Problem: the Media Browser lists .tga and .hdr files (and groups them into
image sequences), but media::is_importable had neither extension and the
still decoder never tried them, so importing them failed.
Fix: enable the image crate's built-in tga and hdr decoders (no new
dependency), add both to STILL_EXTENSIONS and the File > Import dialog, and
pick a still's format by signature with an extension fallback, since Targa
has no signature. Radiance HDR is linear float light and is sRGB-encoded like
OpenEXR (super-white kept); Targa alpha goes through the existing
unlabeled-alpha interpretation.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Problem: `read_exr_channels` kept the exr crate's whole decoded image alive
while it built every f32 plane, and copied each channel through an extra
`Vec` (`values_as_f32().collect()`) before placing it. One decode of a
2560x1440 Blender file with 32 channels briefly held ~320 MB of samples next
to ~470 MB of planes, per concurrently decoded frame.
Fix: consume the decoded image channel by channel, so each channel's samples
are freed (or moved) as its plane is made. A channel that covers exactly the
display window is converted in one pass, and an f32 one (Cryptomatte ranks)
becomes its plane without a copy; other channels are placed into the display
window with bounds-checked indexing, as before. Keeping half channels as f16
and sizing the channel cache by bytes stay open.
Test: a full-window f32 channel's plane is the same allocation; f16 / u32
channels and offset or clipped data windows are placed as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Problem: with shape layers selected, the Tools bar's Fill Options / Stroke
Options (None, Solid Color, Linear Gradient, Radial Gradient, and the
swatch's Alt-click cycle) only changed the defaults for new shapes.
shape.fillStroke had no kind, so a selected layer's fill or stroke could
not be removed or turned into a gradient from the toolbar; it had to be
edited in the Timeline. The swatch also kept showing the tool's kind, not
the layer's.
Fix: shape.fillStroke takes fillType / strokeType. None switches the
layers' Fills and Gradient Fills (or Strokes) off; another kind switches
them on and turns them into that kind, keeping the uid and the
properties both kinds share (blend mode, opacity, width, dashes...), and
gives a group with paths and no such item one (gradients span the
group's paths). It reports the first layer's fillType / strokeType, and
the toolbar shows them. The kind picks in the options popup and the
Alt-click cycle apply to the selected shape layers, one undo step each.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Problem: After Effects 26 imports FBX models into Advanced 3D comps; File > Import rejected
.fbx, so FBX product shots and scenes could not be used.
Fix: a clean-room FBX 7.x reader in effectcraft-model (binary node records with 32/64-bit
offsets and zlib arrays, ASCII syntax) and a scene builder: meshes with normals, UVs and
per-polygon materials, the Model hierarchy with the SDK transform (pivots, offsets, pre/post
rotation, rotation order, geometric transform), Lambert/Phong materials onto
metallic-roughness with embedded or relative textures, and GlobalSettings axes turned to
Y-up. Wired into import, probing, the media browser and the file dialog. Hostile input
(truncation, corruption, huge arrays, deep nesting) fails with an error.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Problem: Mesh Warp keeps its Distortion Mesh in a hidden text parameter,
and the viewer only draws handles for point parameters (Bezier Warp's),
so with Mesh Warp selected there was no grid and nothing to drag: the
effect could not warp anything from the UI.
Fix: with a Mesh Warp effect (or one of its parameters) selected and
layer controls shown, the viewer draws the (rows+1)x(columns+1) grid
over the layer in effect space, with a handle on every vertex
(automation id viewer.meshWarp.<fx>.<row>.<col>). Dragging a vertex
writes its offset to the `mesh` parameter through prop.set, one undo
step per drag; a mesh that does not match Rows/Columns starts from the
undistorted grid, as it renders.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Problem: After Effects 26.3 pastes copied vector artwork as shape layers. Edit ▸ Paste only
knew the app's own clipboard, so SVG markup on the system clipboard was ignored; turning SVG
into shapes needed an imported footage file and Create Shapes from Vector Layer.
Fix: a new `edit.pasteSvg {svg, name?, comp?}` command builds a shape layer from SVG text with
the same conversion as Create Shapes from Vector Layer (paths, fills, strokes, gradients, clip
masks), above the selected layer with the SVG canvas centred in the comp, as one undo step;
oversized, non-SVG or shapeless text is rejected. In the UI, Ctrl/Cmd+V (and the native Edit ▸
Paste) run it when the system clipboard holds SVG. Copying in the app already replaces the
system clipboard with a note, so SVG there is newer than the app's clipboard.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Problem: After Effects 26.3 copies the current composition frame to the
system clipboard; EffectCraft could only write it to a file (Save Frame As).
