Merge main into fix/symbol-instances-are-clicked-and-marquee-touched

This commit is contained in:
dotCooCoo
2026-11-14 19:04:36 -08:00
191 changed files with 9727 additions and 954 deletions
+3 -3
View File
@@ -96,13 +96,13 @@ the two can still win that race.
| `ui.menu.list` / `ui.menu.invoke` | `{command, params}` | the full menu tree / invoke an item |
| `ui.contextMenu.list` | | the canvas context menu for the current selection, flattened like `ui.menu.list` (`path` holds its submenus). `ui.click {x, y, button: "right"}` on the canvas opens it, after selecting the object there unless it is already selected. With the Artboard tool it lists the artboard entries (see `docs/mcp.md`), after selecting the artboard there unless it is already selected. While a menu opened over locked objects is open, it has an `Unlock` submenu with one item for each locked object under the pointer, topmost first and at most 20, labeled with the object's name, which runs `layer.setProps {ids: [id], locked: false}`. A locked group gets one item, and the locked objects inside it get none. Hidden objects, objects on hidden, locked or template layers, and in isolation mode objects outside the isolated group get no item |
| `ui.tool.select` / `ui.tool.list` | `{tool}` | |
| `ui.pointer` | `{events:[{kind: down|drag|up|move|doubleclick, x, y, space?: "doc"|"screen", mods?}]}` | drive the active tool exactly like the mouse: `mods.cmd` held at a press with any tool but a selection tool drags with the selection tool chosen last (Direct Selection with the Pen until one is), and the release gives the tool back as it was; the reply lists `selection`, `key` (the key object, null without one) and `tool`. With the Shape Builder or the Live Paint Bucket and a selection of more than 128 segments, the app finds the regions off the UI thread, so a `move` replies before the region under the pointer is highlighted, and the highlight shows in the first frame after the regions are found. With the Shape Builder and such a selection, the app also merges off the UI thread: the reply to the events that end a click or drag comes once the merge is applied. A click or drag made before the regions are found does not wait for them; the app records it and makes it once they are found. Until then the canvas keeps drawing, with the touched regions highlighted and the busy pointer. A click or drag sent meanwhile, with `mods.cmd` too, and the Shape Builder's arrow keys are made after the merge in the order sent, and the reply to such a request comes once its events are made and any merge they start is applied; a request of `move` events alone is answered at once. A command that can edit, select, save or redraw the documents (an edit, an undo, a selection, a save or an export, a tool option, another tool, another document, closing, a dialog's preview, a menu item, a command of the app's own such as `ui.recolorDialog`, the proof and separations views, a preference, added fonts) waits for the merge and the input sent before it, and runs after them; queries that only read, such as `document.inspect`, run at once on the document before the merge |
| `ui.pointer` | `{events:[{kind: down\|drag\|up\|move\|doubleclick\|tripleclick, x, y, space?: "doc"\|"screen", mods?}]}` | drive the active tool exactly like the mouse (`doubleclick` follows the second click's `up`, and `tripleclick` the third's): `mods.cmd` held at a press with any tool but a selection tool drags with the selection tool chosen last (Direct Selection with the Pen until one is), and the release gives the tool back as it was; the reply lists `selection`, `key` (the key object, null without one) and `tool`. With the Shape Builder or the Live Paint Bucket and a selection of more than 128 segments, the app finds the regions off the UI thread, so a `move` replies before the region under the pointer is highlighted, and the highlight shows in the first frame after the regions are found. With the Shape Builder and such a selection, the app also merges off the UI thread: the reply to the events that end a click or drag comes once the merge is applied. A click or drag made before the regions are found does not wait for them; the app records it and makes it once they are found. Until then the canvas keeps drawing, with the touched regions highlighted and the busy pointer. A click or drag sent meanwhile, with `mods.cmd` too, and the Shape Builder's arrow keys are made after the merge in the order sent, and the reply to such a request comes once its events are made and any merge they start is applied; a request of `move` events alone is answered at once. A command that can edit, select, save or redraw the documents (an edit, an undo, a selection, a save or an export, a tool option, another tool, another document, closing, a dialog's preview, a menu item, a command of the app's own such as `ui.recolorDialog`, the proof and separations views, a preference, added fonts) waits for the merge and the input sent before it, and runs after them; queries that only read, such as `document.inspect`, run at once on the document before the merge |
| `ui.key` / `ui.text` | `{key, shift?, alt?, cmd?}` / `{text}` | synthetic keyboard input |
| `ui.move` | `{x, y}` | screen-point pointer move |
| `ui.click` / `ui.drag` | `{x, y, button?, count?, shift?, alt?, ctrl?, cmd?}` / `{x, y, toX, toY, steps?, count?, button?, shift?, alt?, ctrl?, cmd?}` | egui pointer input; see synthetic input budgets above |
| `ui.wheel` | `{x, y, dy?, dx?, unit?: "line"\|"point", shift?, alt?, cmd?}` | a mouse wheel turn over screen point (x, y): `dy` notches up (+) or down, `dx` sideways. Over the canvas the wheel scrolls, Cmd-wheel (Ctrl on Windows and Linux) scrolls sideways and Alt-wheel (Option on the Mac) zooms about the pointer (with the `ctrlWheelScrollsSideways` preference off, Cmd-wheel zooms too); with the `zoomWithMouseWheel` preference the wheel and Alt-wheel zoom about the pointer, Shift-wheel scrolls up and down and Cmd-wheel (Ctrl on Windows and Linux) sideways. Over a focused numeric field (click it first) each notch steps its value as Up/Down do (Shift: ten, Cmd/Ctrl: a tenth) and the panel stays put; over anything else in a panel the wheel scrolls it. Over the canvas, the wheel preferences in `docs/mcp.md` change what these keys do |
| `ui.set` | `{brightness?, panel?, rulers?, outline?, grid?, smartGuides?, boundingBox?, controlBar?, imageTracePreview?}` | `outline`: View › Outline for the active document only (each document has its own, a document opens in Preview, and `ui.inspect` shows it as `view.outline`; an error without a document). `imageTracePreview`: the Image Trace panel's Preview (`ui.inspect` shows it as `ui.image_trace_preview`) |
| `ui.dialog.set` / `.confirm` / `.cancel` | `{field, value, action?}` or `{fields: [{field, value}…], action?}` | fill and submit the open dialog. `ui.dialog.set` takes the fields the dialog has (`ui.inspect`) other than its `__` fields, the `__` fields that are controls of the dialog (such as `__section` and `__allArtboards` in Save PDF, `__constrain` in Transform Each and the shape dialogs, and `__search` and `__category` in Preferences; Save PDF's and Print's `__savePresetAs`, the preset name, only while Save Preset… asks for it, since OK then saves a preset), `discard` on a dialog with a discard button, `preview` on one with a live preview, the fields that the dialog reads but adds only when a control changes (such as `searchFolder`, or `copy` in the Transform dialogs), and the fields it passes to the command it runs: every parameter of the `command` dialog's command (`ui.paramDialog`), and the `textureImage` that Load Texture… sets in the Glass and Texturizer effects; any other field is an error that lists the dialog's fields, and the dialog stays as it was. Any other `__` field, held or not, holds what the dialog's OK acts on (such as the `__command` and `__params` of a `confirm` or `command` dialog) or the dialog's own state: setting it is the error "the `<kind>` dialog cannot set protected field `<field>`", and the dialog stays as it was. `fields` sets the list (at most 256 items) in order (a missing `value` is `null`) and checks every item first: when one is refused, the call is an error and no field changes. `action: "confirm"` or `"cancel"` on `ui.dialog.set` then confirms or cancels the same dialog in the same request, as `ui.dialog.confirm` and `ui.dialog.cancel` do, and the reply is `{dialog, result}`: the dialog after the fields were set and the reply to confirm (`null` for cancel); any other `action` is an error and no field changes |
| `ui.dialog.set` / `.confirm` / `.cancel` | `{field, value, action?}` or `{fields: [{field, value}…], action?}` | fill and submit the open dialog. `ui.dialog.set` takes the fields the dialog has (`ui.inspect`) other than its `__` fields, the `__` fields that are controls of the dialog (such as `__section` and `__allArtboards` in Save PDF, `__constrain` in Transform Each and the shape dialogs, and `__search` and `__category` in Preferences; Save PDF's and Print's `__savePresetAs`, the preset name, only while Save Preset… asks for it, since OK then saves a preset), `discard` on a dialog with a discard button, `preview` on one with a live preview, the fields that the dialog reads but adds only when a control changes (such as `searchFolder`, or `copy` in the Transform dialogs), and the fields it passes to the command it runs: every parameter of the `command` dialog's command (`ui.paramDialog`), and the `textureImage` that Load Texture… sets in the Glass, Texturizer and Rough Pastels effects; any other field is an error that lists the dialog's fields, and the dialog stays as it was. Any other `__` field, held or not, holds what the dialog's OK acts on (such as the `__command` and `__params` of a `confirm` or `command` dialog) or the dialog's own state: setting it is the error "the `<kind>` dialog cannot set protected field `<field>`", and the dialog stays as it was. `fields` sets the list (at most 256 items) in order (a missing `value` is `null`) and checks every item first: when one is refused, the call is an error and no field changes. `action: "confirm"` or `"cancel"` on `ui.dialog.set` then confirms or cancels the same dialog in the same request, as `ui.dialog.confirm` and `ui.dialog.cancel` do, and the reply is `{dialog, result}`: the dialog after the fields were set and the reply to confirm (`null` for cancel); any other `action` is an error and no field changes |
| `ui.screenshot` | `{path?, data?}` | capture the window (PNG), written to `path`; `data: true` sends it back as `pngBase64` too. The inline PNG is capped at 64 MiB: with a `path` the file is still written and `pngBase64` is omitted when the image is larger, while a path-less `data: true` request fails if either the raw RGBA image or the encoded PNG exceeds the cap. Needs a presented frame: with the screen locked or the window minimized/covered it fails after ~8 s with an explanatory error |
| `ui.render` | `{path?, scale?, artboard?, data?, traceViews?}` | render an artboard headlessly (PNG with the pixels of `app.export {format: "png", background: "white"}` at the same scale; `artboard`: its 0-based index, default the first; `traceViews: true`: Image Trace objects draw in their View, as on the canvas, instead of their tracing result): written to `path`, else (or with `data: true` as well) sent back as `pngBase64` |
| `ui.render` | `{path?, scale?, artboard?, data?}` | render an artboard headlessly (PNG with the pixels of `app.export {format: "png", background: "white"}` at the same scale; `artboard`: its 0-based index, default the first): written to `path`, else (or with `data: true` as well) sent back as `pngBase64` |
@@ -123,7 +123,7 @@ the two can still win that race.
