* Make wasm wrapper not depend on editor on native
* Introduce pkg-native
* Fix tests
* Fix fmt
* Correct span and make cargo fmt work inside editor_commands block
* Review
* Fix lint
* Allow using any graphics type for fill
* Adapt gradient/fill property panels and tools to handle List<T>
* Make the initial gradient transform covers the target's bounding box
* Introduce AnyGraphicListDyn to avoid combinatorial explosion
* Allow using any graphics type for stroke paint
* Add FIll node migration
* Add `for_each_vector_list_mut` instead of `set_paint_attribute`
* Adapt paint flow to read and write attributes instead of legacy Fill/Stroke.color
* Fix responsibilities of paint related node input setters
* Fix Morph by storing List<Graphic> for attributes rather than the concrete types
Store fill/stroke paints as List<Graphic> so Color/Gradient transitions
do not hit set_attribute_value_dyn's type-mismatch fallback to default paint.
* Preserve paint attributes in vector editing ops
* Enhance the clarity between direct and chained fill gradients
* Update demo arts
* Consolidate Fill node gradient appearance inputs
* Fix after the cubic review
* Revert "Consolidate Fill node gradient appearance inputs"
This reverts commit 9622feb20196e2c4da99e98ca95dcc2e34c2c98e.
* Replace AnyGraphicListDyn with generic paint connectors on the Fill and Stroke nodes
* Canonicalize paint attribute storage to List<Graphic> with a single write helper
* Fix Solidify Stroke missing fills stored in the legacy style
* Fix Solidify Stroke producing invisible strokes for legacy-only stroke colors
* Fix the initial gradient transform ignoring the bounding box's vertical extent
* Step paint at the morph midpoint instead of dropping it for unmixable pairings
* Remove migration-stage comments
* Clarify the initial gradient transform helper's doc and name
* Correct the bake_paint_transforms doc and prune dead tolerance arms
* Delete the unused Gradient::lerp
* Fix a comment typo
* Thread network paths through the legacy gradient bake so nested fills migrate
* Restore the legacy fill fallback in Expand Fill and Stroke
* Read gradient stops from the node's own input in the Fill properties panel
* Use the transform input constant instead of a hardcoded index
* Remove the unreachable wired color fallback in the Fill properties solid branch
* Narrow the fill overlay redraw check to actual fill inputs
* Coalesce the fill setter's graph runs into a single dispatch
* Position the Gradient Value node inserted for gradient stops
* Bake backup gradient placement during migration
* Persist pending gradient bakes so unfinished migrations retry on reopen
* Migrate the backup gradient's type and spread method
* Harden the gradient-migration pass against document switches and stale bakes
* Keep the Fill properties UI for layerless and nested Fill nodes
* Leave a wired gradient transform input connected instead of overwriting it
* Refuse to start a gradient chain ahead of existing layer content
* Decode a Fill node's gradient through one shared reader
* Nudge a degenerate bounding box so the Fill gradient transform stays invertible
* Broadcast Fill and Stroke paint with a single attribute-column pass
* Fix Morph stepping the target's stroke in near the source instead of the target
* Tidy conventions: clippy get_first, comment periods, sentence-case test messages
* Reattach the gradient orientation doc to its function
* Re-save demo artwork
---------
Co-authored-by: Keavon Chambers <keavon@keavon.com>
* Revamp the syntax to #[hard(a..b)] and #[soft(a..b)] bounds on node definitions
* Address review feedback: error on empty bounds ranges and remove redundant count clamps
* Fix NaN transform from Repeat Array with a count of 1
* Remove redundant manual clamps already enforced by #[hard(...)] bounds
* Move clipboard code to its dedicated handler
* Adapt clipboard format to work with resources
* Restore visibility, lock, and collapse state when pasting layers
---------
Co-authored-by: Keavon Chambers <keavon@keavon.com>
* Migrate legacy gradients via an isolated measurement pre-pass
Convert legacy bounding-box-relative Fill::Gradient values to absolute
space on document load by measuring each fill's evaluated geometry,
rather than walking each layer's primary flow. The pre-pass scans the
whole root network so fills on secondary branches (e.g. a Copy to Points
instance) and fills in hidden, disabled, or orphaned branches are all
found, redirecting the document export to each fill in turn to force its
branch to evaluate and reading its geometry back from the inspect result.
With every legacy gradient converted before any render, the legacy
bounding-box render path is removed: the ATTR_GRADIENT_LEGACY marker, the
renderer's legacy brush branch, and fill_to_graphic_list's bbox bake all
go away. Gradients nested inside subgraph node networks are out of scope
and warned about.
* Address review: harden gradient migration against stalls and document switches
- Advance the queue on dispatch early returns (missing export or channel
send error) so a failure can't leave the migration permanently stuck
with the document unable to render.
- Tag the migration with its DocumentId: cancel-and-restart when a
different document's migration is requested, and only apply a
measurement to the active document's in-progress migration, so
switching documents mid-migration can't stall or cross-apply.
* Remove nested network non-migration warning
* Cleanup
* Also add migrations for Fill node backup colors
* Apply migrations to the demo art