Fix: new command comp.copyFrame {comp?, time?, scale?} renders the frame as
Save Frame As ▸ File does and emits Event::CopyImage (pixels not
serialized); the egui frontend puts it on the clipboard after checking the
buffer size. In the Composition menu (translated) and the viewer's context
menu; drivable by id over the control channel / MCP.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Problem: After Effects 26 imports STL models besides glTF and OBJ; we
refused .stl files, and the Media Browser and import dialog did not list
them as models.
Fix: a pure-Rust STL reader in effectcraft-model (binary and ASCII,
detected from the contents; several ASCII solids and polygon loops;
VisCAM/SolidView and Materialise Magics facet colours as plain
materials; Z-up turned to the model's Y-up; faceted normals from the
winding). Truncated, empty or non-finite input is an error, never a
crash. .stl is now a model extension for import, probing, the Media
Browser, the import dialog and the system report.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Problem: dropping a PSD/PSB file on the window imported it straight away.
The drop handler always passed `importAs` (Settings > Import > Default Drag
Import As) to `file.import`, and the Import Photoshop File form only opens
when `importAs` is missing, so the dialog that File > Import > File...
shows never appeared for drops.
Fix: a drop that includes a Photoshop file leaves `importAs` out, so the
form asks (other drops keep passing the preference). For dropped files the
form's Import Kind starts at Default Drag Import As; File > Import keeps
starting at Composition. The engine already falls back to the preference
for `drag` imports without `importAs`.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Problem: selection_hashes split the Selection on every comma before
stripping quotes, so a name containing a comma (Blender allows a material
called `Green, Glossy`) was cut into two names that match nothing, quoted
or not; only its `<hex>` hash selected it.
Fix: a small tokenizer splits at commas outside quotes. A name in double
or single quotes stays whole, a backslash takes the next character
literally (`Green\, Glossy`), and unquoted lists, wildcards and `<hex>`
hashes parse as before. The CPU effect and the GPU path share the function.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Problem: with the project's Time Display Style set to Frames, the
current-time display showed frame numbers but the Timeline ruler kept
labelling its ticks in seconds:frames (`00:15f`). The label format was
hard-coded and never read the project settings.
Fix: a `ruler_label` helper prints the frame number in Frames mode the
same way the current-time display does (`{:05}`, counted from Start
Numbering Frames At), and keeps seconds:frames for the other styles.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
They ran on every push to main and on any PR touching Cargo.toml/Cargo.lock,
which is nearly every PR. See craftrules release/playbook.md §2.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Settings > Appearance > UI Scale offered only 75, 100, 125, 150, 175 and
200 %, so a size between 75 and 100 wasn't possible. It is now the shared
numeric field (Kind::Int(75, 200, "%")): type or drag any whole
percentage. The engine already stored the scale as a clamped integer and
the UI applies it with set_zoom_factor, so the range, the default, live
preview, Cancel and saved preferences are unchanged; the preset list
goes.
Numeric settings now apply a typed number once it is entered (Enter or
leaving the field), not at every keystroke: typing 85 used to apply the
partial 8, clamped to 75, and resize the whole interface before the 5.
Dragging still applies as it moves.
Tests: the UI Scale test types 85 a key at a time (nothing applies and
the field keeps the keyboard until Enter), Cancel, 50 and 300 clamp to
75 and 200, 90 with OK, and a new app with the saved preferences starts
at 90 %. The GPU Information transaction test types 125.
The stop colour now uses widgets::srgb_color_button, the After Effects-style
picker every colour input opens (#638): hex field, HSB/RGB fields, OK and
Cancel, sRGB as the project stores colours. The stop keeps its alpha.
The dialog Escape handler only stepped aside for egui popups, and this
picker isn't one, so in a modal dialog Escape with the picker open closed
the dialog too. The picker now marks the pass it was open in
(color_picker::open_recently), and the handler leaves Escape to it: Escape
restores the colour the picker opened with and closes only the picker.
Tests: ui_gradient_editor clicks inside the picker's colour field over the
modal editor, types a hex code, closes the picker by clicking outside it
and saves the stop; a new test checks that Escape and the picker's Cancel
restore the stop's colour and keep the editor and its other edits. That
test fails without the Escape change. ui_color_hex still passes.
