Commit Graph
66 Commits
Author SHA1 Message Date
bflatastic bd80829a27 Batch r2-1 review: a paint keeps its own opacity through unbind and rebind
color_alpha is the part of a solid paint's opacity that is its typed colour's alpha, not a flag:
binding a colour variable takes only that part out, unbinding puts the variable's alpha back in,
and REST import records it. A 50% paint bound to a 50% variable stays at 25% bound or not.
2026-10-24 10:30:49 +07:00
bflatastic 70536a6265 Merge #948: Keep a solid paint's own opacity when binding a colour variable (@Silvertree2010) 2026-10-24 09:21:42 +07:00
BFlat 8167f13a6a M12: Boards: a collaborative whiteboard file kind (#1229)
Board files (File > New board) with sticky notes, shapes with text, connectors that follow what they join (straight or elbow, arrowheads, labels), marker, highlighter and eraser, and stamps; a board toolbar and a contextual bar; the Design panel hidden on boards, Prototype and Dev Mode refused there. Boards are ordinary FigCraft files built from ordinary layers, so saving, undo, export, MCP and live sessions work unchanged. File tabs, recent files and Open files show a pen nib for design files and a note for boards (original icons drawn in code).
2026-10-24 04:18:55 +07:00
NetterSilverandClaude Haiku 5.5 b788b55286 Keep a solid paint's own opacity when binding a colour variable
Fixes #923

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-20 20:43:51 +02:00
bflatastic b82369201d M3: duplicating a main component makes an instance (#892)
Duplicate and Alt-drag copied a main component as a second main. Within
its file the copy is now an instance of it (components::copy_in_file,
shared by edit.duplicate and the Alt-drag copy), as Copy then Paste
already was. A variant duplicated in its set stays a new variant.
2026-10-20 05:44:46 +07:00
bflatastic f22e2ed825 Merge main into #870 2026-10-20 02:04:33 +07:00
BFlat 3ed4578112 Merge pull request #851 from MortenHusted/fix/effect-unknown-keys
M3: effects refuse keys they don't read instead of dropping them
2026-10-20 00:46:45 +07:00
BFlat 409de1a762 Merge pull request #871 from justinjohn0306/fix/809-scalar-bound-edits
M3: explicit edits of bound scalar fields detach their variable (#809)
2026-10-20 00:10:09 +07:00
justinjohn0306andClaude Opus 5.5 6a163721dc M3: explicit edits of bound scalar fields detach their variable (#809)
An explicit edit of a bound width, height, opacity, gap, padding, corner
radius, stroke weight, arc or text field reported ok, but normalize wrote
the variable's value straight back. Now the edit sticks and that field's
binding is dropped, as paint edits already do.

- `styles::bound_before` / `detach_edited`: compare each layer-level
  bound field's stored value before and after the edit (`bound_field`),
  so keys written unchanged keep their bindings and fields moved along
  with another (proportions, min/max) lose theirs. A binding the edit
  itself made (`boundVariables`) stays. List and component property
  bindings are left to their own rules.
- Hooked into node.set, transform.resize, object.opacityKeys, text.style
  and `Session::preview` (canvas drags of the selected layers).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-19 17:33:29 +05:30
MortenHustedandCodex 0108a7c2f5 M3: copied components paste as instances within their file
Co-Authored-By: Codex (GPT-6) <noreply@openai.com>
2026-10-19 13:34:40 +02:00
MortenHustedandClaude Opus 5.5 6f90717e71 M3: effects refuse keys they don't read instead of dropping them
node.set refuses unknown properties, but the effects inside it were read
by serde, which drops keys it doesn't know. A drop shadow given
offsetX/offsetY (the names variable bindings use for its offset) was
stored with the default offset [0, 4] and the call said ok. effect.add
and effect.set dropped them the same way.

The three now refuse what the effect doesn't read and say what it
takes: "effects[0]: unknown properties: offsetX, offsetY (a dropShadow
takes blend, color, offset, radius, showBehind, spread, type, visible)".
A key counts as read when setting it to null makes the effect fail to
parse or come out different, so serde's aliases count as read without a
list of them here.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-18 23:53:05 +02:00
MortenHustedandClaude Opus 5.5 c7261cb907 M3: an explicit paint edit detaches the variable bound to what it changed
fill.set, stroke.set and node.set fill/fills/stroke/strokes reported ok
while normalize put the bound colour straight back, and node.set
{fills: []} brought the bound paint back. They now drop a paint's
colour, opacity or stop binding where the edit changed that value, as
selection.setPaint already did (its comparison moves to
lists::rebind_edited_paint, which also lets stop bindings follow a
removed stop). boundVariables applies last in props::apply, so a
binding given with the paints in one call stays.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-18 14:14:38 +02:00
skylar taylor-barrick c7acd1a81d chore: clear the rustc 1.96 clippy lints that fail xtask ci
Unrelated to the security changes in this branch; `main` already failed
`clippy --workspace --all-targets -- -D warnings` without them. Eight
clippy::nonminimal_bool sites across seven files, each taking clippy's own
suggested rewrite.

