[dcl.init.ref] References are bound to objects and functions, not expressions

[sdkrystian]

There are several occurrences in the wording saying that references are bound to expressions, which is incorrect:

[dcl.init.ref]/5.1 sentence 1

[...] the reference is bound to the initializer expression lvalue in the first case and to the lvalue result of the conversion in the second case

[dcl.init.ref]/5.3 sentence 3

[...] the reference is bound to the resulting glvalue

[dcl.init.ref]/5.4.2

The temporary materialization conversion is applied and the reference is bound to the result

Obviously, the reference should be bound to the object or function that the expression refers to, so this wording is technically incorrect.

Proposed fix

In each case where a reference is said to bind to an expression, it will be replaced with "the object or reference to which the initializer refers", or a variant thereof.

[cpplearner]

I don't think the standard has a definition for "bound to". I usually understand it as a synonym of "initialized with".

[sdkrystian]

@cpplearner I understand it as "set to refer to". The standard does use it all over the place when referring to references, so it may be useful to define it.

[sdkrystian]

@cpplearner Here's some wording to clarify for what bound means, perhaps @zygoloid could look at it to verify that its good.

A reference is said to be bound to an object or function E when it is initialized to refer to E.

[RealLitb]

I think if this is changed as proposed, the definition of "binds directly" should also be changed to be in terms of objects and functions. Just so that if "binds directly to X" is true, then "binds to X" is true aswell.

EDIT: to clarify - we say "the argument expression binds directly" in ranking certain overload resolution scenarios. I don't propose changing any text, just pointing out this.