LWG parts of P2996R13 Reflection for C++26

[eisenwave]

Closes #7930 once the core part is also complete.
Also closes cplusplus/papers#1668.

[jensmaurer]

Forward-looking comment for the library part:

The section labels for the library should be "meta.refl.*" (because shorter without loss of expression) and for the <meta> header, it should be "meta.syn" (because globally consistent where we did not screw up).

Also, "define_aggregate" could maybe become "def.aggr".

[frederick-vs-ja]

I think we should add // see \ref{initializer.list.syn} for consistency. Currently, such comment is added whenever <initializer_list> is included in the synopsis.

[tkoeppe]

@eisenwave Do you already want review feedback, or is this still in "draft"?

[eisenwave]

@tkoeppe you can at FIXMEs and TODOs where I'm unsure how to proceed and give me feedback on those; it's not ready for a full review though. I will progress more towards the end of the week on this.

There's also #7987 if you want to sink a few hours or days into reviewing a PR.

[tkoeppe]

Thanks!

[eisenwave]

Hmm, it has occurred to me that this is a pretty big change.

Everybody expected it and no one is surprised, but when you're 2000 3000+ lines into the diff, it really sinks in ...

[katzdm]

2000 3000+

I was gonna point it out if you didn't..

[jensmaurer]

You need to use \caret for ^ in "usual" source. You can use "^" in codeblocks with no problem.

[jensmaurer]

meta.reflection.operators: Any details we could help with?

[eisenwave]

You need to use \caret for ^ in "usual" source.

Good to know.

You can use "^" in codeblocks with no problem.

No, you cannot. Single carets are fine, but two carets are being treated as a Unicode escape or something, and they seem to have special meaning even inside our codeblock environment. Not sure how to address that.

meta.reflection.operators: Any details we could help with?

No, I got that table to work. The table renders, but it gets put on a separate page, and I'm probably just using the wrong table environment. Not sure which one to pick here.

Anyhow, I'm going to do a linear read-through and fix any obvious mistakes, and then mark this as ready-for-review.

[jensmaurer]

You can use "^" in codeblocks with no problem.

No, you cannot.

Well, there are plenty of uses of ^^ inside codeblocks here, and they work just fine:

https://github.com/cplusplus/draft/pull/8008/files

However, the codeblock environment kicks back into "regular" mode in code comments, so you need to use \caret\caret inside comments.

[eisenwave]

Well, there are plenty of uses of ^^ inside codeblocks here, and they work just fine:

You're right. I think I've had some unbalanced \begin and \end, and it miraculously broke nothing other than double carets, or maybe that's just the first thing that broke. I don't know.

Anyhow, the PR is now ready for review. I have not fixed anything up because I'm under the impression that will only make things harder to review during the first pass.

There are 59 FIXMEs to resolve.

[tkoeppe]

Thank you very much! I've moved both the core and the lib changes into #8008 for further review!