Every rewrite is pure boolean algebra with the negation kept outside, so the
float comparisons and their NaN behaviour are untouched: !A && !B becomes
!(A || B), and !opt.is_some_and(f) becomes opt.is_none_or(!f). Two sites needed
explicit parentheses the suggestion omitted, where the rewritten term is the left
operand of an && and || binds looser.

designcraft cleared the same lint in bc840130; this is figcraft's equivalent.
2026-10-17 13:35:44 -06:00
bflatastic adde034a7c Merge main into #603 2026-10-17 19:40:01 +07:00
bflatastic 99d097636e Merge fix/all-f5 into batch 12 2026-10-17 16:59:57 +07:00
bflatastic e912a2efe8 #655: fill.set / stroke.set read index with params::opt_index (#675) 2026-10-17 16:38:35 +07:00
bflatastic 6f8c2a657e Merge fix/680 2026-10-17 16:37:27 +07:00
bflatastic 01ce23ca76 #680: page refusal names the calling command (edit.copy / edit.cut) 2026-10-17 15:58:23 +07:00
bflatastic 3353d2d5d3 #647: edit.pasteDesignLayers, named neutrally; design data capped at 256 MB before decoding, data accepts wrapped or unpadded base64 2026-10-17 14:34:57 +07:00
NetterSilverandClaude Haiku 5.5 0e25cf4c26 Refuse pages in edit.copy and edit.cut
Found while testing crates/engine/src/cmd/edit.rs:499 (copy_to_clipboard, called from copy at :536 and cut at :542)

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-16 13:47:07 +00:00
NetterSilverandClaude Haiku 5.5 a911ab9614 Reject negative or fractional index in fill.set and stroke.set
Found while testing crates/engine/src/cmd/edit.rs:1180

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-16 08:29:55 +00:00
JimKnoxxandClaude Opus 5.5 ab10e37d43 M1: paste layers copied in Figma
Figma puts copied layers in the clipboard's HTML as a base64 .fig canvas
between (figma) markers. edit.pasteFromFigma reads it with the .fig import,
turns the layers into the session clipboard and pastes them as edit.paste
does (selected frame, in place or view centre), with the import's warnings.
The desktop app reads the clipboard's HTML (arboard) and Paste prefers a
Figma copy; the browser's paste event carries plain text only.

Tested on copies written by our own .fig export; not yet on real copies.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-16 08:20:49 +02:00
Edivan TeixeiraandClaude Opus 5.5 8cccc48516 node.set: apply auto layout keys to the frames and name the rest; resolve fontStyle like text does
- Several ids: auto layout keys go to the frame-like layers, the others get the remaining keys
  and are named in `warnings`. Refused only when no target can have auto layout.
- fontStyle is never refused: it draws in the nearest face, as a text layer resolves it.
  `warnings` says so when that face's weight or slant differs from what the name asks for.
- set_properties passes `warnings` on to the agent.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-15 19:43:52 -03:00
bflatastic db8bf6ba33 Merge main into #560; normalise the replacement where findReplace reads it
Main's case-insensitive replace loop (from an earlier PR) took the place of the
loop this change edited, so the replacement text is normalised once, where
the parameter is read.
2026-10-16 02:31:34 +07:00
MarekSpodBiedry 843b00b22f text: scripted text stores CRLF, CR, U+2028 and U+2029 as LF like the editor (#427) 2026-10-15 17:20:58 +02:00
bflatastic a95f95ae37 #361: a replacement that only changes the case replaces every occurrence; the text's chars are collected once 2026-10-15 18:54:50 +07:00
bflatastic 6215cae877 Merge main and the community batch into #361
# Conflicts:
#	crates/engine/src/cmd/edit.rs
2026-10-15 18:52:02 +07:00
bflatastic 3c52dad774 Merge #353: engine: long digit strings no longer overflow object.opacityKeys (@midasmatterhorn) 2026-10-15 17:54:23 +07:00
bflatastic 5a57388b8d Merge #344: doc: name counters stop at u64::MAX instead of overflowing (@midasmatterhorn) 2026-10-15 17:54:21 +07:00
bflatastic 022821211f Merge #358: Replace every occurrence in edit.findReplace when replacement contains find (@Silvertree2010) 2026-10-15 17:53:06 +07:00
bflatastic 03716da731 Merge #384: Reject negative or fractional index in effect.set, export.update and export.remove (@Silvertree2010) 2026-10-15 17:53:04 +07:00
NetterSilverandClaude Haiku 5.5 87b9a3a9ce Reject negative or fractional index in effect.set, export.update and export.remove
Found while testing crates/engine/src/cmd/edit.rs:256 (effect.set), :305 (export.update), :317 (export.remove)

