Is there a reason not to use SetVector here? This seems to be what it was
designed for...


On Wed, Aug 22, 2012 at 10:05 PM, Douglas Gregor <[email protected]> wrote:

> Author: dgregor
> Date: Thu Aug 23 00:05:18 2012
> New Revision: 162429
>
> URL: http://llvm.org/viewvc/llvm-project?rev=162429&view=rev
> Log:
> array_pod_sort on the addresses of declaration pointers leads to
> inconsistent ordering of results; instead, use use SmallPtrSet to
> eliminate duplicates.
>
> Modified:
>     cfe/trunk/lib/AST/CXXInheritance.cpp
>
> Modified: cfe/trunk/lib/AST/CXXInheritance.cpp
> URL:
> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/AST/CXXInheritance.cpp?rev=162429&r1=162428&r2=162429&view=diff
>
> ==============================================================================
> --- cfe/trunk/lib/AST/CXXInheritance.cpp (original)
> +++ cfe/trunk/lib/AST/CXXInheritance.cpp Thu Aug 23 00:05:18 2012
> @@ -25,13 +25,11 @@
>    assert(NumDeclsFound == 0 && !DeclsFound &&
>           "Already computed the set of declarations");
>
> +  llvm::SmallPtrSet<NamedDecl *, 8> KnownDecls;
>    SmallVector<NamedDecl *, 8> Decls;
>    for (paths_iterator Path = begin(), PathEnd = end(); Path != PathEnd;
> ++Path)
> -    Decls.push_back(*Path->Decls.first);
> -
> -  // Eliminate duplicated decls.
> -  llvm::array_pod_sort(Decls.begin(), Decls.end());
> -  Decls.erase(std::unique(Decls.begin(), Decls.end()), Decls.end());
> +    if (KnownDecls.insert(*Path->Decls.first))
> +      Decls.push_back(*Path->Decls.first);
>
>    NumDeclsFound = Decls.size();
>    DeclsFound = new NamedDecl * [NumDeclsFound];
>
>
> _______________________________________________
> cfe-commits mailing list
> [email protected]
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to