Avoid ambiguity of placeholder on nested parameter #7454
Conversation
source/declarations.tex
Outdated
| @@ -1779,7 +1779,8 @@ | |||
| \opt{\grammarterm{type-constraint}} \keyword{auto}. | |||
| The placeholder type shall appear | |||
| as one of the \grammarterm{decl-specifier}{s} in | |||
| the \grammarterm{decl-specifier-seq} or | |||
| the \grammarterm{decl-specifier-seq} or, | |||
| in case of function pointer type, | |||
There was a problem hiding this comment.
I'm not convinced this is a good change. At the very least, reference-to-function types should be included, too.
There was a problem hiding this comment.
I see your point.
I will try to a better way, unless you have suggestions.
There was a problem hiding this comment.
Proposed a new way, not sure whether for p5 it was really needed to repeat the same note.
jensmaurer
added
the
changes requested
Signed-off-by: Mauro Russo <[email protected]>
mrussoLuxoft
force-pushed
the
clarify__dcl_spec_auto_general__3
branch
from
e589a2a to
afc0ac3
Compare
jensmaurer
removed
the
changes requested
|
Hm... I think this entire change is unnecessary. We clearly say "the type of a parameter-declaration of a function declaration" (not "function type"). A nested function type is not a function declaration. We had situations where we said "in", and those were considered ambiguous whether applying to the top level or to nested levels. We believe we've addressed this by using "of", which we believe linguistically excludes nested levels. The standard is not a tutorial. |
|
ok, you are saying that, despite we have 'IN' during the clarifying wording "... The placeholder type shall appear as one of the [decl-specifier]s IN the [decl-specifier-seq] ...", If, in addition, the term 'THE' in the aforementioned wording "... in THE [decl-specifier-seq] ..." is a non-ambiguous reference to the one directly related to the original [parameter-declaration], then ok to me. |
|
For the "IN" decl-specifier-seq, "of" would work well (and be slightly better), but (as you correctly observe), there is no recursion going on via the decl-specifier-seq. For "THE decl-specifier-seq", I think there is none in sight here where "THE" could refer to other than the one of the parameter-declaration. |
This PR is related to the Issue #7453 proposing multiple changes to clarify the text in [dcl.spec.auto.general]
However, a PR has been created for each specific change.
In [dcl.spec.auto.general]-p2 and p5, an additional text about the case of function type, may avoid the general interpretation of using a placeholder type for nested parameters.