[cfe-commits] r162768 - in /cfe/trunk: include/clang/AST/Stmt.h lib/AST/Stmt.cpp

[Chandler Carruth]

On Tue, Aug 28, 2012 at 11:08 AM, John McCall <rjmccall at apple.com> wrote:

> On Aug 28, 2012, at 11:05 AM, Chad Rosier wrote:
> > On Aug 28, 2012, at 10:48 AM, Dmitri Gribenko wrote:
> >> On Tue, Aug 28, 2012 at 10:43 AM, Chad Rosier <mcrosier at apple.com>
> wrote:
> >>> +std::string AsmStmt::generateAsmString(ASTContext &C) const {
> >>> +  StmtClass stmtClass = getStmtClass();
> >>> +  if (stmtClass == Stmt::GCCAsmStmtClass)
> >>> +    return static_cast<const GCCAsmStmt*>(this)->generateAsmString(C);
> >>> +  if (stmtClass == Stmt::MSAsmStmtClass)
> >>> +    return static_cast<const MSAsmStmt*>(this)->generateAsmString(C);
> >>> +  llvm_unreachable("unknown asm statement kind!");
> >>> +}
> >>
> >> It is conventional to use cast<> in such cases.
> >
> > I based these changes off of John's work in r125183.  I don't claim to
> be an expert on the use of casts, but I'm guessing John used the
> static_cast to elide the run-time checks.  John, please correct me if I'm
> wrong…
>
> I don't remember, but it's a bit different in macro-metaprogrammed code.
>  Here you might as well use cast<>.


Better yet, use dyn_cast and remove the raw getStmtClass...

if (GCCAsmStmt *GCCAS = dyn_cast<GCCAsmStmt>(this))
  GCCAS->generateAsmString(C);
if (MSAsmStmt *MSAS = dyn_cast<MSAsmStmt>(this))
  MSAS->generateAsmString(C);
llvm_unreachable(...);
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20120828/a11c1038/attachment.html>