| Background Save / Export | `app.save`, `app.export {path}`, `file.save`, `file.saveAs`, `file.saveCopy`, `file.exportForScreens {folder, …}` | with the `backgroundSave` / `backgroundExport` preferences on (the default) the desktop app writes the file from a snapshot on a worker thread: the call returns `{path, background: true, status}` at once (`file.exportForScreens`: `{background: true, status}`, and the files are written into the folder from a snapshot taken at the call; an `app.export` given an option the format doesn't read: also `warnings`, naming the option), `ui.inspect` lists the jobs still running in `background` (labels such as `"Saving a.vectorcraft"`), and the document counts as saved (as of the snapshot) once its file is written. Quit and Close wait for them. A job that fails reports the error in the status bar only (`ui.inspect` → `ui.status`), a path outside `--automation-write-root` included. With the preferences off (`prefs.set`), saves return once written. `app.save {path?, wait: true}` and `app.export {path, wait: true}` wait for the jobs already running, write the file before they reply whatever the preferences are set to, and return its result or the error. When the save asks first (the question before replacing a file that opened with losses, or on macOS and Linux the file dialog for an untitled document without `path`), the reply is `{pending}` or `"cancelled"` as before, and the save that runs after the answer uses the preferences. A `wait` that is not `true`, `false` or `null` is refused with "`wait` must be true or false", and nothing is written |
| `engine.execute` `file.exportForScreens` | `{}` / `{…document.exportForScreens params}` | without params opens `exportForScreens` on the settings the document last exported with (`document.exportSettings`). While the Artboard tool is in use, the artboards it has selected (its `selected` option, the rows the Artboards panel highlights) are checked instead, and `select` is `range`, or `all` when every artboard is selected. Fields: `tab` (`artboards` \| `assets` \| `layers` \| `selection`; OK reads Export Layer or Export Selection on those), `layers` ([layer id…] checked), `selectionMode` (`one` \| `each`), `name` (the file name pattern of `document.exportForScreens`), `counterStart`, `folders` (`none` \| `rows` \| `format` \| `item` \| `itemFormat`), `exportPreset` (the preset applied last), `__savePreset` (the name typed to save one), `select` (`all` \| `range` \| `full`), `range` ("1-3, 5"), `boards` (a bool per artboard; checking them sets `select` and `range`), `includeBleed`, `folder` (desktop; default the Desktop, else home), `openLocation`, `subfolders`, `prefix`, `preset` (`""` \| `mobile` \| `density`), `formats` (rows `{format, scale: "2x" \| "100w" \| "100h" \| "72ppi", suffix, quality?, folder?}`), `settings` (`{png \| png8 \| jpg \| webp \| gif \| svg \| pdf: {…options}}`, the gear's Format Settings; `__settings` names the format shown). `ui.dialog.confirm` exports (a bad range or no folder keeps it open); on the web OK reads "Download" and the files download (several as one `.zip`). With params it runs `document.exportForScreens`: no `folder` returns the files (or with `zip: true` one zip) as `dataBase64`; with `openLocation` the first file is shown in the file manager. When `warnings` is not empty, an export into a folder or a download shows the first in the status bar as Save As shows a save's notes |
| Export for Screens' Assets tab, Asset Export panel | `ui.dialog.set` `tab: "assets"`, `assets: [asset id…]`; `window.panel {panel: "assetExport"}` | on the Assets tab of `exportForScreens` the `assets` field lists the checked asset ids (all of them when it opens); `ui.dialog.confirm` runs `document.exportForScreens {assets}` (nothing checked keeps it open) and OK reads "Export Asset". The Asset Export panel acts through `assets.*` and shares the format rows with the dialog (`assets.settings.set`). The dialog has the format presets menu (☰). In the dialog, `exportPreset` names the preset applied last (`ui.dialog.set` of it applies the preset it names, as picking it does; a name that is not a preset's is an error and leaves the dialog as it was), `__exportPresets` lists the presets, and `__savePreset` holds the name being typed after Save Preset…. The panel menu applies presets through `assets.settings.set {exportPreset}` and holds Save As Preset…, Save Changes, Delete Preset, Import Preset… and Export Preset…. Import and Export use picked files. Its Create Sub-folders and Open Location after Export run `assets.settings.set {subfolders}` and `{openLocation}`; its Format Settings… opens the dialog on the Assets tab with `__settings` set to the first row's format. |
| `engine.execute` `file.exportSelection` | `{multiple?}` | File › Export Selection…: collects the selected objects as one asset (`assets.add`; one per object with `multiple: true`) and opens `exportForScreens` on its Assets tab with just them checked → `{assets}`. `document.exportSelection` with params still exports the selection in one go, without the dialog; with no params or `{}` the save panel picks the file and its format's options dialog follows, as after Export As, and the call returns `null` while the dialog is open (OK returns what `file.exportAs` returns, without `format` or `bounds`). `file.exportAs {path, exportSelection: true, …options}` writes the selection without a dialog |
| `engine.execute` `file.exportSelection` | `{multiple?}` | File › Export Selection…: collects the selected objects as one asset (`assets.add`; one per object with `multiple: true`) and opens `exportForScreens` on its Assets tab with just them checked → `{assets}`. `document.exportSelection` with params still exports the selection in one go, without the options dialog (without `path`, the save panel picks the file first); with no params or `{}` the save panel picks the file and its format's options dialog follows, as after Export As, and the call returns `null` while the dialog is open (OK returns what `file.exportAs` returns, without `format` or `bounds`). `file.exportAs {path, exportSelection: true, …options}` writes the selection without a dialog |
| CSS Properties panel | `window.panel {panel: "cssProperties"}`; `engine.execute` `css.copy` / `css.exportFile` | the panel shows `css.selection` of the selection (after Generate CSS, while nothing is selected, `css.generate`); its menu holds the options the commands take (`units`, `position`, `dimensions`, `unnamed`, `rasterize`). `css.copy {scope?, …options}` copies the CSS to the clipboard → `{css, rules}`; `css.exportFile {path?, scope?, …options}` writes `css.export` to `path` (else a picked file; the web downloads it) with the pictures of rasterized art next to it → `{path, rules, images, warnings}`, and the status bar shows `warnings` as Export As shows an export's notes |
| `ui.dialog.*` on `importPdf` | `{field: "page" \| "allPages" \| "range" \| "cropTo" \| "password", value}` | opening (`app.open`, `file.open`) a PDF, `.ai` or `.ait` with a password, or one with several pages unless it is a `.ai` or `.ait` with its editing data or carries its VectorCraft document (those open whole), or placing (`file.place` without `page`) any of them with several pages or a password, opens the Import PDF dialog (`mode`: `open` or `place`; `pageCount`). While `__locked` is true, set `password` and `ui.dialog.confirm` to unlock it (a one-page file then opens or is placed). Then set `allPages: false` and `range` ("2-3, 5") to open some pages, or `page` to place one, and `cropTo` (`bounding`, `art`, `crop`, `trim`, `bleed`, `media`); `ui.dialog.confirm` opens or places them (placing keeps the other `file.place` params, such as `link`, `template` and `at`, in `__place`). `document.open {pages, cropTo, password}` and `file.place {page, crop, password}` do the same without the dialog |
+4 -3
View File
@@ -245,7 +245,8 @@ private editing data. VectorCraft:
group is tagged once and paints its marks with one fill and one stroke, which the graph keeps as that series'
paint; Graph Data then edits it. Other graphs (pie and scatter graphs, graphs with designs or
customizations, records of later versions) come in as plain groups, paths and type, with a note. Saving a graph to
`.ai`, in either form, writes its art without its data or settings, with a warning.
`.ai`, in either form, writes these records around its art when the reader above reads them back as the same graph,
so the file opens again with its data and settings; any other graph is written as its art, with a warning.
Legacy `.ai` patterns in the version 7 form (a list of `@` and `&` layers, `_ &` for the tile's background)
open as pattern swatches, with a fill's offset, scale, rotation, reflection and shear applied before its
matrix. Legacy `.ai` linked images in
@@ -290,8 +291,8 @@ Names in the editing data that aren't UTF-8 (layers, objects, artboards, colors,
| HP-GL/2 `.plt`, `.hpgl`, `.hgl`, `.hpg` | — | yes (pen moves, pens and widths, line types and attributes, circles, arcs, wedges and Bézier curves, filled and edged rectangles, wedges and polygons, encoded polylines, scaling and rotation; labels, symbol mode, raster images and hatching aren't drawn) | yes (`.plt`: each visible path once along its line, a pen per color or a pen table, type as outlines, clipping applied, HP-GL or HP-GL/2 with pen colors and encoded polylines) | recognized by its first instructions after any printer escapes, under its own extensions or one no other format reads; `crates/cad/src/hpgl/tests.rs`, `crates/cad/src/hpgl/write/tests.rs` (plots read back by the reader), `tests_hpgl.rs`, `import_fuzz.rs` |
| DWG | open, export | no | no | no open spec; DXF instead |
| EMF / WMF | open, export | yes | yes | an oversized embedded PNG or JPEG is refused from its header, before any pixels are decoded (#1776) |
| PSD | open, place, export | yes (open: layers as layers, or the merged image with `layers: false`; place: merged image; every mode and depth; `.psd` and `.pdd`) | yes (layered, max editability) | placing as layers and a PSD import options dialog missing; the embedded ICC profile's tags are 4-byte aligned (#1584) |
| PSB | open | yes (layers as layers, or the merged image) | no | |
| PSD | open, place, export | yes (open: layers as layers, or the merged image with `layers: false`; place: merged image; the merged image comes in clipped by the file's clipping path; every mode and depth; `.psd` and `.pdd`) | yes (layered, max editability) | placing as layers and a PSD import options dialog missing; the embedded ICC profile's tags are 4-byte aligned (#1584) |
| PSB | open | yes (layers as layers, or the merged image; the merged image comes in clipped by the file's clipping path) | no | |
| PNG / JPEG / GIF / BMP / TIFF / Targa | open, place, export | yes (PNG: `.png` and `.pns`; BMP: `.bmp`, `.dib` and `.rle`; Targa: `.tga`, `.vda`, `.icb` and `.vst`, raw or run-length encoded, with a 16-bit file's one-bit alpha) | yes; TIFF is also selectable in Export for Screens and Asset Export with its existing options | Targa: `tests_tiffbmp.rs`, `tests/links.rs`; TIFF: grayscale with an alpha channel or other extra channels reads (#1466 in part), and so do files that store each channel as its own plane (PlanarConfiguration 2), CMYK ones with their ink amounts, and 1-bit files compressed with CCITT Group 3 (Compression 2 and 3, one- and two-dimensional rows); the ICC profiles embedded in JPEG and TIFF have 4-byte aligned tags (#1584); PNG export marks its colours as sRGB (`embedIcc`, an `sRGB` chunk; no ICC profile is embedded, #2221) |
| WebP | export (Export for Screens) | yes | yes (lossless only; lossy written lossless with a warning) | |
| JPEG XL `.jxl` | — | yes (first frame; 16-bit and floating-point samples as 8 bits, alpha kept; CMYK refused) | no | decoded with jxl-oxide; the size is checked from the header before decoding; container Exif resolution sizes opened and placed images; tests in `crates/engine/src/cmd/fileio/tests_jxlread.rs` and `crates/engine/tests/import_fuzz.rs` |
+1
View File
@@ -522,6 +522,7 @@ macOS cursor delivery (#2292) is narrowed to a native host adapter using the exi
## G2 Photoshop-style raster effects and the Effect Gallery
- **Done:** Effect › Texture › Texturizer › Load Texture… (stored grayscale `textureImage`, Loaded Texture in the Texture menu), matching Glass.
- **Done:** Effect › Artistic › Rough Pastels › Load Texture… ([raster effects reference](https://helpx.adobe.com/illustrator/desktop/special-effects-styles/apply-filter-effects/raster-effects-reference.html)): the image's gray levels are the paper surface, stored with the effect, listed as Loaded Texture in the Texture menu, with Scaling, Relief, Light Direction and Invert acting on it.
- **Done:** the Effect Gallery's filter thumbnails and in-dialog preview ([Filter Gallery overview](https://helpx.adobe.com/photoshop/desktop/effects-filters/get-started-with-filters/filter-gallery.html)). Each thumbnail is an original landscape drawing passed through that filter at its defaults; the preview area shows the same drawing through the visible stack, or through the filter being browsed when the stack is empty. The canvas preview and the one-undo OK are unchanged (`dialogs::effect_gallery::tests::gallery_shows_thumbnails_and_an_in_dialog_preview`). All 57 Photoshop-style filters are in the catalogue. Measured in
[effects-parity.md](effects-parity.md).
- **Done:** the Effect Gallery (#2883): Effect › Effect Gallery… browses the Artistic, Brush Strokes, Distort, Sketch, Stylize and Texture filters, edits each filter's settings, builds an ordered stack (add, reorder, hide, remove) with a live canvas preview, and OK applies it as separate appearance effects in one undo step (`dialogs::effect_gallery::tests::stack_ok_is_one_undo_step`, `stack_preview_is_one_interaction`).
+15 -2
View File
@@ -32,6 +32,13 @@ every clause such as "all" or "what follows it". Type as a text object is टा
in `crates/ui-egui/src/i18n/tests.rs` checks the save, swatch, crop, resolution and path terms
and the point unit in every row.
Translate a label by its sense in the interface, not by its first dictionary sense. A ruler is
रूलर, never शासक (a sovereign); a text character is कैरेक्टर, never चरित्र (a personality); a
vector point is पॉइंट and a printer's mark is मार्क, never अंक (a digit); fills are फ़िल्स, never
भरता (the verb). In a minimum, desired and maximum range, Desired is वांछित, never देसी (local),
and a range message keeps both comparisons: minimum ≤ desired ≤ maximum. The same test rejects
these senses, and others found in the catalog, in every row whose English names the word.
A variable is वेरिएबल (plural वेरिएबल्स) in every sense: a data-merge variable and the
variable library, a variable font, a variable stroke width and an environment variable. It is
never चर, परिवर्तनीय or वैरिएबल. Joining paths, anchor points, point type or the lines of a
@@ -61,7 +68,10 @@ checks every `@msg` row's lists.
| Artwork | आर्टवर्क |
| Path / Anchor point / Handle | पाथ / एंकर पॉइंट / हैंडल |
| File path on disk | फ़ाइल पाथ |
| Fill / Stroke | फ़िल / स्ट्रोक |
| Fill / Fills / Stroke | फ़िल / फ़िल्स / स्ट्रोक |
| Point (vector) / Printer's marks | पॉइंट / मार्क्स |
| Ruler / Ruler origin | रूलर / रूलर ओरिजिन |
| Minimum / Desired / Maximum | न्यूनतम / वांछित / अधिकतम |
| Gradient / Mesh / Blend | ग्रेडिएंट / मेश / ब्लेंड |
| Pattern / Swatch | पैटर्न / स्वॉच |
| Type (text object) / Crop / Resolution | टाइप / क्रॉप / रिज़ॉल्यूशन |
@@ -89,7 +99,10 @@ percentage of the application's entire interface.
On 2026-11-13 the catalog has 8,782 rows, including 3,014 `@msg`, 47 `@msgplural` and 41
`@plural` rows. A wording pass on that date corrected 193 rows outside the plug-in
messages for meaning, dropped clauses, English left in place and the save, swatch, crop,
resolution, path and type terms. It is not a native proofread.
resolution, path and type terms. A second pass on that date corrected 85 short labels and
messages that used a word's unrelated sense, such as देसी (local) for Desired, शासक (a
sovereign) for a ruler and उच्च न्यायालय (a high court) for High Contrast. It is not a native
proofread.
The comparison covers release `v0.21.1` (`f6aa4309`) and current `main` (`3c7d9e1f`). Their
non-English catalogs have no missing Hindi keys except the names deliberately kept in English.
+1 -1
View File
@@ -31,7 +31,7 @@ Chinese by `imachus`). Georgian and the original Gujarati catalog are recorded a
| English | `en` | source, 100% | all | Latin | — | full | — |
| Simplified Chinese (Mandarin) | `zh-hans` | 5,102 rows, 791 of them `@msg` (counted 2026-10-21): menus, `tl!` literals and messages complete when counted | all (3,554 `@msg` and 48 `@msgplural` rows, counted 2026-11-13: the untranslated report lists 55 missing messages for it, which open #7116, #7211, #7330, #7378 and #7405 translate) | CJK UI font from craft-fonts (BIZ UDPGothic) when embedded, else a system font; IME with marked text at the caret; vertical type partial ([type-parity.md](type-parity.md)) | no | full | — (review) |
| Spanish | `es` | 4,009 / 4,020 (99.7%) | all (604 messages) | Latin | no | full | — (review) |
| Hindi | `hi` | 8,782 rows (2026-11-13): 3,014 `@msg`, 47 `@msgplural`, 41 `@plural`; the untranslated report on main `f793044f83` (2026-11-14) lists 676 missing rows for it: 608 in its own section and 68 that every checked language lacks | menus, dialogs, tooltips and scanned messages; contributor-approved loanword terminology; agent wording pass | UI and document text shape with HarfRust; installed desktop Devanagari fallback, or a `Deva` craft-fonts face; web font still needed | contributor terminology review; 2026-11-13 wording pass on messages and longer labels, with a glossary-term test; wider proofread welcome | complete catalog, machine-assisted draft | reported missing rows, native review, layout and web font |
| Hindi | `hi` | 8,782 rows (2026-11-13): 3,014 `@msg`, 47 `@msgplural`, 41 `@plural`; the untranslated report on main `f793044f83` (2026-11-14) lists 676 missing rows for it: 608 in its own section and 68 that every checked language lacks | menus, dialogs, tooltips and scanned messages; contributor-approved loanword terminology; agent wording pass | UI and document text shape with HarfRust; installed desktop Devanagari fallback, or a `Deva` craft-fonts face; web font still needed | contributor terminology review; 2026-11-13 wording pass on messages and longer labels, and a word-sense pass on short labels, with a glossary-term test; wider proofread welcome | complete catalog, machine-assisted draft | reported missing rows, native review, layout and web font |
| Gujarati | `gu` | 4,673 rows (2026-11-14): original 4,286 unchanged; all 387 additions contributor-reviewed and approved | menus and dialogs; status/error messages partial | installed Gujarati fonts on desktop; optional `Gujr` craft-fonts fallback; no Gujarati face in the pinned CI/release fonts or web startup list | original catalog reviewed by a native speaker; all 387 later additions approved by the contributor, with 11 wording corrections | contributor-reviewed catalog | remaining messages and a bundled Gujarati web font |
| Arabic | `ar` | 0 | none | document text: bidi layout and Paragraph Direction; interface: no right-to-left layout, no Arabic shaping in egui | no | none | 25–40 |
| French | `fr` | 9,166 rows, 3,436 of them `@msg` and 48 `@msgplural` (counted 2026-11-13): the untranslated report lists 178 missing messages for it | all | Latin | no | full | — (review) |
+170 -34
View File
@@ -647,8 +647,8 @@ sampler.
