[lex.icon] Itemize extended integer choice #6470
Conversation
Eisenwave
force-pushed
the
itemize-extended-integer-choice
branch
from
5c62015
to
5be71e7
Compare
|
Related follow-up CWG issue: cplusplus/CWG#404 |
source/lex.tex
Outdated
| \item If all types in the list for the \grammarterm{integer-literal} | ||
| are unsigned, it is unsigned. | ||
| \item If the list contains both signed and unsigned types, | ||
| it may be signed or unsigned. |
There was a problem hiding this comment.
In order to make the standard as readable as possible, I feel that the antecedent for "it" should be the immediately preceding noun that matches grammar-wise. In this last bullet, it seems that "it" refers to the list. (Those in the know understand that a list is not supposed to be signed or unsigned, but the point of the standard is to convey knowledge, not to assume the reader has knowledge.)
(This is obviously violated in the status quo quite a bit, so this is more an aspirational comment for new edits.)
If you want to avoid the repetition, introduce "T" for the type we're talking about.
Given the editorial overlap with CWG/issues#404, I'm wondering whether the bulletization (and possibly slightly larger massagings) should be left to the CWG issue.
There was a problem hiding this comment.
If you want to avoid the repetition, introduce "T" for the type we're talking about.
Sounds like a good plan, I've done that.
Given the editorial overlap with CWG/issues#404, I'm wondering whether the bulletization (and possibly slightly larger massagings) should be left to the CWG issue.
I've thought about it too, and don't think it's strictly necessary. The CWG issue is just about putting implementation-defined somewhere in the wording. The diff suggested in the core issue makes sense before and after the editorial changes here, so the two are somewhat independent.
Eisenwave
force-pushed
the
itemize-extended-integer-choice
branch
from
5be71e7
to
56306c1
Compare
|
New version with the suggested changes. |
|
OK, leaving this to the resolution of the CWG issue, but I've made a note there to incorporate this editorial change. |
This edit slightly improves readability of the three possible choices for the signedness of the extended integer type.
Furthermore, it conceptually separates the last
If ...sentence from the previous ones, which is good, because otherwise it reads like just another bullet in a list.Furthermore, the last bullet in the list is abbreviated to an
Otherwise,bullet, since the list is exhaustive anyway.