Problem was caused by unneeded call of CDDM_calc_normals when
creating derived mesh for object which duplicated mvert array.
It shouldn't happen unless object has got shapekeys or deformation
modifiers on the top of stack.
Previous fix wasn't correct. It made things working, but it removed
advantage of using PBVH for speed because derivedMesh used to be
recreated on each stroke step.
This commit fixes non-multires sculpting for Solid shading,
wireframe and render wouldn't work correct -- it'll require
a bit more tricky things because DM created for base mesh
doesn't points to mesh->mverts anymore, so sculpting and
drawing modules are using different verts arrays.
Added callback to drawMappedFaces which checks if two faces have got equal draw options.
After discussion with Brecht we found it's nicest solution for now:
- Disabling VBOs in edit mode for this case wouldn't be nicer for this case -
some additional flag stored in DM should be added in this case.
- Adding new callback in DM isn't nicer that this solution.
- Handling face selection in drawobject would lead to duplicated code
which is also not nice.
Hopefully, this callback could handle all cases in the future.
Also, Brecht mentioned current VBO implementation isn't perfect, so maybe
when we'll redesign this area dealing with edit mode wouldn't be so tricky.
This patch attempts to clean up and document the GPU buffers
code. There are a few bug fixes as well.
Patch reviewed here: http://codereview.appspot.com/4631052/
Summary:
* Bugfix: make GPU_buffer_copy_normal convert from shorts to floats
correctly, also fixed the use of cached face normal CustomData.
* Bugfix: changed the `mat_nr' field of GPUBufferMaterial from char to
short.
* Changed color buffer setup to not alloc a temporary copy of color
data, just passes the MCol data in directly.
* Changed the GPU buffer pool code to make clearer what operates
specifically on the global pool.
* Lots of refactoring for GPU_drawobject_new; should operate mostly
the same (except got rid of one unecessary allocation), just split
into more functions and without macros now.
* Converted some #defines into enumerations.
* Made some stuff private, pulled out of header file.
* Deleted unused function GPU_buffer_pool_free_unused().
* Removed GPU_interleaved_setup and related #defines. (I think this
was used for editmode VBOs, but those were disabled.)
* Added lots of comments.
* Added a few comments in the code signed `--nicholas' to note places
where I am unsure about design or usage, would be good to address
these better.
* Code formatting changed to be more consistent with the rest of
Blender.
* Renamed some fields and variables to be more consistent with
Blender's naming conventions.
* Renamed some fields and variables to use more descriptive names,
e.g. renamed `redir' to `mat_orig_to_new'.
* Removed print outs with DEBUG_VBO -- don't feel too strongly about
this one, just not used elsewhere in Blender, could be easily added
back if others disagree though.
* Moved the PBVH drawing code down to the bottom of the file, before
was sitting in the middle of the other VBO code
Committed changes from previous weeks, biggest changes are:
* Canvas can now have multiple "surfaces" that each can have specific format, type and settings.
* Renewed UI to support this new system.
* Aside from old "image sequence" output format, Dynamic Paint can now work on vertex level as well. Currently vertex paint and displace are supported.
* Integrated vertex level painting with Point Cache.
* Added viewport preview for Point Cache surfaces.
Due to massive amount of changes, old Dynamic Paint saves are no longer supported. Also some features are temporarily missing or may not work properly.
Removed the DerivedMesh face iterators (they sucked).
This should make subsurf faster. Also sped up multires
a bit (not sure if it's strictly correct, need to look
at it later).
- Constructive modifiers are enabled by default in sculpt mode.
- There's option to disable all constructive modifiers in the "Options"
panel of toolbox in sculpt mode,
- Use one column in options panel to make strings easier to read
- No modifiers would still be applied on multires
- made EXPANDED_AGRP take bAnimContext as an argument.
- remove unneeded NULL check drawFacesColored functions.
- comment some vars which are set but not used.
======================
All this work with sculpting on armatured/deformed mesh allowed to
implement sculpting on non-locked keys very easy -- just use the same
approach of propagating offsets from deformed PBVH to "sculpting layer".
- If key is locked, then old logic would be used.
- If there's multires modifier enabled, sculpting would happen on multires.
Multires interpolation is considerably better
now, though it still has a problem with occasionally
producing little random tangent spikes. Still, it's
far better then it was.
Also fixed a bug in dissolve faces.
into account that some tools use normals for things other than display. Now
we properly initialize vertex normals at flat faces too.
Also fixed a normal refresh issue, and deduplicated CDDM/mesh normal
calculation code.
Coded a new modifier, "Precision UV Interpolation",
that triangulates, subdivides, then uses brecht's mean
value interpolation to interpolate face data.
Textures on ngon faces get interpolated a bit nicer, in
other words (though concave cases, e.g. 'N', don't work very well).
weighted normals as the render engine, and the render engine will copy
normals from the mesh rather than always recalculating them.
Subsurf/multires still use regular vertex normals, but they are expected
to be sufficiently high resolution to not need this.
This means that normal maps displayed in the viewport actually match the
render engine exactly and don't have artifacts due to this discrepancy.
It of course also avoids unexpected surprises where your render normals
look different than your viewport normals.
Subversion bumped to 4 for version patch to recalculate normals.
Patch by Morten Mikkelsen, with some small changes.
Fixed view selected code. Made MDisps->disps use the
special allocation library I wrote as a
temporary fix to MDeformGroup->dw problems.
Why was the vgroup solution reused? It's really slow,
and its hard to tie in something like mempool or memarena.
The hackish allocator I wrote really needs to go, eventually,
or be folded into guardedalloc.
Also fixed a customdata bug with modifiers.
Mirror modifier no longer uses bmesh, but is now
100% CDDM. This is faster; the sluggishness was
driving me crazy. Still need to deal with UVs,
but that shouldn't be hard.
Fixed join meshes, and an annoying modifier bug (making modifiers
not work in editmode). Also fixed a tesselation bug.
Also got edge slide to work right, yay! Dunno why I couldn't get
it working right before; took me twenty minutes to fix.
the two files mikktspace.h and mikktspace.c. These are standalone files
which can be redistributed into any other application and regenerate the
same tangent spaces. The implementation is independent of the ordering
of faces and the vertex ordering of faces.
- use NULL rather then 0 where possible (makes code & function calls more readable IMHO).
- set static variables and functions (exposed some unused vars/funcs).
- use func(void) rather then func() for definitions.
