Re: [Pango] Calls to Pango::Layout without effect

[Matthias Kaeppler <matthias finitestate org>]

Bob Caryl wrote:
> I may be all wet here, but because the MessageDialog displays a lot of 
> things such as a graphic indicating the type of MessageDialog, a 
> separator bar graphic, a couple of buttons (at least) along with the 
> message it displays, I'd say that the pango context (which you don't 
> seem to be using) you're getting is *not* for the child widget that 
> displays the message in the MessageDialog window.  I'd also say that get 
> a pango layout pointer you are creating using the Glib::ustring that 
> you're passing to the constructor function is not going to affect the 
> way the MessageDialog ultimately displays it.

That sounds reasonable; I figured it would be something like this. 
However, since Nautilus also uses the pango layout returned by the 
dialog, I still had hope it would work.

>  I believe that there is a 
> child widget that displays that text, and it will probably have its own 
> pango layout; hence, maybe you could somehow use 
> Gtk::Container::get_children() and parse that list to find the GtkLabel 
> that is ultimately displaying the message.  Again, I'm guessing based on 
> my cursory perusal of the MessageDialog header and source files for 
> gtkmm-2.6.2.

That's a good idea, I will look into this. Reading the MessageDialog 
source code could also come in handy I guess. However, I don't want to 
write code which relies on MessageDialog internals and which could 
possibly be subject to change one day (which would render my code 
useless). Have to be careful here :)

> Oh yeah, the Gtk::MessageDialog::set_secondary_text method has a second 
> parameter that defaults to turning the use of markup text off.

Okay, thanks Bob.

> Bob Caryl

Best Regards,
Matthias Kaeppler