Talk:cpp/feature test

Could it be possible to link to the individual papers that introduced or bumped a given feature-test macro? Right now there's this page, which has the macros and the values (but not the papers), or pages like https://en.cppreference.com/w/cpp/compiler_support/23 which have the paper but not the macros.

Sounds like a good idea, but good lord would that be a lot of work --Ybab321 (talk) 12:21, 5 December 2022 (PST)

I hope this is quite doable. In fact, it is a bit surprising that that was not done from the very beginning. A good starting point could be this page (and SD-6), — replicate that data in these tables.

We should decide the position of Paper(s) column:

1) | Macro name | Feature | Value | Header | Std | Paper(s) | (i.e. at the end)

or

2) | Macro name | Paper(s) | Feature | Value | Header | Std | (i.e. second column)

or_else.) --Space Mission (talk) 16:40, 5 December 2022 (PST)

I'm indifferent, but if I had to vote, I would go with the former --Ybab321 (talk) 10:00, 6 December 2022 (PST)

Yes, and it is slightly easier to add stddoc to the end of a table's string. So, WIP.) --Space Mission (talk) 10:58, 6 December 2022 (PST)

The attribute-token order on this page has no_unique_address in the sequencing as:

maybe_unused

no_unique_address

nodiscard

noreturn

This matches the "_" handling in other names in other tables.

But the Compiler Features Dump code has the ordering:

COMPILER_ATTRIBUTE_ENTRY(maybe_unused)

COMPILER_ATTRIBUTE_ENTRY(nodiscard)

COMPILER_ATTRIBUTE_ENTRY(noreturn)

COMPILER_ATTRIBUTE_ENTRY(no_unique_address)

not matching the tables. In order to use join on the output, I locally moved the no_unique_address line in the code.

Marklmi (talk) 09:44, 7 April 2023 (PDT)