search

cplusplus / draft

C++ standards drafts
http://www.open-std.org/jtc1/sc22/wg21/
5.62k stars 740 forks source link

Stop talking about fields #5544

Open jwakely opened 2 years ago

jwakely commented 2 years ago

[intro.memory] has a note that says "It is not safe to concurrently update two bit-fields in the same struct if all fields between them are also bit-fields of nonzero width." The term fields is not defined in the standard.

[support.srcloc.class.general] has a comment saying "source location field access" which is novel in the library clauses. We could say "access functions" or "member access" or just get rid of the comment. It has very little value. Either way, I don't think it needs to include "source location" in that comment or the one above the constructors. We know which class we're looking at.

[mem.res.pool.options] says "specified in this field" twice, which should be member.

[time.cal.ymd.overview] and adjacent subclauses uses "field-based time point" which I think means the mathematical meaning of the word, i.e. a set that supports arithmetic. It's arguable if this actually helps.

[time.hms.overview] says "The individual duration fields always return non-negative durations". I'm not sure what this means. It seems to be referring to the non-static data members h, m, and s (which have types that are specializations of duration) but what does it mean for them to "return" something? Does it mean the corresponding member functions? This could be improved.

All uses of "field" in iostreams, locales, and std::format are OK, it's talking about widths and padding of text fields, which seems fine. This isn't confusing, and matches the use of "field" in ISO C.

[diff.cpp20.utilities] says "bit fields" which should be hyphenated.

[diff.dcl] refers to "field initializers" which isn't defined in ISO C or ISO C++.

jwakely commented 2 years ago

Oh and a footnote that says "Thus, the valarray template is neither a matrix class nor a field class." This is also the mathematical meaning.

C++98 was much better, it only had one anachronistic field: "The result of applying the offsetof macro to a field that is a static data member or a function member is undefined." That wording is no longer present, but we've introduced more fields elsewhere.

Eisenwave commented 11 months ago

At the time of writing, the status is

Done

Not done

jwakely commented 11 months ago

[time.cal.ymd.overview] and adjacent subclauses uses "field-based time point" which I think means the mathematical meaning of the word, i.e. a set that supports arithmetic. It's arguable if this actually helps.

This is actually using the terminology explained in https://howardhinnant.github.io/date/date.html#Overview which comes from N3344. But we shouldn't be using that terminology without defining it or adding a bibliography reference.