A few points I forgot to mention:

Replace the two flags, EHOnly and DisposeCannotThrow, passed to pushCaptureCleanup with a single flag ForCopyHelper.

Call MarkFunctionReferenced to mark the function __builtin_operator_new or __builtin_operator_delete will call as referenced.

Define functions inline in the header file.

ping

Address review comments.

Address review comments.

Modify getBlockVarCopyInits and setBlockVarCopyInits to get and set both the copy expression and the boolean flag that indicates whether the expression can throw or not.

Change the value type of BlockVarCopyInits to PointerIntPair<Expr *, 1, bool>. The boolean flag indicates whether the copy expression can throw. Serialize/deserialize the copy expression and the boolean flag and add a regression test to test/PCH.

Since BlockVarCopyInits is a map with key VarDecl *, I think we want to add a flag to VarDecl (NonParmVarDeclBits) that indicates whether the copy expression can throw or not. Or is there a reason to store the bit in BlockDecl::Capture instead?

I thought it was okay to skip the work done by ResolveExceptionSpec in IRGen as long as the exception specifications that are needed have already been resolved in Sema. But calling Sema::canThrow in Sema::CheckCompleteVariableDeclaration and storing the result in BlockDecl::Capture is clearly the better solution since it doesn't introduce the complexity introduced in the updated patch.

Remove a stale comment and add an assertion to check the destructor's exception specification has been resolved.

Address review comments.

Address review comments.

Address review comments.

Address review comments.

I just wanted to make sure that this doesn't have the same problem as https://reviews.llvm.org/D34556. Is that correct?

Address review comments.

Use llvm::sort.

• Produce a note that tells users where the class extension conforming to the protocol containing the non-escaping method is declared.
• Add a class method that has the same name as the instance method and check that no spurious warnings are issued.

Thanks, I'll update the patch and commit it today.

Add a test case for an interface conforming to a non-escaping protocol.

Document the meaning of enterByrefCleanup's parameters.

ping

Set the BLOCK_FIELD_IS_WEAK bit of the flag passed to the call to enterByrefCleanup in EmitAutoVarCleanups and fix test case test/CodeGenObjC/blocks.m.

Check if destructors are accessible from InitListChecker's constructor. Add test cases for designated initializer and brace elision.

Merge pushBlockObjectRelease and enterByrefCleanup.

Address review comments.

Note that in order to destruct C++ objects, I'm using pushDestroy which pushes a NormalCleanup when exception handling isn't enabled. This is different from PushDestructorCleanup which always pushes a NormalAndEHCleanup, but I think using pushDestroy is more correct since I don't think we need to do a cleanup on the EH path when exception isn't enabled.

Add a comment that explains the meaning of BLOCK_IS_NOESCAPE to the docs. Rename function needsCopyDispose to needsCopyDisposeHelpers.

Fix a bug where the capture cleanups weren't pushed when BlockDecl::doesNotEscape() returns true.

ping

In D49303#1165902, @rjmccall wrote:

Please test that we still copy captures correctly into the block object and destroy them when the block object is destroyed.

ping

ping

Address review comments.

ping

ping

Implement the new rule defined here: http://wg21.link/p0968r0#2227. Produce a diagnostic if a class is initialized with aggregate initialization and one of its element's destructor is not accessible.

Yes, that is the case. It doesn't change overloading resolution, it is just producing a diagnostic.

I mean ToT clang (without my patch applied) seems to select the non-sized host version 'T::operator delete(void*)'.

The original comment said "call sized device delete even though host has preferable non-sized version", but it seems that the non-sized host version 'T::operator delete(void*)' is being called in the IR, not the sized device version of delete. Is that a bug in overload resolution?

@jlebar, is the change I made to call-host-fn-from-device.cu correct?

LGTM