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-15 01:08:36 +00:00
NetterSilverandClaude Haiku 5.5 2aa3ca3c45 Make edit.findReplace replace case-insensitive matches
Found while testing crates/engine/src/cmd/edit.rs:962-977

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 22:26:37 +00:00
NetterSilverandClaude Haiku 5.5 c02e864873 Replace every occurrence in edit.findReplace when replacement contains find
Found while testing crates/engine/src/cmd/edit.rs:975-981

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-14 21:22:29 +00:00
midasmatterhorn 65e52aa7b9 engine: long digit strings no longer overflow object.opacityKeys
opacity_from_digits multiplied and added on every character of its input
string before the one-or-two-digit check, so "9999999999" overflowed u32 on
the way (panic in debug builds, wrap in release). The scan now steps n with
checked_mul/checked_add and stops (None) at the first character the number
cannot take, so a long digit string is refused like any other bad input.

Test: opacity_digits (engine keymap) covers the ten-digit string and the
command's BadParams refusal.
2026-10-14 23:22:41 +03:00
midasmatterhorn 245837ff27 doc: name counters stop at u64::MAX instead of overflowing
Document::next_name summed max + 1 over numbers parsed from existing layer
names, and bump_name did the same for the duplicate-rename path: a document
(or peer, or node.set) holding a layer named "Frame 18446744073709551615"
overflowed both (panic in debug, wrap to a wrong name in release). next_name
now collects the numbers it sees and, when the counter is spent, picks the
lowest free number (mirrors new_style_id); bump_name only bumps when the
number can step, else falls to its existing " 2" branch.

Tests: next_name_never_overflows (doc) and
duplicating_a_layer_named_past_the_counter_does_not_overflow (engine panel
behaviour).
2026-10-14 22:52:15 +03:00
RealtaNuaandClaude Opus 5.5 73b2845ba4 fix(engine): Alt-drag copies snap to their original; Duplicate repeats the step
- A copy Alt-dragged off a layer snaps to that layer too: its edges,
  centres and equal spacing (Targets::add_box with the originals' box).
  Before, the originals were left out of the snap targets, so a copy
  couldn't line up with or space evenly from the layer it came from.
- Duplicate (Cmd+D) repeats the last duplicate's page offset while its
  copies are what it duplicates: Alt-drag a copy, or Duplicate and move
  the copy, then each Cmd+D steps out the next one as far again. The
  step is measured when Duplicate runs. It lasts as long as the
  selection (undo reuses ids), copies placed by an auto layout set no
  step, and explicit dx/dy win.
- measure_tests: an Alt-dragged copy between its original and another
  layer now snaps to equal spacing (45/45 instead of 50/40).

Tests: duplicate_spacing_tests (snap to the original, equal spacing,
repeat after moving a copy and after an Alt-drag, only for the last
copies, undo then a new layer, rotated frame, auto layout).

Refs #256

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-10-14 18:03:48 +07:00
Euisuh Jeong d30255bf62 fix(engine): preserve the underlying repeat command (#212)
edit.repeat no longer records itself as the last command, so repeating keeps applying the command it repeats instead of recursing, and it refuses when there is nothing to repeat. Fixes #181.
2026-10-14 03:26:16 +07:00
bflatastic 29e31c2054 M1: rotation origin and place-with-Enter shortcuts
The last two documented shortcuts the parity table listed as missing, and a stale toast.

