If there are images need to be saved, such as the model turnaround reference sheet and texture images, the document saving may be slow because of image encoding to png, now, all the images are pre encoded when the first load, so the document saving could be speed up by avoid the cost of image encoding.
Usually one pose consists of one frame, however, sometimes multiple frames per one reference sheet for a serial of action could be very useful, such as a sprite sheet. Multiple frames per pose is different with one motion, one motion could contains multiple poses. Currently, the duration of one frame is fixed to 0.042s, it's based on the 24 frames per second calculation.
This commit include lots of changes for pose editor.
Before, we use buttons to represent the transform controller for bones;
Now, we use nodes and edges to represent the bones, just like the nodes and edges for representing the guide spheres for mesh generating. There are advantages by doing this. Firstly, The edit UI for both mesh and pose are unified, secondly, it is possible to set a reference sheet for pose editing now, this is very important.
This new pose editor is inspired by the Eadweard Muybridge's work.
Refactor outcome. Move processes of outcome to separated files, and leave the outcome class as clean as possible.
Normal export in glb format is been fixed.
Ambient occlusion baker is been temporarily removed.
The normals haven't fix yet, this commit is mainly preparing for the next coming commits about animation export.
The glTF text format abandoned for the binary format.
Add a crc64 hash check for history keeping to avoid repeated snapshot.
Besides, before, there is a logic error in the undo implementation, which recover the most recent snapshot from undo stack, so user need undo twice and at the second undo get the expected result. This incorrect behavior got fixed.
Many source file name and data type are prefixed by "skeleton" before, now all are refactored.
The class MeshResultContext represents the outcome of mesh generation before, now we have material, pose and animation, so we just rename it as Outcome.
huxingyi