Re: sort functions with user_data

[Tim Janik <timj gtk org>]

On Wed, 11 Oct 2000, David Benson wrote:

> I've been wanting user_data-bearing equivalents of the
> following functions for a long time:
> 
>    qsort
>    g_slist_sort
>    g_list_sort
>    g_tree_new
> 
> What would people think of a patch that added the following functions?
> I'd be happy to cook up such a patch...

while i can understand the need for user_data in these functions,
*_full is a very bad misnomer.
*_full variants are provided for functions that need a destroy
notifier for certain data items, originally added for language bindings,
theoretically, if we used closures all over the place, we'd not have
a single _full function.
*_full variants are _not_ "some function version that alowes specification
of more details", with that scheme we'd end up with _fuller, _fullest,
_full_full, _full_and_extra_full etc, uglyness at its best.
go for _udata or something else.

> 
> - Dave
> 

---
ciaoTJ