Stream: t-compiler/wg-nll

Topic: seme-integration


nikomatsakis (May 16 2018 at 14:53, on Zulip):

really feeling torn about how much work to do integrating SEME =)

nikomatsakis (May 16 2018 at 14:54, on Zulip):

it's 'basically done' conceptually

nikomatsakis (May 16 2018 at 14:54, on Zulip):

but I have to juggle a few things to get it integrated

nikomatsakis (May 16 2018 at 14:54, on Zulip):

I guess I will back up and see what is fastest route to get a perf measurement

nikomatsakis (May 16 2018 at 14:54, on Zulip):

I was thinking @pnkfelix that while https://github.com/rust-lang/rust/pull/50593 is a pretty big win, perhaps most interesting is to profile afterwards

nikomatsakis (May 16 2018 at 14:54, on Zulip):

and see what the blockers are

nikomatsakis (May 16 2018 at 14:54, on Zulip):

many of them may still apply

nikomatsakis (May 16 2018 at 14:55, on Zulip):

that is, before doing work to integrate SEME, I suppose it behooves me to analyze where time is going — it may be going into parts that SEME wouldn't help with at all!

nikomatsakis (May 16 2018 at 14:55, on Zulip):

/me feels stupid for not realizing that earlier

nikomatsakis (May 16 2018 at 14:56, on Zulip):

profiling is such a pain in the neck though

Last update: Nov 21 2019 at 13:10UTC