| Age | Commit message (Collapse) | Author |
|---|
|
used locally.
Will be used by link/append code.
|
|
Switch from a fixed stack to a linked list,
since guessing the maximum possible size may fail with invalid/overlapping strips.
|
|
|
|
Many checks for out->rect_float being [non-]NULL are done back-to-back.
Combining them into a single check for slightly more efficient code and
less code clutter for easier readability/understanding.
Differential Revision: https://developer.blender.org/D2097
|
|
And use __func__ macro.
|
|
|
|
Applying cflags globally can be problematic especially with extern, intern libs.
Now flags from target named will be used when defined,
allowing for developers to define flags for modules they maintain.
Convention is CMAKE_CFLAGS_${UPPERCASE_TARGET_NAME}, (CXXFLAGS for C++).
eg: CMAKE_CFLAGS_BF_BLENDER, CMAKE_CFLAGS_MAKESDNA, CMAKE_CXXFLAGS_CYCLES_KERNEL
On Linux run `make help` for full list of names, MSVC shows these in the solution.
|
|
Accessing `bpy.app.binary_path_python does search, then swaps its getset with the string it finds.
This caused a freed pointer to be stored in bpy.app's dictionary.
Fix by using the same string for get/set access.
|
|
|
|
Thanks to Kévin Dietrich for spotting driver leak,
checked other uses of PyDict_SetItem and found more.
|
|
No need to re-convert from C strings to PyUnicode.
|
|
Recently the snap to edges has been changed to work with coordinates 2d. Thus, the lambda of the snap pointon on edge becomes different from the actual lambda in a 3d edge.
The solution to correct the lambda was this:
```
if (is_persp) {
const float fac = depth_a / (depth_a + depth_b);
lambda *= (1.0f + (fac - 0.5f) * (1.0f - lambda));
````
...
But the lambda was still not very accurate.
Now the function uses the `dist_squared_ray_to_seg_v3` to get the coordinate of the snap point on the edge. And then tests that point.
The accuracy returned to normal values.
|
|
|
|
|
|
from D2104: reference all repeated (and strange) equations (example: mul_m4_m4m4(pmat_local, pmat, obmat)) in the function `precalc_project.
This is useful for maintenance.
|
|
Error in recent snap changes.
|
|
Caused by typo in rBd2312602125a452e6562a76ab91779943c67396d
|
|
|
|
Current code changes the selection then then runs ED_armature_sync_selection
to copy to the bones rootsel flag from the parents tipsel.
Use this logic for entering edit-mode too.
|
|
There are still issues with overlapping geometry,
however some of the issues reported are are causing problems,
or fail entirely with Carve too.
|
|
Fixes possible missing update of proxy_from pointer before using
it in relations builder.
|
|
This reverts commit 47d0d9cca4d0c3ccbdc368e97fc24652379fd368.
Reverting the commit. Not only it did not solve all the cases of proxy popping,
but also broke real cases with single proxy involved.
|
|
This solves bug when same object is instanced from multiple bases.
|
|
The snapping functions when performed in the perspective view,
have some problems in the threshold (a distortion) and in the clip plane (the normal is incorrect).
These problems can be only observed when making the snap to edges or to vertices (nearest to ray function).
This patch propose a totally different solution.
The idea is to project the edges of bvh nodes and test the 2d projection of the snap element.
For this it used the BLI_bvhtree_walk_dfs function.
It is important to pay particular attention also to the changes in `ED_transform_snap_object_project_view3d_ex`
|
|
D1875 by @lichtwerk
|
|
icon' from non-ID icon draw code.
Now using generic icon rendering system in that case too, instead of custom code
which was only handling 'deferred' custom file-loaded icons.
|
|
- icon_id from ID and PreviewImage were not guaranteed to be in sync.
- PreviewImage one was not reset on file read.
- Through RNA e.g., it was possible to ensure an ID icon via its preview image,
which was running code designed for custom previews/icons system, instead
of generating correct 'auto ID icon'.
|
|
This reverts commit f12204196fb1ee985ab9745cf9c70877601145f9.
Campbell, sorry. have to revert this for the time being.
We've missed some very important bits, such as:
- FFmpeg is usually linked against OpenJPEG
- OIIO needs OpenJPEG as well.
For FFmpeg issues we can either disable OpenJPEG there (since
we don't really use it), or bump FFmpeg to version 3.1.1 which
can use either of OpenJPEG 1.5 or 2.1.
For OIIO we do need OpenJPEG support (otherwise Cycles will
not be able to use j2k/j2c textures) and currently there is
NO solution to make OIIO working with OpenJPEG 2.1.
According to Matthias Fauconneau (aka mfv) Larry is working
on the patch to get OIIO work with OpenJPEG 2.1, but it'll
take some time still.
I've tried to look into support of some sort of build system
flag and do ifdefs, but it all becomes quite nasty, especially
with bundled OpenJPEG bumped to 2.1.
Surely such an update is something we'll have to apply to
but at this exact moment it causes quite some pain for all
developers.
Suggest to wait for until OIIO supports OpenJPEG 2.1 and then
go with the updates for real.
|
|
|
|
Stream handling has changed so this required changes to how files & memory are accessed.
|
|
Quality was defaulting to lowest and no codec was set.
|
|
Fix from @cheleb
|
|
Avoids decoding the SDNA string every undo step.
|
|
This was printed to the stdout, however the error case wasn't checked or well supported.
Also, errors decoding SDNA would sometimes call exit(1).
|
|
- Move last-hit index out of SDNA struct
(allows for access by multiple threads).
- Replace O(n^2) search with hash lookup in DNA reconstruction.
|
|
Idea looked good, but we have too much custom situations here (some half-fake-sub-ID
being copied with their 'owner', animdata, etc.), let's let datablock copy functions
handle that themselves.
Also allows to safely call BKE_id_expand_local from all copy functions now (only when
copying linked data).
|
|
This one is already called by matching copy functions, no need to call it twice!
|
|
|
|
image/material/texture/world.
As said in previous commits, did not touch to copy functions for now, due to ntree issues...
|
|
ntree is once more time a PITA - actually, all sub-IDs that get copied together with the 'main' one are
(shapekey was workedaround, as was animdata/action issue, but nodes are more touchy).
For now, better not to touch that, needs careful check & rethink.
|
|
Did not changed ntree copy for now, this func is a mess of expand/non-expand/bmain/non-bmain cases...
Keep it for later!
|
|
functions of actions, brushes and particles.
This greatly simplifies said code, once again no change expected from user PoV.
|
|
functions of obdata
(armature, mesh, curve, mball, lattice, lamp, camera, and speaker).
This greatly simplifies said code, once again no change expected from user PoV.
|
|
datablock in its usercount handling.
Reason is, typically those actions datablock usercounts have already been adjusted in BKE_animdata_copy()
(called by generic iddata copy function).
Think this needs to be reworked a bit too, there are way too much ways to copy animdata currently,
it's rather confusing. But not the goal here, so for now we'll live with the hack!
|
|
preview image.
Also cleaned-up/synchronized code across al ID types that support preview.
|
|
will be used by both make_local() and copy() datablock functions.
Note that this new func make 'extern' all IDs used by localized datablock,
not only refcounted ones as it used to be in each type's functions (with a few exceptions).
|
|
Missed from c5b2f12b
|
|
|
|
|
|
|
