r183283 - Model temporary lifetime-extension explicitly in the AST. Use this model to

[Richard Smith]

On Wed, Jun 5, 2013 at 4:43 AM, Timur Iskhodzhanov <timurrrr at google.com>wrote:

> 2013/6/5 Richard Smith <richard-llvm at metafoo.co.uk>:
> > +/// Determine the declaration which an initialized entity ultimately
> refers to,
> > +/// for the purpose of lifetime-extending a temporary bound to a
> reference in
> > +/// the initialization of \p Entity.
> > +static const ValueDecl *
> > +getDeclForTemporaryLifetimeExtension(const InitializedEntity &Entity,
> > +                                     const ValueDecl *FallbackDecl = 0)
> {
> > +  // C++11 [class.temporary]p5:
> > +  switch (Entity.getKind()) {
> > +  case InitializedEntity::EK_Variable:
> > +    //   The temporary [...] persists for the lifetime of the reference
> > +    return Entity.getDecl();
> > +
> > +  case InitializedEntity::EK_Member:
> > +    // For subobjects, we look at the complete object.
> > +    if (Entity.getParent())
> > +      return getDeclForTemporaryLifetimeExtension(*Entity.getParent(),
> > +                                                  Entity.getDecl());
> > +
> > +    //   except:
> > +    //   -- A temporary bound to a reference member in a constructor's
> > +    //      ctor-initializer persists until the constructor exits.
> > +    return Entity.getDecl();
> > +
> > +  case InitializedEntity::EK_Parameter:
> > +    //   -- A temporary bound to a reference parameter in a function
> call
> > +    //      persists until the completion of the full-expression
> containing
> > +    //      the call.
> > +  case InitializedEntity::EK_Result:
> > +    //   -- The lifetime of a temporary bound to the returned value in a
> > +    //      function return statement is not extended; the temporary is
> > +    //      destroyed at the end of the full-expression in the return
> statement.
> > +  case InitializedEntity::EK_New:
> > +    //   -- A temporary bound to a reference in a new-initializer
> persists
> > +    //      until the completion of the full-expression containing the
> > +    //      new-initializer.
> > +    return 0;
> > +
> > +  case InitializedEntity::EK_Temporary:
> > +  case InitializedEntity::EK_CompoundLiteralInit:
> > +    // We don't yet know the storage duration of the surrounding
> temporary.
> > +    // Assume it's got full-expression duration for now, it will patch
> up our
> > +    // storage duration if that's not correct.
> > +    return 0;
> > +
> > +  case InitializedEntity::EK_ArrayElement:
> > +    // For subobjects, we look at the complete object.
> > +    return getDeclForTemporaryLifetimeExtension(*Entity.getParent(),
> > +                                                FallbackDecl);
> > +
> > +  case InitializedEntity::EK_Base:
> > +  case InitializedEntity::EK_Delegating:
> > +    // We can reach this case for aggregate initialization in a
> constructor:
> > +    //   struct A { int &&r; };
> > +    //   struct B : A { B() : A{0} {} };
> > +    // In this case, use the innermost field decl as the context.
> > +    return FallbackDecl;
> > +
> > +  case InitializedEntity::EK_BlockElement:
> > +  case InitializedEntity::EK_LambdaCapture:
> > +  case InitializedEntity::EK_Exception:
> > +  case InitializedEntity::EK_VectorElement:
> > +  case InitializedEntity::EK_ComplexElement:
> > +    llvm_unreachable("should not materialize a temporary to initialize
> this");
> > +  }
> > +}
>
> gcc 4.6.3 gives the following warning on this code:
> In function ‘const clang::ValueDecl*
> getDeclForTemporaryLifetimeExtension(const clang::InitializedEntity&,
> const clang::ValueDecl*)’:
> tools/clang/lib/Sema/SemaInit.cpp:5196:1: warning: control reaches end
> of non-void function [-Wreturn-type]
>

Thanks, looks like Benjamin Kramer fixed this already (thanks!).

Why do we still have this g++ warning enabled?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130605/6bd523c4/attachment.html>