Issue 4288: The Constraints: element in [time.clock.system.members] is probably wrong

Title: The Constraints: element in [time.clock.system.members] is probably wrong
Status: new
Section: [time.clock.system.members]
Submitter: Jiang An

Created on 2025-06-19.00:00:00 last changed 3 days ago

Messages

msg14898 (view)

Date: 2025-07-15.15:24:16

Proposed resolution:

This wording is relative to N5008.

Modify [time.clock.system.members] as indicated:
```
using system_clock::rep = unspecified;
```
-1- ~~Constraints:~~ system_clock::duration::min() < system_clock::duration::zero() is true. [Note 1: This implies that `rep` is a signed type. — end note]

msg14881 (view)

Date: 2025-07-15.00:00:00

[ 2025-07-15; Reflector discussion ]

The discussion revealed a preference to not insert the additional "The implementation shall ensure that", because its not really needed, since this is just a normal implementation requirement that falls out of the specification.

msg14880 (view)

Date: 2025-07-15.15:24:16

Currently, the Constraints: element in [time.clock.system.members] for the member typedef `system_clock::rep` imposes a requirement for the implementation, without establishing any condition for user code. Perhaps it's wrong to use a Constraints: element there.

This wording is relative to N5008.

Modify [time.clock.system.members] as indicated:
```
using system_clock::rep = unspecified;
```
-1- ~~Constraints:~~The implementation shall ensure that system_clock::duration::min() < system_clock::duration::zero() is true. [Note 1: This implies that `rep` is a signed type. — end note]

History
Date	User	Action	Args
2025-07-15 15:24:16	admin	set	messages: + msg14898
2025-07-05 17:08:59	admin	set	messages: + msg14881
2025-06-19 00:00:00	admin	create