| `inspect_document` | `{depth?, childLimit?, ids?, kind?, limit?, offset?}` | Artboards (with `color`, `#rrggbb`, for one with a background color), layer tree (ids, kinds, bounds, paint), selection, history, tool. A type node's `fill`, `stroke` and `strokeWidth` are the paint its characters show (its first run's); fills and strokes of the type object itself come as `objectFill`, `objectStroke` and `objectStrokeWidth`. Paint comes as a label (`"#ff0000"`, `"None"`, `"G1 40% (#ff9999)"` for a swatch at a tint); `paint.proxies` gives it as an object. |
| `inspect_ui` | `{}` | UI state. Remote mode only. |
| `select_tool` | `{tool}` | `selection`, `directSelection`, `pen`, `rectangle`, `ellipse`, `polygon`, `star`, `lineSegment`, … |
| `pointer_gesture` | `{events:[{kind,x,y,mods?,backquote?,pressure?,tiltX?,tiltY?,rotation?,holdMs?}], tool?, mods?, backquote?}` | `kind` is one of `down`, `drag`, `up`, `move`, `doubleclick`. `backquote` is the held state for each event; the top-level value supplies its default, and a `backquote` that is not a boolean is an error. Events go through the same path as the mouse. Headless, a `move` replies after the tool's highlight under the pointer is ready, including the Shape Builder's and the Live Paint Bucket's regions of a large selection, which the desktop app finds off the UI thread and highlights after the reply. The release of a Shape Builder click or drag over such a selection replies once the merge is applied, headless and in the desktop app, which computes the merge off the UI thread. In the desktop app, a gesture sent before then is made after the merge, and its reply comes once it is made; a command that can edit, select, save or redraw the documents (a preference, the proof and separations views), sent before then, waits for the merge and runs after it. |
| `pointer_gesture` | `{events:[{kind,x,y,mods?,backquote?,pressure?,tiltX?,tiltY?,rotation?,holdMs?}], tool?, mods?, backquote?}` | `kind` is one of `down`, `drag`, `up`, `move`, `doubleclick`. `backquote` is the held state for each event; the top-level value supplies its default. Events go through the same path as the mouse. Headless, a `move` replies after the tool's highlight under the pointer is ready, including the Shape Builder's and the Live Paint Bucket's regions of a large selection, which the desktop app finds off the UI thread and highlights after the reply. Remote, the events wait until the keys and other input sent before them (`press_key`, `type_text`) have reached the app, so a key pressed between two calls of one drag applies before the release. When that input is dropped because the document, a dialog or the focus changed first, the call fails and its events are not sent. The release of a Shape Builder click or drag over such a selection replies once the merge is applied, headless and in the desktop app, which computes the merge off the UI thread. In the desktop app, a gesture sent before then is made after the merge, and its reply comes once it is made; a command that can edit, select, save or redraw the documents (a preference, the proof and separations views), sent before then, waits for the merge and runs after it. |
| `pointer_gesture` | `{events:[{kind,x,y,mods?,backquote?,pressure?,tiltX?,tiltY?,rotation?,holdMs?}], tool?, mods?, backquote?}` | `kind` is one of `down`, `drag`, `up`, `move`, `doubleclick`, `tripleclick` (`doubleclick` follows the second click's `up`, and `tripleclick` the third's). `backquote` is the held state for each event; the top-level value supplies its default, and a `backquote` that is not a boolean is an error. Events go through the same path as the mouse. Headless, a `move` replies after the tool's highlight under the pointer is ready, including the Shape Builder's and the Live Paint Bucket's regions of a large selection, which the desktop app finds off the UI thread and highlights after the reply. The release of a Shape Builder click or drag over such a selection replies once the merge is applied, headless and in the desktop app, which computes the merge off the UI thread. In the desktop app, a gesture sent before then is made after the merge, and its reply comes once it is made; a command that can edit, select, save or redraw the documents (a preference, the proof and separations views), sent before then, waits for the merge and runs after it. |
| `pointer_gesture` | `{events:[{kind,x,y,mods?,backquote?,pressure?,tiltX?,tiltY?,rotation?,holdMs?}], tool?, mods?, backquote?}` | `kind` is one of `down`, `drag`, `up`, `move`, `doubleclick`, `tripleclick` (`doubleclick` follows the second click's `up`, and `tripleclick` the third's). `backquote` is the held state for each event; the top-level value supplies its default. Events go through the same path as the mouse. Headless, a `move` replies after the tool's highlight under the pointer is ready, including the Shape Builder's and the Live Paint Bucket's regions of a large selection, which the desktop app finds off the UI thread and highlights after the reply. Remote, the events wait until the keys and other input sent before them (`press_key`, `type_text`) have reached the app, so a key pressed between two calls of one drag applies before the release. When that input is dropped because the document, a dialog or the focus changed first, the call fails and its events are not sent. The release of a Shape Builder click or drag over such a selection replies once the merge is applied, headless and in the desktop app, which computes the merge off the UI thread. In the desktop app, a gesture sent before then is made after the merge, and its reply comes once it is made; a command that can edit, select, save or redraw the documents (a preference, the proof and separations views), sent before then, waits for the merge and runs after it. |
| `draw_path` | `{points \| d, closed?, fill?, stroke?, strokeWidth?, dash?, cap?, join?, opacity?, name?, effects?}` | `points` is `[[x,y],…]` or `[{x,y,in?,out?,smooth?},…]`. `d` is SVG path data. The options are below. |
| `draw_shape` | `{shape, …geometry, fill?, stroke?, strokeWidth?, dash?, cap?, join?, opacity?, name?, effects?}` | `rectangle`/`ellipse`: `x,y,width,height` (plus `radius` for corners). `polygon`: `cx,cy,radius,sides`. `star`: `cx,cy,radius1,radius2,points`. Polygons and stars take `rotation` in degrees clockwise (0: a corner or point straight up). `line`: `x1,y1,x2,y2`. |
| `set_paint` | `{fill?, stroke?, strokeWidth?, ids?}` | Applies to the selection (or `ids`) and becomes the default for new art. Returns `{applied, objects}`: `objects` is the number of objects painted, as the `objects` of the `paint.setFill`, `paint.setStroke` and `stroke.set` replies (the largest of them): a selected group counts the objects inside it. When it is 0, a `note` says that only the paint for new art changed. |
@@ -1127,8 +1127,9 @@ follows as after Export As: Save PDF without its Artboards section, SVG Options
options, EPS Options, DXF Options without Export Selected Art Only, Text Export Options with Selection Only on, or an
exporter plug-in's parameters. Other formats are written at once. OK writes the selection alone, cropped to it, as
`document.exportSelection` writes it, with the chosen options. While a dialog is open the call returns `null`, and
OK returns what `file.exportAs` returns, which has no `format` or `bounds` (Text Export Options: `{path}`). To
write without a dialog, pass `path` or any option. `file.exportAs {path, exportSelection: true, …options}` writes the selection the same way.
OK returns what `file.exportAs` returns, which has no `format` or `bounds` (Text Export Options: `{path}`). Any
option skips the options dialog, but without `path` the save panel still picks the file. To write without a
dialog, pass `path`. `file.exportAs {path, exportSelection: true, …options}` writes the selection the same way.
Raster exports default to the document's raster effects resolution. An explicit `scale` sets pixels per
point (72 times that value in ppi); `options.ppi` wins over either scale option. CLI `run --export` and
@@ -1735,6 +1736,16 @@ one opens as before and draws it as `rgb`:
→ invalid parameters for `effect.setParams`: `channel` must be one of rgb, red, green, blue
```
Radial Blur's Blur Center, `centerX` and `centerY`, must each be a number from 0 to 1: the share of the object's bounds
across from its left edge and down from its top edge (0.5 each, the object's center, by default). The effect turns or
zooms about that point, which moves and scales with the object. `effect.apply` and `effect.setParams` refuse any other
value themselves in the same way, and a document saved with one draws it at the nearest edge:
```json
{"id": "effect.setParams", "params": {"index": 0, "params": {"centerX": 1.5}}}
→ invalid parameters for `effect.setParams`: `centerX` must be a number from 0 to 1
```
A number outside the range the effect documents is still stored as given and clamped when the effect draws. A
plug-in effect's parameters are resolved against its manifest instead. `effect.setParams` as a step of
`command.batch` is checked when the step runs, against the effects the steps before it added, and the batch then
@@ -3227,6 +3238,13 @@ already walks Image Trace. Short agent entry points:
unchanged. No undo step is added when the symbols are already in order. The panel menu's Thumbnail View, Small List
View and Large List View change only how the panel shows the symbols: thumbnails, or rows with a thumbnail and the
name (smaller rows in Small List View).
- Panel views: `ui.panelView {panel, view?}` sets the view of the Swatches, Graphic Styles or Symbols panel, as the
view items of its panel menu (and the Swatches panel's list and thumbnail buttons) do. `panel` is `swatches`,
`graphicStyles` or `symbols`. The Swatches panel takes `smallThumb`, `mediumThumb` (the default), `largeThumb`,
`smallList` or `largeList`; the Graphic Styles and Symbols panels take `thumbnail` (the default), `smallList` or
`largeList`. Without `view` the command answers the current one. Each panel keeps its own view for the session, and
the document does not change. An unknown panel or view is an error, and then nothing changes. It answers
`{panel, view}`.
- The Brushes panel menu's Select All Unused: `brush.unused` (a query) lists the brushes no stroke paints with, in
library order, and the panel selects them. Strokes in the art (opacity masks included), symbols, graph designs,
pattern tiles (with the tiles that pattern editing's Cancel puts back) and graphic styles count, and so do the
@@ -3343,32 +3361,48 @@ already walks Image Trace. Short agent entry points:
`dataset.import`, adds no undo step when it leaves the variables and data sets as they were. The library
is XML whose `variableSets` element holds `variableSet` elements, each with a `variables` element of `variable`
elements (`varName`, `trait`) and a `sampleDataSets` element of `sampleDataSet` elements (`dataSetName`). Every
data set is deleted, and every variable of the document is deleted with its bindings and values, except a text
variable the library also declares: it keeps its bindings, so selecting one of the library's data sets changes the
objects bound to it. Then the library's other variables whose `trait` is `textcontent` are defined as text
variables, unbound. In each data set, the element named
after a text variable gives its value: the text of its `p` elements, one paragraph each. An element without a `p`
element gives no value. Variables of other kinds are left out with their values: the reply lists them in `skipped`
(`name`, `trait`), and its `warning` names each `trait` with its count. The reply's `deleted` and `deletedDataSets`
data set is deleted, and every variable of the document is deleted with its bindings and values, except a variable
the library also declares with the same kind: it keeps its bindings, so selecting one of the library's data sets
changes the objects bound to it. Then the library's other variables are defined, unbound, with the kind their
`trait` gives: `textcontent` is text, `visibility` is visibility, `fileref` (a linked file) is image and
`graphdata` is graph. A variable without a `trait`, or with one VectorCraft doesn't know, is defined of unknown
kind and takes no values. In each data set, the element named after a variable gives its value:
- text: the text of its `p` elements, one paragraph each; an element without a `p` element gives no value;
- visibility: `true` or `false`, in any case;
- image: a file's absolute path (from `/`, `\\` or a drive's root) or a `file:` URL of one, whose `%XX` escapes
are decoded; the file isn't read, so a library can name files that aren't there;
- graph: a `data` element. Its optional `propertyRow` holds a corner cell and then the series names, and each `row`
of its `values` holds the row's category in a first `value` whose `key` is `name`, then a `value` per series. A
value that is empty, marked `type="string"` or not a number is a blank. The data is read as the CSV
`graph.setData` takes, at most 256 KiB of it. Graph data with more than 10,000 rows or 256 series, which a
graph doesn't hold, is refused, by `variable.saveLibrary` too. `variable.saveLibrary` also refuses graph data
with neither rows nor series, which loads back as no value.
Visibility, image and graph values are read without the white space round them, and an element holding only white
space gives no value. The reply's `variables` lists the library's variables, and `deleted` and `deletedDataSets`
list the variables and data sets the load deleted. While the variables are locked, a library that would delete or
define a variable fails with "variables are locked". The reply is `{variables, deleted, deletedDataSets, names,
skipped, warning?}`. A library larger than 8 MiB, one whose elements nest more than 256 deep, and one with more
define a variable fails with "variables are locked". The reply is `{variables, deleted, deletedDataSets, names}`. A
library larger than 8 MiB, one whose elements nest more than 256 deep, and one with more
than 256 variables or 1,000 data sets are refused. Its document type declaration may hold an external identifier,
comments, processing instructions and up to 64 entity declarations whose values are plain text (no `<`, `&` or
`%`). A declaration that holds anything else, and entity references that expand the library past 8 MiB, are
refused. So are a variable or data set declared twice and a value for a variable the library doesn't declare.
