RE: [gtkmm] textbuffer's signal_erase

[Joshua A Gagnon <gagnon WPI EDU>]

The only documentation I've been able to find is from the gtkmm site.
http://gtkmm.org/gtkmm2/docs/reference/html/classGtk_1_1TextBuffer.html#a80
This link is all I can find about the signal, which is why I asked a
question here. I do know from experimentation that signal_erase is called
everytime text is erased in any way. Delete, backspace, cut, highlighting
then typing something else, all cause it to fire. That's pretty much all I
know about it.

On Wed, 13 Aug 2003 Murray Cumming Comneon com wrote:

> > From: gagnon WPI EDU [mailto:gagnon WPI EDU]
> > I've been trying to add undo/redo support to a project of
> > mine. I used a
> > signal handler connected to signal_insert() and it worked
> > fine to add any
> > insertions to the undo/redo stack. When I tried to do the same with
> > signal_erase() I ran into a problem. From what I can tell, the text is
> > deleted before the signal handler is called, so that the two iterators
> > passed into the handler are at the same spot.
>
> I'm not sure exactly how this signalled is meant to behave. Could you point
> me to some documentation that says what it should do?
>
> > Now matter how
> > I delete the
> > text, or how much I delete, my program thinks I deleted "". Is this
> > stupidity on my part or a bug in gtkmm?
>
> It might be a bug, in gtkmm or in GTK+. First we need to find out what it
> should do.
>
> I believe that Daniel Elstner implemented some undo/redo stuff with TreeView
> in regexxer, but maybe only in the cvs version. You might take a look.
>
> Murray Cumming
> murrayc usa net
> www.murrayc.com
>
>
>