[std] might and could: re-review for complicated cases

[jensmaurer]

[meta.rqmts] p5

that instantiation would yield a different result if performed in a context where T were a complete type

"yield a different result" -> "the validity or meaning of the instantiation depends on the completeness of T"; run by CWG / LWG.

[time.zone.info.sys]

A correct offset for the \tcode{time_zone} off daylight saving time
can be obtained by querying the \tcode{time_zone} with a \tcode{time_point}
that returns a \tcode{sys_info} where \tcode{save == 0min}.
There is no guarantee
what \tcode{time_point} (if any) returns such a \tcode{sys_info}
except that it is guaranteed not to be in the range \range{begin}{end}
(if \tcode{save != 0min} for this \tcode{sys_info}).

[tkoeppe]

[basic]: "A literal type is one for which an object can possibly be created within a constant expression."

[tkoeppe]

[expr]: "Historically, lvalues and rvalues were so-called because they can appear on the left- and right-hand side of an assignment" -- this is borderline incorrect; we should look for a better way to explain this historic perspective.

[tkoeppe]

[expr]: "and can therefore be defined as deleted." This sounds uncomfortably ambiguous with the meaning of "and therefore you the user can define it as deleted". The previous wording wasn't quite so readily interpretable that way. Maybe extra words like "and can therefor end up as being defined as deleted".

[tkoeppe]

Placeholder topic for noteworthy "good rewordings".

---

"These special member functions are implicitly defined as usual, and might therefore be defined as deleted." => "... as usual, which can result in them being defined as deleted."

---