[PATCH] D113158: ADT: Fix const-correctness of iterator adaptors

[Duncan P. N. Exon Smith via Phabricator via llvm-commits]

dexonsmith marked an inline comment as done.
dexonsmith added inline comments.


================
Comment at: llvm/unittests/ADT/IteratorTest.cpp:104
+  EXPECT_EQ(true, canAssignFromInt(*I));    // int *
+  EXPECT_EQ(true, canAssignFromInt(*IC));   // int *const
+  EXPECT_EQ(false, canAssignFromInt(*CI));  // const int *
----------------
mehdi_amini wrote:
> dexonsmith wrote:
> > I expected this test to fail (and the proof below not to compile) after https://reviews.llvm.org/D112981, but locally it seemed fine. It appears that `const ReferenceT` is equivalent to `int&` when `ReferenceT==int&`... which confuses me a bit, but so be it.
> Right this always confuses me as well, this is  "reference is a const reference, but not a reference to const", so the "const" applies to the reference, but a reference is always "const" (as in `int *const`).
> This is why we have `std::remove_reference_t` in generic code right?
> 
Right, right; thanks!

In which case, https://reviews.llvm.org/D112981 only adds `const` to by-value returns (which are super rare, outside of proxies).


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D113158/new/

https://reviews.llvm.org/D113158