Fork me on GitHub
#cljs-dev
<
2015-08-29
>
crisptrutski14:08:45

@dnolen: welcome back! saw you’re back in release cutting mode - would you mind taking a look at http://dev.clojure.org/jira/browse/CLJS-1435?

dnolen14:08:28

@crisptrutski: please don’t mark self-host tickets as critical

dnolen14:08:37

nothing around self-hosting is considered critical

crisptrutski14:08:55

my thinking was that with things like planck and http://clojurescript.net, and such a generic issue with self hosted - quite possible for new comers to run into

crisptrutski14:08:38

but guess it’s up to those ambassadors to warn about the sharp edges simple_smile

dnolen15:08:35

@crisptrutski: self hosting is a completely new thing, issues are expected, few people are using this stuff for production work

dnolen15:08:42

it’s not critical, and won’t be for a long time

dnolen15:08:32

@crisptrutski: the patch for CLJS-1400 is interesting but it doesn’t understand the intent of the original patch which was perf.

crisptrutski15:08:10

potentially just the :shadow value is required with name?

dnolen15:08:21

System/identityHashCode is fast. So we need some similar efficient way to get a some what unique identifier for the symbol.

crisptrutski15:08:23

would a combination of those two be sufficient?

dnolen15:08:01

I think shadow is a map, you don’t want to hash a collection

dnolen15:08:09

but that idea is headed in the right direction

dnolen15:08:28

maybe hash-combine the hash and shadow depth

crisptrutski15:08:38

ok thanks for the tips, i’ll try it out simple_smile

crisptrutski15:08:18

off for now, but updated the patch