Fork me on GitHub
#shadow-cljs
<
2019-02-15
>
thheller09:02:12

you gotta be kidding me ... what year is this 馃槢

thheller09:02:12

> MongoDB's JIRA is currently down for maintenance.

馃槥 5
anmonteiro13:02:23

@thheller I opened https://github.com/thheller/shadow-cljs/pull/439 earlier today. Happy to discuss it here

anmonteiro13:02:29

no pressure to review it though

thheller13:02:51

@anmonteiro looks straightforward. will merge later. the entire goog loader stuff has changed a lot and doesn't work at all anymore with the new closure-library release. we'll see what else needs changing 馃槈

anmonteiro13:02:22

this one is present currently in the old goog release though

anmonteiro13:02:30

we鈥檙e on 1.10.516

thheller13:02:23

yeah I don't know why I even exposed those methods. probably never tested them.

thheller13:02:44

in all my experiements the only thing that made an actual difference in initial loading time was <link rel="preload" ...>

thheller13:02:34

you can always interact with the mm instance directly if you need

thheller13:02:50

(.prefetchModule shadow.loader/mm "foo") should work

anmonteiro13:02:31

would that be amenable to advanced compilation?

anmonteiro13:02:58

I didn鈥檛 do that because they鈥檙e not exported through 鈥渆xportSymbol鈥

thheller13:02:27

from where are you calling the functions? from inside code that was :advanced compiled or from HTML?

anmonteiro13:02:58

I was calling it from the HTML but I won鈥檛 need to do it

mhuebert15:02:06

Random question about externs inference: if there is a (.val ^js snapshot) in the codebase, does that mean 鈥渧al鈥 will never be renamed anywhere, or only for things that the compiler thinks might have the same type as snapshot?

mhuebert15:02:22

My recollection is that it would be universal

thheller15:02:14

it depends. If Closure can identify the type it can be renamed, if not it won't be. so generally in CLJS code nothing will ever be renamed if the name exists in the externs

mhuebert18:02:04

ok. so the ^js hint there does not mean that val will be added to the inferred externs

thheller19:02:02

no, it will be added