`variable.saveLibrary {path?}` (the panel menu's Save Variable Library…) writes the document's text variables and
every data set as a variable library in the layout `variable.loadLibrary` reads: a `variable` element with
`varName` and `trait="textcontent"` for each text variable, and a `sampleDataSet` element with `dataSetName` for
each data set, holding an element named after each text variable it has a value for, with one `p` element per line
of the value. Loading the file gives back the same text variables, data set names and text values. Variables of
other kinds are left out with their values: the reply lists them in `skipped` (`name`, `kind`), and its `warning`
names each kind with its count. Saving is refused for a document without a text variable, for a text variable whose
name isn't an XML name (it starts with a letter or `_` and holds only letters, digits, `-`, `_` and `.`), for a
data set whose name or text value holds a control character, for a data set name with white space at either end,
for a name longer than 255 characters or used twice, and past `variable.loadLibrary`'s caps (256 variables, 1,000
data sets, 8 MiB). Without `path` the reply is `{xml, variables, datasets, skipped, warning?}`;
with it, the file is written and the reply is `{path, variables, datasets, skipped, warning?}`.
refused. So are a variable or data set declared twice, a value for a variable the library doesn't declare, and a
value that doesn't fit its variable's kind.
`variable.saveLibrary {path?}` (the panel menu's Save Variable Library…) writes the document's variables and every
data set as a variable library in the layout `variable.loadLibrary` reads: a `variable` element with `varName` for
each variable, with the `trait` and `category` of its kind (`textcontent` and `&ns_flows;`, `visibility` and
`&ns_vars;`, `fileref` and `&ns_vars;`, `graphdata` and `&ns_graphs;`; neither for a variable of unknown kind), and
a `sampleDataSet` element with `dataSetName` for each data set, holding an element named after each variable it has
a value for. A text value is written with one `p` element per line, a visibility value as `true` or `false`, an
image value as its path (a path from the root of a Unix file system as a `file:///` URL), and a graph value as a
`data` element, with a blank category marked `type="string"`. Loading the file gives back the same variables with their kinds, data set names and values; graph
data comes back in the CSV form Graph Data writes. Saving is refused for a document without a variable, for a
variable whose name isn't an XML name (it starts with a letter or `_` and holds only letters, digits, `-`, `_` and
`.`; a variable of unknown kind has no values, so its name isn't checked), for a name or value that holds a control
character, for a name with white space at either end, for a name longer than 255 characters or used twice, for an
image value that isn't a file's absolute path, and past `variable.loadLibrary`'s caps (256 variables, 1,000 data
sets, 8 MiB). Without `path` the reply is `{xml, variables, datasets}`; with it, the file is written and the reply
is `{path, variables, datasets}`.
- Batch: `action.batch {commands, source?: folder|datasets, folder?, destination?: none|saveAndClose|folder,
destinationFolder?, overrideOpen?, overrideSave?, overrideExport?, includeSubdirs?, stopOnError?}` plays an
action on every file in a folder or every data set (the Actions panel menu's Batch…). Override Save and Override
@@ -3671,8 +3705,20 @@ already walks Image Trace. Short agent entry points:
`ui.imageTracePresetDialog {params?, rename?}` opens the Image Trace panel's Save as New Preset… or Rename… dialog.
- Live Paint: `livePaint.make`, `livePaint.merge`, `livePaint.fill {group?, point}`, `livePaint.strokeEdge`,
`livePaint.info {group?|id?}`, `livePaint.release`, `livePaint.expand`, and `livePaint.gapOptions {ids?, detect?, size?,
color?, closeGaps?}` (Object › Live Paint › Gap Options). Make gives each covered face the fill of its front-most
covering source, including compound paths with shared curved boundaries. Expand keeps those face fills.
color?, closeGaps?}` (Object › Live Paint › Gap Options). Make gives each covered face the fill rows that draw of
its front-most covering source that has one, including compound paths with shared curved boundaries, and each edge
the stroke rows that draw of its source. A row draws when it is shown and has paint and an opacity above 0; a
stroke also needs a weight above 0, and a source whose opacity is 0 has no rows that draw. The rows keep their order
and settings (opacity, blending mode, effects, overprint, and a stroke's weight, caps, joins, dashes and alignment),
gradients keep their fit on the source, and the face or edge takes the source's opacity and blending mode. Where a
path is cut into several edges, each edge runs the way the path does, its dashes go on from where it starts along
the path (Align Dashes to Corners is turned off on it), the arrowheads stay only at the path's own ends, and a width
profile or brush stays only on an edge that runs the whole path. A gradient applied along a stroke starts again on
each cut edge, and an inside or outside stroke is drawn centered on an edge that is an open piece of a closed path.
When `livePaint.gapOptions` or `livePaint.merge` cuts an edge into pieces, each piece takes the edge's paint, with
its rows fitted to its part of the edge in the same way. A face or edge given no such row has one unpainted row. Rows that do not draw stay with the stored sources, which
`livePaint.release` returns unchanged. The source's object effects and the strokes' pen pressure are not carried.
Expand keeps those face fills.
`livePaint.expand` keeps each face and edge that draws, with all its fill and stroke rows: it is shown, its opacity
is above 0, and a fill or stroke row is shown, has paint and an opacity above 0 (a stroke also a weight above 0),
whichever row is on top. A painted lower row under an unpainted top row keeps the part; a part whose only painted
@@ -3745,6 +3791,31 @@ already walks Image Trace. Short agent entry points:
`livePaint.strokeEdge` when an edge is within 4 screen pixels of the pointer and `livePaint.fill` otherwise, and
Shift-click runs `livePaint.fill`. With both off, a click runs nothing. With `highlight` off, the face or edge under
the pointer is not outlined.
A double-click runs the command a click there runs with `scope: "connected"`, and a triple-click with
`scope: "samePaint"`; Alt held runs nothing, and so does Cmd/Ctrl (`mods.cmd`), whose presses go to the selection
tool Cmd lends. With Space held (`mods.space`, the view tools), a press, a
double-click or a triple-click runs nothing, and a move outlines no face or edge. On a face, `connected` fills the face and every face reached from it
across edges that draw no stroke (an edge that is hidden or fully transparent, or whose strokes are each hidden,
without paint, 0 pt wide or fully transparent); a stroked edge, a hidden face, a gap
that Gap Detection closes and the group's outside stop it. `samePaint` fills every face
of the group, joined to it or not, whose fill is the `match` paint. On an edge, `connected` paints the edge and the
edges joined to it end to end that draw with the stroke paint `match`, and `samePaint` paints every edge of the
group that draws with that stroke paint; the stroke's weight is not compared. An edge draws with the paint of its
topmost stroke that is shown, with paint, above 0 pt and not fully transparent, and with none when the edge is fully
transparent or has no such stroke. Each click of a double-click or triple-click runs as it
arrives and is its own undo step. `match` is the paint the face or edge had before the first of the double-click's
two clicks or the triple-click's three that painted it, or before the double-click when the triple-click's double-click
reached it, so the clicks before the last do not change what it matches;
a first click that made the Live Paint group counts when the face it filled is the clicked face. Hidden faces and edges are not painted. An edge takes the
paint and weight in its topmost stroke that is shown, or in a new stroke above the others when it has none. These rules
for what faces and edges join and what paints match are VectorCraft's own. The help page describes the gestures:
https://helpx.adobe.com/illustrator/desktop/paint-and-fill/learn-painting-basics/paint-with-the-live-paint-bucket-tool.html.
`livePaint.fill {scope}` and `livePaint.strokeEdge {scope}` take `scope` as `"one"` (the default), `"connected"` or
`"samePaint"`; any other value is an error and changes nothing. `match` is a paint in the model JSON that
`document.node {id}` writes for an appearance's paints (`{"type": "none"}` for none), taken only with a scope that compares paints, and
defaults to the face's fill or the stroke paint the edge draws with. `livePaint.fill` refuses a `scope` other than `"one"` with
`points`. Both replies add `count`, the number of faces or edges the click reaches, including those that already had the
paint.
With the Live Paint Bucket selected and no drag in progress, Left and Right without Shift, Alt or Cmd/Ctrl (which
keep their shortcuts) run `paint.setFill {swatch, ids: []}` (or
`paint.setStroke` when the Stroke proxy is in front) with the color swatch before or after the active paint in the
@@ -3761,7 +3832,8 @@ already walks Image Trace. Short agent entry points:
objects; otherwise it returns an error and does not merge the objects.
A Live Paint command that changes nothing records no undo step, keeps the redo steps, leaves a saved document saved
and replies as usual: `livePaint.fill` (a click, or a Bucket drag) when every face it reaches already has the paint,
`livePaint.strokeEdge` when the edge already has the paint and the width, `livePaint.gapOptions` (and the Gap Options
`livePaint.strokeEdge` when the topmost shown stroke of each edge it reaches already has the paint and the width,
`livePaint.gapOptions` (and the Gap Options
dialog's OK) when every group already has the options and `closeGaps` finds no gap, and `livePaint.make` or
`livePaint.merge` given one Live Paint group alone, which they select. Such a fill or stroke also leaves the recent
colors and the last color as they are. `livePaint.gapOptions` makes a group's faces again, with new ids, only when it adds a path or when Gap Detection is turned on or off or given a new size while it is on.
@@ -5223,7 +5295,7 @@ The Attributes panel (`window.panel {panel: "attributes"}`, Cmd+F11) reads `attr
objects differ) and `recentUrls`. `attributes.set {overprintFill?, overprintStroke?, showCenter?, imageMap?, url?,
note?, ids?}` sets them in one undo step; `path.setFillRule {rule: "nonZero"|"evenOdd"}` sets the fill rule of the
paths and compound paths, and `path.reverse {reversed?}` makes subpaths run counter-clockwise (true) or clockwise
(false). When every subpath of a path turns alike, its strokes' width profiles (left and right swapped) and pen
(false). When every subpath of a path or compound path turns alike, its strokes' width profiles (left and right swapped) and pen
pressure, tilt, bearing and rotation turn with it, so the strokes draw as before. SVG export wraps an object with a URL in `<a xlink:href>`, and SVG import reads `<a href>` links back.
```json
@@ -5646,6 +5718,25 @@ used, as a transform's does.
{"name":"run_command","arguments":{"command":"object.transformEach","params":{"scaleH":50,"scaleV":50,"strokes":false}}}
```
## Scale and Reflect on a turned bounding box
`object.scale` and `object.reflect` with a named axis act along the selection's bounding box: after a rotation, `sx`
stretches along the box's own width and `sy` along its height, and `"vertical"` and `"horizontal"` are the box's own
axes, until Object › Transform › Reset Bounding Box squares the box to the page. Both take `basis?: "box"|"page"`:
`"box"` is the default, and `"page"` acts along the page's axes. Any other `basis` is an error, and nothing changes.
`origin` is still a page point, and a numeric `axis` is still a line on the page at that angle. Selections whose
objects stand at different angles, and anchors selected with Direct Selection, use the page's axes; with `copy` or
with `objects: false`, the whole objects or their pattern tiles transform, along the box. The journal entry
records the `basis` used, so a recorded action replays in the same frame; a saved step without `basis` (an action or
journal entry recorded before it) plays along the page's axes, as it did when it was recorded. `object.transformAgain` repeats the transform's page matrix, also after Reset Bounding Box. The Scale
and Reflect dialogs, the Transform and Properties panels' flips, the Scale tool and the Reflect tool's default axis
and Shift steps follow the box the same way; a Reflect axis drawn with the tool is the line drawn on the page.
```json
{"name":"run_command","arguments":{"command":"object.scale","params":{"sx":200,"sy":100}}}
{"name":"run_command","arguments":{"command":"object.reflect","params":{"axis":"vertical","basis":"page"}}}
```
## Free Distort on a turned bounding box
`object.distort` maps the corners of `from` (TL, TR, BR, BL) onto `corners`, which are page coordinates. With
@@ -5855,7 +5946,8 @@ commands send `to: "key"` while the selection has a key, else `"artboard"` for o
anchor points, else the Align To choice. The Distribute commands send `to: "key"` while the selection has a key, else
the Align To choice, and while there is a key object or a key guide they send the panel's Distribute Spacing value as
`spacing`. Edit › Keyboard Shortcuts lists them under Align, with no default shortcut, and
`shortcuts.set {id, shortcut}` gives one a shortcut.
`shortcuts.set {id, shortcut}` gives one a shortcut. The Object › Distribute menu lists the eight Distribute commands,
and each item runs its command.
With the Artboard tool chosen, the same buttons and UI commands, and the Control bar's Align and Distribute buttons,
run `artboard.align`, `artboard.distribute` and `artboard.distributeSpacing` on the artboards the tool has selected.
@@ -6226,8 +6318,9 @@ Color on (HDR or Automatic Color Management) and macOS windows are converted by
again. In a Wayland session, and where no profile is set or the profile is not an RGB display profile of at most 1 MiB,
the canvas stays sRGB. The app reads the display again every 10 seconds and when the window moves to another display.
Exports, `ui.render`, the document, panels and thumbnails are not converted; with Proof Colors on, the proof is drawn
first and then converted. The art is converted before it is drawn over the converted paper, so the blend of a
translucent object or an antialiased edge with the paper can differ slightly from a conversion of the blended color.
first and then converted. Translucent art and the antialiased edges of art over an artboard's paper or the
transparency grid are blended with it in sRGB, and the blend is converted; over the pasteboard and in Outline the art
is converted on its own.
`ui.inspect` returns `displayProfile`: `{enabled, display, source, profile, from, applied, reason,
osManaged}`, where `source` is `profile`, `srgbProfile` (a profile that shows sRGB unchanged), `system`, `none`,
`pending` or `unsupported`, `applied` says whether the canvas is converted, and `reason` says why not.
@@ -6535,7 +6628,8 @@ report warn about it).
Commands take and return lengths in points (the distances of `object.path.offsetPath`, `object.path.simplify`,
`object.path.splitIntoGrid` and `text.setStyle` `size`/`leading` also take a string with a unit, `"5 mm"`). A length
string takes the same arithmetic as the desktop app's fields, where `%` is a share of the value before it:
`"200mm*50%"` is 100 mm and `"50pt+25%"` is 62.5 pt. The units
`"200mm*50%"` is 100 mm and `"50pt+25%"` is 62.5 pt. A length string with more than 128 `+ - * /` signs is not read
as a length. The units
only change what the UI shows and how it reads typed numbers. Preferences ▸ Units has
four: General (rulers, positions and sizes, the Info panel, dialog distances, canvas measurement labels), Stroke
(weights, dashes), Type (font size, leading, baseline shift) and East Asian Type (no field uses it yet). General is the active document's units
@@ -7256,6 +7350,14 @@ is mixed with middle gray by its alpha.
