`expect` messages should follow our guidance

[fee1-dead]

It's very likely that this turns into a LLM honeypot issue, but...

#96033 added official guidance on messages for .expect four years ago. Many .expect messages we write were either before the guidance existed or did not know of this guidance.

I like the reasoning on the guidance, and I think we should be consistent. So here's a list of files containing user-facing doc .expect messages that can be fixed (E-easy):

• library/alloc/src/boxed.rs (Update expect messages in library/alloc/boxed.rs and library/alloc/string.rs to follow the style guide #159882)

• library/alloc/src/string.rs (Update expect messages in library/alloc/boxed.rs and library/alloc/string.rs to follow the style guide #159882)

• library/alloc/src/collections/binary_heap/mod.rs (I'm just not liking the "why is the test harness" message) (@SzilvasiPeter)

• library/alloc/src/collections/vec_deque/mod.rs

• library/alloc/src/ffi/c_str.rs (@asder8215)

• library/alloc/src/vec/mod.rs

• library/core/src/primitive_docs.rs (@pluiee)

• library/core/src/result.rs (@pluiee)

• library/core/src/fmt/mod.rs

• library/core/src/ptr/non_null.rs

• Not listed: std, etc.

Please do not claim this issue, comment below and claim a subdirectory to work on would make it possible for multiple people to work on :D

What is more interesting is that there are many pre-existing .expect messages in our library code that use the expect message wrongly, e.g. Poll::Ready(self.0.take().expect("Ready polled after completion")) in impl<T> Future for Ready<T>, it would be nice if our messages change to be better (to say something like "Ready should not be polled after completion" instead), but that is more user facing and takes more time to decide IMO:

• TODO: list out user-facing expect messages

[ChrisDenton]

it would be a theoretical breaking change to change these panic messages

Panic messages, like debug output, are not covered by our stability guarantee. Anyone depending on precise output should expect to be broken.

[pluiee]

@rustbot claim

[fee1-dead]

@pluiee please specify a smaller scope that you want to make PRs on so that this issue can be worked on by multiple people concurrently :)