Summary:
Made objects update happening from multiple threads. It is a task-based
scheduling system which uses current dependency graph for spawning new
tasks. This means threading happens on object level, but the system is
flexible enough for higher granularity.
Technical details:
- Uses task scheduler which was recently committed to trunk
(that one which Brecht ported from Cycles).
- Added two utility functions to dependency graph:
* DAG_threaded_update_begin, which is called to initialize threaded
objects update. It will also schedule root DAG node to the queue,
hence starting evaluation process.
Initialization will calculate how much parents are to be evaluation
before current DAG node can be scheduled. This value is used by task
threads for faster detecting which nodes might be scheduled.
* DAG_threaded_update_handle_node_updated which is called from task
thread function when node was fully handled.
This function decreases num_pending_parents of node children and
schedules children with zero valency.
As it might have become clear, task thread receives DAG nodes and
decides which callback to call for it.
Currently only BKE_object_handle_update is called for object nodes.
In the future it'll call node->callback() from Ali's new DAG.
- This required adding some workarounds to the render pipeline.
Mainly to stop using get_object_dm() from modifiers' apply callback.
Such a call was only a workaround for dependency graph glitch when
rendering scene with, say, boolean modifiers before displaying
this scene.
Such change moves workaround from one place to another, so overall
hackentropy remains the same.
- Added paradigm of EvaluaitonContext. Currently it's more like just a
more reliable replacement for G.is_rendering which fails in some
circumstances.
Future idea of this context is to also store all the local data needed
for objects evaluation such as local time, Copy-on-Write data and so.
There're two types of EvaluationContext:
* Context used for viewport updated and owned by Main. In the future
this context might be easily moved to Window or Screen to allo
per-window/per-screen local time.
* Context used by render engines to evaluate objects for render purposes.
Render engine is an owner of this context.
This context is passed to all object update routines.
Reviewers: brecht, campbellbarton
Reviewed By: brecht
CC: lukastoenne
Differential Revision: https://developer.blender.org/D94
Summary:
This completly changes the way modal numinput is handled. Now, edited expression is a string, which then gets unit- and py-evaluated to get a float value.
We gain many power and flexibility, but lose a few "shortcuts" like '-' to negate, or '/' to inverse (if they are really needed, we still can add them with modifiers, like e.g. ctrl-/ or so).
Features:
- units (cm, ", deg, etc.).
- basic operations from python/BKE_unit (+, *, **, etc.), and math constants and functions (pi, sin, etc.).
- you can navigate in edited value (left/right key, ctrl to move by block) and insert/delete chars, e.g. to fix a typo without having to rewrite everything.
- you can go to next/previous value with (ctrl-)TAB key.
- As before, hitting backspace after having deleted all leading chars will first reset the edited value to init state, and on second press, the whole "modal numinput" editing will be cancelled, going back to usual transform with mouse.
Notes:
- Did not touch to how values are shown in header when modal numinput is not enabled (would do that in another commit), so this is still quite inconsistent.
- Added back radian support in BKE_unit.
- Added arcminute/arcsecond to BKE_unit.
(those unit changes affect all angle UI controls, btw, so you can now enter radians or longitude/latitude values when in degrees units).
Related to T37600.
Reviewers: brecht, campbellbarton, carter2422
Reviewed By: brecht, campbellbarton, carter2422
Thanks everybody!
Differential Revision: http://developer.blender.org/D61
Currently we don't properly support snapping + axis-constraints, but
there are cases where its handy to project-snap for eg, and constraint
to an axis so re-enable this.
The problem here was that the hack used to fix T24551 in r32863
would cause the camera to jump back to its pre-transform position
after motionpaths were recalculated. This made it seem like a keyframe
wasn't inserted by AutoKeying (even though it was).
- when 3 verts are selected ensure the normal is flipped on the side of existing vert normals.
also use the most distant 2 verts to define the tangent.
- when 2 vertices are selected, the normal wasn't aligned with the vert normal.
This adds vertex snapping capabilities for curves. Snaps to all control points
of other objects, and visible + selected control points and handles in curve
edit mode.
Reviewed By: brecht
Differential Revision: http://developer.blender.org/D3
confusion, grid snap is now the default as it seems to be the most wanted and easy to use mode.
Absolute grid snapping happens in a somewhat generic function 'applyGridAbsolute', which could also be used for objects and other transforms later on. It is conceptually similar to the 'project' snapping
option, in that it calculates a delta vector for each element on top of the overall transform, which places each node on the grid.
Node transform now uses the top-left node corner for TransformData->loc. The transform center is still the average of node centers, so that scaling and rotation works nicely.
snapGrid*** functions have been renamed to snapGridIncrement*** to distinguish better between incremental and absolute grid snapping.
It wasn't enabled in snapping code from the beginning it seems,
but from quick tests snapping for mballs works just fine.
Maybe we could drop out check for edit object type now?
Added two options to a header of FCurve editor:
- Normalize which makes it so every individual
curve is fit into -1..1 space.
- Auto-normalize, which probably is to be called
"Lock" which "locks" curve normalization scale.
This is useful to prevent curves from jumping
around when tweaking it.
It's debatable whether it need to be a button to
normalize curves n purpose only, and it's fully
depends on animator's workflow.
Here during Project Pampa we've got Francesco
who get used to auto-renormalization and Hjalti
who prefers locked behavior.
Docs are to be ready soon by Francesco.
Thanks Brecht for the review!
Remove usages of ANIM_unit_mapping_apply_fcurve in favor of
runtime scale factor apply.
There're still calls to ANIM_nla_mapping_apply_fcurve are
hanging around, they're the next t be cleaned up!
since enough bmesh operations can also take advantage of direct index lookups on verts/edges/faces.
developers note:
- EDBM_index_arrays_init/ensure/free -> BM_mesh_elem_table_ensure/init/free
- EDBM_vert/edge/face_at_index -> BM_vert/edge/face_at_index
- EDBM_uv_element_map_create/free -> BM_uv_element_map_create/free
- ED_uv_element_get -> BM_uv_element_get
