Date: Tue, 13 Jan 2026 17:30:25 -0800
On Tuesday, 13 January 2026 16:55:24 Pacific Standard Time Ryan P. Nicholl via
Std-Proposals wrote:
> Of course, it's possible to implement std::latch using things like
> SYS_futex, so this isn't impossible to implement per se, only impossible to
> implement using the C++ standard library without relying on undefined
> behavior.
How does futex make it possible while notify_one/all doesn't? Wouldn't it run
into the same issue?
If the thread doing the wait successfully set a "contended" state and entered
wait, then it won't wake up until a notification happens. Why would it destroy
the object before then?
Std-Proposals wrote:
> Of course, it's possible to implement std::latch using things like
> SYS_futex, so this isn't impossible to implement per se, only impossible to
> implement using the C++ standard library without relying on undefined
> behavior.
How does futex make it possible while notify_one/all doesn't? Wouldn't it run
into the same issue?
If the thread doing the wait successfully set a "contended" state and entered
wait, then it won't wake up until a notification happens. Why would it destroy
the object before then?
-- Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org Principal Engineer - Intel Data Center - Platform & Sys. Eng.
Received on 2026-01-14 01:30:52