- **Alt+R: rotation origin** (both platforms). `transform.rotationOrigin {on?}` shows (or hides) a
  handle at the selection's rotation origin: a ring with a cross through it, drawn in code. Drag it
  to move the origin (it snaps to the selection box's corners, edge middles and centre within the
  snap distance); double-click it to put it back at the centre; Escape hides it before it deselects,
  and Alt+R again hides it too. `transform.setRotationOrigin {x, y} | {reset: true}` does the same
  from the CLI, MCP and the control channel. The corner rotation handles, the rotation field
  (`node.set {rotation}` on the selection: each layer's point under the origin stays put, so flipped
  layers work too), Rotate 90/180 and `transform.rotate` without `cx`/`cy` turn about it. The
  origin is stored on the document state per selection, in the first layer's own coordinates, so it
  follows the layers as they move and turn; it is dropped when the selection changes and is not
  saved in the file (the public documentation doesn't say it persists; pinned in the parity row).
- **Enter places an object with the chosen tool.** With a shape or frame tool chosen, Enter on the
  canvas, or Enter again on the toolbar's chosen tool (after picking it from the toolbar with the
  keyboard), places what a click with the tool makes (100x100; lines 100 long; sections 400x300, the
  existing click sizes, now one `ToolId::click_size`) centred in the view, selected, as one undo
  step; the canvas then takes the keyboard. New `tool.place {tool?, x?, y?}`; MCP's headless `key
  Enter` does the same. The documented crosshair that the arrows move before placing is not built.
- **Keyboard box selection toast** (and its screen-reader text) now names both ways to size the box:
  "Shift+arrows or Cmd+arrows" (Ctrl on Windows and Linux). The changed key is renamed in every
  catalog with its translation updated.
- **Shortcuts panel labels:** a label too long for its column was measured in a smaller font than
  it is drawn in, so long ones (several Spanish rows) ran into their keycaps; it is now measured in
  its own font, ends with "..." before the keycaps and shows in full on hover. The new Tools row
  reads "Place a shape or frame (its tool chosen)" with an Enter keycap.
- **Parity table:** the two rows move from Missing to bound (Alt+R -> `transform.rotationOrigin`;
  Enter -> a Tools-tab gesture row, the keymap's Enter staying `select.children`); counts 129
  documented, 128 bound, 1 moved, 0 missing, 2 kept by a platform, 48 pinned.
- New command labels in Spanish (`tool.place`, `transform.rotationOrigin`,
  `transform.setRotationOrigin`), new UI strings in English and Spanish (es-419 overlays "Enter").
- Docs: ui-parity.md (selection and transform rows, shortcut counts, header, revision row),
  mcp.md (the new commands).
2026-10-12 07:53:24 +07:00
bflatastic 8d5f8a0978 M1: keyboard shortcut parity suite and bindings
- A shortcut parity table (`crates/engine/src/shortcut_parity_tests.rs`): one row per shortcut in
  the public documentation (help-centre feature articles: keyboard layouts, text properties,
  alignment, frames, components, prototypes, plugins, keyboard navigation), with the macOS and the
  Windows/Linux keys, what handles it here (keymap command, held key/gesture listed in the panel,
  vector edit tool key, or missing), and a note in our own words. Rows the documentation is silent
  on pin today's keys, so every default binding is in the table.
  - 129 documented rows: 126 bound (110 before), 1 moved off a documented conflict (Run last
    plugin on Windows: the docs also give Ctrl+Alt+P to pixel preview; the plugin keeps
    Ctrl+Alt+Shift+P), 2 missing (moving the rotation origin with Alt+R; placing an object with
    Enter after choosing a tool from the toolbar), 2 kept by a platform (Ctrl+Space with some CJK
    input methods; F6 in browsers). 48 undocumented bindings pinned.
  - Tests: every row fires its command on both platforms and sits in the right shortcuts-panel tab;
    every default binding has a row; no binding uses a chord the OS keeps (Cmd+H/M/Q/Tab/Space,
    Ctrl+Space, screenshot keys on macOS; Alt+Space/Tab/F4, Ctrl+Esc, Ctrl+Alt+Delete, Linux
    workspace and VT keys); macOS keycaps show Cmd/Option glyphs only and Windows/Linux keycaps
    Ctrl/Alt words only; vector edit tool keys; counts pinned for the docs.
- Newly bound: A (Frame tool), Shift+' (pixel grid; Cmd+' kept, the articles disagree),
  Alt+Cmd+\ / Ctrl+Alt+\ (others' cursors in a live session), Ctrl+F6 (toolbar focus on
  Windows/Linux), Shift+Cmd+U (link text; Cmd+K kept), Cmd/Ctrl+] and [ (list levels while
  editing), Cmd/Ctrl+Backspace (ungroup), Alt+Space / Ctrl+Space (keyboard box selection; F8 kept,
  Cmd/Ctrl+arrows draw the box out), Ctrl+Shift+A / Ctrl+Alt+Shift+A (suggest auto layout),
  Ctrl+Alt+Cmd+K / Ctrl+Alt+Shift+K (go to main component).
- New commands: `view.multiplayerCursors {on?}` (View menu item, menu breadth 207/207),
  `view.pan {dx, dy, big?}` (arrows pan the view by 20 screen px, 200 with Shift, when nothing is
  selected; the canvas and MCP `key` both use it), `text.listIndent` / `text.listOutdent`.
- Behaviour: Control held while dragging suspends snapping on macOS too (Cmd/Ctrl already did);
  the presenter takes Space and N for the next frame, C for comments (Shift+C kept) and F for full
  screen, after the prototype's own key triggers; the paint bucket's tooltip shows Shift+B (B still
  works).
- Keyboard layouts: shortcuts follow the character typed, named as on a US keyboard. A symbol
  another layout types with Shift (`/` is Shift+7 on German keyboards) now counts without the
  Shift (canvas, text editing and F6 navigation share one lookup). Characters egui has no key for
  already arrive as the key at that position.
- Keyboard shortcuts panel: new rows for panning with the arrows, suspending snapping, moving along
  the toolbar, and the presentation keys, in the tabs the documentation groups them in (en/es
  strings).
- Docs: ui-parity.md gets a Keyboard shortcuts section with the counts; target-app-parity.md,
  ROADMAP.md, gaps.md (gap 4 landed list), mcp.md, parity-checklist.md (regenerated).
2026-10-12 06:49:02 +07:00
bflatastic 626a9eac08 M1: layers and assets panels behaviour suite and fixes
A behaviour suite for the layers panel and the Assets panel written from the public documentation
(table rows with one-line derivations; rows the documentation is silent on say so and pin today's
behaviour), and the fixes it found.

