Revert write! and writeln! to late drop temporaries

[dtolnay]

Closes (on master, but not on beta) #99684 by reverting the write! and writeln! parts of #96455.

argument position	before #94868	after #94868	after #96455	after this PR	desired (unimplementable)
write!($tmp, "…", …)	⸺late	⸺late	early⸺	⸺late	⸺late
write!(…, "…", $tmp)	⸺late	⸺late	early⸺	⸺late	early⸺
writeln!($tmp, "…", …)	⸺late	⸺late	early⸺	⸺late	⸺late
writeln!(…, "…", $tmp)	⸺late	⸺late	early⸺	⸺late	early⸺
print!("…", $tmp)	⸺late	⸺late	early⸺	early⸺	early⸺
println!("…", $tmp)	early⸺	⸺late	early⸺	early⸺	early⸺
eprint!("…", $tmp)	⸺late	⸺late	early⸺	early⸺	early⸺
eprintln!("…", $tmp)	early⸺	⸺late	early⸺	early⸺	early⸺
panic!("…", $tmp)	early⸺	early⸺	early⸺	early⸺	early⸺

"Late drop" refers to dropping temporaries at the nearest semicolon outside of the macro invocation.

"Early drop" refers to dropping temporaries inside of the macro invocation.

[rustbot]

Hey! It looks like you've submitted a new PR for the library teams!

If this PR contains changes to any rust-lang/rust public library APIs then please comment with @rustbot label +T-libs-api -T-libs to tag it appropriately. If this PR contains changes to any unstable APIs please edit the PR description to add a link to the relevant API Change Proposal or create one if you haven't already. If you're unsure where your change falls no worries, just leave it as is and the reviewer will take a look and make a decision to forward on if necessary.

Examples of T-libs-api changes:

• Stabilizing library features
• Introducing insta-stable changes such as new implementations of existing stable traits on existing stable types
• Introducing new or changing existing unstable library APIs (excluding permanently unstable features / features without a tracking issue)
• Changing public documentation in ways that create new stability guarantees
• Changing observable runtime behavior of library APIs

[rust-highfive]

r? @kennytm

(rust-highfive has picked a reviewer for you, use r? to override)

[Mark-Simulacrum]

r=me, but we may want a T-libs-api FCP here? It seems like a minimal fix to the regressions, not sure if this was already discussed in some capacity. Will leave that up to you, @dtolnay :)

[dtolnay]

I think they'll agree with me that regressing async-std's write macro is worth fixing. I don't think anyone knows any other way to fix it today, without language change, sadly.

There is room for someone to suggest (see table at top) reverting even more of the macros to "late" than necessary, but:

• we already can't do that for println and eprintln, because "early" is better and people have been relying on it in those macros for a long time, see [regression] println borrow check errors #96434;
• doing it for print and eprint does not accomplish better "consistency" than this PR;
• the approved FCP in Make write/print macros eagerly drop temporaries #96455 was kicked off with the comment "I think everything other than the 1st and 3rd rows are an obvious improvement" and that's still the case.

(@rust-lang/libs-api)

@bors r=Mark-Simulacrum

[bors]

📌 Commit ffab6bf has been approved by Mark-Simulacrum

It is now in the queue for this repository.

[joshtriplett]

@dtolnay We discussed it in the meeting yesterday, and yeah, the revert seems reasonable to avoid a regression.