If the "Framing" mode is set to extend,
the camera frustrum changes when you
resizes the blenderplayer window.
Before this patch, there were no way to
control which part of the framing you want
to extend (vertical, horizontal or arbritary).
Now:
If the camera sensor fit is set to HORIZONTAL,
the horizontal field of view doesn't change.
If set to VERTICAL, the vertical fov doesn't change.
If set to AUTO the old behaviour takes place, arbitrarly
showing more of the horizontal or vertical field of view
depending on the aspect ratio of the window.
Test file:
https://svn.blender.org/svnroot/bf-blender/trunk/lib/tests/gameengine/framing_extend.blend
Bugfix supported by NF-UBC Nereus Program as part of the development
of OceanViz/NereusViz
Problem is that the scanfill result gives poorly shaped triangles, which is fine
for OpenGL drawing but does not work well for unwrapping. Another problem was
that it was doing random float offsets to avoid some degenerate cases, but this
meant that unwrapping a part of a mesh or the whole thing could give different
results for a UV island. Now UV unwrap does own triangulation of ngons.
window, the game would stop drawing in the first and mess up the OpenGL state of
the second.
Also fixes glPushAttrib/glPopAttrib getting out of sync in some cases.
Error in shadowbuffer with halo effect. It had render error when you look straight into
the lamp, with something in front of lamp.
The code that was sampling the shadowbuffer exited too soon (for ray going perpendicular to buffer it should take at least 1 sample).
Code from 90ies. And still got a bug fix. Whoa :)
Tiny tweak in "Any key modifier" code. It now refuses to accept events when you hold
any key, and then press another key - unless that has key-modifier set.
This enables Houdini style "Hold Spacebar - Leftmouse" view moves.
* Index was assigned after increment, leading to NULL pointer access later on when looking up socket from list
* Copying default_value requires a valid NULL pointer for clean check
* Was using default_values as parameters instead of sockets, void pointers just passed through without warning ...
It's more like an ownership thing, but it's not so much
clear who actually is an owner of temporary scene, so
for now just ensure Render does not have link to it
after strokes are rendered.
There're also suspicious thing related on why separate
RE_free_database is needed for freestyle renders, but
that's also another story.
To make the type-dependent socket linking in this template a bit more manageable, there are now separate functions that generate "items" for a particular node type consisting of the socket index, name and possibly additional properties required (currently only node groups vs. all other types). This is still nowhere near flexible enough to be used as a generic template all node systems, but works for now.
In order to make this usable as a generic tool we will have to devise a way of storing, comparing, applying node settings *outside of actual node instances*. Then each node needs to tell how sockets are generated based on these properties. For the current nodes this would be far too complicated ...
* Remove difflib's quick_ratio() call, from quick tests looks it is four times slower than ratio(), and yet it gives the same results...
* Moved code to find all po's in a dir into its own func (will be used by i18n addon tools too).
With extreme narrow scaled editors, the slider/mask code in View3d could deliver
zero sized or invalid window matrices.
Needs confirmation from Sergey if it works :)
This means main database is no longer pollutes with
temporary scene and objects needed for freestyle
render.
Actually, there're few of separated temporary mains
now. Ideally it's better to use single one, but it's
not so much trivial to pass it to all classes. Not
so big deal actually.
Required some changes to blender kernel, to make it
possible to add object to a given main, also to
check on mesh materials for objects in given main.
This is all straightforward changes.
As an additional, solved issue with main database
being infinitely polluted with text blocks created
by create_lineset_handler function.
This fixes:
- #35003: Freestyle crashes if user expands objects in FRS1_Scene
- #35012: ctrl+f12 rendering crashes when using Freestyle
internal improvement to editmesh_bvh.c
- optionally pass cage-coords as an arg, rather then calculating the coords in BKE_bmbvh_new(),
since all callers already have coords calculated.
- de-duplicate coords creation function from knife and bmbvhm, move into own generic function: BKE_editmesh_vertexCos_get()
Now there is a single BVH traversal code with #ifdefs for various features.
At runtime it will then select the appropriate variation to use depending if
instancing, hair or motion blur is in use.
This makes scenes without hair render a bit faster, especially after the
minimum width feature was added. It's not the most beautiful code, but we can't
use c++ templates and there were already 4 copies, adding 4 more to handle the
hair case separately would be too much.
MinGW-w64 headers (where _stat is actually _stati64), but since we
recommend a specific compiler build it's OK for now. Also tweaked other
places where _wstat is used.