================
@@ -329,6 +329,14 @@ static void emitVectorOverload(const OverloadContext &Ctx,
StringRef ElemType,
});
}
+/// Emit a dependent-size vector template overload for the given element type.
+static void emitLongVectorOverload(const OverloadContext &Ctx,
+ StringRef ElemType) {
+ emitOverload(Ctx, ElemType, [](StringRef ET) {
+ return ("vector<__detail::enable_if_t<(N > 4), " + ET + ">, N>").str();
----------------
bogner wrote:
I can't find the text right now, but I believe the TC57 wording is intended to
be that the limit is implementation defined. DXC happens to choose 1024 but I
don't think we need to follow suit here - from an implementation strategy
perspective we're capable of legalizing whatever size until we run out of
memory.
Omitting the upper limit in the template makes sense to me.
https://github.com/llvm/llvm-project/pull/216143
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
