[temp.spec] Make "specialization" a proper definition & adjust definition. #3248
Conversation
|
|
Yeah, I had trouble thinking of a way to describe aliases, since instantiation produces no entity, rather, replacement occurs. Another option would be to stomp out all references to "alias specializations", and replace them with something to the effect of "template-id where the template-name denotes an alias template", or something similar. |
jensmaurer
added
the
changes requested
|
The Travis CI check failed. Maybe there's a space at the end of a line or so. Please fix that. |
source/templates.tex
Outdated
| instantiated or explicitly specialized\iref{temp.expl.spec}. | ||
| A \defn{specialization} is a class, variable, function, or class member that is either | ||
| instantiated\iref{temp.inst} from a template or produced by an | ||
| explicit specialization\iref{temp.expl.spec} of a template. |
There was a problem hiding this comment.
I am having second thoughts about "produced by ...". In fact, an explicit specialization is a class etc. right there; it doesn't obviously "produce" one. (Oh, and "produce" sounds awkward anyway.)
maybe "... or an explicit specialization of a template."
@zygoloid ?
There was a problem hiding this comment.
Hmm, are the "from a template" additions correct? For cases like:
template<typename T> struct X {
void f() {}
};
... is X<int>::f() instantiated from a template, or from a member of a template?
There was a problem hiding this comment.
Also, "is either instantiated from a template or an explicit specialization of a template" is missing a verb. Or what an explicit specialization? (If this is "or is an explicit specialization", what does that mean for members of templates:
template<typename T> struct Y {
struct Z {};
};
template<> struct Y<int>::Z {};
template<> struct Y<float> { struct Z {}; };
... which of Y<int>::Z and Y<float>::Z are specializations? Do we really mean that the former is a specialization and the latter is not?)
There was a problem hiding this comment.
@zygoloid I think both of these could be addressed by saying:
A specialization is a class, variable, function, or class member that is either instantiated from a template, or a member thereof, or is an explicit specialization of a template, or member thereof.
This seems a little clunky though.
There was a problem hiding this comment.
Correct, but the variable itself is not
There was a problem hiding this comment.
Well, it was certainly intended that variable templates are also templated entities. Maybe we should simply say that a variable is an entity, then.
There was a problem hiding this comment.
A variable template is a template though, so it is a templates entity anyways.
There was a problem hiding this comment.
So, we're good with "templated entity", and all we need is an updated pull request.
jensmaurer
removed
the
changes requested
Specialization is the second most common term that is defined and used in [temp]. The definition in [temp.spec] p4 should be italicized to make it a proper definition. Also, the phrase " function, or class member that is [...] explicitly specialized" is incorrect: templates are explicitly specialized, not classes, functions etc (in some cases non-templates too, but rarely).
Additionally, a specialization of a alias template (such as the use here: http://eel.is/c++draft/temp.alias#2) is meaningless for non-class types. Consider the following:
A specialization is a class, function, variable, or class member. The type denoted (if you could even say that, since its replaced) by the simple-template-id
A<void>is none of these.