The modern curve used g = 2^(2c), so Contrast 100 crushed 0.25 to about 8/255 while AE is about 40/255. Brightness is now a pin-black-and-white power toward white or black, then a milder S-curve g = 1 + 0.53 c, fitted to kasper9n's AE grey ramps. Legacy linear mode is unchanged. The GPU kernel matches the CPU.
Six effectcraft-engine tests failed on a Windows machine whose working directory
and home folder are on different drives:
- rq_tests::add_configure_render and tests_stubs::{add_output_module_encodes_twice,
pre_render_imports_and_replaces, post_render_actions_import_and_set_proxy,
log_notify_and_overflow_settings} used "/tmp/..." output paths. On Windows that
has no drive letter, so the render queue resolves it against the home folder's
drive (C:) while std::path::absolute, which the tests compared with, uses the
working directory's (D:). They only passed where both are on one drive, as on CI.
They now build their outputs with tests_stubs::tmp_out, a path under the OS temp
directory that is absolute on every platform and unchanged on Unix.
- footage_check::relative_footage_is_found_next_to_the_project compared strings:
Windows joins the stored "media/clip.png" onto the project's folder and leaves
mixed separators. It now compares the paths by components.
Test-only: the render queue stores the path it writes to, so the product is
consistent. effectcraft-engine --lib: 497 passed (6 failed before).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Settings > Previews > GPU Information was fixed text ("Compositing: CPU, N
threads... Display: egui on wgpu."): it never named a GPU, even while the GPU
compositor was running, so a card looked undetected. The System Report's GPU
line only said whether the compositor started.
sysinfo::gpu_summary builds the graphics lines from the session: the compositor's
adapter, or the reason compositing is on the CPU, and the adapter the window
itself got. The dialog uses it.
This does not explain why a particular card isn't used for compositing; that
needs the reporter's System Report (Help > System Report).
Tests: engine gpu_summary_names_the_adapter_or_says_why_there_is_none and UI
gpu_information_names_the_adapter_and_why_the_gpu_is_off (fails with the old
text). The six engine tests that fail on this Windows machine (relative footage
and render-queue paths) fail on main too.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Time > Pixel Motion Blur fetched the layer at its shutter times and averaged the decoded
frames as they were, so fast motion rendered as doubled copies and the accepted Vector
Detail parameter changed nothing. Build a TimewarpPlan instead: every shutter instant is
decoded between the two source frames it falls between along a measured motion field — the
same block-matching and in-between synthesis Time > Timewarp uses, with a search radius
wide enough for full-frame displacements and the vector grid tightened from Vector Detail
0 (one vector per 32 px) to 100 (one per 2 px) — and the instants average into the blur.
CPU and GPU share the plan (fx_time renders it with its Timewarp step loop, declining to
the CPU when there is no host), so the existing GPU parity cases for the effect compare
the new path. Falling back to the old average is kept for layers without a time host.
A render integration test moves a textured block 12 px between source frames, renders
the straddling shutter, and asserts the output is neither the still nor the plain frame
overlay, lays trail coverage along the block's path, and changes with Vector Detail;
without the synthesis it fails at the overlay assertion.
The Timeline's "Edit Gradient…" on a gradient property (Gradient Fill /
Gradient Stroke ▸ Colors, Gradient Overlay) only painted its label. It now
opens a Gradient Editor (prop.editGradient): opacity stops above a preview
bar and colour stops below. Click to add a stop (it takes that track's
own value there), drag or type a Location to move it, set its opacity or
colour (egui's colour picker), Delete removes it (never the last of a
track, never while typing). The tracks stay in order while editing and the
selection follows its stop.
The editor works on a draft. OK sets the property once with prop.set at
the composition time it opened, snapped to the frame, with
timeBase: "comp" (offset and stretched layers get the key where it was
opened): one undo step, and nothing at all when the draft is unchanged, so
OK adds no keyframe. Cancel and Escape change nothing; Escape with the
picker open closes only the picker. On an animated gradient only the
keyframe at that time changes; when its stop counts no longer match a
neighbouring keyframe's (gradients only blend with equal counts) the
editor says so. The draft is saved in the typed form, so one colour stop
and each stop's RGBA survive.
prop.editGradient checks its target is a gradient property before the UI
opens anything, so agents get an error for a wrong one. New Gradient
color_at / opacity_at sample one track each (sample() multiplies opacity
in, which would apply it twice to a new stop).
Tests: ui_gradient_value (fails on main: the value had no automation id),
ui_gradient_editor (11), gradient_editor and keyframe unit tests.
