Fork me on GitHub
#cljs-dev
<
2016-12-22
>
ambrosebs00:12:13

does (when (seq e) ) get optimised?

ambrosebs00:12:17

running some tests on my changes, but there's a preliminary diff if anyone has spare eyes https://github.com/frenchy64/clojurescript/pull/3/files

ambrosebs02:12:24

pprint-table-tests fails on Spidermonkey on master for me. Is that expected?

thheller05:12:42

;;; str
[], (str 1), 1000000 runs, 96 msecs
[], (str nil), 1000000 runs, 80 msecs
[], (str "1"), 1000000 runs, 86 msecs
[], (str "1" "2"), 1000000 runs, 123 msecs
[], (str "1" "2" "3"), 1000000 runs, 160 msecs

;;; bstr
[], (bstr/str 1), 1000000 runs, 0 msecs
[], (bstr/str nil), 1000000 runs, 0 msecs
[], (bstr/str "1"), 1000000 runs, 0 msecs
[], (bstr/str "1" "2"), 1000000 runs, 0 msecs
[], (bstr/str "1" "2" "3"), 1000000 runs, 0 msecs

thheller05:12:54

I think I win at this benchmark

thheller05:12:45

@dnolen I want to create a new JIRA issue to fix a glaring oversight with regards to str performance but there is not benchmark for it yet

thheller05:12:03

should I create a seperate patch for the benchmark itself so it can be run against the "old" impl

thheller05:12:14

or just bundle the benchmark with the new patch?

thheller05:12:02

(the new benchmark is 0 because it is dead code eliminated, while the current str impl never is)

thheller05:12:17

;; str
[], (str 1), 1000000 runs, 55 msecs
[], (str nil), 1000000 runs, 39 msecs
[], (str "1"), 1000000 runs, 44 msecs
[], (str "1" "2"), 1000000 runs, 257 msecs
[], (str "1" "2" "3"), 1000000 runs, 334 msecs

;;; bstr
[], (bstr/str 1), 1000000 runs, 13 msecs
[], (bstr/str nil), 1000000 runs, 8 msecs
[], (bstr/str "1"), 1000000 runs, 11 msecs
[], (bstr/str "1" "2"), 1000000 runs, 173 msecs
[], (bstr/str "1" "2" "3"), 1000000 runs, 210 msecs

thheller05:12:29

for :simple on node.js

thheller05:12:34

still a quite big gain

thheller06:12:15

nvm I guess, the patch is so straightforward .. no need to make it any more complicated

dnolen14:12:17

@ambrosebs failing under SM is not a high priority (if you’re using a nightly or something)

ambrosebs15:12:10

@dnolen actually it was Nashorn that failed http://pastebin.com/CQkHLh67

ambrosebs15:12:17

doh, I don't think I actually tested with master. Sorry, still finding my feet.

ambrosebs15:12:17

yea, no it was the latest master.

dnolen16:12:11

@ambrosebs Nashorn we care about even less

dnolen16:12:18

it has tons of problems

dnolen16:12:37

(wrt. JS specification, standard behavior)

ambrosebs16:12:16

so I seem to have installed the 2 least useful envs 😛

ambrosebs16:12:01

@dnolen here's the patch for some? and ^boolean tags in the analyzer http://dev.clojure.org/jira/browse/CLJS-1878

r0man17:12:02

I'm trying out the latest closure compiler with ClojureScript. It's available on Maven. But where do I get the org.clojure/google-closure-library dependency from? Is this repo public?

r0man17:12:40

Or how would I build this artifact with the latest Closure library myself?

r0man17:12:23

Nevermind, I think I found it: script/closure-library-release/closure-library-release.sh