In the afternoon today I’ve been having another look at GTK+ 3 and I have to admit that albeit the look and feel of the resulting applications is quite nice on Windows 10 (yes, haven’t downgraded to Windows 7 just yet), the interface editor, GLADE 3.x, is still as unstable as I remembered it from trying it a few years back:
I managed to see GLADE segfault a few times and creating a project file that would always result in subsequent segfaults. Eventually I had to manually fix the project in a text editor (it’s just XML) in order to get it back to a working state.
That’s not a good enough experience so I think I will only make marginal use of it in the future, if any.
Regarding GTK+ 2, however, I made quite some progress with the MinGW SDK and runtime testing under Windows: During my regression testing with the IECHost client compiled against GTK+ version 2.24.31 I haven’t been able to reproduce the 2 major bugs that have so far frustrated me the most. This means that there are good chances I will be providing a runtime installer for GTK 2.24.31 that will be usable for all of my GTK+ applications: TAPClean FE, The Tast Ninja Construction Kit and Integrator 2012, the GUI clients for DC2N4-LC, C2NEmu, and IECHost, to name a few. This means I will not have to provide binary packages for my GTK+ applications that include a working, and possibly application specific, runtime.
Essentially, I’m over the moon today 🙂