Commit Graph

66005 Commits

Author SHA1 Message Date
059b7a81e2 Cycles: Implement constant fold for the ConvertNode.
This way socket type conversions (such as color to float, or float to vector) do not stop the folding process.
Example: http://www.pasteall.org/pic/show.php?id=96803 (selected nodes are folded).
2015-12-23 21:48:19 +01:00
1a2b5d9a8b Fix a few warnings with Apple LLVM 7.0.2. 2015-12-23 20:39:03 +01:00
93c75bf992 Fix T47051: cycles viewport textured shadeless draw bug. 2015-12-23 20:38:58 +01:00
eb6a8e4321 Update Link to the Blender Store. 2015-12-23 17:53:36 +01:00
083d6d63c8 Fix weld edges into faces splicing verts that shared a face
This could happen with degenerate faces.
2015-12-24 03:36:06 +11:00
6732da8340 Fix weld edges into faces - eternal loop
Would happen with complex edge-nets mixed with faces.
2015-12-24 03:36:00 +11:00
b9dea0f941 Remove BLI_buffer calloc option
This wasn't working reliably (after clear for example), and wasn't used anywhere.
2015-12-24 01:52:54 +11:00
ac782ba31c Fix BLI_buffer_reinit not clearing static flag 2015-12-24 01:51:45 +11:00
9c7d2ff62f Fix error in BLI_buffer_reinit
alloc_size was set incorrectly causing arrays not to be resized correctly.
2015-12-24 01:38:59 +11:00
548eb9eb4b Cycles: Sort tiles in rendering order at construction time
This commit modifies the TileManager to sort render tiles once after tiling the image,
instead of searching the next tile every time a new tile is acquired by a device.

This makes acquiring a tile run in constant time, therefore the render time is linear
w.r.t. the amount of tiles, instead of the quadratic dependency before.

Furthermore, each (logical) device now has its own Tile list, which makes acquiring
a tile for a specific device easier.
Also, some code in the TileManager was deduplicated.

Reviewers: dingto, sergey

Differential Revision: https://developer.blender.org/D1684
2015-12-23 13:14:36 +01:00
9e6a22c7e0 Fix bevel RNA enum/boolean mixup 2015-12-23 22:37:24 +11:00
36f916f200 Math Lib: clamp closest_to_line_segment_v# when segment has no length
For a line this makes sense but segments should clamp,
avoids assert in edge-rip.
2015-12-23 22:08:15 +11:00
b06f004d5b Fix issue w/ transform orientation & vert selection
Accessing selected verts as an array failed when only
some of the selected verts were in the selection-history.
2015-12-23 21:31:22 +11:00
ee9d60b033 Remove feof check for file reading
This wasnt correct since we're not reading past the stream,
though it worked on Linux and some Windows systems.
2015-12-23 17:45:46 +11:00
e74a023309 Disable calculating manipulator while transforming
Gives ~10% speedup
2015-12-23 16:35:13 +11:00
44b593ae2d Add BM_vert_edge_pair utility function 2015-12-23 16:08:23 +11:00
b51f730b0c Fix error in BM_vert_is_edge_pair
Returned true for verts with a single edge.
2015-12-23 16:08:20 +11:00
b3e94791d3 Typo in last commit 2015-12-23 15:11:54 +11:00
21b91bc764 Select Linked: only use seam delimit for faces
Delimiting on seams was only ever intended for face selection,
Previously this option didn't work for vertices,
now it's fixed the defaults aren't right for vertex/edge select-linked.

Add a workaround that bypasses operator-defaults - since this is such a rare case.
2015-12-23 15:08:58 +11:00
071a69fc95 Fix T47003: OpenGL draw missing selection highlight when using hidden wire. 2015-12-22 20:56:36 +01:00
eabf79e40a Get rid of OMP in MOD_build.
Reasons:
  - Only parallelized piece of code gives little local speedup (code block only about 25% quicker even on 1M polys cube).
  - No gain nor loss using new BLI_task system.
  - At 10% of build, parallelized piece of code is only about 5% of total func runtime (run-time explodes as build proportion increases).

See no point in adding (in utmost best optimistic case, in real use-case, when  depsgraph will likely already fire several evaluations in parallel,
speedup would be even smaller) 1% speedup here at the cost of threading complexity...

