Re: Glade & new GTK+ 2.5.x features

[Damon Chaplin <damon karuna uklinux net>]

On Wed, 2004-12-01 at 20:16, Bill Haneman wrote:
> Damon Chaplin wrote:
> 
> >On Wed, 2004-12-01 at 17:11, Bill Haneman wrote:
> >
> >>>> <widget class="GtkLabel" id="label1">
> >>>>    <property name="label" markup="&lt;b&gt;">label1</property>
> >>>> </widget>
> >>        
> >>This looks really ugly to me - storing what is essentially an XML tag as
> >>content! 
> >
> >Yes, it isn't ideal. But that's what we are currently doing for markup
> >anyway.
>
> Yes, but the suggested changes in your email are IMO even uglier because 
> they put the escaped text in an attribute.  For in-line markup (what you 
> are doing now), escaping the chars makes sense, but for global markup 
> IMO it does not.

OK, scratch that idea. I've thought of a better way:

  <widget class="GtkLabel" id="label1">
    <property name="label" font="bold,italic,x-large">label1</property>
  </widget>

I think that is fairly clean.

Though I also like Owen's idea of a GtkWidget "font" property. But I
think a PangoFontDescription only supports absolute sizes so we'd have a
problem with the sizes like "x-large".


> >> <>Why not
> >>
> >> <widget class="GtkLabel" id="label1">
> >> <property name="label">label1</property>
> >> <markup>
> >> <b>
> >> .. [etc.]
> >> </markup>
> >> </widget>
> >>
> >> i.e. any tag could be places in the <markup/> element.
> >
> Damon:
> 
>     I don't think it really warrants major changes to the XML structure like
> 
> >> <>that.
> >
> I don't see it as that much bigger a change that some of the other 
> proposals, and it has the advantages of both allowing arbitrary global 
> markup tags (i.e. extensible) and avoiding the escaping of the markup tags.

It does look pretty tricky to implement though. I wouldn't really fancy
it.

Damon