================
Comment at: lib/Sema/SemaExpr.cpp:1940
@@ +1939,3 @@
+ const TemplateArgumentListInfo *TemplateArgs) {
+ CXXMethodDecl *MD = dyn_cast<CXXMethodDecl>(DC);
+
----------------
David Majnemer wrote:
> auto?
done
================
Comment at: lib/Sema/SemaExpr.cpp:1949
@@ +1948,3 @@
+ SourceLocation Loc = NameInfo.getLoc();
+ unsigned DI = diag::warn_found_via_dependent_bases_lookup;
+ DiagnosticBuilder DB = S.Diag(Loc, DI) << NameInfo.getName();
----------------
Richard Smith wrote:
> This diagnostic seems inappropriate: it says the member was "found by lookup
> into dependent bases", but we don't know that. All we know is it wasn't found
> by normal lookup. Maybe rephrase it to only suggest that the name was not
> found by normal lookup?
What do you think of this diagnostic text?
warning: use of undeclared identifier 'StartWindowProc'; lookup into
dependent bases of class template 'Derived' is a Microsoft extension
I'm trying to make it phrase well for -pedantic-errors.
================
Comment at: lib/Sema/SemaExpr.cpp:1952
@@ +1951,3 @@
+
+ if (MD->isInstance()) {
+ DB << FixItHint::CreateInsertion(Loc, "this->");
----------------
Richard Smith wrote:
> This isn't exactly correct. You should also check `S.CXXThisTypeOverride`; if
> that's set, then member accesses with `this->` are OK even though we're not
> in a member function. (This happens when parsing a trailing return type or
> default initializer.)
I wasn't trying to be this general, but I went ahead and implemented it with
some extra tests.
http://reviews.llvm.org/D4079
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
