Date
2017-03-15.00:00:00
Message id
9026

Content

[ 2017-03-01, Kona ]

SG1 recommends: Close as NAD

There has not been much demand for it, and it would usually be difficult to deal with thread_local destructor races. It can be approximated with a condition variable wait followed by an unconditional join. Adding it would create implementation issues on Posix. As always, this may be revisited if we have a paper exploring the issues in detail.