Previously, the global storage of these types either used a GHash or a
blender::Map. VectorSet is preferrable to GHash because it's type safe,
clearer, and faster. It's preferrable to Map because the key doesn't
have to be duplicated and because iteration is faster.
This PR moves these registered types to VectorSet, just like the
node, node socket, and node tree types. Note that none of these types
use RAII for allocation, so freeing is still done manually.
Testing was manually interacting with each of these systems, including
with addons that register their own types.
Pull Request: https://projects.blender.org/blender/blender/pulls/133778
This was caused by the subpass input workaround for non-tilebased
GPU using `texelFetch` on an `image`. This was supported before
the cleanup 9c0321ae9b.
But is against the GLSL specification and was removed inside the
cleanup.
Using `imageLoad` instead of `texelFetch` fixes the crash.
However rendering seems to be broken for other reasons.
Currently the "Show All" layer operator operator only enables visibility
for layers only. If the Grease Pencil object has layers that are children
to a Layer Group that is hidden, these layers would not be revealed
to the user.
This pull request modifies the grease_pencil_layer_reveal_exec function
to iterate through Layer Groups present in the object and sets them visible.
Pull Request: https://projects.blender.org/blender/blender/pulls/133613
10/12bpp movie writing code assumes that input image is properly
floating point for these cases. That is always true for regular
rendering, however VSE can produce 8bpp images, if there
are no float/HDR strips in there. Make these convert into a float
image in the render pipeline when needed.
Pull Request: https://projects.blender.org/blender/blender/pulls/133802
Was just getting the default name, but if we also have a bone selected
in weightpaint mode, these should be named after the bone (otherwise the
armature modifier does not deform accordingly).
To resolve, basically do the same we do for meshes in
`ED_wpaint_ensure_data`.
Pull Request: https://projects.blender.org/blender/blender/pulls/133804
This adds another operator mode for changing brush radius and opacity while the
modal primitive operator is running. It's based on the `wm.radial_control`
operator but only implements simple mouse control for these two properties.
Pull Request: https://projects.blender.org/blender/blender/pulls/133675
In renderdoc the debug stack got corrupted when render graphs where
reused. The previous usage didn't clear the stack. This PR clears
the debug stack when render graphs are reset.
MSVC does this by default, and is a good way to see build progress
indication, when using msbuild. clang-cl also got that option back
in 2018, but it is off by default.
Pull Request: https://projects.blender.org/blender/blender/pulls/133810
Box select is broken for all node editors. It ignores selection modes and always uses the mode `Set a new selection`.
The operator `NODE_OT_select_box` works fine, but clicking anywhere in the node editor calls `NODE_OT_select` first, so we need to make sure it has no undesired side effects when it is not run.
Pull Request: https://projects.blender.org/blender/blender/pulls/133736
This patch adds support for drag and dropping colors (from color buttons) into
the Compositing, Shading and Geometry node trees.
Additional support was added for dragging and dropping colors with an Alpha
component (which was previously ignored), both in the context of the Node Editor
and for Color Buttons in general. This handles cases like drag and dropping a
color from an RGB to an RGBA button, which recreates the color with a default
Alpha value of 1.0.
Pull Request: https://projects.blender.org/blender/blender/pulls/129026
Changes to the compositor Normal node does not propagate and the
compositor result is not updated. That's because the node is a special
case since its output socket value is used as its input property, which
is not tracked by the node update code.
Apparently, a similar node exists in shader nodes, and a workaround is
implemented for that node in the node updater, so we just need to extend
that check to include the compositor node as well.
Pull Request: https://projects.blender.org/blender/blender/pulls/133803
When developing and testing add-ons it can be useful to get to the
folder containing such add-on.
When Developer Extras is enabled, show a folder icon next to the
extension's path to quickly open it using the system file browser.
Also display it in the list of duplicate add-ons, if any, so user can
access them more conveniently.
Pull Request: https://projects.blender.org/blender/blender/pulls/128474
Refactor the keyframe copy-paste code to prepare for supporting action slots.
This is a non-functional change, as the slots are stored but otherwise completely ignored. This way the refactor to the new data structure can be reviewed & tested before actually changing the behaviour.
Related: #129690
Pull Request: https://projects.blender.org/blender/blender/pulls/133497
There were three methods with succinct but slightly unclear names:
- Action::slot_move()
- Channelbag::channel_group_move()
- Channelbag::fcurve_move()
The ambiguity is due to other functions with similar names that e.g.
move fcurves *between* Channelbags, etc. Whereas these methods only move
items to other positions within the array they're already in.
