[expr.prim.req.type] Clarify example comment re: validity of name only #4457
Conversation
source/expressions.tex
Outdated
| typename S<T>; // required valid~(\ref{temp.names}) \grammarterm{template-id}; | ||
| // fails if substitution failure occurs for the second template parameter of \tcode{S} |
There was a problem hiding this comment.
Thanks for the cross-reference.
"substitution failure" feels too unspecific here, because it could mean substitution failure while instantiating all of S<T, X> ("substitution for" is not really "substitution into"). What we mean is substitution failure into the template-head. Suggestion, a little more tailored to the example:
"fails if \tocde{T} has no nested type \tcode{type}"
There was a problem hiding this comment.
"substitution failure" feels too unspecific here, because it could mean substitution failure while instantiating all of
S<T, X>("substitution for" is not really "substitution into").
Okay, I can change it to say "substitution into".
What we mean is substitution failure into the template-head. Suggestion, a little more tailored to the example:
"fails if \tocde{T} has no nested type \tcode{type}"
Well, it also fails if 0 cannot be implicitly converted to T::type.
|
(needs squashing) |
The comment that a class template specialization is required can be taken as meaning that more is required than the validity of the template-id (which is all that the requirement really checks). Additionally, the specific implications of the validity requirement is spelled out.
hubert-reinterpretcast
force-pushed
the
better_concept_type_req_example_comment
branch
from
81152b2
to
bb1fbd8
Compare
The comment that a class template specialization is required can be taken as meaning that more is required than the validity of the template-id (which is all that the requirement really checks).