Re: PangoLayoutLine and language bindings

[Owen Taylor <otaylor redhat com>]

On Fri, 2003-05-16 at 09:30, Gustavo J. A. M. Carneiro wrote:
> A Qui, 2003-05-15 ās 18:17, Gustavo J. A. M. Carneiro escreveu:
> >   I was thinking I will need to access individual PangoLayoutLine's
> > inside a PangoLayout from pygtk.  However, this is not implemented in
> > pygtk.  Since PangoLayoutLine is not GType-registered, I can understand
> > why.
> > 
> >   So, I propose:
> > 	1. Add reference counting to PangoLayoutLine;  However, as I was
> > writing this, I realized this is already implemented :-)
> > 	2. Register it as boxed type, G_TYPE_PANGO_LAYOUT_LINE, with the copy
> > function being pango_layout_line_ref() and the free function being
> > pango_layout_line_unref().
> > 
> >   So, can I send a gtk+ patch for this?
> 
>   Any problem committing this patch, then?

It's never OK to commit a patch simply because you didn't get a
response. Certainly not in one day.

I'm a little hestitant about doing anything to make PangoLayoutLine 
more language binding friendly because that whole portion of Pango 
is ugly and difficult to use and difficult to language bind. I'm
not sure that you can do anything interesting with a PangoLayoutLine
without accessing line->runs, which gets you completely out of
language binding territory. (At least without a lot of custom
written API for the language binding.)

Why do you think you need to access PangoLayoutLine from Pango?

Regards,
                                    Owne