[PATCH] D65022: [Sema] Always instantiate the initializer of a variable template with undeduced type (8.0 regression)
Richard Smith - zygoloid via Phabricator via cfe-commits
cfe-commits at lists.llvm.org
Fri Jul 19 15:03:30 PDT 2019
rsmith added a comment.
Well, this restores an incorrect behaviour: we're not permitted to substitute into the initializer of the variable template here.
Can you look into fixing this properly, by instantiating the type when forming the MemberExpr? If not, taking this to fix the regression seems OK, but please file a bug for the introduced misbehaviour.
CHANGES SINCE LAST ACTION
More information about the cfe-commits