{"name":"run_command","arguments":{"command":"effect.apply","params":{"effect":"distort.glass","params":{"distortion":10,"textureImage":{"path":"/tmp/texture.png"}}}}}
```
Texturizer (`texture.texturizer`) and Rough Pastels (`artistic.roughPastels`) take `params.textureImage` the same
way, as the surface that Scaling, Relief, Light Direction and Invert act on. With `texture: "loaded"` and no stored
texture they draw Canvas.
```json
{"name":"run_command","arguments":{"command":"effect.apply","params":{"effect":"artistic.roughPastels","params":{"relief":12,"textureImage":{"path":"/tmp/texture.png"}}}}}
```
## Linked images
A placed image linked to its file (`file.place {path}`, Link on) records the file's absolute path, its size,
@@ -7578,6 +7680,12 @@ bounds; `clip` puts it in a clip group of the old bounds when it is larger). The
except for a placed document, which opens in a tab. With the preference `editOriginalSystemDefaults` (Preferences ›
File Handling › Use System Defaults for Edit Original, on by default) off, Edit Original opens a raster image in the app
that made it, PhotoCraft, where it is installed, and anything else in the system's default app (#5236).
The UI command `links.copyInfo {ids?}` (the panel menu's Copy Info) copies the file names of images and placed
documents `ids` to the clipboard, one per line in the order given, and returns `{text, names}`. Without `ids` (or with
`ids: null`, as for the other `ids` commands) it copies the selected ones, and the rows outside the layer tree picked in
the panel, in the order `links.list` gives them; the panel menu passes the selected rows in the order the panel shows
them. An embedded image gives the name its row shows. An id that is not an image or a placed document, a
malformed `ids`, or nothing to copy is an error, and the clipboard is left as it was.
```json
{"name":"run_command","arguments":{"command":"links.list","params":{"show":"missing"}}}
@@ -8220,6 +8328,16 @@ flattened picture every PSD file carries: Bitmap, Grayscale, Duotone, Indexed, R
A document with transparency keeps it. `file.place {path}` links the file like any image (`links.update` reads it
again when it changes); Multichannel files are refused.
A PSD or PSB file that names a clipping path (the saved path its clipping path resource names, in the
[published PSD file format specification](https://www.adobe.com/devnet-apps/photoshop/fileformatashtml/)) places, and
opens as its merged image (`layers: false`, a file with no layers, or a file whose layers can't be read), inside a clip group whose clipping
path is that path, under the even-odd rule, so the image shows only inside it. The clipping path is an ordinary path
that can be edited or released. `file.place` answers `image`, the id of the image in the clip group. The
[import options help page](https://helpx.adobe.com/illustrator/desktop/add-and-import-files/import-from-other-apps/photoshop-import-options.html)
says that a file imported as one flattened image keeps the document's clipping path when it has one. `links.update`
and `links.relink` read the file's pixels again and leave the clip group and its clipping path as they are. Files that name no clipping path, or name a path
they don't have, come in as a plain image.
`document.open` converts a PSD or PSB file's layers to objects. In
[PSD import options](https://helpx.adobe.com/illustrator/desktop/add-and-import-files/import-from-other-apps/photoshop-import-options.html),
the option that converts layers to objects keeps as much of the layer structure as it can, merges the layers of a
@@ -9483,6 +9601,24 @@ not kerned). Off by default and saved only when on. One undo step.
{"name":"run_command","arguments":{"command":"text.setFormat","params":{"proportionalMetrics":true}}}
```
## H or V Style and Roman Italics
With Show East Asian Options on, the OpenType panel menu has H or V Style and Roman Italics
([the help page](https://helpx.adobe.com/illustrator/using/formatting-asian-characters.html)). They are the character
feature tags `hkna` and `ital` of `text.setStyle` and `text.setRangeStyle {features, featureChanges}`
([feature list](https://learn.microsoft.com/en-us/typography/opentype/spec/featurelist)). `hkna` draws kana with the
forms the font makes for horizontal type, and in vertical type the shaper applies `vkna`, the forms for vertical type,
in its place. `ital` draws half-width alphanumerics with the font's italic forms. A font without the feature draws the
characters unchanged. Both are on or off only: `hkna=2` and `ital=0` are refused as unknown features, and `vkna` is
not a tag of its own. A menu item is checked when every character the panel acts on has it; choosing it turns the
feature on for all of them, or off when it is checked, each character keeping its other features, in one undo step.
Reset OpenType Features turns both off. Character styles carry them, and Character Style Options › OpenType Features
has an on, off or unset switch for each. Live SVG text writes them in `font-feature-settings`.
```json
{"name":"run_command","arguments":{"command":"text.setRangeStyle","params":{"id":12,"start":0,"end":6,"featureChanges":["hkna","ital"]}}}
```
## Variable fonts
`text.setFormat {fontAxes: {"wght": 650}}` (the selected type, or `ids`) and `text.setRangeStyle {id, start, end,
+1 -1
View File
@@ -159,7 +159,7 @@ Gujarati interface follow-up (2026-11-14): the catalog and zero/one plural rule
| M1 | Selection, transform, layers, MCP (common command tools, strict keys, session resources and version-aware result hints) | ✅ mostly done (document-tab artwork drag with lossless resource transfer and one target Undo; malformed MCP app replies discard the connection without replay, #1761; bounded control socket waits and connection-slot recovery, #1726; bounded synthetic control input, #1709; no-op anchor Align keeps Undo, Redo and saved state (#1180 follow-up); transform reference point snaps to anchors/centres; rotated objects keep a rotated bounding box and their angle; Free Transform, Scale and Reflect along a rotated box pending; the Transform fields take one Tab each, the W/H link after them) | 4–6 |
| M2 | Drawing tools + smart guides | ✅ mostly done (#3380 deletion creates independent selected path pieces and keeps bounds-fitted gradients in place; endpoint highlighting remains open; Polygon/Star exact-size count persistence, with repeated-drag, tool-switch and Cancel regression coverage; Curvature point markers without direction handles (#1236, in part); Direct Selection segment click/Delete preserves neighboring edges, #1278; Flare, Reshape, Live Corners widget dragging and per-corner radii and kinds landed, on any path's corners; Shaper rough-shape recognition and orientation steps, the remaining Pen modifier nuances; Shaper Groups with editable sources and separate fill/stroke edits landed; custom ruler origin with drag preview, active-artboard reset, relative X/Y fields and native round-trip (#1212); construction widget/refinements pending); a shape just drawn moves by its centre point and turns from the zone outside a corner with its tool still active (#5477); Tangent and straight-segment perpendicular snapping use saved Smart Guides Geometric Guides preferences and Control bar quick access (#2537), including the existing Snap to Glyph and Alignment Guides toggles (#6164); no View submenu; Copy and Cut with the Pen or Curvature tool take the anchors a Cmd/Ctrl-lent Direct Selection picked (#6123) | 12–18 |
| M3 | Paint & appearance (swatches, color, gradient, stroke, appearance, transparency, styles) | ✅ done (M3.7–M3.98): swatches, groups, libraries, tints, Lab spots and Registration; Eyedropper copies transformed character dimensions (#6404) and styles highlighted text only (atomic transparency rejection); Eyedropper reads visible locked and outside-isolation sources (#1251); Color Picker, Color panel, Color Guide, Color Themes, Edit Colors and Recolor Artwork (screen-bounded Edit chip list, #3416); gradients that follow every transform, the annotator, linked stops, freeform gradients, focal points, gradients on strokes and Expand; visible strokes when paint is added to inherited None/zero-width art (#1253); stroke geometry, arrowheads, dashes fitted to corners, stroke on type, width profiles and the Width tool (Shift-double-click enables adjoining adjustment in Width Point Edit; finite-arithmetic follow-up regressions and full CI passed ([Validation evidence](target-app-parity.md#evidence-what-ships-today-by-area))), Scale Strokes and preview bounds; Appearance targeting, container appearance, target circles, Expand Appearance; knockout, isolation, blend accuracy, CMYK blending, overprint, Attributes, masks and clipping, Flatten Transparency; graphic styles, links and libraries; CSS library export preserves fractional linear and radial gradient stop positions. Left: freeform gradients in SVG/PDF, mixed spot/process gradients in SVG and PDF gradients mixing a spot color with RGB or Lab stops export as stops or process colors; PDF export writes dithered gradient fills with spot stops, and Perceptual spot gradients with RGB or Lab stops other than white or more than 4,096 stops on their curve, as images in process colors; confirm the Scale Strokes & Effects default; Recolor model selection is available through the engine and MCP only (#2520); the dialog retains its existing controls; Library panels' lists (swatch, Color Book, brush, symbol and graphic style libraries) fill a taller panel above its footer (#2992; scores and estimates unchanged); Graphic Styles list fills a taller panel above its footer (#2992; scores and estimates unchanged); Swatches list fills a taller panel above its footer (#1638; `the_list_fills_a_tall_panel_above_its_footer`; scores and estimates unchanged); | 4–8 |
| M4 | Files (native, SVG, PDF, raster, Export for Screens, clipboard interop) | Explicit editing-data import keeps oversized capped-page artboards (#6865; estimates unchanged); Save refusal continuation (#6146): ordinary Save offers Save As; unchanged format keeps options, changed format opens its Save options; cancellation/automation isolation; focused UI regressions passed, native acceptance pending; PDF authored ActualText preserved once in editable import (#6165; first-glyph line/style, bounded input; estimates unchanged); PDF export keeps harakat and zero-advance Noto Arabic dot glyphs in one compatible batch per layout line, with per-glyph vertical offsets (#6309). `crates/pdf/src/tests/realtext.rs` (`mark_text_*`) checks decoded cluster mappings, mixed Arabic/Latin/digits, contour placement against outline export, vertical columns, differing path rotations and PDF/X-1a/3/4. Editable-text import preserves authored ActualText separately (#6165), with cross-line/style layout limits; separately styled or differently rotated clusters and external viewer search/copy remain outside this export-only evidence. `.ai`/EPS type from the text document keeps its paragraphs' tab stops (#6147); `.ai`/EPS type from the text document keeps auto leading per line, its hyphenation options, and may show more of the last lines the file sets in an area frame of its own, when growing the frame by its lines' deepest descent plus 0.05 pt shows them (#6137); ✅ done (M4.14–M4.98 on 2026-10-05): Save As/Save a Copy/Revert/templates, Data Recovery and background save, SVG import and export fidelity (units, text, symbols, filters, SVGZ, Preserve Editing, SVG Options), shared guarded PDF writer and windowed reader ([pdf-backend.md](pdf-backend.md)), PDF import (colours, layers, masks, editable text, security) and export (presets, marks, bleed, ICC and output intent, raster effects, subset fonts, PDF/X-1a, X-3 and X-4, PDF layers, overprint, page thumbnails, fast web view, PDF 1.3), Place and Links, Package, EPS and DXF in and out (EPS legacy mesh corners, vector opacity masks and raster page comparison verified; closed SPLINE flags round-trip and circular paths export as native CIRCLE/ARC, #1299), EMF/WMF, TIFF/BMP/Targa/PSD/GIF/PNG-8 export, Print and print presets, slices, Save for Web, TIFF in Export for Screens and Asset Export, clipboard flavours, File Info; AF.1–AF.4 add Affinity import (native layers, artboards, vectors, text and images from Affinity 1–3 files; the embedded preview only as a warned fallback; AF.4 expands current `.af` fixtures and every-board round-trip validation); EPS Save and Save As keep an opened EPS as EPS with its page frame and complete editable VectorCraft document (2026-10-26; G9 import fidelity unchanged). Left: DWG (use DXF), PSD placement as layers, overprint read back from PDF, Affinity export, polish; SVG CSS classes survive export and repeated import in all four styling modes (#2480); current-main integration retains event attributes and scripts; SVG CSS classes survive export and repeated import in all four styling modes (#2480); current-main integration retains event attributes and scripts; PDF import: glyphs inside a marked-content span with an empty `ActualText` come in as outlines in their fill and stroke, adding no text (#6971; scores and estimates unchanged); | 3–6 (existing scope; Affinity export unestimated) |
| M4 | Files (native, SVG, PDF, raster, Export for Screens, clipboard interop) | Rename artboards in Export for Screens (#7275); Explicit editing-data import keeps oversized capped-page artboards (#6865; estimates unchanged); Save refusal continuation (#6146): ordinary Save offers Save As; unchanged format keeps options, changed format opens its Save options; cancellation/automation isolation; focused UI regressions passed, native acceptance pending; PDF authored ActualText preserved once in editable import (#6165; first-glyph line/style, bounded input; estimates unchanged); PDF export keeps harakat and zero-advance Noto Arabic dot glyphs in one compatible batch per layout line, with per-glyph vertical offsets (#6309). `crates/pdf/src/tests/realtext.rs` (`mark_text_*`) checks decoded cluster mappings, mixed Arabic/Latin/digits, contour placement against outline export, vertical columns, differing path rotations and PDF/X-1a/3/4. Editable-text import preserves authored ActualText separately (#6165), with cross-line/style layout limits; separately styled or differently rotated clusters and external viewer search/copy remain outside this export-only evidence. `.ai`/EPS type from the text document keeps its paragraphs' tab stops (#6147); `.ai`/EPS type from the text document keeps auto leading per line, its hyphenation options, and may show more of the last lines the file sets in an area frame of its own, when growing the frame by its lines' deepest descent plus 0.05 pt shows them (#6137); ✅ done (M4.14–M4.98 on 2026-10-05): Save As/Save a Copy/Revert/templates, Data Recovery and background save, SVG import and export fidelity (units, text, symbols, filters, SVGZ, Preserve Editing, SVG Options), shared guarded PDF writer and windowed reader ([pdf-backend.md](pdf-backend.md)), PDF import (colours, layers, masks, editable text, security) and export (presets, marks, bleed, ICC and output intent, raster effects, subset fonts, PDF/X-1a, X-3 and X-4, PDF layers, overprint, page thumbnails, fast web view, PDF 1.3), Place and Links, Package, EPS and DXF in and out (EPS legacy mesh corners, vector opacity masks and raster page comparison verified; closed SPLINE flags round-trip and circular paths export as native CIRCLE/ARC, #1299), EMF/WMF, TIFF/BMP/Targa/PSD/GIF/PNG-8 export, Print and print presets, slices, Save for Web, TIFF in Export for Screens and Asset Export, clipboard flavours, File Info; AF.1–AF.4 add Affinity import (native layers, artboards, vectors, text and images from Affinity 1–3 files; the embedded preview only as a warned fallback; AF.4 expands current `.af` fixtures and every-board round-trip validation); EPS Save and Save As keep an opened EPS as EPS with its page frame and complete editable VectorCraft document (2026-10-26; G9 import fidelity unchanged). Left: DWG (use DXF), PSD placement as layers, overprint read back from PDF, Affinity export, polish; SVG CSS classes survive export and repeated import in all four styling modes (#2480); current-main integration retains event attributes and scripts; SVG CSS classes survive export and repeated import in all four styling modes (#2480); current-main integration retains event attributes and scripts; PDF import: glyphs inside a marked-content span with an empty `ActualText` come in as outlines in their fill and stroke, adding no text (#6971; scores and estimates unchanged); | 3–6 (existing scope; Affinity export unestimated) |
