Summary:
Uses some magic pseudo-random which is actually a
texture coordinate hashing function.
TODOs:
- Dither noise is the same for all the frames.
- It's different from Floyd's dither we've been
using before.
- Currently CPU and GPU dithering used different
implementation. Ideally we need to use the same
dither in CPU.
Reviewers: brecht
Reviewed By: brecht
Differential Revision: http://developer.blender.org/D58
The trackpad would give isolated mouse pan events with a distance smaller than
one line or character. For other 2D views this is all accumulated in floats,
but in the text editor it only keeps track of integers, and the small mouse pan
events got lost due to rounding.
Now it keeps track of sub-line or sub-character scroll offsets to avoid this.
* world to object matrix is calculated for view vector as well, reuse
that.
* if no orientation exists use world negative axis instead of object
negative axis for gravity.
there's also the issue of mirroring, it may be wise to invert the
gravity effect there but it may be better to have artist feedback on
that.
When the entire label was clipped, the right aligned number component
could be clipped on its left hand side too.
So "Foo: 1.00" could be clipped to ".00"
Add Selected to Active Group presented a menu with groups to choose from, while
this seemingly inverse operation did not and used all groups of the active
object.
Now both operators present a menu, with as first option "All Groups".
So, after hours of wild hunt, I think both CMake and Scons builds work again.
There were several small issues hiding here and there, esp. for scons.
*WARNING* the --all-static is currently officially broken (it was already quite useless in fact).
I’m seriously considering dropping it completely, having a script compiling a chain
of static libs in a complete working way for *two* different build systems is just
too much work (static libs are *really* a nightmare at link time).
Builders who want to generate static Blender (to share it e.g.) will have to handle things by hand for now.
Main other points changed:
* Many changes to generated settings for scons (notably, I had to hack around the fact that the short ld -l
option does not work with lib names like 'Iex-2_1', which is not that surprising).
* Opencollada remains the only lib built statically (can't get it to build dyn currently).
* All other libs, even ILMBase ones, are now fully dynamic ones.
* Some general cleanup with CMake definitions of built libs.
* Git-libs do not clone again and again the repo anymore!
There were several issues with how bounding box and texture space
are calculated:
- This was done at the same time as applying modifiers, meaning if
several objects are sharing the same curve datablock, bounding
box and texture space will be calculated multiple times.
Further, allocating bounding box wasn't safe for threading.
- Bounding box and texture space were evaluated after pre-tessellation
modifiers are applied. This means Curve-level data is actually
depends on object data, and it's really bad because different
objects could have different modifiers and this leads to
conflicts (curve's data depends on object evaluation order)
and doesn't behave in a predictable way.
This commit moves bounding box and texture space evaluation from
modifier stack to own utility functions, just like it's was done
for meshes.
This makes curve objects update thread-safe, but gives some
limitations as well. Namely, with such approach it's not so
clear how to preserve the same behavior of texture space:
before this change texture space and bounding box would match
beveled curve as accurate as possible.
Old behavior was nice for quick texturing -- in most cases you
didn't need to modify texture space at all. But texture space
was depending on render/preview settings which could easily lead
to situations, when final result would be far different from
preview one.
Now we're using CV points coordinates and their radius to approximate
the bounding box. This doesn't give the same exact texture space,
but it helps a lot keeping texture space in a nice predictable way.
We could make approximation smarter in the future, but fir now
added operator to match texture space to fully tessellated curve
called "Match Texture Space".
Review link:
https://codereview.appspot.com/15410043/
Brief description:
http://wiki.blender.org/index.php/User:Nazg-gul/GSoC-2013/Results#Curve_Texture_Space
sockets passed to the nodetree.links.new function, otherwise this could
lead to invalid connections between different node trees! The
ntreeAddLink function has asserts for this, but this condition needs to
be checked in the RNA method calling it.
The problem here was that the hack used to fix T24551 in r32863
would cause the camera to jump back to its pre-transform position
after motionpaths were recalculated. This made it seem like a keyframe
wasn't inserted by AutoKeying (even though it was).
The problem is that the animation curves are first before object updates,
where BKE_pose_rebuild and proxy synchronize will undo the animation when the
proxy is evaluted for the first time after loading.
This is somewhat of a workaround but it's also useful to have the scene updated
once when opening in background mode, and it makes rendering from the UI and
background more similar and so more predictable to give the same result.
When you drag past the smallest size of the list, it switches back to auto-size mode.
This commit makes wider the gap between smallest size and auto-size activation,
previously it could switch to auto-size a bit too easily...