[dcl.init.ref] instructs performing direct initialization after applying a user-defined conversion

[RazvanAM]

[dcl.init.ref] paragraph 5, bullet 4, sub-bullet 1 specifies the following:

If T1 or T2 is a class type and T1 is not reference-related to T2, user-defined conversions are
considered using the rules for copy-initialization of an object of type “cv1 T1” by user-defined
conversion (9.4, 12.4.1.4, 12.4.1.5); the program is ill-formed if the corresponding non-reference
copy-initialization would be ill-formed. The result of the call to the conversion function, as
described for the non-reference copy-initialization, is then used to direct-initialize the reference.
For this direct-initialization, user-defined conversions are not considered.

The highlighted wording (which requires performing direct initialization after applying the user-defined conversion) was added in CWG1604. Given that "user-defined conversions are not considered" for the next step of the initialization, I believe that requiring a specific type of initialization does not have any effect on the outcome.

I think there is another issue with the quoted text above, namely:

[...] The result of the call to the conversion function [...]

I believe that the user-defined conversion referred to here could use either a converting constructor or a conversion function, not just a conversion function.

Thank you!

[RazvanAM]

Also, the last sentence from the quoted text:
[...] For this direct-initialization, user-defined conversions are not considered.
seems to be superfluous, as the type that results from [over.match.copy] or [over.match.conv] (henceforth referred to as T3) is:

• T1 or a type derived from T1 (in case T1 is a class type; [over.match.copy]), or
• T1 or a type that can be converted to T1 using a standard conversion sequence (in case T1 is not a class type; [over.match.conv])

Only the first case (T1 is a class type) is of interest (otherwise user-defined conversions are excluded by default in the second stage of the initialization). T3 is reference-related to T1, and [dcl.init.ref] paragraph 5 only considers user-defined conversions in case the reference type and the initializer type are not reference-related. Therefore explicitly disallowing user-defined conversions in the second stage of the reference initialization is not required.

I believe the restriction was unnecessary even considering the state of the standard when the resolution for CWG1604 was proposed (the resolution for CWG1604 added this restriction).

[jensmaurer]

Editorial meeting 2021-04-30: Send to CWG.