Revision 1dbece8 by Antonio Vazquez (greasepencil-experimental, greasepencil-object, greasepencil-refactor, soc-2019-npr, temp-gpencil-eval) March 24, 2019, 11:12 (GMT) |
Merge branch 'master' into greasepencil-object Conflicts: source/blender/blenloader/intern/versioning_280.c |
Revision d819361 by Jacques Lucke (builtin-simulation-nodes, functions, functions-experimental-refactor, particle-solver-dev, simulation-tree) March 24, 2019, 10:31 (GMT) |
simplify allocating tuple in prepared buffer |
Revision dc171b2 by Jacques Lucke (builtin-simulation-nodes, functions, functions-experimental-refactor, particle-solver-dev, simulation-tree) March 24, 2019, 09:41 (GMT) |
non recursive method to find required sockets |
March 24, 2019, 03:02 (GMT) |
Merge branch 'master' into sculpt-mode-features |
March 22, 2019, 15:16 (GMT) |
Automasking System: Topology automasking initial implementation This commit adds a general automasking system and the topology automasking option (not topology falloff, that can be added later). It should be faster and with fewer memory errors that the last version I made. It is designed to work with every brush and every stroke mode, but for now, I only added support for draw and grab to avoid changing every brush until the system is more polished. The main purpose of adding this in such an early stage of development is to check if the cursor is working correctly. The purpose of the vertex preview in the cursor was to show the active vertex for automasking, so now I can test if the whole system is behaving as it should. Limitations: - Only meshes for now (no dyntopo or multires). I still need to think about what to do with dyntopo. I don't know if brushes that generate new topology are going to work with this. - It only stores a binary mask (a vertex can be active or not). This could be changed if the future to support topological falloff or other more complex types of automasking. - As it is now, brushes that use stroke space update the active vertex dynamically, setting it to the closest vertex to the brush location each sample. This often causes masking errors with small brushes. I'm considering changing it to always check topology using the initial active vertex, but it is going to be much slower. |
March 22, 2019, 14:35 (GMT) |
Clang-format: make format fixes and performance improvements for Windows. Fixes by Ray Molenkamp. |
Revision e77a8b9 by Jacques Lucke (builtin-simulation-nodes, functions, functions-experimental-refactor, particle-solver-dev, simulation-tree) March 22, 2019, 10:57 (GMT) |
initial support for calling subprograms |
Revision cf72da1 by Jacques Lucke (builtin-simulation-nodes, functions, functions-experimental-refactor, particle-solver-dev, simulation-tree) March 22, 2019, 10:52 (GMT) |
fail early when node inserter does not work |
Revision 0efee91 by Jacques Lucke (builtin-simulation-nodes, functions, functions-experimental-refactor, particle-solver-dev, simulation-tree) March 22, 2019, 09:51 (GMT) |
rename tree type |
Revision 95ccb24 by Jacques Lucke (builtin-simulation-nodes, functions, functions-experimental-refactor, particle-solver-dev, simulation-tree) March 22, 2019, 09:21 (GMT) |
implement custom search terms for nodes |
Revision 46d35d0 by Antonio Vazquez (greasepencil-experimental, greasepencil-object, greasepencil-refactor, soc-2019-npr, temp-gpencil-eval) March 22, 2019, 08:05 (GMT) |
Merge branch 'master' into greasepencil-object |
Revision 24d89e1 by Jacques Lucke (builtin-simulation-nodes, functions, functions-experimental-refactor, particle-solver-dev, simulation-tree) March 21, 2019, 17:44 (GMT) |
better file naming |
Revision ad79af4 by Jacques Lucke (builtin-simulation-nodes, functions, functions-experimental-refactor, particle-solver-dev, simulation-tree) March 21, 2019, 17:40 (GMT) |
typo |
March 21, 2019, 17:16 (GMT) |
Clang-format: make format is now multithreaded. |
March 21, 2019, 16:12 (GMT) |
Clang-format: Wrap brace after Objective-C declaration Makes those behave similar to C/C++ functions. |
March 21, 2019, 16:09 (GMT) |
Clang-format: half-indent access modifier We definitely shouldn't use a regular indent for it. The half one (which is a single space in this case) might not be so well visible, but kind of helps reading that it is inside a class. |
March 21, 2019, 16:08 (GMT) |
Clang-format: Indent the constructor initializer Makes it easier to read what is an initialzier and what is the code, adding some space to the initialization. |
March 21, 2019, 16:04 (GMT) |
Clang-format: Don't line break after class One can think of a class as a structure with different default visibility. Point is: in C++ code it will be weird to see different rules for classes and structures. |
March 21, 2019, 16:03 (GMT) |
Clang-format: Make it easier to tweak running command Allows to compare behavior of various clang-format versions. |
Revision c6e4c29 by Jacques Lucke (builtin-simulation-nodes, functions, functions-experimental-refactor, particle-solver-dev, simulation-tree) March 21, 2019, 11:17 (GMT) |
implement concept of implicit conversions |
|