Allow while let chains on all editions

[est31]

PR #132833 has stabilized if and while let chains on edition 2024, with the reasoning that editions before 2024 have bad scoping wrt the else clause (#124085).

This argument applies to if let chains, but while doesn't require the rescoping behaviour of edition 2024. So we can actually stabilize it on all editions.

See also: #140198

[rustbot]

r? @BoxyUwU

rustbot has assigned @BoxyUwU.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[est31]

Note: if this is merged, one needs to adjust rust-lang/edition-guide#337 . If this is NOT merged, one needs to adjust rust-lang/reference#1740 on the other hand, as there is only an "Edition differences" notice for if chains.

[BoxyUwU]

This needs a lang FCP right?

[est31]

Yes, a FCP is needed or at least a statement from them that no FCP is needed :). @traviscross has nominated #140198, maybe it makes sense to discuss this one in the same meeting as well?

[joshtriplett]

@est31 If you're on an older edition, is there currently a clear error message if you try to use let chains, that tells you it's available in the Rust 2024 edition and newer?

That would help make sure people know what works in what edition, rather than just getting a parse error.

[tmandry]

@rfcbot fcp merge

[rfcbot]

Team member @tmandry has proposed to merge this. The next step is review by the rest of the tagged team members:

• @joshtriplett
• @nikomatsakis
• @scottmcm
• @tmandry
• @traviscross

Concerns:

• bugs (Allow while let chains on all editions #140204 (comment))

Once a majority of reviewers approve (and at most 2 approvals are outstanding), this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up!

cc @rust-lang/lang-advisors: FCP proposed for lang, please feel free to register concerns.
See this document for info about what commands tagged team members can give me.

[traviscross]

@rfcbot reviewed

[rfcbot]

🔔 This is now entering its final comment period, as per the review above. 🔔

[est31]

@joshtriplett

If you're on an older edition, is there currently a clear error message if you try to use let chains, that tells you it's available in the Rust 2024 edition and newer?

The error message says that the feature is unstable. Which is suboptimal of course. The way that #132833 was implemented is that it removed the gating for let chains on newer editions, so the feature is still there and unstable. Back when I filed the PR, this was a necessity as the compiler hadn't been on 2024 yet (and it makes heavy use of let chains). Now that it is on 2024, it's a different situation.

Once clippy has updated their rustc to one with the stabilization, and we know precisely which syntax is restricted to newer editions (which is what this PR and #140198 are about), I think we can introduce a proper error like there is for async await, remove the feature gate, and mark the feature as stabilized. One might even be able to use the machinery added by #115677, not sure. It might not make it into 1.88.0 any more but I think that's not a big problem.

[est31]

Hmmm doing a check, the bugs are still there with while let chains:

• Broken MIR ICE with let chains #104843 with -Z lint-mir and s/if/while/ ICEs on 2021, passes on 2024
• Temporary does not live long enough in let chains #103476 (comment) with s/if/while/ errors on 2021, passes on 2024

So it turns out that it's not as simple as I've thought... we'll probably need to add this as a concern to the FCP or maybe even cancel. I'm sorry for the back and forth.

As good news though, if I modify the examples to use match and if let guards, they don't reproduce the error/ICE on edition 2021, so we won't need #140198 I guess?

[traviscross]

@rfcbot concern bugs

Let us know what you find there with more investigation.

[bors]

☔ The latest upstream changes (presumably #140324) made this pull request unmergeable. Please resolve the merge conflicts.