- Engine suite `crates/engine/src/panels_behaviour_tests.rs` (6 tests, about 50 rows): reordering
  and moving layers into frames, groups and auto layout, refused drops, renaming several layers,
  show/hide and lock rules, pages and dividers, Assets folders, hidden-from-publishing names, drops.
- Panel suite `crates/ui-egui/src/layers_behaviour_tests.rs` (5 tests, driven through egui
  headless): click / Shift range / Cmd-Ctrl toggle, Alt-click expand, Collapse layers, sweeping
  across eyes and padlocks, the Rename layers dialog, dragging pages, dividers.

Fixes:
- Several layers dropped together in the layers panel land together at the drop point (layers
  from apart split around it before).
- Collapse layers (Alt+L) keeps the rows leading to the selection open (it closed everything).
- The eye and padlock no longer vanish when the pointer reaches them (the row lost its hover to
  them, so clicks could miss); dragging across eyes or padlocks sets every row passed in one undo
  step. Layers inside a locked frame show a faint padlock.
- Several layers selected + Cmd/Ctrl+R / F2 / Rename open a Rename layers dialog (Match, Rename to,
  Current name / Number up / Number down buttons, start value, preview) instead of renaming the top
  row only.
- `object.batchRename`: `$n` counts up and `$N` down (they were swapped), `$nnn` pads, `start`
  sets where numbers start/stop, `regex: true` matches a regular expression with `$1`..`$9`,
  `` $` `` and `$'`, and codes work in the replacement too. Returns the new names.
- Pages: dragged up and down the list to reorder; a blank page named with a leading hyphen is a
  divider (drawn as a line, never opened, holds no layers); duplicates are named "Copy of ...";
  the last page that isn't a divider can't be deleted. `page.list` reports `divider`.
- Assets panel: components filed in folders by page, the frames holding them and the slash parts
  of their names (folders fold away; open while searching), tiles show the last name part;
  `component.list` gives each component's `folder` and `label` for agents.
- Publishing: a leading `.` or `_` keeps a component (style, collection) out of a library even
  after it was published, and removing the prefix brings it back (the previous publish's choice
  won before).

The engine gains the `regex` crate (already in the lock file through env_logger) for the
documented regular-expression matching.
2026-10-12 03:47:34 +07:00
bflatastic e917d3c2e6 M1: frames, groups, sections and masks behaviour suite and fixes
**A frames, groups, sections and masks behaviour suite written from the public documentation, and fixes for every divergence it found.** Each documented rule (frames vs. groups when resized, constraints, Cmd/Ctrl-resize, clip content, Frame selection, grouping and ungrouping, group bounds and opacity, X/Y in groups, click targets, drawing a frame or section around layers, dropping into and out of frames, where sections may go, resizing sections, top-level frames as screens and exports, Use as mask, what a mask covers, alpha / vector / luminance masks, groups as masks, clicks on masked-out areas, a mask group's box and exports) is a table-driven case with the result the rule implies and a one-line derivation; where the documentation is silent the case says so and pins today's behaviour.

