Re: [evolution-patches] gal, accessiblity, etable, handle the columns change (add/delete/reorder)

[Larry Ewing <lewing ximian com>]

On Mon, 2003-12-01 at 10:29, Mike Kestner wrote:
> On Mon, 2003-11-24 at 00:25, Yuedong Du wrote:
> 
> > The patch fix bug 49594.
> > +static void
> > +eti_header_structure_changed (ETableHeader *eth, AtkObject *a11y)
> > +{
> > +                                                                                
> <snip>                                     
> > +        /* Init to ETI_HEADER_UNCHANGED. */
> > +        state = g_malloc0 (sizeof (gint) * n_cols);
> > +        prev_state = g_malloc0 (sizeof (gint) * prev_n_cols);
> > +        reorder = g_malloc0 (sizeof (gint) * n_cols);
> > +	g_return_if_fail (state && reorder && prev_state);
> 
> This could possibly leak two of the three blocks, although if mallocs
> are failing the wheels have already come off.  Also, typically gal uses
> g_new0 instead of g_malloc0, but that's not a big deal.  If you really
> want to guard against malloc failures, you should g_free state,
> prev_state, and reorder before returning.
> 

g_malloc* can never fail, they will allways abort with g_error if malloc
returns NULL, so testing the return value is pointless.  If you need
them glib provides allocators that can fail using the g_try_malloc*
functions.

--Larry