[dcl.stc] Define mutable subobject and make use of keyword consistent #3901
Conversation
|
Please use \keyword, not \tcode, when referring to keywords. That also helps the index. |
| @@ -6938,7 +6938,7 @@ | |||
| \item | |||
| an invocation of an implicitly-defined copy/move constructor or | |||
| copy/move assignment operator | |||
| for a union whose active member (if any) is mutable, | |||
| for a union whose active member (if any) is \keyword{mutable}, | |||
There was a problem hiding this comment.
We say "const" (no special decoration), so I think we shouldn't make this case a \keyword. Unless we say "declared \keword{mutable}" or so.
There was a problem hiding this comment.
The motivation for this was to keep the distinction between objects and members (i,e. a mutable member corresponds to a mutable subobject).
| A member subobject\iref{intro.object} corresponding to a non-static | ||
| data member declared with the \keyword{mutable} specifier is a \defnadj{mutable}{subobject}. |
There was a problem hiding this comment.
Should "mutable subobject" be transitive? That is, should a (non-const) subobject of a mutable subobject also be a mutable subobject? The use of "mutable subobject" in [basic.type.qualifier]/1.1 in the definition of "const object" appears to assume transitivity -- but maybe it means "or a non-mutable direct subobject of a const object"
|
I would prefer that we define what it means for a non-static data member to be "mutable" (no |
zygoloid
added
the
changes requested
wg21bot
added
the
needs rebase
We use the term "mutable subobject" to refer to objects corresponding to members declared
mutable, but we don't define this anywhere. This defines it in [dcl.stc], and additionally changes adds \tcode to "mutable" when referring to a member declaredmutable, since it's a keyword in those contexts (this is used for the majority of references to members declaredmutable).