**Fixes (the suite failed on 7 tests before them)**
- **Masked-out areas don't take clicks**: hit testing and Select layer skip what a visible mask hides (the nearest mask below a layer; for a group or frame used as a mask, what its layers draw), so a click there reaches what's behind. Checked only once something is hit, and skipped in files without a mask (`Document::has_masks`, worked out once after each change), so hit testing costs what it did.
- **A mask group's box and exports are the mask's**: a group's bounds and export area count a mask for the layers it masks (layers under the mask still count), so the selection box, W/H, snapping and PNG/SVG/PDF exports crop to the mask instead of the whole masked layer (`Document::box_children`).
- **Sections stay on the page or in sections**: moving a section into a frame or group (`node.reparent`, refused at the root in `Document::reparent`), dragging one over a frame (it lands in the section around the frame, or on the page), grouping or framing a section, creating a section in a frame and pasting one into a selected frame (it goes on the page) no longer put sections where they can't be.
- **Drawing a frame or section around layers puts them in it**: the layers fully inside the drawn box (its siblings; visible and unlocked; a frame never takes a section; not in auto layout) move into it, in their order, without moving on the page; one undo step.
- **Resizing a section leaves its contents** where they are on the page (no constraints in sections), from a field, a handle (also the left/top edges) or `transform.resize` with any anchor.
- **Cmd/Ctrl while resizing a frame leaves its layers as they are**: #133 (merged in) keeps their places; this branch also keeps their sizes (the frame is resized without running constraints, so a left & right or scale child no longer stretches and is only moved back).
- **X/Y are measured in the frame, not the group**: groups and boolean groups have no coordinate space of their own, so a layer in a group shows (Design panel, `node.get`, `inspect_document`) and takes (`node.set`, `node.create` with a group parent) X/Y in the frame the group is in, or on the page (`Document::frame_relative`). Rotated groups too.

Already there and now covered by the suite: constraints when a frame resizes, groups scaling their layers, clip content (drawing, hits, export size), Frame selection's size and place, grouping/ungrouping keeping world transforms (rotated groups too), nested groups, empty groups going away, group opacity on the composite, click / Cmd-click / double-click targets (frames in sections count as top-level frames), dropping by the pointer and pulling out, group layers staying in their group when dragged, top-level frames as prototype screens and export boards, Use as mask (bottom layer, "Mask group", at the topmost layer's place, single-layer toggle), masks covering every layer above them, hidden masks, alpha / vector / luminance masks, a group as a mask, masks in SVG and PDF exports.

No new commands or UI strings (no i18n changes); each edit stays one undo step.
2026-10-12 00:57:43 +07:00
BFlat cc30a261e7 M1: layers panel keyboard (box selection, F2 rename, context menu) and stroke styles (#135)
## What

**Layers panel from the keyboard** (two of the three open items of area 3, plus long-list QA)
- **F2 renames** the selected layer (a second key for `object.rename`, next to Cmd/Ctrl+R) from the canvas, after a click on a row, or from a layer row that has the keyboard (F6 mode, where the canvas shortcuts are off, so the layers branch of `a11y::keys` runs it). When the rename ends the row gets the keyboard back (`a11y::refocus_row`); Escape in the field cancels and stays in the panel instead of jumping to the canvas.
- **Shift+F10 opens the context menu** (`view.contextMenu`, a new command that only asks the UI): at the focused layer row when the layers panel has the keyboard (F6 or a clicked row), else on the canvas just below the selection (the middle of the canvas with nothing selected). The menu takes the keyboard like the other menus opened from the keyboard (`a11y::popup`), and Escape gives it back to the row or the canvas. The Menu key can't be supported: egui 0.36 has no key for it.
- **Keyboard box selection** (`select.area`, F8): for keyboard-only users, a selection box appears around the selection (or in the middle of the view); the arrows move it, Shift+arrows move its far corner (10 screen pixels a press), Enter selects what it covers exactly like a marquee (Shift+Enter toggles, Cmd/Ctrl+Enter takes nested layers), Escape or a click drops it. A toast and the canvas's screen-reader description say the box's size and the keys. With a box (`{x, y, width, height, add?, deep?}`) the command selects directly, for agents. The marquee's rules (top-level layers, frames only when covered, the children of the frame the box starts in, the focus view, Shift toggling) moved into `Session::marquee_within` / `marquee_selection`, shared by the pointer marquee and the command.
- **Long lists (10k layers)**: a test that a layer near the bottom of a 10k-layer list is revealed (parents expanded, list scrolled) and that F6 lands the keyboard on its row and the arrows walk on from it; the UI bench (`perf_ui_frames`) gained rows for expanding every layer, arrow keys down the expanded list, revealing a layer at the bottom and collapsing everything. Release, 10k layers: first frame with every layer expanded 1.22 ms, an arrow key 1.04 ms, reveal 0.30 ms, collapse all 0.37 ms, idle 0.17 ms.

**Stroke styles** (area 6)
- What the public documentation describes for strokes was mostly already in: colour styles apply to strokes from the stroke section's style picker (`style.apply {target: "stroke"}`, linked, following `style.update`, one undo step, `style.detach {which: "stroke"}` keeping the colour), and the stroke weight binds to number variables. A test now covers that path.
- Missing piece, added: **each side's stroke weight binds to a number variable on its own** (`strokeTopWeight`, `strokeRightWeight`, `strokeBottomWeight`, `strokeLeftWeight`): `node.set` keys (setting or binding one side turns on individual strokes with the other sides at the weight), the stroke-weight variable scope, the stroke override group for instances, and REST `boundVariables` import. The Design panel's per-side fields are bindable, two to a row like the corner radii so a variable's name fits.

