From: Richard Dale Date: 2005-04-07T08:14:41+09:00 Subject: Re: QT for Win32? Kujawa, Greg wrote: > I took your advice and took the second route. I built the smoke sources on > my FreeBSD box and copied them over to my Windows box. Then I ran qmake. > Past that when I ran make I encountered a slew of errors stating "This > conversion requires a reinterpret_cast, a C-style cast" which blew up the > make process. At this point I will just explore Tk and GTK2 on my Windows > box a bit more and stick with FreeBSD for my Qt development work... Oh dear, I was hoping it would be easier than that. People have ported the Smoke library to Windows in the past to use with the PerlQt bindings, so it must be possible. -- Richard > Kujawa, Greg wrote: > >> I really want to experiment with Qt some, since I have used the GTK >> and Tk Ruby bindings and want to compare them all before choosing one >> for most of my GUI work. I guess I can hop over to my FreeBSD or Red >> Hat boxes to try things over there, but my client workstations are >> mostly Win32 and they would be the end users :-( > Yes, that sounds the best plan. The nice thing about Qt is that you can > develop on Linux, and deploy on Win32 - with ruby you wouldn't even need > to recompile. > > If you have a look at the instructions for building QtRuby on Mac OS X, > using qmake and extconf.rb instead of automake/autoconf it should be very > similar. The Smoke library generation is driven by a list of headers in > smoke/qt/header_list, which are fed into a perl tool called 'kalyptus'. > The code to do that works fine on Mac OS X, but I don't think it would on > Windows. > > So one way to get a port working, would be to generate the Smoke sources > on FreeBSD or Red Hat, and build them on Windows using qmake. If you have > errors from a specific header in the 'header_list' file, then either fix > the compile error by hand, or remove the header and retry. I would expect > that using extconf.rb on Windows to work ok, and build a .DLL instead of a > shared lib or Mac OS X bundle. > > -- Richard