[PATCH 2/4] New data structure for cgraph_summary introduced.

[Markus Trippelsdorf]

On 2014.11.13 at 15:11 +0100, mliska wrote:

Just two remarks:

> +template <class T>
> +class GTY((user)) cgraph_summary <T *>
> +{
> +public:
> +  /* Default construction takes SYMTAB as an argument.  */
> +  cgraph_summary (symbol_table *symtab, bool ggc = false): m_ggc (ggc),
> +    m_insertion_enabled (true), m_symtab (symtab)
> +  {
> +    cgraph_node *node;
> +
> +    FOR_EACH_FUNCTION (node)
> +    {
> +      gcc_assert (node->summary_uid > 0);
> +    }
> +
> +    m_map = new hash_map<int, T*, summary_hashmap_traits>(13, m_ggc);
> +
> +    m_symtab_insertion_hook =
> +      symtab->add_cgraph_insertion_hook
> +      (cgraph_summary::symtab_insertion, this);
> +
> +    m_symtab_removal_hook =
> +      symtab->add_cgraph_removal_hook
> +      (cgraph_summary::symtab_removal, this);
> +    m_symtab_duplication_hook =
> +      symtab->add_cgraph_duplication_hook
> +      (cgraph_summary::symtab_duplication, this);
> +  }
> +
> +  /* Destructor.  */
> +  virtual ~cgraph_summary ()
> +  {
> +    destroy ();
> +  }

>From https://gcc.gnu.org/wiki/CppConventions:

Constructors and destructors are often much larger than programmers
expect. Prefer non-inline versions unless you have evidence that the
inline version is needed.

>...
> +  inline T* operator[] (int uid)
> +  {
> +    T **v = m_map->get (uid);

The inline keyword is redundant for members inside a class definition.
Please drop it.

-- 
Markus