Re: problem of split-cursor

[Owen Taylor <otaylor redhat com>]

On Thu, 2003-07-03 at 06:27, miaocheng wrote:
>    I use red hat linux 8.0, the version of XFree86 is 4.0.3 ,  the 
> desktop manager is Gnome and Gnome's version is 2.0

Actually, Red Hat 8.0 came with XFree86-4.2.0...

>    I encounter a problem about split-cursor. Uyghur is one kind of Arbic 
> and written from right to left .  the Uyghur's characters are adopted on 
> Unicode standard 3.1 and puted in Arabic Presention A , B . When I input 
> character of Uyghur in a  GtkEntry or GtkTextBuffer , two cursors appear 
> . one is at the left end of line , the other is at the right end of line 
> . According to some reference , this is called split-cursor. But this 
> don't conform to our tradition . In my zone , when people inputed Uyghur 
> , they would like to look at a cursor glittering in character insertion 
> position . I put 'gtk-split-cursor = 0' in my /etc/gtk-2.0/gtkrc to 
> remove the split-cursor.  When I input Uyghur in a  GtkEntry or 
> GtkTextBuffer which contain only one line Uyghur characters, the cursor 
> at insertion position of inputed Uyghur character, in the left end of 
> line , disappear and the cursor at the other end of line is still 
> glittering . the visible cursor is located in position that has not any 
> relation with where to insert inputed Uyghur character . So I want to 
> make the cursor at insertion position visible and the other cursor 
> disappear . In other word , the cursor at the left end of line glitter 
> ,at the same time the right cursor disappear . Please tell me how to 
> realize it . which functions do I modify in Gtk+ widght or Pango  ? 
> which library , Gtk's widget , Pango or both, control the cursor show . 

For those not on i18n xfree86 org, this was asked there a couple of 
weeks ago, I responded to use gtk-split-cursor, but apparently it
didn't work. 

The possible answer I just sent there is that the second group of
the Uyghur keyboard map is not named 'arabic' or 'hebrew' so GTK+
doesn't recognize it as a a RTL group.

(http://bugzilla.gnome.org/show_bug.cgi?id=116626 describes a more
robust approach that could be implemented in GTK+)

>Another question , where I can find speciafications that demonstrate
the 
> Pango library how to work on the hierarchy of GTK+ in detail and 
> relation within Pango functions .

I don't understand this question. There are API docs for Pango at

 http://developer.gnome.org/doc/API/2.0/pango/

The way that GTK+ uses Pango is best determined by looking at the
GTK+ source code. 

Regards,
						Owen