An authored layer's effect buffer spanned only its painted content — the glyph
bounding box for text, the drawn bounds for shapes — while the size the stack
measures from (effect_bounds) is the whole comp-sized layer rectangle. Tiling
and displacement effects computed correct output for pixels outside the painted
rectangle and then wrote them into the buffer that was too small, so CC Tiler
repeated only within the text box and CC Smear lost its liquid trail at the
glyph edge (#658). Buf::cover now grows the stack's input to the layer's
rectangle with transparent margins — what a sized source's buffer already is —
applied by run_effect_stack whenever an enabled video effect will run, and
natively on the GPU by reallocating and copying the texture. The placement and
the sized-source paths are unchanged; two regression tests (CC Tiler spans the
region, CC Smear trails past the text bounds) fail without the growth.
* Release prep: bump to 0.8.0 and add a Linux riscv64 release build
- release.yml: new `linux-riscv64` job, ported from GridCraft. It cross-compiles on
ubuntu-24.04 against riscv64 libraries from ports.ubuntu.com (the Linux job's X11/Wayland/GL
-dev packages plus ALSA), ships effectcraft-<v>-linux-riscv64.tar.gz, and smoke-tests
effectcraft-cli under qemu-riscv64. The draft-release job now needs it, so its tarball is
uploaded and covered by SHA256SUMS.txt.
- packaging/linux/package.sh: cross-compiles when CROSS_ARCH/CROSS_TARGET/CROSS_COMPILE are set
(cargo --target, the cross strip, the EMULATOR smoke test), as GridCraft's does.
- Version 0.7.0 -> 0.8.0 (cargo xtask version set), Cargo.lock refreshed.
- README and docs/releasing.md list the riscv64 tarball.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Fix FreeBSD CI: pointer_wrap test harness, wasm32 target, clippy as_chunks
- pointer_wrap tests: the harnesses dropped run_ui's FullOutput with the font atlas upload still
in textures_delta, and epaint 0.36's drop check panics on that in debug builds. Every
pointer_wrap test failed on every platform; FreeBSD is the only CI that runs them. Clear the
delta, as the other UI tests do.
- freebsd.yml: install wasm32-unknown-unknown in the VM. Without it the host plug-in test's build
of examples/plugins/posterize-bands failed (E0463, no std) and the test skipped silently; now
it builds and runs.
- media/convert.rs: chunks_exact(2) -> as_chunks::<2>() (clippy chunks_exact_to_as_chunks,
new in Rust 1.99); same pairs, same dropped remainder.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Importing an animated GIF as a movie (bfbdadd) made it preview and export
again, but nothing pinned it: the item must probe as a GIF movie and its
rendered frames must follow the file's own per-frame delays, not hold one
still. Add a host integration test that imports a red/green/blue GIF and
asserts each third of its duration renders the matching colour, so a
regression to still-frame import is caught.
Animated GIFs came in as one still frame in 0.7.0; bfbdadd3 (after that
release) probes them as movies. This end-to-end test imports a GIF file
through the wired session into a comp and checks the layer shows each of
its frames at its time, the path the viewer and exports share.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QiAJeFttZp83Fh7YuzAR5f
file.import stores a path as given, so footage imported by the CLI or an
agent as `clip.mp4` was saved relative, and opening the project looked it
up from the working directory. The CLI, started in the project's folder,
found it; the desktop app, started from the Finder, Explorer or a launcher
in another folder, showed and rendered neither its picture nor its sound,
while the Project panel still listed the saved duration.
On open, a relative footage path that doesn't exist from the working
directory but does next to the project file is resolved there (After
Effects also looks for footage relative to the project). Paths that exist
as they are, and absolute ones, are kept.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QiAJeFttZp83Fh7YuzAR5f
A text or shape layer's effect bounds are comp-sized (as in After Effects),
but its buffer surrounds only its content, and each effect grows the buffer
itself when it needs room. CC Smear and CC Tiler never did, so a smear was
cut off where the glyphs end and tiles filled only the text's box.
CC Tiler now covers the whole layer rectangle; CC Smear grows by how far it
moves pixels (Reach x From->To), at most to the layer bounds, so footage
layers, whose buffer is their bounds already, are unchanged. CPU and GPU
alike.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QiAJeFttZp83Fh7YuzAR5f
Every viewer request spawns a job on the frame pool. A pool thread blocked
inside a render (waiting on a parallel loop's join) runs other pool work
meanwhile, rayon's injected jobs included: the spawn of the next request
started a whole frame render on top of the waiting one. Dragging makes a
new revision, and a new viewer frame, every UI frame, so renders nested
without bound on every thread, each holding its buffers (a large image's
full-resolution copies), until all memory was used. Holding the layer still
makes no new revisions, which is why only moving it did this.
