[std] Update references to [container.requirements.general]

[JohelEGP]

The entirety of [container.requirements.general] is a list of items of the form

• X denotes ...

yet there are many references to this subclause, which probably need updating. For example

strings.tex-\tcode{InputIterator} is a type that qualifies as an input
strings.tex:iterator\iref{container.requirements.general}.
strings.tex-

[jwakely]

N.B. this is a regression. In C++20 all of [container.reqmts] and [container.requirement.pre] and [container.requirements.general] was one big subclause called [container.requirements.general], so the refs were correct.

Looking at it now, I feel like we could have done better with the new subclause names. We currently have this structure under 24.2:

1. Preamble [container.requirements.pre]
2. General containers [container.gen.reqmts]
   2.1 General [container.requirements.general]
   2.2 Containers [container.reqmts]
   2.3 Reversible container requirements [container.rev.reqmts]
   2.4 Optional container requirements [container.opt.reqmts]
   2.5 Allocator-aware container requirements [container.alloc.reqmts]
3. Container data races [container.requirements.dataraces]
4. Sequence containers [sequence.reqmts]

The preamble is fine, and the new structure is fine, but the names could be improved. "24.2.2.1 General [container.requirements.general]" is actually just two paragraphs introducing some placeholder names for the following subclauses.

Maybe this would be better:

1. Preamble [container.requirements.pre]
2. General containers [container.requirements.general]
   2.1 Introduction [container.intro.reqmts]
   2.2 Containers [container.gen.reqmts]
   2.3 Reversible container requirements [container.rev.reqmts]
   2.4 Optional container requirements [container.opt.reqmts]
   2.5 Allocator-aware container requirements [container.alloc.reqmts]
3. Container data races [container.requirements.dataraces]
4. Sequence containers [sequence.reqmts]

This would not have broken the xrefs, because anything referring to [container.requirements.general] (including material outside the standard, like old proposals and Stack Overflow answers, etc.) would still be referring the the entire subclause containing all those requirements. Those refs could be more precise, pointing to one of the new subclauses like [container.alloc.reqmts], but they wouldn't be wrong (as they are now, because they just point to the two paragraphs introducing placeholder names).

[jwakely]

@tkoeppe is it too late to rename those stable tags for C++23?

[tkoeppe]

We're still in the process of submitting the DIS, and I expect it to come back at least once, so I'm happy to add a few more changes for the next round!

[jwakely]

OK, I'll prepare a PR. It might take a few hours to make sure I've done it correctly ...

[tkoeppe]

No worries, I haven't heard back from ISO yet!