warn toolchain on different platform

[frbimo]

Fixes
#2520

Testing
cargo test

[frbimo]

@kinnison kindly reminder*

[kinnison]

There's only one tiny niggle around what I feel might be a wasted clone() otherwise this looks excellent. I'm especially grateful you added a test for this.

[frbimo]

@kinnison ptal

[kinnison]

This looks good to me. One final chance for @rbtcollins to have a look in case the approach or wording isn't to his taste, otherwise I'll look to merge it fairly soon.

[rbtcollins]

Should this ask interactively perhaps?

…

On Wed, 28 Oct 2020, 15:54 Daniel Silverstone, ***@***.***> wrote: ***@***.**** approved this pull request. This looks good to me. One final chance for @rbtcollins <https://github.com/rbtcollins> to have a look in case the approach or wording isn't to his taste, otherwise I'll look to merge it fairly soon. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2534 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AADZ7XUJ3TR2MZT7MPT7FGTSNAWD3ANCNFSM4S2TJAZQ> .

[kinnison]

I'm loathe to add an interactive prompt to a part of rustup which has never been interactive before now. If we think interactivity makes sense then we have to detect if there's a tty on stdin and stuff like that, otherwise it'd be a pain for anyone who has automated this in CI or similar. Also we simply don't know for sure so warning non-interactively is probably enough.

[kinnison]

I rebased to clean up the conflict and am minded to merge this as I remain of the opinion that interactive prompting doesn't seem ideal to me.

[kinnison]

The CI badness is snapcraft being odd.