Re: [gtkmm] Glib::IOChannel

[Christopher Clark <clark compudata-systems com>]

Murray Cumming Comneon com wrote:
> > From: Christopher Clark [mailto:clark compudata-systems com] 
> >   I've been working through the tutorial, and when I found that the 
> > Monitoring I/O page was out of date, I wanted to try and get the demo 
> > running anyway with a Glib::IOChannel.
>
>
> I believe that it's up-to-date now. Please submit a bug if something in the
> book is wrong, with a link to the book online, showing exactly which part
> you mean.

  I was looking at the example in:

http://www.gtkmm.org/gtkmm2/docs/tutorial/html/ch17s02.html

  Which says to use Gtk::Main::input.connect, which doesn't appear to 
exist in the current headers.  However, if I use Glib::signal_io() 
(thanks, Ross) it works fine.

> >   This is what I think the logic should look like (called 
> > from within 
> > ExampleApp::ExampleApp):
> >
> > Glib::RefPtr<Glib::IOChannel> ifile =
> >   Glib::IOChannel::create_from_file("testfifo", "r");
> > Glib::RefPtr<Glib::IOSource> isource = 
> > ifile->create_watch(Glib::IO_IN);
> > isource->connect(SigC::slot(*this, &ExampleApp::on_input_read));

This works, and gets rid of an object.

Glib::RefPtr<Glib::IOChannel> ifile =
  Glib::IOChannel::create_from_file("testfifo", "r");
Glib::signal_io().connect(SigC::slot(*this, &ExampleApp::on_input_read),
                          ifile, Glib::IO_IN);

  Expansion of this into a revised section in the tutorial is left as 
an exercise for the reader.

--
Christopher Clark <clark compudata-systems com>
Pongidae, and proud of it.

  Praeterea censeo Carthaginem esse delendam.
                                                             -- Cato sr.