[clang] [clang][docs] open details of C++{17, 14, 11} implementation by default (PR #193141)

[Corentin Jabot via cfe-commits]

cor3ntin wrote:

@AaronBallman I'm not even sure what we are talking about

- There are core issues we should support as DR
- There are core issue resolved through papers
- There are core issues that CWG decided were design-y in nature and were later solved, or made irrelevant by a paper
- There are papers approved as DR that do not solve core issues
- There are papers approved as DR that solve multiple or many issues.
- There are papers approved as DR that solve multiple or many issues and also change things that are not the subject of a core issue - which is the case for P0588.

I don't think "well, we are never gonna implement all of that" is a good criterion for not listing papers.

So yes, i think it makes sense to keep tracking papers, especially knowing that a paper often arise when core decides an issue is too complicated or impactful to be listed by the normal issue process.

The functional consequences of P1787 are fairly minor - it was designed as a wording cleanup exercise, even though
it does affect lookup in unusual cases (and it a lot of unrelated changes). (I do think we have a bunch of issues that have been or would be solved by implementing parts of P1787)

However the consequences of not implementing P0588 is that our lambda implementation is not conforming and we have a bunch of user-reported issues because people keep having issue with our lambda implementation. Hiding that information is not helping anyone.



https://github.com/llvm/llvm-project/pull/193141