Re: Making it easier to change a font.

["Marco Scholten" <mscholtn xs4all nl>]

On Fri, 4 Feb 2005 13:32:09 -0700, <pks timing com> wrote:

> Marco Scholten writes:
>  >
>  > We could do cool stuff like packing some labels and entry's in a  
> table and
>  > then call Gtk::Table::modify_font() to change all the labels and  
> entry's
>  > at once.
>
> I don't think it's that great for "non bin" containers such as table
> (especially if modify_*() is used internally) or great care should be
> taken not override settings the user made.
>
> Imagine a two column table with the left column using one font and
> the right column using another (because programmer took care to make
> it so) and something coming along calling modify_font() on what holds
> the table which propagates the modification to the table, which in turn
> changes the appearance of the labels ...

I don't think that something like that would come along unless you  
explicitly call modify_font on the table or one of its parent containers,  
or can you name another situation where this would happen?
Remember that you would still be able to make columns use different fonts  
like you always did, this would simply add functionality whithout adding  
API.

> You have the same issue with Gtk::Bin like containers, but your  
> suggestion
> makes more sense (to me) for those.  I don't really think of tables,
> boxes, layouts and stuff as widget's that "have (or should have)" fonts.
>
>  > Perhaps this can be done for more functions then just modify_font.
>
> if the function had a separate signature such as modify_all_*(), like
> show() vs show_all(), none of it would be problem ... somebody calling
> the "all" version of function should be aware of repercussions, or would
> soon be :)

That would then require you to call m_Button.modify_all_font() wich would  
still require you to know that a button is in fact a container holding a  
label.
From a users point of view, why should changing a font on a button be  
different from changing a font on a label?