15e "[email protected]"] 1 2]
Working on warning the user cljs-side http://dev.clojure.org/jira/browse/CLJS-1445

alexmiller16:10:05

There is a clj ticket for this too

alexmiller16:10:34

It is known and not ok imo :)

bronsa16:10:04

@alexmiller: can you close CLJ-1773? (it's just marked as resolved)

fdserr16:10:31

@alexmiller: will search for the ticket, thanks. BTW, I don't get why it'd be bad to be able to have variadics in protocols (one day)?

alexmiller16:10:06

protocols create Java interfaces as well as Clojure functions and that creates some complications

alexmiller16:10:25

potentially resolvable complications

fdserr16:10:41

alright; do you see use cases come often?

alexmiller16:10:59

it comes up occasionally

alexmiller16:10:03

I guess 1661 is the one. there's some weird history some of which is on the mailing list. I marked it triaged but that ticket needs some work though.

fdserr16:10:25

It is not evident if someone is onto it, I'll be checking maybe, missing Java a bit ^^

alexmiller16:10:53

the ticket description needs work to better capture the problem and approach

alexmiller16:10:57

feel free to do so

alexmiller16:10:11

prob good to check whether the patch still applies too

bronsa16:10:43

@alexmiller: tbh I never understood the position clojure has wrt java varargs. there is bytecode annotation for varargs so clojure could handle java varargs automatically rather than as a single arg (and in the same vein, compile vararg protocol methods into java varargs)

bronsa16:10:55

unless there's some impl detail I'm forgetting

alexmiller16:10:20

I've never talked to Rich about it, but my guess is that all those ends were just never tied down

fdserr16:10:48

Well, a warning would be cool until the odds are settled =]

stuartsierra18:10:21

@bronsa: By the way, what's your status with regard to tools.reader releases? I put the "alpha" label on tools.namespace 0.3.0-alpha1 because it depends on tools.reader 0.10.0-alpha3

bronsa18:10:37

@stuartsierra: I just forgot to release 0.10.0

bronsa18:10:49

thanks, I'm launching the hudson job now

stuartsierra18:10:06

@bronsa: hah! No worries, thanks simple_smile

bronsa18:10:03

0.10.0 is released simple_smile