Re: Upcoming GLib changes

[Owen Taylor <otaylor redhat com>]

Trog <trog@gtk.org> writes:

> On 30-Jun-99 Tim Janik wrote:
> > On Wed, 30 Jun 1999, Trog wrote:
> > 
> >> 
> >> > GString
> 
> I don't disagree with anything you've said, given the position you
> are saying it from. I do, however, disagree with your position. It is
> desirable to use strings as a mechanism for reducing (or eliminating)
> buffer overflow problems. That is why most people do a strings
> implementation. GLib does not provide that - and your position is
> that it shouldn't aim to. This, unfortunately, also makes the strings
> implemenation virtually worthless, as I don't think that anyone will
> use it as a consequence of that position.

The current GString implementation is meant to be used where buffer
overflows are a problem - when one is taking bits of strings, adding
characters, etc.  When you are just passing around strings, overflow
is not a problem, so one might as well pass a char *.

There are a few missing functions that we need to add to make GString
really useful - the one I tend to miss is:

 g_string_insert_n (GString *string, gchar *new, gint n);
 
But I don't think we need to wrap every single thing that you can do
with a char * with a GString.  The conversions are easy enough.

Conversion from GString => gchar *
  
 string->str

Conversion from gchar * => GString

 g_string_new (str);

> Even gtk+ hardly uses the current implementation.

It's not like GTK+ is very heavy on string manipulation.
A lot (though not all) of the places that could 
benefit from GString do use it.

Regards,
                                        Owen