| M5 | Performance | 🟡 background render + caches + MT done; `vectorcraft-cli bench` and `vectorcraft-cli perf` (budget suite); opaque solid-fill knockout paths draw once in hard-edge/supersampled raster export, with analytical-AA compositing unchanged; file format v2 opens 3× faster (50k paths: 722 → 244 ms); raster effects (glows, shadows, blur, feather) no longer force the whole frame single-threaded (filtered offscreen per effect, verified equal to the single-threaded reference); Craquelure/Mosaic Tiles/Stained Glass distance searches optimized with pixel-invariance checks and the `texture_drag_perf` probe (#1257); effect-heavy demos need a clean-machine benchmark; dirty-region rendering, GPU backend spike pending | 15–25 |
| M6 | Path operations (Pathfinder, Shape Builder, offset…) | ✅ mostly done (live compound shapes: Make, Alt-click a Shape Mode, member modes, Release, Expand (#4522); single-result Pathfinder names and SVG IDs retained (#1293); Divide open-cutting-path geometry, appearance and history verified (#1307; existing segment cap remains); Simplify preview/counts and rewritten-path component selection verified (#1272; the slider opens at the Auto-Simplify level; a Direct Selection part of a path simplifies on its own); Compound appearance and member fill-rule edits verified (#1197); Boolean precision on almost-horizontal edges fixed and the property tests made deterministic; Shape Builder gap detection, large-offset bug open); Shape Builder Gap Detection shares Live Paint’s spatial gap finder and temporary merge bridges (#2526); Custom is capped at 72 pt in the existing options dialog | 3–6 |
| M7 | Type (point/area/path, editing, styles, OpenType, threading, glyphs) | Spell check engine: `text.spelling.check` and `text.spelling.suggest` on craft-proof and the custom dictionary and Ignore All commands and Auto Spell Check with its right-click suggestions, Ignore All and Add to Dictionary, and Spelling › Auto Spell Check in the context menu of type, and the Edit Custom Dictionary dialog (#4936; the Check Spelling dialog remains); Right-aligned area type sets the text after a line's last tab against the right edge, the text before it at its stops (#6147); A line a soft return ends is justified as a wrapped line is (#6137); Indian-language pattern hyphenation preserves syllables in both composers (#4947); 🟡 Hanging punctuation keeps its published amounts under Glyph Scaling, and Standard burasagari counts a hung line start as room; Fullwidth `!` / `?` keep an upright fallback in vertical type; Hyphenation… and Justification… per paragraph (#2310, #2755), Auto Leading also driving the Character panel's Leading, the Alt+↑/↓ leading nudge and Fill with Placeholder Text; Document-point character controls after object/group scaling (#1034; optional font-size preservation remains separate), Character/Paragraph Styles (explicit Auto-leading resets and local leading overrides), per-paragraph paragraph attributes, mixed paragraph alignment indication (headless regressions, full integration CI and visual review pass), Area Type Options (with Width/Height, Auto Size and Shrink Text to Fit; overflow reported to agents and vertical Align), type areas resized by their handles and reshaped with Direct Selection, threaded text, Fit Headline, Glyphs (feature-valued alternates and unencoded glyphs applied in one Undo step while keeping Arabic joining, #5514), OpenType panel (including font-defined Stylistic Sets), Find Font (Document, Recent and System replacement sources, #5217), installed system fonts in every font menu (fonts registered with Windows outside its font folders, the XDG and Flatpak host font folders on Linux, and fonts installed while the app runs, listed when it comes back to the front; searchable; the Character and Properties menus show each family's sample of the selected text, preview it live under the pointer or ↑/↓, show recently applied families first (#3515), filter by script and kind, list a family's styles, and star favourites; the Glyphs panel and Find Font's Replace With use the same menu without the live preview) and found by name on open/import (full-name aliases and Missing Fonts searches verified, #1258; Windows font-service QA pending), variable fonts' named instances as styles, CJK names in the UI, input methods (marked text in place, candidate window at the caret; Japanese IME), Text Wrap (offset, invert, both sides of an object; follows edits), Type on a Path effects and Options with draggable start/centre/end brackets (slide, flip), tab stops + Tabs panel (Above Text floats it over the selected text, its ruler at the canvas zoom so stops line up with the columns, #729), inline symbols in text (beyond parity), Hebrew/Arabic bidirectional text, Paragraph Direction, Every-line/Single-line Composer per object (ragged text too) done; paint-only Arabic harakat keep joining with GDEF-mark ownership and Insert Kashidas tatweels on the base letter's paint (`tests::harakat`); tab leaders, spell check, vertical type pending | 45–60 |
+23 -8
View File
@@ -258,8 +258,8 @@ dominate; 3D, graphs and automation are specialist). They sum to 105. Feature de
| Layers, artboards, document setup | 5 | 78% | Selected-artboard deletion is consistent across panel actions (all-selected retains artboard 1); history prunes removed indices while preserving valid multiselection; remaining: 30.0's artboard locking and colours and managing artboards from the canvas, Layers row menu commands beyond Collect for Export and menus on guide and blend spine rows (click-to-fit objects is done, #1227) | 5–8 | estimated (lowered from 82%: target moved); Artboard deletion keys keep artwork and the first of an all-selected set; Artboard Options' Show Center Mark and Show Cross Hairs draw on the canvas only (exports and print leave them out), regression-tested; Artboard Options' Show Video Safe Areas and Video Ruler Pixel Aspect Ratio (per artboard, native format) are done ([video display options](#evidence-what-ships-today-by-area)), score unchanged |
| View & navigation | 3 | 70% | multiple windows and the rest of Arrange (Cascade, Float in Window, Float All in Windows) | 9–14 | estimated; Window › Arrange › Tile and Consolidate All Windows done (score/hours unchanged); document-tab artwork hover/copy and lossless transfer tested (2026-10-22; score/hours unchanged); cached canvas reprojection preserves scale during resize, rotation and DPI changes (regression and contributor original-file evidence below; scores/hours unchanged); cached canvas dimensions match the renderer limit before cache-key creation, view construction and worker submission (regression evidence below) |
| File formats | 6 | 75% | `.ai` files open with a flattened or wrong layer structure in real cases (#951), no writing of Illustrator's editing data (#868), FreeHand/CorelDRAW/DWG/HEIC import, 3D export; see [file-format-parity.md](file-format-parity.md); EPS legacy mesh/opacity-mask and raster comparison follow-up verified PDF marked-text batching covered by M4 evidence (#6309); cross-line/style ActualText layout (#6165) and styled-cluster limits remain.| 10–20 | estimated (lowered from 85%: open issues); DXF control-point SPLINE import closure regression-tested (#1299 follow-up after #1320); DXF export's Include Placed Images and Include Fill Hatches options, off in CNC / CAM (`images_and_hatches_are_left_out_independently_and_the_cnc_preset_leaves_out_both`; score, gaps and hours unchanged); PDF writer metadata, reader streaming and immutable backend checks in `crates/pdf/src/tests.rs`, `crates/pdf/src/tests/reader_streaming.rs` and `xtask/src/pdf_deps.rs`; PDF Place and clipboard paste retain live text (synthetic regressions); EPS ordinary Save retains the page frame and editable document (2026-10-26); `.ai`/EPS editing data checked against a user's files (2026-10-30): case setting (`ate::tests`, `tests_native::type_from_the_text_document_keeps_its_case_setting`), placed-file placements (`tests_native::a_small_placed_file_goes_where_it_draws_closer_to_the_page`), pages cropped to the art (`tests_native::type_on_the_copies_of_a_page_cropped_to_the_art_is_read_once`), mesh corner weights and state (`ai::mesh::tests::corners_with_handle_weights_of_zero_are_read`, `ai::tests::a_gradient_mesh_takes_the_transparency_written_after_it`), type on a path spacing (`tests_native::type_round_a_circle_is_spaced_as_the_file_lays_it_out`; the distance from the path under Align to Path is measured from the top of each character's em box, half the height of the top of its H or the bottom of its font's bounding box (`tests_native::aligned_type_lies_as_far_from_its_path_as_the_file_lays_it_out`), and the fit under Gravity does not reproduce the file's layout), blend state (`ai::tests::a_blends_transparency_follows_its_record`); scores and estimates unchanged. PDF ActualText preserved in editable import (#6165; bounded synthetic coverage below); score/hours unchanged. Legacy `.ai` type keeps its manual kerning pairs (`Tk`, `TK`), discretionary hyphens (`T+`, `T-`) and hyphenation language (`XL`) when it opens, and legacy `.ai` saves write manual kerning as `Tk` and soft hyphens as `T+` (`import::ai::tests::legacy_manual_pair_kerning_is_kept`, `legacy_ai::tests::manual_kerning_and_soft_hyphens_are_written_and_read_back`); score, gaps and hours unchanged. |
| File formats | 6 | 75% | `.ai` files open with a flattened or wrong layer structure in real cases (#951), a saved `.ai`'s editing copy without live effects on objects other than paths, compound paths and groups that qualify (at most a fill and a stroke, a non-zero fill rule when filled, no large canvas) and whose appearance and effects the art-style writer supports, and with live objects as their art (#868), FreeHand/CorelDRAW/DWG/HEIC import, textures, map art and shading in 3D export; see [file-format-parity.md](file-format-parity.md); EPS legacy mesh/opacity-mask and raster comparison follow-up verified PDF marked-text batching covered by M4 evidence (#6309); cross-line/style ActualText layout (#6165) and styled-cluster limits remain.| 10–20 | estimated (lowered from 85%: open issues); DXF control-point SPLINE import closure regression-tested (#1299 follow-up after #1320); PDF writer metadata, reader streaming and immutable backend checks in `crates/pdf/src/tests.rs`, `crates/pdf/src/tests/reader_streaming.rs` and `xtask/src/pdf_deps.rs`; PDF Place and clipboard paste retain live text (synthetic regressions); EPS ordinary Save retains the page frame and editable document (2026-10-26); `.ai`/EPS editing data checked against a user's files (2026-10-30): case setting (`ate::tests`, `tests_native::type_from_the_text_document_keeps_its_case_setting`), placed-file placements (`tests_native::a_small_placed_file_goes_where_it_draws_closer_to_the_page`), pages cropped to the art (`tests_native::type_on_the_copies_of_a_page_cropped_to_the_art_is_read_once`), mesh corner weights and state (`ai::mesh::tests::corners_with_handle_weights_of_zero_are_read`, `ai::tests::a_gradient_mesh_takes_the_transparency_written_after_it`), type on a path spacing (`tests_native::type_round_a_circle_is_spaced_as_the_file_lays_it_out`), blend state (`ai::tests::a_blends_transparency_follows_its_record`); scores and estimates unchanged. PDF ActualText preserved in editable import (#6165; bounded synthetic coverage below); score/hours unchanged. Legacy `.ai` type keeps its manual kerning pairs (`Tk`, `TK`), discretionary hyphens (`T+`, `T-`) and hyphenation language (`XL`) when it opens, and legacy `.ai` saves write manual kerning as `Tk` and soft hyphens as `T+` (`import::ai::tests::legacy_manual_pair_kerning_is_kept`, `legacy_ai::tests::manual_kerning_and_soft_hyphens_are_written_and_read_back`); score, gaps and hours unchanged. |
| File formats | 6 | 75% | `.ai` files open with a flattened or wrong layer structure in real cases (#951), a saved `.ai`'s editing copy without live effects on objects other than paths, compound paths and groups that qualify (at most a fill and a stroke, a non-zero fill rule when filled, no large canvas) and whose appearance and effects the art-style writer supports, and with live objects as their art (#868), FreeHand/CorelDRAW/DWG/HEIC import, 3D export; see [file-format-parity.md](file-format-parity.md); EPS legacy mesh/opacity-mask and raster comparison follow-up verified PDF marked-text batching covered by M4 evidence (#6309); cross-line/style ActualText layout (#6165) and styled-cluster limits remain.| 10–20 | estimated (lowered from 85%: open issues); DXF control-point SPLINE import closure regression-tested (#1299 follow-up after #1320); PDF writer metadata, reader streaming and immutable backend checks in `crates/pdf/src/tests.rs`, `crates/pdf/src/tests/reader_streaming.rs` and `xtask/src/pdf_deps.rs`; PDF Place and clipboard paste retain live text (synthetic regressions); SVG CSS classes survive export and repeated import in all four styling modes (#2480); current-main integration retains event attributes and scripts; EPS ordinary Save retains the page frame and editable document (2026-10-26); `.ai`/EPS editing data checked against a user's files (2026-10-30): case setting (`ate::tests`, `tests_native::type_from_the_text_document_keeps_its_case_setting`), placed-file placements (`tests_native::a_small_placed_file_goes_where_it_draws_closer_to_the_page`), pages cropped to the art (`tests_native::type_on_the_copies_of_a_page_cropped_to_the_art_is_read_once`), mesh corner weights and state (`ai::mesh::tests::corners_with_handle_weights_of_zero_are_read`, `ai::tests::a_gradient_mesh_takes_the_transparency_written_after_it`), type on a path spacing (`tests_native::type_round_a_circle_is_spaced_as_the_file_lays_it_out`), blend state (`ai::tests::a_blends_transparency_follows_its_record`); scores and estimates unchanged. PDF ActualText preserved in editable import (#6165; bounded synthetic coverage below); score/hours unchanged. Save refusal continuation (#6146): `READ_ONLY_LOCATION`, Save As on ordinary UI Save only; regression execution pending; score/hours unchanged. Legacy `.ai` type keeps its manual kerning pairs (`Tk`, `TK`), discretionary hyphens (`T+`, `T-`) and hyphenation language (`XL`) when it opens, and legacy `.ai` saves write manual kerning as `Tk` and soft hyphens as `T+` (`import::ai::tests::legacy_manual_pair_kerning_is_kept`, `legacy_ai::tests::manual_kerning_and_soft_hyphens_are_written_and_read_back`); score, gaps and hours unchanged. |
