[dcl.init] should this 'declarator' be the 'initializer'8

[Heuristack]

Hi, I got confused when reading [dcl.init]

A declarator can specify an initial value for the identifier being declared.

Should it be like

An initializer can specify an initial value for the identifier being declared.

Noticed that the BNF reflecting the relationship between a declarator and an initializer,

\begin{bnf}
\nontermdef{init-declarator}\br
    declarator initializer\opt
\end{bnf}

FYI, @sdutoit, thanks!

[tkoeppe]

I suppose a "declarator" (English) isn't the same as a "declarator" (grammar production). The former is more general. Although [dcl.dcl]p4 seems to use the term in the narrow sense.

[jensmaurer]

@tkoeppe, I believe you're confused here.

[jensmaurer]

[dcl.decl] p1 clearly distinguishes "declarator" and "initializer", and I fail to see a use of an English-level "declarator" anywhere. (Some of these words are not \grammarterm'ed, but probably should be.)

The note in [dcl.decl] p3 muddies the waters a bit, though.

Back to the original issue: I think this should be "An init-declarator can specify an initial value ..." In this context, an initializer always specifies an initial value, so it doesn't fit the "can" in the sentence.

[Heuristack]

I don't think #1616 is a proper fix for this issue and proposed #1617.

In this context, an initializer always specifies an initial value, so it doesn't fit the "can" in the sentence.

I think "can" here means an initializer is capable of specifying an initial value, and it "can" be used to do that or not, since it is optional (for a declarator).