Remove the unhelpful let binding diag comes from FormatArguments

[chenyukang]

Fixes #114374

[rustbot]

r? @fee1-dead

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

[fee1-dead]

Not entirely sure if this is what should happen.

r? compiler

[b-naber]

What if we have nested format_args expressions?

Do we only care about the case where we assign the result of format_args!? If so, why can't we just check for stmt being an assignment where the rhs expr is_format_args_item?

[chenyukang]

the if self.span == expr.span.source_callsite() guarantee we find out the expr we're intrested in, the code goes here means a borrow(mostly binding) happens, I think use span to find expr is mostly enough for suggestions.

[b-naber]

What does your branch output on something like this: let x = format_args!("a {} {} {}.", 1, format_args!("b{}!", 2), 3)? Can you maybe add this to your test?

[chenyukang]

The output is:

error[E0716]: temporary value dropped while borrowed
 --> tests/ui/borrowck/issue-114374-invalid-help-fmt-args.rs:5:13
  |
5 |     let x = format_args!("a {} {} {}.", 1, format_args!("b{}!", 2), 3);
  |             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^- temporary value is freed at the end of this statement
  |             |
  |             creates a temporary value which is freed while still in use
6 |     bar(x);
  |         - borrow later used here
  |
  = note: the result of `format_args!` can only be assigned directly if no placeholders in it's arguments are used
  = note: to learn more, visit <https://doc.rust-lang.org/std/macro.format_args.html>
  = note: this error originates in the macro `format_args` (in Nightly builds, run with -Z macro-backtrace for more info)

I added it to the test case.

[b-naber]

If we already special-case for FormatArguments, can't we also provide a more helpful diagnostic, which explains why this fails, instead of just getting rid of the binding suggestion? Note that we also still have this note: = note: consider using a let binding to create a longer lived value, so it might make sense to special case this.

[chenyukang]

Good point!
I add a special note for format_args!, I'm not sure whether the words are accurate enough, seems it's hard to explain 😂 , so I added a link to the document.

[b-naber]

I'm sorry that I didn't respond. I missed the notifications for your replies, so thought you hadn't replied.

@bors r+ rollup

[bors]

📌 Commit 1f107b1 has been approved by b-naber

It is now in the queue for this repository.