Type-check `break value;` even outside of `loop {}`.

[kennytm]

Fix #43162, fix #43727.

[rust-highfive]

r? @arielb1

(rust_highfive has picked a reviewer for you, use r? to override)

[rust-highfive]

r? @arielb1

(rust_highfive has picked a reviewer for you, use r? to override)

[kennytm]

(The break rust; thing is following #43162 (comment). It can be reverted if considered inappropriate.)

[kennytm]

(The break rust; thing is following #43162 (comment). It can be reverted if considered inappropriate.)

[zackmdavis]

It can be reverted if considered inappropriate.

I don't think the type-checker of a production compiler is a good place for easter eggs.

(For comparison, Python has import this and from __future__ import braces, but it feels significant that those are tucked away in the standard library, rather than affecting the semantics of the language itself.)

[zackmdavis]

It can be reverted if considered inappropriate.

I don't think the type-checker of a production compiler is a good place for easter eggs.

(For comparison, Python has import this and from __future__ import braces, but it feels significant that those are tucked away in the standard library, rather than affecting the semantics of the language itself.)

[kennytm]

@zachreizner Not saying the feature should stay (I may delete those in 1 or 2 days r+'ed 🎊)... but the change does not affect the semantics of the language itself, since entering that branch already indicates the code is wrong and is going to fail soon anyway (which is also the reason why I've added an extra test case to run-pass to ensure normal code is not affected).

[kennytm]

@zachreizner Not saying the feature should stay (I may delete those in 1 or 2 days r+'ed 🎊)... but the change does not affect the semantics of the language itself, since entering that branch already indicates the code is wrong and is going to fail soon anyway (which is also the reason why I've added an extra test case to run-pass to ensure normal code is not affected).

[arielb1]

Generally r=me, but we need signoff from @aturon for all new features.

[arielb1]

Generally r=me, but we need signoff from @aturon for all new features.

[arielb1]

I think we could also go with signoff from our release manager, @brson

[arielb1]

I think we could also go with signoff from our release manager, @brson

[aturon]

@bors: r+

for as long as we can get away with it

[aturon]

@bors: r+

for as long as we can get away with it

[bors]

📌 Commit 3cb23a7 has been approved by aturon

[bors]

📌 Commit 3cb23a7 has been approved by aturon

[killercup]

This is awesome. I'll wait for the first nightly that includes this and pin all my projects to it!

[killercup]

This is awesome. I'll wait for the first nightly that includes this and pin all my projects to it!

[jD91mZM2]

Does this only work outside of loops? Or is it a breaking change?

[jD91mZM2]

Does this only work outside of loops? Or is it a breaking change?

[bors]

⌛️ Testing commit 3cb23a7 with merge 9868352...

[bors]

⌛️ Testing commit 3cb23a7 with merge 9868352...

[bors]

☀️ Test successful - status-appveyor, status-travis
Approved by: aturon
Pushing 9868352 to master...

[bors]

☀️ Test successful - status-appveyor, status-travis
Approved by: aturon
Pushing 9868352 to master...