Looks like you run formatter on the whole file, maybe narrow down its scope a little? For example, only for the touched function?

A few comments.

@NoQ Do we need to change a DeclaratorDecl field in PointerToMember SVal to something more common, like ValueDecl, to support IndirectFieldDecl as well?
Of course in another patch, this one LGTM.

Looks good to me, apart from a very small cavil introduced in not even your code. Thanks for not leaving pointers to static member functions out of account!

Fix issues pointed by @dcoughlin and rebase patch on master.

Thanks for your comments, Devin! You were right about the list of path specifiers construction order, so i fix it. Now the base specifier list is being used for figuring out the correct subobject field. Also this diff changes fallthrough behavior in some cases and renames functions with 'cons' prefix.

Thanks for your comments, Artem! Make function name less ambiguous. Also I take liberty to update analogical function name.

ping

ping

According to dcoughlin suggestion PointerToMember SVal is now modeled as a PointerUnion of DeclaratorDecl and bump pointer-allocated object stored DeclaratorDecl and immutable linked list of CXXBaseSpecifiers.

Null pointer-to-member dereference is now modeled as UndefinedVal. Fixing this also releases us from loading from nonloc SVal.
I delete an assertion suppression in ExprEngine::performTrivialCopy because I can't reproduce its violation.