| File formats | 6 | 75% | `.ai` files open with a flattened or wrong layer structure in real cases (#951), a saved `.ai`'s editing copy without live effects on objects other than paths, compound paths and groups that qualify (at most a fill and a stroke, a non-zero fill rule when filled, no large canvas) and whose appearance and effects the art-style writer supports, and with live objects other than grouped column, stacked column, line and area graphs the `.ai` reader reads back whole as their art (#868), FreeHand/CorelDRAW/DWG/HEIC import, textures, map art and shading in 3D export; see [file-format-parity.md](file-format-parity.md); EPS legacy mesh/opacity-mask and raster comparison follow-up verified PDF marked-text batching covered by M4 evidence (#6309); cross-line/style ActualText layout (#6165) and styled-cluster limits remain.| 10–20 | estimated (lowered from 85%: open issues); DXF control-point SPLINE import closure regression-tested (#1299 follow-up after #1320); PDF writer metadata, reader streaming and immutable backend checks in `crates/pdf/src/tests.rs`, `crates/pdf/src/tests/reader_streaming.rs` and `xtask/src/pdf_deps.rs`; PDF Place and clipboard paste retain live text (synthetic regressions); EPS ordinary Save retains the page frame and editable document (2026-10-26); `.ai`/EPS editing data checked against a user's files (2026-10-30): case setting (`ate::tests`, `tests_native::type_from_the_text_document_keeps_its_case_setting`), placed-file placements (`tests_native::a_small_placed_file_goes_where_it_draws_closer_to_the_page`), pages cropped to the art (`tests_native::type_on_the_copies_of_a_page_cropped_to_the_art_is_read_once`), mesh corner weights and state (`ai::mesh::tests::corners_with_handle_weights_of_zero_are_read`, `ai::tests::a_gradient_mesh_takes_the_transparency_written_after_it`), type on a path spacing (`tests_native::type_round_a_circle_is_spaced_as_the_file_lays_it_out`), blend state (`ai::tests::a_blends_transparency_follows_its_record`); scores and estimates unchanged. Graphs saved to `.ai` in either form open again as live graphs (`ai_editing::tests::graphs_are_written_with_their_records_and_open_live`, `legacy_ai::tests::graphs_are_written_with_their_records_and_open_live`, `cmd::graph::tests::a_drawn_graph_saved_as_ai_opens_as_a_graph_unless_the_save_warns`); other graphs are still written as their art (`graphs_the_records_cant_keep_are_written_as_their_art`); score and hours unchanged. PDF ActualText preserved in editable import (#6165; bounded synthetic coverage below); score/hours unchanged. Legacy `.ai` type keeps its manual kerning pairs (`Tk`, `TK`), discretionary hyphens (`T+`, `T-`) and hyphenation language (`XL`) when it opens, and legacy `.ai` saves write manual kerning as `Tk` and soft hyphens as `T+` (`import::ai::tests::legacy_manual_pair_kerning_is_kept`, `legacy_ai::tests::manual_kerning_and_soft_hyphens_are_written_and_read_back`); score, gaps and hours unchanged. |
| File formats | 6 | 75% | `.ai` files open with a flattened or wrong layer structure in real cases (#951), a saved `.ai`'s editing copy without live effects on objects other than paths, compound paths and groups that qualify (at most a fill and a stroke, a non-zero fill rule when filled, no large canvas) and whose appearance and effects the art-style writer supports, and with live objects other than grouped column, stacked column, line and area graphs the `.ai` reader reads back whole as their art (#868), FreeHand/CorelDRAW/DWG/HEIC import, 3D export; see [file-format-parity.md](file-format-parity.md); EPS legacy mesh/opacity-mask and raster comparison follow-up verified PDF marked-text batching covered by M4 evidence (#6309); cross-line/style ActualText layout (#6165) and styled-cluster limits remain.| 10–20 | estimated (lowered from 85%: open issues); DXF control-point SPLINE import closure regression-tested (#1299 follow-up after #1320); PDF writer metadata, reader streaming and immutable backend checks in `crates/pdf/src/tests.rs`, `crates/pdf/src/tests/reader_streaming.rs` and `xtask/src/pdf_deps.rs`; PDF Place and clipboard paste retain live text (synthetic regressions); SVG CSS classes survive export and repeated import in all four styling modes (#2480); current-main integration retains event attributes and scripts; EPS ordinary Save retains the page frame and editable document (2026-10-26); `.ai`/EPS editing data checked against a user's files (2026-10-30): case setting (`ate::tests`, `tests_native::type_from_the_text_document_keeps_its_case_setting`), placed-file placements (`tests_native::a_small_placed_file_goes_where_it_draws_closer_to_the_page`), pages cropped to the art (`tests_native::type_on_the_copies_of_a_page_cropped_to_the_art_is_read_once`), mesh corner weights and state (`ai::mesh::tests::corners_with_handle_weights_of_zero_are_read`, `ai::tests::a_gradient_mesh_takes_the_transparency_written_after_it`), type on a path spacing (`tests_native::type_round_a_circle_is_spaced_as_the_file_lays_it_out`), blend state (`ai::tests::a_blends_transparency_follows_its_record`); scores and estimates unchanged. Graphs saved to `.ai` in either form open again as live graphs (`ai_editing::tests::graphs_are_written_with_their_records_and_open_live`, `legacy_ai::tests::graphs_are_written_with_their_records_and_open_live`, `cmd::graph::tests::a_drawn_graph_saved_as_ai_opens_as_a_graph_unless_the_save_warns`); other graphs are still written as their art (`graphs_the_records_cant_keep_are_written_as_their_art`); score and hours unchanged. PDF ActualText preserved in editable import (#6165; bounded synthetic coverage below); score/hours unchanged. Save refusal continuation (#6146): `READ_ONLY_LOCATION`, Save As on ordinary UI Save only; regression execution pending; score/hours unchanged. Legacy `.ai` type keeps its manual kerning pairs (`Tk`, `TK`), discretionary hyphens (`T+`, `T-`) and hyphenation language (`XL`) when it opens, and legacy `.ai` saves write manual kerning as `Tk` and soft hyphens as `T+` (`import::ai::tests::legacy_manual_pair_kerning_is_kept`, `legacy_ai::tests::manual_kerning_and_soft_hyphens_are_written_and_read_back`); score, gaps and hours unchanged. |
| View & navigation | 3 | 70% | multiple windows and the rest of Arrange (Cascade, Float in Window, Float All in Windows) | 9–14 | estimated; Window › Arrange › Tile and Consolidate All Windows done (score/hours unchanged); document-tab artwork hover/copy and lossless transfer tested (2026-10-22; score/hours unchanged); cached canvas reprojection preserves scale during resize, rotation and DPI changes (regression and contributor original-file evidence below; scores/hours unchanged) |
| View & navigation | 3 | 70% | multiple windows and the rest of Arrange (Cascade, Float in Window, Float All in Windows) | 9–14 | estimated; Window › Arrange › Tile and Consolidate All Windows done (score/hours unchanged) |
| Guides, grids, smart guides, snapping, rulers | 3 | 80% | "intersect" smart guides for a moved selection's anchors (drawn points and dragged anchors snap to intersections, #1658) | 4–7 | estimated (lowered from 85%: target moved); Control bar snapping quick access includes Snap to Glyph and Alignment Guides (#6164); new ruler-guide Alt orientation implemented; global/artboard ruler modes (`view.globalRulers`) and video rulers (View › Rulers › Show Video Rulers, #2849) done; status-bar X/Y uses the active ruler origin and General units; score unchanged |
@@ -616,6 +616,7 @@ since 2026-10-01.
- **Tab stops in `.ai`/EPS type and right-aligned lines with tabs (#6147):** type from the text document keeps its paragraphs' tab stops (key 30: positions and left, center, right and decimal alignments; leaders aren't read), at most 1,024 a paragraph and 2^20 a story, past which the default stops apply with a warning. In right-aligned area type, the text before a line's last tab sits at its stops, measured from the left edge, and the text after it (none when the tab ends the line, with its hyphen when the line is hyphenated) against the right edge; centred and justified lines, point type and right-to-left paragraphs are aligned as a whole as before. Evidence: `ate::tests::a_paragraph_keeps_its_tab_stops`, `ate::tests::a_story_s_tab_stops_are_bounded`, `tests::right_aligned_area_type_sets_the_text_after_its_last_tab_against_the_edge`, `import_fuzz::text_frames_with_hostile_paragraph_attributes_never_panic`. Scores unchanged.
- **`.ai`/EPS graphic styles that paint with a named gradient:** a graphic style library's style (or a file's listed style) whose fill, or the fill or stroke of a style of one paint, paints with one of the file's gradients by name loads with that gradient at the style's angle, fitted to the bounds of each object it is applied to (VectorCraft's own placement; the style stores only the gradient's name and angle). A style naming a gradient the file doesn't define, or painting with a pattern, is still left out. Evidence: `fileio::tests_aiimport::graphic_styles_in_a_ai_library_that_paint_with_a_named_gradient_load_with_it`, `style::tests::a_style_of_one_paint_reads_as_the_fill_and_stroke_its_xd_turns_on`. Scores unchanged.
- **`.ai`/EPS type laid out closer to the file's page (#6137):** type from the text document keeps auto leading as auto, each line taking the paragraph's Auto Leading share (key 7) of the largest type on it ([Line and character spacing](https://helpx.adobe.com/illustrator/using/line-character-spacing.html)); the paragraph's hyphenation options (keys 10 to 15) are read; area type of one frame whose last lines the file sets, but which don't fit here, may get a frame taller from its bottom, by its lines' deepest descent plus 0.05 pt and only when that shows more of those lines and no more than the file has (`vectorcraft_text::fit_source_lines`), with a warning when it does: the frame can stay as it is, or show only some of them; and a line a soft return ends is justified as a wrapped line is, a paragraph's last line keeping its alignment. Evidence: `ate::tests::auto_leading_follows_the_size_of_each_line`, `ate::tests::a_paragraph_keeps_its_auto_leading`, `ate::tests::a_paragraph_keeps_its_hyphenation_options`, `ate::tests::the_last_lines_the_file_sets_in_its_frame_show`, `tests::fit_source_lines_keeps_the_last_lines_the_source_sets`, `tests::typo::soft_return_in_justified_text_spreads_like_a_wrapped_line`. Limits: frames grow only for area type of one frame, horizontal, by its lines' deepest descent plus 0.05 pt; type in a font that isn't installed, and hyphenation in a language that isn't imported, still lay out differently. Scores unchanged.
- **Graphs saved to `.ai` keep their data and settings:** a grouped column, stacked column, line or area graph is written to a `.ai` file, in either form, with the records the `.ai` reader reads (`crates/eps/src/import/ai/read/graph.rs`, #7311): `%_Gs`, its plot, type, settings, axes and table, its series' groups each tagged `Go`, and `%_GS` ([the specification](https://13thmonkey.org/documentation/fonts/AI7FileFormat.pdf), §17 and appendix A). The file opens again with the same live graph. A graph with settings the reader doesn't keep, graph designs, label formatting, more than 10,000 categories or 256 series, or art the reader doesn't attach a graph to is written as its art, with a warning. `crates/eps/src/legacy_ai/tests.rs` and `crates/eps/src/ai_editing/tests.rs` write and read back each kind of graph; `a_drawn_graph_saved_as_ai_opens_as_a_graph_unless_the_save_warns` checks every graph type the graph tools draw. Scores unchanged.
- **Selected Grid Repeat handles (#2691):** `tools/src/repeat_grid.rs` tests place the handles outside the grid at zoom 0.5, 1 and 4, preview rows, columns and spacing, leave no undo step for a drag that ends where it began, and give bounding-box handles priority; a headless canvas regression drags the row handle two tiles down in a zoomed view (3 → 5 rows, the source unchanged, one Undo back) and the column-spacing slider. Scores and hours unchanged.
- **Selected Mirror Repeat handles:** a headless canvas regression creates the default mirror repeat and drags its Top bounding-box handle where the axis crosses it; the repeat grows vertically while its width stays fixed. Existing mirror-axis checks cover center movement and end-handle hit priority. The focused UI mirror tests pass. Scores and hours unchanged.
- **Transform effect options (#885):** Effect › Distort & Transform › Transform… has visible Transform Objects, Transform Patterns and Scale Strokes & Effects checkboxes (`distort.transform` params `objects`, `patterns`, `strokes`; files without them draw as before). With Transform Objects off the geometry stays and makes no copies; Scale Strokes & Effects scales stroke weights, dashes and the fills' and strokes' effect distances by the mean scale, and Transform Patterns moves pattern tiles, each copy `k` by the transformation `k` times. Such objects draw, export and expand through the same per-copy art, so the canvas and Expand Appearance agree (engine `tests_transformfx`, effects `tests_transformfx`, the dialog test). A Transform effect in a fill's or stroke's own effects applies the same options to that fill or stroke: its weight and dashes and its pattern follow the transformation, and with copies each copy is an object of its own; the canvas, export, Expand Appearance and Use Preview Bounds use the same art (effects `tests_transformfx`). The effects listed after it in that fill's or stroke's list keep their own distances and apply to the transformed art and to each copy, as an object's later effects do after the object's Transform effect: a Drop Shadow after it keeps its offset, an Offset Path after it offsets by its own distance, and an SVG filter after it keeps its scale and the origin of its user space. With Transform Objects off, the effects before it keep their distances too. Before, Scale Strokes & Effects scaled a fill's or stroke's later effects by the mean scale, so existing documents with such effects draw differently. Effects apply in the order they are listed ([Appearance attributes](https://helpx.adobe.com/illustrator/desktop/special-effects-styles/apply-graphic-styles/appearance-attributes-in-illustrator.html)); how the two options treat the effects after a fill's or stroke's own Transform effect is VectorCraft's own rule, following that order (effects `tests::rasterfollow::a_fills_own_transform_leaves_the_shadow_after_it`, `tests::rasterfollow::a_fills_own_transform_leaves_the_offset_after_it`, `tests::rasterfollow::a_fills_own_transform_leaves_the_svg_filter_after_it`, `tests::transformfx::a_fills_own_transform_with_scale_strokes_and_effects_leaves_its_later_effects`; engine `plugins::a_transform_after_a_painted_fill_effect_leaves_the_later_shadow_on_every_paint`). Public reference: [Scale objects](https://helpx.adobe.com/illustrator/desktop/manage-objects/reshape-transform-objects/scale-objects.html), [How to apply effects](https://helpx.adobe.com/illustrator/using/effects.html), [Distort and transform effects](https://helpx.adobe.com/illustrator/desktop/manage-objects/reshape-transform-objects/distort-and-transform-effects-in-illustrator.html). Scores and hours unchanged.
