Re: GTK+ 3, Win32 and GtkApplication (DBus woes)

[Kean Johnston <kean johnston gmail com>]

> Yes I'm sorry about that. I innocently assumed it would be properly
> portable as everything in GTK+ is portable, and there was no
> proper notation that the application object was not available
> on win32 or osx (like say, GtkUnixPrint widgets).
For the most part it seems that GTK+'s portability holds true. The trouble 
is in layers beneath it, in this case Gio. I just came across another 
example of strangeness in Gio on Win32. On Windows it is much rarer to have 
things installed in a central location available to everyone to use, many 
applications will ship their own versions of the DLL's in their own private 
directories. This makes reliance on central-installed, compiled files like 
compiled schemas in Gio a problem on Windows.

Other underpinning layers like Pango can have the same problem if not 
compiled properly. For example, you *can* compile it to dynamically load in 
modules and that works really well on Linux but less well on Win32, where 
the modules tend to be built in.

Since I am new to this list I am unsure of the etiquette involved in 
discussing issues in integral underpinnings of GTK+ like Glib/Gio, and I 
know there's a fair amount of cross-pollination of developers, but if 
discussions of Glib/Gio issues, especially as related to GTK+ are unwelcome 
or unproductive here, I'll move my questions to the appropriate place, 
wherever that is.

Kean

PS thank you very much for Glade :) I have some patches for you to get it 
to compile with MSVC that I'll be sending you shortly.