Re: [jirka@5z.com: new version of the default patch]

[Owen Taylor <otaylor redhat com>]

George <jirka@5z.com> writes:

> --FL5UXtIhxfXey3p5
> Content-Type: text/plain; charset=us-ascii
> 
> On Thu, Jan 21, 1999 at 02:54:45PM -0500, Owen Taylor wrote:
> > The flags handling is still not quite right - I think that 
> > only CAN_DEFAULT widgets should get HAS_DEFAULT. 
> > 
> > RECEIVES_DEFAULT widgets that are are not CAN_DEFAULT should
> > just get an "activate" signal when the user hits return.
> > 
> > This will help preserve backwards compatibility with 
> > button-derived widgets that probably assume they'll
> > never get HAS_DEFAULT if they are CAN_DEFAULT.
> > 
> > I don't think that complicates the code much.
> 
> actually it makes the code a very tiny bit simpler in a way
> 
> here's then a patch that does this (has default is only
> set if CAN_DEFAULT and implies drawing of the border)

I think this is ready to go in. I've suggested two small
additional changes to George:
 
 - The patch should disable the "default switching" behavior
   of GTK+, where if a dialog has two CAN_DEFAULT buttons,
   then activating one of them makes the default switch
   there. 

   This was always confusing, seldom helpful, and is even 
   more confusing with the new behavior where we are
   already moving the current default around for different 
   reasons.

 - A few style cleanups need to be made. Things like 
   if ((a == b) && (b == c)).

So, if you have objections to the idea of the patch 
(which, summarized brieflym is that if a button has
the focus, it also gets the default temporarily) 
now is the time to speak up.

Regards,
                                        Owen