Am Sonntag, 10. August 2014, 10:59:24 schrieb Alvaro Soliverez: > The main issue preventing us from a migration to QML is the sheer > number of pages we have in the application. It's simply a lot to > migrate, and QML in KDE is still a moving target. I haven't seen an > application as big as KMyMoney in QML yet.
Very right, migrating all widgets at once is not possible. But can I start with new things I make? Especially list views with custom delegates are easier to write in QML and the user will not notice a difference. Later we can move more and more when needed — or someone wants to. Porting dialogs will probably not make sense in near future. > Next issues are: > - Too much business logic in UI, which should be moved to core classes You are right, but this is not an issue of qml. > - Most data is not using models, much less a MVC approach. I like the MVC (or in Qt it is just MV) approach, it makes a lot of stuff cleaner and easier to understand. Moving some stuff to models can solve the previous problem. > - Customs controls that will have to be migrated to QML You are right again. But at the views mentioned above I rarely need custom controls. > - Developers involved will have to have a grasp of the application > domain (It can't be random people who develop in QML) Sure, you are right. > - It's all or nothing, you can't have QWidgets and QML together. At > least not in a stable release Well, you can have QML inside QWidgets, but not the other way round. Even in a stable release. It can look strange if not done correctly, that's the biggest issue. Actually I smuggled such an dialog into the add-onlinebanking branch: the tan input with chipTan: QWidgets are used to create everything except the chipTan flicker filed which is written in QML. _______________________________________________ KMyMoney-devel mailing list KMyMoney-devel@kde.org https://mail.kde.org/mailman/listinfo/kmymoney-devel
