Re: [patch] fix int-enum preference update in eel

[Alexander Larsson <alexl redhat com>]

On Fri, 17 Jan 2003, Alexander Larsson wrote:

> On 17 Jan 2003, Jan Arne Petersen wrote:
> 
> > Hi,
> > 
> > Here is a fix for a bug in eel-preferences-glade.c.
> 
> Uhm. This looks quite wrong:
> 
> @@ -249,13 +248,17 @@
>  	
>  	option_menu = GTK_OPTION_MENU (glade_xml_get_widget (dialog, component));
>  
> -	map = g_hash_table_new (g_int_hash, g_int_equal);
> +	map = g_hash_table_new_full(g_int_hash, g_int_equal, (GDestroyNotify) g_free, NULL);
>  	value_list = NULL;
>  
>  	for (i = 0; values[i] != -1; i++) {
> -		value = values[i];
> -		value_list = g_slist_append (value_list, GINT_TO_POINTER (value));
> -		g_hash_table_insert (map, &value, GINT_TO_POINTER (i));
> +		value = g_malloc (sizeof (int));
> +		if (value == NULL) {
> +			return;
> +		}
> 
> g_malloc never returns NULL.
> 
> +		*value = values[i];
> +		value_list = g_slist_append (value_list, GINT_TO_POINTER (*value));
> +		g_hash_table_insert (map, value, GINT_TO_POINTER (i));
> 
> This hash table is quite useless, as it maps from "address of some memory 
> that nobody knows about" and the index i. Of course the previous code was 
> bogus too, mapping from "some address on the stack not valid anymore" to 
> the index i. Not to mention that the hash table uses g_int_hash/equal and 
> not the pointer versions.

Oops. Sorry about this, i see i was mistaken about how g_int_hash works.

-- 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 Alexander Larsson                                            Red Hat, Inc 
                   alexl redhat com    alla lysator liu se 
He's a shy native American grifter for the 21st century. She's a radical 
communist magician's assistant with an incredible destiny. They fight crime!