Remove do

[bytbox]

Fixes #10815.

[bytbox]

Wheee!

This needs rebasing of course, but with so many files affected, it'll need rebasing every 12 hours anyway, so I figured I'd go ahead and submit. Any major issues? (It does compile and stuff, so I'm mainly asking about the documentation, which I had to mangle a bit - see 03f9eb1)

[emberian]

Awesome! Looks good to me, r+ with rebase

[bytbox]

Rebased.

[bytbox]

Hmmm... looks like it tried to pull and test an older version than I thought was in this pull request.

[alexcrichton]

I think that a commit snuck in after this was rebased and before it was merged that introduced some do-notation.

[bytbox]

Yeah, that's it: e901c4c. Rebasing and fixing.

[bytbox]

Ok, it's probably good now - running tests on my box, but that's about 4 times as slow as bors, so if you want to optimistically tell bors to retry, it should work.

[brson]

Extreme kudos! 🏄

[bytbox]

@alexcrichton Bors seems to have missed your comment? (Or I just don't understand how it works, but the dashboard isn't updating for the PR.)

[huonw]

@bytbox looks like yet another instance of the confusion caused by github ordering commits chronologically, rather than topologically (that is, the HEAD commit wasn't listed last, and bors only looks at HEAD).

Silly GitHub! :(

[alexcrichton]

Oops, sorry @bytbox! I should have been watching the queue and saw that this didn't go through...

[pnkfelix]

Isn't this going to have a better chance of eventually landing if we do it piecemeal? I.e. instead of trying to atomically land the removal-of-usages and the removal-of-support-in-rustc in one PR, first get in as many removal-of-usages as possible in one PR and do the removal of support-in-rustc in a separate one?

Maybe people really are concerned about do's sneaking in during the lag time between separate PRs. Anyway, just a thought.

[huonw]

Can we just try to deal with that via review? Also the most recent build bootstrapped but failed on android for reasons that may be fixed by #11903.

(Also, I'd guess that people very rarely add dos now that it is specific to procs.)

[bytbox]

Did another rebase/test cycle and confirmed that this is still safe to merge, so (assuming #11903 did in fact fix the android build issue) somebody should poke bors I guess. @alexcrichton