[container.requirements] Simplify Returns specification for try_emplace #7892
Conversation
|
@jwakely , since we already say we do as-if |
|
Why not just Equivalent to: Returns ... except that ... ? |
Eisenwave
force-pushed
the
emplace-hint-returns
branch
from
cb056ae to
806be71
Compare
|
@jwakely that seems better, yes. I've also deleted a bunch of the adjacent specifications because those are inherited when "Equivalent to" is specified. Let me know if that's correct. |
Eisenwave
force-pushed
the
emplace-hint-returns
branch
from
806be71 to
132931e
Compare
There was a problem hiding this comment.
I think the \result elements should remain, those are still useful (and [structure.specifications] p4 does not say that Effects: Equivalent to implies the Result: element).
I prefer the original form of "is a hint ... Implementations are permitted to ignore the hint." rather than "is an ignorable hint..."
Eisenwave
force-pushed
the
emplace-hint-returns
branch
from
132931e to
4c7ff44
Compare
source/containers.tex
Outdated
| @@ -3233,15 +3233,10 @@ | |||
|
|
|||
| \pnum | |||
| \effects | |||
| Equivalent to \tcode{a.emplace(std::forward<Args>(args)...)}, | |||
| Equivalent to: \tcode{return a.emplace(std::forward<Args>(args)...).first}, | |||
There was a problem hiding this comment.
I've just realised that this doesn't work, because a_uniq.emplace and a_eq.emplace return different things. It's not a pair for a_eq, so you can't do .first
So maybe what you had before was better, sorry.
There was a problem hiding this comment.
I must not have been paying attention either; my bad. The latest commit should resolve everything.
Eisenwave
force-pushed
the
emplace-hint-returns
branch
from
4c7ff44 to
a4e485e
Compare
Fixes #7881.