[basic.def.odr] Adds note that x may be part of a qualified id #1996
Conversation
|
Please add the section label to the start of your commit message. Please squash your commits. |
source/basic.tex
Outdated
| @@ -355,7 +355,8 @@ | |||
| \end{itemize} | |||
|
|
|||
| \pnum | |||
| A variable \tcode{x} whose name appears as a | |||
| A variable \tcode{x} whose name appears, possibly as part of a | |||
| \grammarterm{qualified-id}, as a | |||
| potentially-evaluated expression \tcode{ex} is \defn{odr-used} by \tcode{ex} unless | |||
There was a problem hiding this comment.
Would this read better if it said "A variable x whose name appears as a potentially-evaluated expression ex (possibly as part of a qualified-id), is odr-used by ex unless..."?
In general, I'm not convinced this change is a net improvement to start with. In the immediately preceding paragraph, we talk about a "function named by an expression", and qualified-ids are also included there, but not explicitly called out. Calling out qualified-ids in one place, but not another, makes the appearance of a difference where there is none.
There was a problem hiding this comment.
If you have a static variable referenced by Widget::x, is that covered by the term name? A "function named by an expression" imo carries a better implication of optionally being qualified.
I don't expect this pr's proposed wording to be accepted.
There was a problem hiding this comment.
How about this:
A variable
xthat is named by a potentially-evaluated expressionex([expr.prim.id]) is odr-used [...]
This "whose name appears" wording is problematically vague. For example:
int a;
void f(int a) { a = 0 }; // odr-uses ::a?
There was a problem hiding this comment.
I like @zygoloid's suggestion.
ryanhaining
changed the title
ryanhaining
force-pushed
the
master
branch
2 times, most recently
from
384f958
to
53137df
Compare
burblebee
force-pushed
the
master
branch
2 times, most recently
from
bac9908
to
18c6909
Compare
source/basic.tex
Outdated
| A variable \tcode{x} whose name appears as a | ||
| potentially-evaluated expression \tcode{ex} is \defn{odr-used} by \tcode{ex} unless | ||
| A variable \tcode{x} that is named by a potentially-evaluated expression | ||
| \tcode{ex}~(\ref{expr.prim.id}) is \defn{odr-used} by \tcode{ex} unless |
There was a problem hiding this comment.
Can you use the \iref macro here instead?
|
Looks good to me. |
|
I think @zygoloid was still concerned by the potential loss of clarity that it's now no longer clear whether the expression is in its entirety the thing that names |
|
Hmm, perhaps if we switch from passive to active voice:
|
jensmaurer
added
the
needs rebase
tkoeppe
removed
the
needs rebase
zygoloid
force-pushed
the
master
branch
2 times, most recently
from
e3dbfe2
to
1a21a65
Compare
zygoloid
added
the
changes requested
jensmaurer
added
the
needs rebase
|
@ryanhaining Could you please rebase and apply Richard's suggestion? |
|
Yes, sorry you caught me in the middle of two back-to-back trips so might take me a bit, I'm having a really hard time with hotspots so just doing a pull is taking a while. |
|
Okay, ready for review. |
|
I think the rewording of https://wg21.link/p1787r6 improved the text sufficiently to resolve the original issue. Please double-check, and feel free to reopen if you would still like to tweak the wording further. |
Following from https://stackoverflow.com/questions/47952024
I'm sure there's a better way to word this.