Hacker News Reader

Comment by sylware | Hacker News Reader

https://docs.gtk.org/gtk4/migrating-3to4.html

oooof...

I'm a bit excited about the very last point. I always found the GtkTreeModel/GtkCellRenderer stuff hard to use. Hopefully the new system is less confusing.

My first experience with trying to develop against Gtk4 (about a year ago) ran into an issue where none of the example programs in the Gtk repo worked. Apparently the details on how to actually get the program to pick up the now mandatory CSS file were not right? All it did was to convince me to stick with Gtk3 for the meantime.

Is there any plans to support the web dynamic CSS rendering model?

I kind of hope not since that would likely involve bundling almost a full web browser in every GTK app. Basically turning it into another electron clone. One of the nice things about GTK apps is that you can run them on more constrained environments and get a lot of work done.

I have that in checklist form on solvespace github here:

https://github.com/solvespace/solvespace/issues/853

Offtopic shoutout for solvespace, it is amazing! Is separating out the geometry code into a stand alone library on the roadmap at all? Or plans to incorporate the manifold library?

https://github.com/elalish/manifold

If you mean the NURBS code, there are no plans to separate it for now. It does live almost entirely in the src/surf folder and does not depend on the sketch entities or constraint solver.

My hope is to squash the remaining bugs in that code. I'm currently working on an issue with coincident and tangent surfaces.

The constraint solver is available as a library and is currently used in FreeCAD assembly 3 workbench, as well as the Blender CAD Sketcher addon.

Is migrating from 3 to 4 easier or harder than migrating from 2 to 3?

That is a very broad question. For something the size and complexity of gimp with lots of rendering it is still a big task. Broadly speaking I think 3->4 is often easier though.

Reply on news.ycombinator.com