[libc++][format] Adds parser std-format-spec.

[Mordante]

Summary

This implements the generic std.format.spec framework for all types.

The Unicode support will be added in a separate patch.

Implements parts of:

• P0645 Text Formatting

Completes:

• LWG-3242 std::format: missing rules for arg-id in width and precision
• P1892 Extended locale-specific presentation specifiers for std::format

[Mordante]

Reupload, forgot to set the dependencies so CI failed.

[Mordante]

Updates the paper status.
Fixes no-except and no-localization unit tests.

[Mordante]

Rebase.
Mark unit test requiring concepts.
Minor cleanup.

[Mordante]

Tabs -> spaces.

[vitaut]

libcxx/include/__format/parser_std_format_spec.h

132

This should really be a grapheme cluster but {fmt} uses a code point at the moment (grapheme clusterization will be added later). In any case this should be a range (e.g. string_view) and not a single code unit.

138

This will have to change for the same reason.

184–185

This is incorrect because format specs can consist of a single sign specifier. Same for other "End of input" errors.

196–201

I don't think you need to have this check because it will be caught elsewhere anyway. Format specifiers are type-specific and it generally makes more sense to say what is supported rather than what is not. Same for other __parser_no_* cases.

277–279

This is redundant and only adds an overhead without substantially improving diagnostics. The error message is not particularly correct either.

307–308

It might be better to just return a large value and let the error be reported elsewhere (when the ID is actually used). It's both more efficient and gives better diagnostic without referring to internal magic limits.

321

with -> width

373

Note that any nonnegative int32_t value is a valid precision so I'm not sure __number_max is OK here.

[vitaut]

libcxx/include/__format/parser_std_format_spec.h

421

Looks like a stray "without".

471

This is a wrong place to do such validation because format specs may not have }.

551

Did you mean std-format-spec?

556–558

It does now.

582

This isn't correct. std-format-spec doesn't have to be terminated by anything. formatter::parse can be used to parse standalone format specs. } is part of replacement-field and should be handled there.