[intro.memory] What is a "field"?

[morinmorin]

In [intro.memory] (the C++ memory model), there are several occurrences of the word "field":

• p4: if all fields between them are also bit-fields of non-zero width.
• p5: The field a and bit-fields d and e.ee are each separate memory locations, …

While that word is a popular terminology in programming (and in fact it is used in diagnostic messages in some C++ compilers), it is not a defined term in C++.

So would it be better to add a note saying '"field" means a non-static data member'?

[morinmorin]

c.f.
Since the language C is so simple, these wordings are modified to use "member" in C11:

[intro.memory]

• p4
  • C++17: a bit-field and an adjacent non-bit-field are in separate memory locations,
  • C11: A bit-field and an adjacent non-bit-field member are in separate memory locations.
• p4
  • C++17: if they are separated by a non-bit-field declaration.
  • C11: if they are separated by a non-bit-field member declaration.
• p4
  • C++17: if all fields between them are also bit-fields of non-zero width.
  • C11: if all members declared between them are also (non-zero-length) bit-fields,
• p5
  • C++17: The field a and bit-fields d and e.ee are each separate memory locations,
  • C11: The member a, and bit-fields d and e.ee are each separate memory locations,

[jensmaurer]

I agree this is an issue to be fixed, but the proposed fixes don't quite work, because C++ has static data members, and unnamed bit-fields are not non-static data members.

[jensmaurer]

Ok, this refers to two mentions of "field" in notes, not in normative text, so we can be a bit more sloppy with using common, but not-otherwise-defined terms. p5 can easily use "member" (it's referring to a specific non-static data member), but p4 could either:

• just say "member", which is correct, but does not represent the entire rule (but the note is phrased as if it were)
• or say "if all non-static data members and unnamed bit-fields between them are bit-fields of nonzero width." which sounds a bit confused, but is complete and correct.

[morinmorin]

unnamed bit-fields are not non-static data members.

I didn't know that, thanks for the correction!

this refers to two mentions of "field" in notes, not in normative text, so we can be a bit more sloppy with using common, but not-otherwise-defined terms.

Also, thanks for the explanation; now I understand the situation.

p4 could either:

• just say "member", which is correct, but does not represent the entire rule (but the note is phrased as if it were)
• or say "if all non-static data members and unnamed bit-fields between them are bit-fields of nonzero width." which sounds a bit confused, but is complete and correct.

Or put "and unnamed bit-fields" in parentheses to alleviate lengthy appearance slightly?

• …, if all non-static data members (and unnamed bit-fields) between them are bit-fields of nonzero width.