New strings go through `tr(...)` with `en.toml` / `es.toml` entries, and the two new commands have Spanish labels. Shortcuts are only in `keymap.rs`. Nothing redraws while idle (the box draws only when a key moves it). Selecting from the box leaves no undo step; each stroke edit is one.

## Verified
- **Desktop app** over the control channel (`target/debug/figcraft.exe` from this worktree's target dir, `--control <free port>`, separate `APPDATA`, keys sent as real `ui.input` key events, stopped by PID): F2 on a selected layer opens its name field in the layers panel (screenshot), typing and Enter rename it; F6 to the layers panel, Shift+F10 opens the menu at the Avatar row with Copy focused (screenshot), Escape gives the keyboard back to the row, F2 + typing + Enter renames it and the row keeps the keyboard ("Profile"); Shift+F10 on the canvas opens the menu for the selected frame with Copy focused (screenshot); F8 shows a box around the selected frame, Shift+arrows stretch it over a second layer (screenshot with the hint toast), Enter selects both; a rectangle with the colour style "Brand/Accent" on its stroke and its top weight bound to a number variable (screenshot: the style chip in the stroke section, the red stroke thicker at the top, the per-side field showing the variable).
- **MCP**: `figcraft-cli mcp` (headless): created a paint style, `style.apply {target: "stroke"}`, `object.rename`, bound `strokeLeftWeight` to a variable and ran `select.area` over the layer; `inspect_document` shows the name "Outlined card", `styles.stroke: "S:1"` with the style's colour on the strokes, `strokeSides [2, 2, 2, 6]`, `boundVariables {strokeLeftWeight}`, and one undo step per command.
- **Web**: `cargo xtask wasm` passes (part of the gate).
- **Performance**: release UI bench at 10k layers (numbers above).
- `cargo +1.99.0 xtask ci` passes (fmt, clippy, tests, layers, assets, wasm), after merging #131 and #132.

## Tests
- `ui-egui` `layers_keyboard_tests` (3): F2 from the canvas and from a focused row (the row gets the keyboard back, Escape cancels without leaving the panel); Shift+F10 at a focused row, on the canvas below the selection and after a clicked row, the menu taking and giving back the keyboard; F8: the box around the selection, the screen-reader description, Shift+arrows sizing, Enter selecting, arrows moving it without nudging layers, Escape, the middle of the view with nothing selected, a click dropping it.
- `ui-egui` `perf_tests::a_layer_revealed_at_the_bottom_of_10k_layers_gets_the_keyboard_on_its_row`, plus the new bench rows.
- `engine` `select_area_tests` (2): marquee rules through the command (negative sizes, frames, starting inside a frame, deep, add toggling, empty spots), the UI request with no box, bad boxes refused (the junk sweep covers both new commands too).
- `engine` `stroke_style_tests` (2): a colour style on strokes (link, undo, style edits, detach); stroke weight and per-side weights bound to variables following the frame's mode, `node.set` per-side keys, scopes and override groups.