To make this clearer, this PR adds `_to_index` to the end of these
array-oriented methods.
Pull Request: https://projects.blender.org/blender/blender/pulls/133730
The current poll function uses `ED_view3d_context_user_region` [which
falls back to `ED_view3d_area_user_region` if the context region is
flagged `RV3D_LOCK_ROTATION`, so it finds the non-locked region in quad
view]. So in a way it is permissive to do the roll even when we are
within one of the quad regions which is actually locked. This can be
useful if we just want to roll by a specific angle or use one of the
types `V3D_VIEW_STEPLEFT`/`V3D_VIEW_STEPRIGHT` - these are then executed
in the non-locked region.
However for mouse-interacting roll, we should not do this at all when
initiated from a locked region.
To resolve, we could use `view3d_rotation_poll` (instead of
`ED_operator_rv3d_user_region_poll`), but that would prevent us from the
useful scenarios mentioned above. So instead, once we reach the mouse-
interacting code path, check again if we are really in a non-locked
region and only proceed if we are.
Pull Request: https://projects.blender.org/blender/blender/pulls/133643
This commit reworks the RenderReport base class to avoid adding
`--cycles-device` device arguments to non Cycles tests.
This reduces some warnings that can show up with EEVEE and
Workbench tests that accidentally used these arguments.
Pull Request: https://projects.blender.org/blender/blender/pulls/133724
This commit checks the total number of files being processed for
rendering tests and causes it to emit a failure if none are found. This
has the benefit of loudly failing if there is a configuration error
instead of succeeding.
Pull Request: https://projects.blender.org/blender/blender/pulls/133691
This pairs with the recently added Export test and reuses the same
.blend file. While there happens to be an instancing related rendering
test, that happens to cover Import, it's better to have a direct test
for it.
Future improvements to both Export and Import to account for the recent
"hidden layer" visibility bug would be good to do as a follow-up.
Pull Request: https://projects.blender.org/blender/blender/pulls/133782
That version has a bunch of API changes, so by dropping support for older
versions we can remove old compatibility code.
Also, that version is required for OptiX support, so building a fully-featured
Cycles wasn't possible with older OSL anyways.
Pull Request: https://projects.blender.org/blender/blender/pulls/133746
- There's only a few unit conversion options, just test all of them
- Use reference instead of pointer when passing export settings struct
- Organize scaling struct fields to keep similar options together
Pull Request: https://projects.blender.org/blender/blender/pulls/133774
Artists often want to quickly switch back and forth between two or more nodes while compositing.
This patch implements two operators `NODE_OT_viewer_shortcut_set` and `NODE_OT_viewer_shortcut_get` that allow users to map a viewer node to a shortcut. For example, pressing `cltr+1` while a node is selected, assigns that node to the shortcut `1`, creates a viewer node if it has none attached and sets that viewer node to active. Pressing `1` will set the active node with shortcut `1` to active.
Shortcuts are saved in DNA to preserve them after saving/loading blend files.
Limitations:
- Only compositor node tree is supported, because shading editor has no viewer node and geometry nodes viewer works differently.
Pull Request: https://projects.blender.org/blender/blender/pulls/123641
Adding an image object using drag and drop can fail without an apparent reason.
This patch shows an error when this happens so that the user can react accordingly (in this example switch to object mode, add the image and switch back to edit mode).
Pull Request: https://projects.blender.org/blender/blender/pulls/132994
In certain setups, nodes whose inputs are single value and whose outputs
are expected to be single value wrongly return an image. That's because
they wrongly join a pixel operation that operates on images.
To fix this, we split pixel operations by their value types. Single
value sub-trees get compiled into their own pixel operation and none
single value sub-trees get compiled into their own pixel operation.
This might unfortunately break up a long chain of pixel operations if
one of them was single value. This is solvable, but will require more
time and research, so we need to fix the bug first then look into it.
Pull Request: https://projects.blender.org/blender/blender/pulls/133701
Regardless of what mem info reports. We can't move this to the host, so
might as well try because the free memory might not be a reliable predictor
of success.
Pull Request: https://projects.blender.org/blender/blender/pulls/132912
With host mapped memory these can be shared, and we can't get back the
original host pointer unless we make a copy which is inefficient.
Also add asserts to verify this doesn't happen.
Pull Request: https://projects.blender.org/blender/blender/pulls/132912
