Debug code for graphviz output moved to a dedicated file COM_Debug.h/cpp.
The DebugInfo class has only static functions, which are called from a number of places to keep track of what is happening in the compositor. If debugging is disabled these are just inline stubs, so we
don't need #ifdefs everywhere and don't get any overhead.
The graphviz output is much more useful now. DebugInfo keeps track of node names in a static string map for meaningful names. It uses a number of colors for various special operation classes.
ExecutionGroups are indicated in graphviz with clusters.
Currently the graphviz .dot files are stored in the BLI_temporary_dir() folder. A separate dot file is generated for each stage of the ExecutionGroup scheduling, this is intended to give some idea of the
compositor progress, but could still be improved.
enum is an int, and values from CD_PREVIEW_MLOOPCOL (32) onwards
will not give what apparently was expected.
Found with help from Oscurart in #blendercoders while trying to figure
out compile error with CMake/msvc9. Occasionally it is good to heed
the warnings given by compilers!
rendering, in case some script wants to set it in the render_pre callback. In
case of decoupled 3d view layers or local view it will still override this
though.
node.inputs collection, the node.file_slots or node.layer_slots collections should be used. Both of them work, they just provide slightly different properties for use with simple files or multi-layer
EXR.
The FileOutput node design is cumbersome and should be considered broken. It should eventually be replaced, the problem with that is backward/forward compatibility.
Throw a python error if user attempts to use CurveMap without calling
CurveMapping.initialize() first. Added access to the initialize function
to CurveMapping on RNA level.
Thanks to Campbel for the help and remarks!
Disabling display lists for legacy ATI cards since they don't support display lists well.
Also removing an unused variable from the display list rasterizer.
The chunk indices for scheduling chunks based on a given area were calculated incorrectly. This caused chunks at the very border of the render (pixels 256..257) to be omitted, leading to incorrect values
in the Z buffer of the test file, which in turn caused wrong normalization range and the resulting almost-white image.
Also added a dedicated executePixel function for Z buffer to avoid any interpolation of Z values.
don't give up too early finding unique geometry by keeping track of edges with matching vertex hashes and keep checking for unique topology as long as the total of these edges keeps increasing.
- make cmake osx use of -ftemplate-depth match scons.
- use array size within sizeof(), more compact.
- replace AT with __func__ where the function is unique enough.
- BLI_box_pack_2D -> 2d to match other functions.
- rename new mesh normal calculation to mesh.calc_normals_split()
forgetting to change it from RGB when painting images or saving renders.
This makes more sense in the new alpha pipeline, renders no longer contain an
alpha channel that does not match the RGB channels, so saving files as RGBA
should give the expected results when opening them in other applications.