## Not in this PR
- The Menu (application) key: egui 0.36 doesn't deliver it.
- Enter doesn't rename: on the canvas and a focused row it already selects children / activates the row.
- The keyboard box doesn't pan the view when it is moved off screen (zoom and pan keys still work while it shows).
- Shift+F10 does nothing while a toolbar or properties control has the keyboard.
- Progress docs: area 3 95%/78% → 98%/82%, 2–4 h → 1–2 h; area 6 90%/72% → 95%/75%, 1–3 h → 1–2 h; mainstream depth 67.8% → 68.1%; percentages unchanged at this rounding; full parity ~323–621 h → ~322–618 h (target-app-parity, ROADMAP); ui-parity rows (context menus, keyboard box selection, layers from the keyboard); `docs/mcp.md` documents `object.rename` without a name, `view.contextMenu`, `select.area` and stroke styles and variables.
2026-10-11 23:10:41 +07:00
BFlat 894f25c4c4 M1: selection and transform behaviour suite and fixes (#133)
A selection and transform behaviour suite derived from public documentation, with fixes: Alt-resizing shrinks as well as grows; groups, Scale-constrained frames and multiple selections resize each layer along its own axes keeping rotation and flip; several layers flip together past an edge; a flipped layer reads back the typed rotation; resizes land on whole pixels with the pixel grid; presses inside tiny layers move them; locked layers get no handles; a marquee can start on a top-level frame's background; select inverse skips hidden layers. Adds equal-spacing snapping with labelled gaps and Cmd/Ctrl to ignore constraints while resizing a frame.
2026-10-11 20:30:38 +07:00
BFlat 09a3b74466 M8: canvas fixes: boolean hit testing, group drags, frame outlines and slices view, a flaky idle test (#123)
Clicks on a boolean group test its computed result, so a click in a subtracted hole reaches what is behind it. Layers inside a group, boolean group or instance stay in their parent when dragged (one shared rule). View > Frame outlines and Slices now draw. The idle test's flake came from the render worker clearing its in-flight flag before waking the UI; jobs now carry sequence numbers.
2026-10-11 14:25:25 +07:00
BFlat cb5435d5ff M4: stretch and scatter brushes, pen pressure (#120)
Strokes can use a brush: eight presets drawn in code (four stretched along the path, four scattered with spacing, size, rotation and offset jitter from a fixed seed), drawn by the shared stroke-outline code in canvas, PNG, SVG, PDF and outline stroke, with caps on copies and steps. Pointer events carry pen pressure; the Pencil records it as the stroke's width profile, which brushes follow (Windows pens and browser styluses; constant width elsewhere).
2026-10-11 13:43:48 +07:00
BFlat f62cb4b1f6 M4: text on a path and pattern fills (#118)
Text layers can follow a path (start offset, flip, alignment; edited in place; canvas, PNG, SVG, PDF; SVG textPath imports as real text on a path). A new pattern paint tiles a source layer in grid or hex rows with scale, spacing and alignment, live as the source changes (SVG pattern, PDF vector tiling pattern).
2026-10-11 13:25:33 +07:00
BFlat ee82cdd138 M3: extended variable collections, variables import/export, gradient stop and variant bindings (#114)
Library collections can be extended in a consuming file (inherited modes, per-mode overrides with reset, extra variables; library updates keep overrides). Variables export to the REST JSON shape or CSS custom properties and import back (merged by name, validated by the REST reader, one undo step). Gradient stops and variant properties bind to variables. Also fixes clicking a gradient stop adding a new stop.
2026-10-11 10:49:16 +07:00
BFlatandClaude Opus 5.5 3c63e6d7c8 M3: components at scale; Alt-drag to swap instances (#91)
* M3: components at scale; Alt-drag to swap instances

- Synthetic design system (engine `synthetic_ds`): ~300 components (icons; button, input,
  card, nav and badge sets of size x state x type x theme variants), boolean/text/instance
  swap/slot properties, instances nested four deep, text and paint styles, light/dark
  variables bound across it, and 2,000 instances with overrides at every level.
- Propagation: value, size and added/removed-layer edits of mains reach their copies
  directly; other edits rebuild each affected instance once, inner mains first; rebuilds keep
  unchanged layers shared; bound variables no longer touch unchanged layers; auto layout
  re-flows only where sizes or places changed.
- Fixes: values from variables (theme switches, bound radii) are no longer recorded as
  overrides; a detached variable on an instance layer is an override that survives main
  edits; exposed nested instances stay exposed; swapped nested instances keep their layer
  names and the ids/overrides of layers their own main adds; renamed nested instances keep
  their name when detached out.
- Instance swap properties have preferred values. `instance.swapAt` / `library.swapAt` and
  Alt-drop from the Assets panel onto an instance swap it (highlighted while dragging).
- Instances pasted into other files follow their mains (library copies shared, published
  components used from their library, other mains left behind as a missing main).
- Assets panel lays out only visible tiles and renders thumbnails lazily with a cache
  invalidated per changed component; instance panel lists components once per revision.
- `figcraft-cli perf --instances`, UI frame bench and tests in components_scale_tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* M3: the swap perf case drops on a glyph the hit test reaches

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Merge origin/main into feat/components-scale; weighted depth 79%

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-10 21:07:43 +07:00
BFlatandClaude Opus 5.5 fe550fc14a M6: PDF print options (bleed, marks, CMYK) (#92)
PDF export gains print options, saved on PDF export settings and taken
by file.export, file.exportLayers, file.exportFramesPdf, export.add /
export.update, figcraft-cli export and the MCP export tool:

- a bleed (points or a length like 3mm / 0.125in, up to 2 in) that grows
  each page and shows what lies past the frame: its fills reach out to
  the bleed edge and its children are clipped there instead; a print
  page of one frame is trimmed to the frame's box;
- crop and bleed marks: 0.25 pt hairlines outside the bleed in
  registration colour;
- MediaBox / BleedBox / TrimBox;
- CMYK colour: fills, strokes, text, gradient stops, images and
  rasterised effects converted from sRGB with the naive, deterministic
  K = 1 - max(R, G, B) formula; DeviceCMYK without an output intent (no
  CMYK ICC profile can ship in the repository), soft masks stay RGB.

The Design panel shows a PDF row's print settings as a summary line that
opens them; the Export dialog has print options for its PDF preset.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-10 20:50:29 +07:00