Re: [gtkmm] problems with pango

[MHL Schulze t-online de (Martin Schulze)]

Am 2002.10.27 21:10 schrieb(en) Daniel Elstner:
> Am Son, 2002-10-27 um 16.40 schrieb Morus Walter:
> >
> > I would like to make two aditional remarks on the methods I used
> > for low level text rendering:
> >
> > - the method Gdk::Drawable::draw_glyphs, which is used for low level
> >   text rendering, takes a c pointer PangoGlyphString* glyphs instead of
> >   a Pango::GlyphString object.
> >   This is not a problem since one can use the gobj method of
> >   Pango::GlyphStringto get the c pointer from the object, but it looks
> >   a bit inconsistent to me.
> >
> >   void  draw_glyphs (const Glib::RefPtr<const GC>& gc,
> >   const Glib::RefPtr<const Pango::Font>& font, int x, int y,
> >   PangoGlyphString* glyphs)
> >
> > - similar point for pango_shape where no c++ wrapper exists.
> >   Again one can use the c function directly by using the gobj methods.
> >
> >   void pango_shape (const gchar      *text,
> >                     gint              length,
> >                     PangoAnalysis    *analysis,
> >                     PangoGlyphString *glyphs);
>
> OK.  Thanks a lot for reporting the problem.

A wrapper for pango_shape does exist: Pango::Item::shape().

It might look a bit strange that shape() is a member function of
Pango::Item but it really makes sense: You need a PangoAnalysis
object for pango_shape() and the PangoAnalysis is always a rather
internal data member of a PangoItem object. Note that this is
documented in the reference docs (e.g. class description of
Pango::Item)!

There is also a Pango::GlyphString::GlyphString() constructor
that takes some text and a Pango::Analysis object.

Regards,

  Martin