- yes, this does break scripts, but the api is marked experimental.
ED_region_draw_cb_activate() adds a callback to a region type whereas the api made it look like the callback was being added to the region instance.
Use a class method on bpy.types.Space to manage region drawing, eg.
was:
self._handle = context.region.callback_add(draw_callback_px, args, 'POST_PIXEL')
is now:
self._handle = bpy.types.SpaceView3D.draw_handler_add(draw_callback_px, args, 'WINDOW', 'POST_PIXEL')
Mouse coordinates were not mapped correctly for code that allows to use
multiple windows efficiently (mouse over not-active windows).
Apple's high-density display mode works a bit strange, requiring some hacks :/
- Desktop coordinate system (mouse pos and for windows) is as usual
(set by display resolution settings)
- However, the available pixels in a window is always on 'retina' level.
(full screen - 2880 wide, but window can be 1440 or 1920 wide)
In order to get this to work for opengl and Blender, we use internally the
coordinates on pixel level. That means that window positions and sizes have
to mappend in our code.
Once all issues for retinas have been tackled, I'll check on clean API for
it, so you can also use it in future for other high density screens.
- Since 2.50, the .blends have the current file name stored
- In versions up to 2.65.0 this didn't save for files saved from startup
(i.e. files saved without loading a file)
Code now adds the file name after all, except for recovery loads.
Recoded the (2.65.1 version) region scale, which happened on loading files with
different saved size windows. Also scaling window itself was affected.
Old method: scaled region widths based on area/editor scaling factors.
That was leading to too small or too large button regions easily.
New method: region width/height now are in DPI control. Much nicer!
- On changing dpi, buttons remain visually same widths.
- On changing window sizes, the button views and zooms stick to exactly same.
Caveat: people who were using Blender with 'extreme' dpi setting, might find
the layouts slightly differ. Not sure if this is worth version patching...
Todo: overlapping regions that overlap together draw badly. Fix underway.
- Include RNA properties when checking for matches.
- Don't include the context's property store (these are normally set by the UI code and not accessible by a script author)
Note: added CTX_data_dir_get_ex() which has options for returning different members from the context.
* Cleaning up the conversion code to avoid a per-face material conversion. Materials are now stored in buckets and only converted if a new material is found. This replaces some of Campbell's earlier work on the subject. His work wasn't as thorough, but it was much safer for a release.
* Shaders are only compiled for LibLoaded materials once. Before they could be compiled twice, which could really slow things down.
* Refactoring the rasterizer code to use a strategy design pattern to handle different geometry rendering methods such as immediate mode, vertex arrays and vertex buffer objects. VBOs are added, but they will be disabled in a following commit since they are still slower than vertex arrays with display lists. However, VBOs are still useful for mobile, so it's good to keep them around.
* Better multi-uv support. The BGE should now be able to handle more than two UV layers, which should help it better match the viewport.
- Exported via Inkscape a 16 and 32 pixel bitmap version
- Use these as mipmap levels for OpenGL texture drawing.
- Changed code to get right sizes for drawing icons - better than last week's method.
Todo:
- Custom icons don't work yet (old one)
- Missing icons in the svg
- The .sh script for inkscape needs changed to support this
(now do manual saving)
Also fixed CTX_FOO_ macros when building without i18n, those were kinda wrong.
And hid i18n ui section in userpreferences when built without its support too.