Note that since later code uses hashes, I don't think it's easy to thread it, so think we can leave with it for now.
2015-12-22 20:13:50 +01:00
b176206b94 View3D: measure smoothview angle comparing quat angles
Rolling the view registered as no angle-difference.
2015-12-23 05:23:04 +11:00
3e903794f4 View3D: Rotate around selected support for view-roll 2015-12-23 05:23:04 +11:00
642b77ef0e Fix T31605: Nupad ignores Rotate around selection
Add support for smooth-view orbiting a point besides the view center.
2015-12-23 05:04:52 +11:00
0241e28049 Fix error orbiting out of camera + orbit-selected
Would use the wrong view center (that wasn't update from the camera).
2015-12-23 04:09:15 +11:00
c86395c02d View3D: use c99 compound literals for smoothview params
Smooth-view functions took many arguments which were often NULL,
now take struct instead.
2015-12-23 02:50:50 +11:00
887829e3d8 Cleanup: Remove comments, now that we have the dedicated constant fold functions, it's obvious. 2015-12-22 14:02:57 +01:00
377b52be2e Cycles: Constant fold for the Gamma Node. 2015-12-22 13:53:13 +01:00
ba82981a2f Rename mesh operators
- use 'Intersect' prefix for boolean and regular intersection.
  hints that they use the same basic logic with different modes.
- 'split by edges' while correct - isn't very descriptive.
2015-12-22 17:12:32 +11:00
6d10bf7c44 Fix bad selection after symmetrize
Weld verts now keeps original faces so callers don't need to keep track of newly created face data.
2015-12-22 16:41:01 +11:00
a9c881f6a3 BMesh: store stackdepth as an index
Avoids -1 all over.
2015-12-22 16:41:01 +11:00
Dalai Felinto
ee2b583a05 Cycles-Bake: Fix Uninitialised value created by a heap allocation
This was causing some random black bakes. It was introduced in of of the commits replacing calloc by malloc. Things should be better now.

This error was only noticeable when baking the selected objects (not when baking from selected to active).
2015-12-21 15:31:48 -02:00
39d11352e0 UI: fix minor glitch copying small float value
Would copy 0.0 when button displayed 0.0000001.
2015-12-21 19:26:15 +11:00
5fef3c3326 BLI_storage: Split text/binary version of mem-from-file funcs
Fix T47022, caused by own commit de0119d08
2015-12-21 18:16:14 +11:00
46af314bd9 CMake: disable warning 4146
This is warning on noisy (false positives).
2015-12-21 14:54:36 +11:00
10cf7499e7 Fix T47023: Boolean w/ empty mesh didn't transform 2015-12-21 14:41:05 +11:00
580067769c Cleanup: remove unused file 2015-12-21 13:13:52 +11:00
32be51dc66 Cleanup: warnings (msvc)
Part of patch D1670 by @LazyDodo
2015-12-21 13:02:38 +11:00
6f22494140 NewDepsgraph: Fix typo tagging wrong flag.
Tagging NodeTree->flag instead of NodeTree->id.flag, not sure if this fixes something, but...
2015-12-20 23:06:18 +01:00
fdcec62dad Cleanup: invalid comment & style 2015-12-21 00:14:41 +11:00
008e7701c5 Buildbot: Cleanup unused code
- Remove deprecated/unused builders
- Remove unused SCons OSX slave configuration
- Remove SCons slave logic, it is not giving error about unknown building
  system used for the slave.
2015-12-20 15:44:53 +05:00
db6da8af85 Buildbot: Make sure optiosn override are passed after the config file 2015-12-20 15:32:59 +05:00
f01714506a Buildbot: Disable CUDA binaries for 32bit OSX slave
Funny thing, 32bit OSX slave was always building 64bit CUDA kernels
and nobody never noticed this..
2015-12-20 15:19:02 +05:00
3566d0d3a6 Buildbot: Update master config file 2015-12-20 14:49:54 +05:00
2fff2f651e Buildbot: Disable some unsupported features for 32bit OSX CMake slave 2015-12-20 14:22:34 +05:00
5a5b540356 Buildbot: Fix wrong architecture passed to CMake 2015-12-20 14:12:43 +05:00
101341845c Buildbot: Use proper archive name for CMake OSX slave 2015-12-20 14:12:04 +05:00
03d9f9a619 Buildbot: initial work to enable CMake OSX slave 2015-12-20 13:59:04 +05:00
7ac5f2fef6 Fix T47016: curve tilt value in N (redo) panel to much precise with mouse dragging.
Let's just use dafaults of angle props here. :)
2015-12-19 16:58:30 +01:00
6703c1df4e Fix shadowing var in own previous commit. 2015-12-19 11:04:56 +01:00