A spawn now drains the queue, and one that lands inside a render on its
thread returns at once; the render below takes the queued jobs when it is
done. Each pool thread renders one frame at a time.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QiAJeFttZp83Fh7YuzAR5f
The Scrub helper and the real-properties test discarded FullOutput, so the
first pass' font-atlas texture delta panicked on drop in debug builds —
which is how the FreeBSD CI gate failed. Consume the delta the way the
rest of the crate's egui tests do.
Preserve legacy selection repair, modifier routing, viewer lifecycle and workspace persistence through the shared docking integration. Qualify the complete workspace gate by refreshing the 18 stale Audio Spectrum hashes for the existing linear renderer, with an independent synthetic-tone regression.
`effectcraft-cli mcp` had no way to keep an agent's file access to a folder;
the flags were unknown options. Both now take a folder (repeatable, as in
PhotoCraft): with either given, files are read only under a read or write root
and written only under a write root (none: no writes). Defaults are unchanged.
automation::confine::Confinement checks every engine command the backend runs
(execute_command, command_run, open/save_project, each engine.batch step), in
both headless and bridge mode, before it runs: the file parameters of the
commands that read or write files must be absolute paths inside the roots
(symbolic links resolved through the longest existing ancestor, `..` refused),
as must any other absolute path in a command's parameters. Commands that write
where they aren't told are checked against the target they would write: Save /
Increment and Save the open project's file, renderQueue.render every queued
output (renderQueue.list). Scripts, Execute File, Reveal in Finder, Open Recent
by index, `$N` file-path references in a batch, render_frame/screenshot paths
outside a write root and the bridge's UI-driving tools are refused while
confined. docs/mcp.md describes it.
Test: a confined headless server opens and saves inside its roots and is refused
outside them, for a relative path, `..`, a batch step, a `$N` path, Save to a
read-only root's project, the Render Queue's default output there, a symbolic
link out of the write root, and scripts; the unconfined server still writes
anywhere.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QiAJeFttZp83Fh7YuzAR5f
The file dialogs were blocking rfd::FileDialog calls made inside eframe's frame,
that is inside winit's event handler. On macOS the dialog runs a nested modal
event loop there; a file dragged over the window while it was open reached
winit's handler again, which refuses re-entry with a panic across the
Objective-C boundary, aborting the app (Import with Cmd+I, then a drag).
Hooks gains file_dialog_async: a host that sets it shows a command's file dialog
(Import, Open, Save As, Save Frame As, presets, guides, templates, Collect Files,
Relink...) without blocking, and the command runs with the answer on a later
frame (menus::dialog_answered, drained at the start of each frame); a cancelled
dialog runs nothing. The macOS app implements it with rfd::AsyncFileDialog, a
sheet on the window answered when it closes, outside the frame. Other platforms
keep the blocking dialogs (winit queues their re-entrant events), as do the few
panel buttons that pick a file in place (Settings paths, shortcut import/export,
effect file pickers).
Test: with a non-blocking hook, Save As and Save Frame As return at once without
the blocking dialog, the save happens when the answer arrives from another
thread, and a cancelled dialog writes nothing. The macOS sheet itself needs a
manual check (Cmd+I, drag a file over the window, pick or cancel).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QiAJeFttZp83Fh7YuzAR5f
widgets::keep_focus (Project panel rename and comment, Render Queue, Essential
Graphics and Timeline fields) selected the whole text whenever it saw its field
without the keyboard focus, not just on the frame the editor opened. Anything
that took the focus for a frame (an input method, another handler) then
re-selected the name, so the arrow keys seemed to do nothing and Backspace erased
the whole name, as reported on Linux. The editor now remembers the pass it was
last shown: it selects all only when it opens and otherwise just takes the
keyboard back with the caret where it was.
The harness does not reproduce the report's Linux behaviour itself (arrows reach
the field there); the tests cover the path that did: a unit test drops the focus
for a frame and checks the caret survives (and that reopening selects all), and a
kittest renames a Project item and a Timeline layer with Left / Right / Up /
Down / End and Backspace, editing part of each name.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QiAJeFttZp83Fh7YuzAR5f
