LWG4174 How does [atomics.order] p3 apply when then modification is an initialization?

[jwakely]

LWG would like SG1 to consider this issue please
https://cplusplus.github.io/LWG/issue4174

[gonzalobg]

SG1 conclusion (cc @ogiroux):

• We agree that LWG4174 is a defect.
• We expects an SG1 paper proposing a fix for this issue both for atomic and for atomic_ref.