[std] Some uses of "Equivalent to `expression`" violate specification style

[eisenwave]

The style guidelines say

For "Equivalent to", a plain expression should appear only if it is of type void; for value-returning functions, use return expression; (see below).

However, this rule is followed very inconsistently. For example, https://eel.is/c++draft/istream.assign#1 uses an equivalent to swap(rhs) in operator=, without return. If you want a more recent example, see yield_value in https://eel.is/c++draft/coro.generator.promise#4

I haven't checked every use to figure out how much is broken, but it doesn't look too bad overall.

[jwakely]

The style guidelines say

For "Equivalent to", a plain expression should appear only if it is of type void; for value-returning functions, use return expression; (see below).

However, this rule is followed very inconsistently. For example, https://eel.is/c++draft/istream.assign#1 uses an equivalent to swap(rhs) in operator=, without return.

There is no bug here, I think you've misunderstood the rule, or the rule doesn't match how the standard is actually written.

Saying return swap(rhs); would be wrong.

We could change it to say "Equivalent to: swap(rhs); return *this; but I see no benefit to doing that. The current wording is fine, and the form is used widely.

If you want a more recent example, see yield_value in https://eel.is/c++draft/coro.generator.promise#4

Again, there's nothing wrong here.

[eisenwave]

There is no bug here, I think you've misunderstood the rule, or the rule doesn't match how the standard is actually written.

It's more plausible to me that the rule doesn't match the standard.

To be clear, I don't feel strongly about this; we should just align the specification guidelines with the standard, whichever side needs changing. You think that the standard is fine, so I guess we ought to update the guideline then.