Stream: t-lang

Topic: sealed rust design meeting


nikomatsakis (Mar 23 2020 at 16:05, on Zulip):

Hi there @T-lang and others -- sealed rust design meeting, features @James Munns, starting now (Zoom)

Dropbox paper

boats (Mar 23 2020 at 16:06, on Zulip):

joining soon :)

James Munns (Mar 23 2020 at 17:01, on Zulip):

Hah! I forgot to plug in my machine, and it died right at the end :)

RalfJ (Mar 23 2020 at 17:02, on Zulip):

I think we got the gist of it :D

James Munns (Mar 23 2020 at 17:02, on Zulip):

In case you couldn't hear my last point:

Feel free to reach out if you have ideas. We're coming up with the plan as we go, based on ours and others experiences. If something sounds wrong, let me know, because it's either a bug in our plan, or a bug in our communication about the plan :)

RalfJ (Mar 23 2020 at 17:02, on Zulip):

the thing I was going to say: if the sealed compiler does fewer optimizations, it could work with a simpler operational semantics. like, it might not need Stacked Borrows. that would simplify the spec. but there is also a risk here of forking the language, because fewer code would then be UB in "sealed rust" than is UB in "stable rust", which is not good. no conclusion here, just food for thought. ;)

James Munns (Mar 23 2020 at 17:02, on Zulip):

But, thanks again for having me to chat, happy to follow up whenever.

RalfJ (Mar 23 2020 at 17:04, on Zulip):

@James Munns thanks for being so communicative, what you said sounds great to me :)

nikomatsakis (Mar 23 2020 at 20:16, on Zulip):

@RalfJ interesting thought re: simpler op-sem, I hadn't considered that, although I agree that it does seem like not a clear win

Last update: Jun 05 2020 at 23:10UTC