Re: [AD] set_window_close_button_callback_hook_button()

[Ben Davis <bnd25@xxxxxxxxxx>]

On Friday 29 November 2002 12:21 pm, you wrote:
> > The close button that killed the program was already broken.
>
> Then I don't see any problem with breaking an already broken feature.

No, but you also broke programs that don't use the feature (i.e. programs 
that do register a callback).

> > The first need not be called, since it would be "enabled" by default (but
> > greyed out because there's no callback). This would be consistent with
> > the older versions where it was enabled and had broken functionality.
>
> So the first function is superfluous ?

In a sense. It provides a means for disabling the close button but leaving 
the callback registered for future use. Just in case someone was using it.

> > The program was already broken.
>
> Yes, but at runtime. With 4.1.5 it is at compile time. What do you prefer ?

You obviously want me to say compile time. However, by breaking programs at 
runtime, we avoid breaking programs that were not originally broken (i.e. 
programs that registered a callback). Noting that the breakage is hardly 
fatal (disabled close box), and the fix is hardly obscure (RTFM and then 
register a callback), I'm going to say runtime. :)

Ben