Repository navigation
P2615 Meaningful exports CWG2443 #1236
Description
Activity
This was discussed during the EWG 6/23/2022 Telecon, and the following poll was taken:
EWG Approves of the direction of Core's proposed solution to CWG2443 as a Defect Report.
SF F N A SA 7 6 0 0 0 Result: Consensus
This will be forwarded to CWG, pending feedback from a not-present-for-the-discussion implementer.
CWG telecon 2022-07-01: Reviewed; needs a re-review due to a late small wording update.
An objection was raised against removing the normative requirement "An exported declaration that is not a module-import-declaration shall declare at least one name." As agreed with the EWG chair, the concern will be discussed in the upcoming EWG telecon.
P2615R0 Meaningful exports (Davis Herring)
- changed the title
[-]CWG2443 Meaningless template exports[/-][+]P2615R Meaningful exports CWG2443[/+]on Jul 2, 2022 - changed the title
[-]P2615R Meaningful exports CWG2443[/-][+]P2615 Meaningful exports CWG2443[/+]on Jul 2, 2022 Time + Chair error resulted in this not getting seen at the 7/7 EWG telecon, so I'll ensure it is scheduled 1st thing on the 7/21 meeting.
CWG telecon 2022-07-15: Approved P2615R0 pending EWG approval.
Re-discussed during the EWG 7/21/2022 Telecon, and the following poll was (re)taken:
SF F N A SA 1 3 4 2 2 Result: Not Consensus
Due to the procedural concerns, this is replacing the previous vote on this topic. During the meeting, a replacement proposal was promised in the next mailing or so to solve this core issue.
3 remaining items
Davis tells us he needs to change the paper to implement EWG guidance.
We need to address a few more cases:
static_assert(...); // as discussed, accept asm("..."); // as discussed, accept template void f(); // as discussed, reject template<> void f<int>(); // as discussed, reject A() -> A<0>; // treat like template and reject? ; // treat like static_assert and accept? [[...]]; // treat like static_assert and accept?
The potentially attributed null statement may have the 'export' keyword, despite it being meaningless.
SF F N A SA 5 1 1 1 1 Result: Consensus
A deduction guide declaration inside of an
export {}should be permitted, despite not having meaning.SF F N A SA 7 2 0 0 1 Result: Consensus
exporton a deduction guide declaration (no {}) should be permitted, despite not having meaning.SF F N A SA 6 2 1 0 2 Result: Consensus
In its 2022-05-20 telecon, CWG agreed with the wording for CWG2443 as shown here:
https://wiki.edg.com/pub/Wg21telecons2022/CoreWorkingGroup/cwg2443.html
Allowing (almost) any kind of declaration in the declaration-seq of an export-declaration is a design change that needs EWG approval.