Hello. So I think this should be possible to do for the 1.34 release (beta cut on 2019-02-28, release on 2019-04-11). I'll post some more here later.
First some notes on the current open issues. As far as I can tell all blocking issues are marked either P-High or NLL-Sound of those issues the current state is:
|Issue number||Blocks migrate||Assigned To||Notes|
|#55748||Yes||felix||Maybe fixed by #55937? This doesn't appear to be an issue on nightly.|
|#56254||Yes||matthew||to me, needs a decision|
|#57170||Yes||matthew||Fixed by #57202 - but I'm trying a new (cleaner) approach|
|#54105||?||niko||Is this user visible?|
|#54940||?||niko||Obscure, no easy fix|
|#46901||No||felix||We should have some idea what we're allowing for 2 phase borrows|
|#54943||No?||matthew||I will try to fix as part of #46702|
|#56993||No||niko||I will try to fix as part of #46702|
|#57374||No||lqd/niko||Not an issue in migrate mode|
The current state on the 2018 edition is maybe not quite what people think it is. On the 2018 edition:
-Zborrowck=astdoes not override either of these
I propose that we do the same thing on the 2015 edition, meaning that we will always run the NLL borrow checker and always use two-phase borrows.
Since a lot will change from that I would like to split the work across PRs for the following
I've reassigned #54943 to you.
A few questions/comments, @Matthew Jasper:
I just planned to do them together, although thinking about it some more, I'm not sure if they need to be.
- why do you put "no" in terms of blocks migrate for those two issues? Basically because 2018 code is already exposed?
They don't affect migrate mode.
Basically I suspect they are non-issues, but I think I would treat them mildly differently --
I'd like to chime in and say I'm super happy to see NLL come to 2015. Thanks for your work! :heart:
(I took the liberty of allocating an issue number, #57804, for this worthwhile task.) Thank you for looking into this @Matthew Jasper !