@@ -744,6 +745,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- **Document Setup modal handoff:** Document Setup → Edit Artboards applies valid settings, closes the dialog and activates the Artboard tool. Pointer-click tests cover successful handoff, invalid settings, OK, Cancel and Undo in `dialogs/document_setup.rs`.
- **Document Setup › Color:** a Color tab shows the document color mode (RGB Color or CMYK Color), a Convert to CMYK or Convert to RGB box for a mode change (cleared: the colors keep their numbers), and the RGB and CMYK profiles assigned to the document, with a working-space entry for none. OK runs `edit.assignProfile` and then `file.documentColorMode` with the dialog's other changes in one `command.batch`, and the document's changes are one undo step. An unchanged Color tab adds no color commands, and Cancel, or OK with nothing changed, records no undo step. A Color tab field with a value of the wrong kind is an error, and the dialog stays open. The tab is VectorCraft's own layout; File › Document Color Mode and Edit › Assign Profile… stay as they are. Tests in `dialogs/document_setup.rs`.
- **Canvas through the display's profile on Windows and X11:** Preferences › Performance › Use Monitor Profile for Canvas (on by default) converts the canvas's art, artboard paper and transparency grid from sRGB into the ICC profile of the display that shows most of the main window, after any proof, as the help page describes the monitor's profile translating the document's colors for the monitor ([the help page](https://help.adobe.com/archive/en_US/illustrator/cs4/illustrator_cs4_help.pdf#page=151)). Windows gives the display's default profile, and a display with Advanced Color on is left to Windows ([Microsoft Learn](https://learn.microsoft.com/en-us/windows/win32/wcs/advanced-color-icc-profiles)); X11 gives colord's profile for the output or the `_ICC_PROFILE` property. macOS converts the window itself, and a Wayland session stays sRGB. `ui.inspect` reports `displayProfile`. Tests in `color/src/cms/display.rs`, `ui-egui/src/display_profile.rs` and `canvas::tests::the_canvas_is_drawn_through_the_display_profile_while_the_preference_is_on`.
- **Translucent art blended before the display profile:** with Use Monitor Profile for Canvas converting, translucent art and the anti-aliased edges of art over an artboard's paper, a colored artboard or the transparency grid are blended with it in sRGB and the blend is converted, so the monitor's profile translates the document's colors as they are composited ([the help page](https://help.adobe.com/archive/en_US/illustrator/cs4/illustrator_cs4_help.pdf#page=151)). The frame rendered whole, the part an edit renders again, the render worker's frame, the pan fill and the other windows' panes do the same; over the pasteboard and in Outline the art is converted on its own, and with no profile to convert into nothing changes. Tests `display_profile::tests::translucent_pixels_over_the_paper_are_blended_before_they_are_converted`, `the_transparency_grid_is_blended_cell_by_cell`, `the_backdrop_is_found_through_the_view` and `canvas::tests::translucent_art_over_the_paper_is_blended_before_the_display_profile`.
- **Out-of-gamut warning after a profile change:** the out-of-gamut warning and its closest-color correction in the Color panel and Color Picker ([the help page](https://helpx.adobe.com/illustrator/using/adjusting-colors.html)) check the color against the document's assigned CMYK profile, else the working CMYK space, and update when Edit › Assign Profile… or Edit › Color Settings… changes that space or a loaded profile changes, without a change of color. Which CMYK space the warning uses is VectorCraft's own rule. Tests: `gamut_warning_follows_a_profile_change` and `gamut_warning_follows_a_profile_loaded_again` in `dialogs/color_picker/tests.rs`.
- **Assign Profile and the working spaces:** Edit › Assign Profile… tags the document and leaves the working spaces of Edit › Color Settings… as they are, in Color Settings, the Separations panel and the settings saved for the next start. Colors show and convert through the profiles assigned; a change in Color Settings keeps the profiles the document in front is assigned in use, and removing an assigned profile in use puts the working space back in its place. The profiles in use follow the document in front: a new, opened or reverted document, a switch to another tab, a close, Undo and Redo of Assign Profile, a failed batch, each step of a batch, Export All Open and the update of placed documents put that document's profiles in use, and the working spaces where it is assigned none or no document is open. With Window › Arrange › Tile or floating document windows, each document behind the front one draws its art, paper and transparency grid through its own assigned profiles, and loading or removing a profile redraws the documents assigned it; placed documents inside it still draw through the front document's profiles. How documents behind the front one draw is VectorCraft's own rule. View › Proof Setup › Working CMYK simulates the working CMYK space of Color Settings in Proof Colors and Separations Preview, also in a document assigned another CMYK profile. Public references: [Changing the color profile for a document and About color working spaces](https://help.adobe.com/archive/en_US/illustrator/cs3/illustrator_cs3_help.pdf#page=150) and [Soft-proof presets](https://help.adobe.com/archive/en_US/illustrator/cs3/illustrator_cs3_help.pdf#page=134). Tests in `engine/tests/color_settings.rs`.
- **New Document's More Settings and Save Preset:** More Settings opens a pop-up over New Document with only Number of Artboards, the artboard layout, Change to Right-to-Left Layout, Spacing and Columns; Done, Escape or a click outside closes it and keeps the values. Save Preset turns the top of Preset Details into a Save Document Preset heading with the preset's name, Cancel and Save, and hides the document's name, Close and Create until the preset is saved or canceled; nothing in that mode creates a document. Save Preset's default name is the first "Preset N" no saved preset has. VectorCraft's own layout. Harness tests in `dialogs/new_document.rs` cover both.
@@ -807,6 +809,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- Batch… over data sets offers None, Save and Close and Folder as the destination, as for a folder source. Save and Close saves the open document over its own file after each data set, so the file keeps only the last data set, unless a recorded Save step that Override Save points into the destination folder first makes each data set's file the document's own. A recorded Save, Save a Copy, Save as Template or export step that writes a file and that the batch does not point into the destination folder (destination None, Override Save or Override Export off, or a copy, template, PDF, EPS, DXF, Save for Web or Save for Office Documents step) also writes the same file for every data set. In either case the dialog shows a warning and `action.batch` returns a `warning`. Unless a recorded Save or Save As step writes a file and makes it the document's own first, `action.batch` refuses Save and Close over the data sets of a document that was never saved, was converted from an older version or was restored by Data Recovery.
- Export for Screens: the dialog's Example line shows the path of the file the export writes first, sub-folder included, named as the export names it, for the first chosen artboard, asset, layer or object (`document.exportForScreens`).
- Export presets: the Asset Export panel menu lists the export presets of Export for Screens, with Import Preset… and Export Preset…, which read and write `.vcexportpresets` files (`document.exportPresets.import`, `document.exportPresets.export`). An imported preset whose name is taken gets a number. In the panel, a preset applies to the shared export settings (`assets.settings.set {exportPreset}`), and Save Preset… saves them. With Create Sub-folders off, the panel's Export writes every file into the folder itself, mobile and density rows included.
- Export for Screens artboard names can be double-clicked to rename them in grid or list view (#7275). Enter, a click outside the field or switching to Assets commits the name; Escape cancels the edit. The dialog stays open, checked artboards stay selected and the rename is undoable. The typed name goes to the artboard that was double-clicked even when the artboards change order while it is typed; the field closes when its artboard is deleted or another document becomes active. Evidence: `dialogs::tests_screens::double_clicking_an_artboard_name_in_export_for_screens_renames_it`, `dialogs::tests_screens::switching_export_tabs_commits_an_artboard_name`, `dialogs::tests_screens::an_artboard_name_follows_its_artboard_while_it_is_typed` and `dialogs::tests_screens::an_artboard_name_stays_with_its_document`.
- Asset Export panel menu: Create Sub-folders (on by default) and Open Location after Export are check items that set the dialog's two checkboxes (`assets.settings.set {subfolders, openLocation}`), and the panel's Export follows them, including a folder per format or item that Export for Screens remembers (3D assets' mesh files go in the folder itself); Format Settings… opens Export for Screens on the first row's format settings ([Asset Export panel](https://helpx.adobe.com/illustrator/desktop/save-and-export/export-files-to-different-formats/collect-and-export-assets-from-asset-export-panel.html)). Tests: `panels::tests_asset_export`.
- Export for Screens is laid out as the reference app's dialog: on the left the Artboards and Assets tabs (thumbnails with their checkboxes, number and name, in a grid or a list), Clear Selection and Prefix; on the right Select (All, Range, Include Bleed, Include Artboard Fill, Full Document), Export to, Open Location after Export, Create Sub-folders by Scale or by Format (on by default), Export PDFs As (Single File writes the chosen artboards as the pages of one PDF named after the document, `pdfs: "single"`), and the format rows with the iOS and Android buttons, the format presets (☰) and Format Settings; under them, how many items are selected and how many files the export writes. File name patterns, the Layers and Selection sources and per-item folders stay available through `document.exportForScreens` (scripts and MCP). Tests: `dialogs::tests_screens` (layout, iOS/Android, Clear Selection, sub-folders, Single File) and `fileio::tests_screens::single_file_pdfs_hold_every_chosen_artboard`.
- Save for Web: Preview in Browser writes each preview's page and images to a folder of its own inside `VectorCraft Save for Web` in the system's temporary folder. A later preview, from the same app or another one running, does not overwrite an earlier one. Each preview first removes the preview folders there that are an hour old or older and, on Linux, macOS and FreeBSD, belong to the current user; it leaves other files and folders alone. The folders a preview makes there are open to their owner only on those systems. When `VectorCraft Save for Web` is a link or is not a folder, or on those systems belongs to another user or can be written by other users, the preview goes to a new folder of its own in the temporary folder, which a later preview removes once it is an hour old (`file.saveForWeb.browser`).
@@ -817,8 +820,8 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- Variables: Import Data Sets… (`dataset.import`) reads a UTF-16 file without a byte-order mark, instead of refusing it, when at least half of its two-byte units hold a NUL byte; it uses the byte order that reads its line breaks.
- Graphs: the Graph Data window's Import Data reads a UTF-16 text file without a byte-order mark by the same rule as Import Data Sets…, instead of filling the cells with NUL characters. It reads a file cut short by one byte, or with an unpaired surrogate, as it reads one after a byte-order mark.
- Variables: Import Data Sets… (`dataset.import`) defines the variables a file's header names that the document lacks, unbound. `#Shown` defines a visibility variable, `@Photo` defines an image variable, and a header without a prefix defines a text variable. A header that is a variable's exact name keeps naming it, and any other `%` header (graph data files) is refused.
- Variables: Load Variable Library… in the Variables panel menu (`variable.loadLibrary`) replaces the document's variables and data sets with a variable library's, read from its XML file, in one undo step: every data set is deleted, and every variable is deleted with its bindings and values, except a text variable the library also declares, which stays bound so the library's data sets change its objects (#4136); then the library's other text variables are defined, unbound, and its data sets take the text of each value's `p` elements, one paragraph each. Variables of other kinds are left out with a warning that names their kinds.
- Variables: Save Variable Library… in the Variables panel menu (`variable.saveLibrary`) writes the document's text variables and data sets as a variable library's XML file, which Load Variable Library reads back with the same text variables, data set names and text values. Variables of other kinds are left out with a warning that names their kinds.
- Variables: Load Variable Library… in the Variables panel menu (`variable.loadLibrary`) replaces the document's variables and data sets with a variable library's, read from its XML file, in one undo step: every data set is deleted, and every variable is deleted with its bindings and values, except a variable the library also declares with the same kind, which stays bound so the library's data sets change its objects (#4136); then the library's other variables are defined, unbound: text, visibility, image (linked file) and graph variables, and variables without a known `trait` as unknown. Its data sets take text values from each value's `p` elements, one paragraph each, visibility values from `true` or `false`, image values from a file's absolute path or `file:` URL, and graph values from a `data` element of series names and rows.
- Variables: Save Variable Library… in the Variables panel menu (`variable.saveLibrary`) writes the document's variables of every kind and its data sets as a variable library's XML file, which Load Variable Library reads back with the same variables, kinds, data set names and values.
- JSON-lines control channel with real egui pointer and keyboard injection.
- Variables: a type object holds one text variable. `variable.bind` and Make Text Dynamic refuse a second one and name the text variable the object has; a visibility variable can still share the object, and a document saved with two text variables on one object opens with both.
- Synthetic input validates complete gestures before queue mutation: at most 256 drag steps, 16 clicks and 1,024 pending events; up to 64 control requests per frame, with later requests retained for the next frame (#1709).
@@ -910,6 +913,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- Hindi (2026-11-06): Hindi (`hi`): 7,877-row machine-assisted catalog, zero/one singular plurals, desktop font fallback and optional Devanagari craft-fonts fallback; all 55 translation tests pass; native review and a web font remain; scores and estimates unchanged. [Terminology and limitations](localization-hi.md).
- Gujarati (2026-11-14, #6613): `gu` has 4,673 catalog rows, zero/one singular plurals, installed desktop font fallback and support for optional `Gujr` craft-font faces. The original 4,286 rows remain unchanged; all 387 additions are contributor-reviewed and approved, with 11 wording corrections. Status/error messages remain partial; the pinned fonts and web startup list have no Gujarati face, so web selection remains unavailable until one is bundled. Current-head validation is recorded in the PR; scores and estimates unchanged.
- Hindi wording (2026-11-13): rows outside the plug-in messages keep their English's meaning, quantities and clauses, and use the catalog's save, swatch, crop, resolution, path and type terms; a test checks those terms and the point unit in every row; native review remains; scores and estimates unchanged.
- Hindi word senses (2026-11-13): 85 labels and messages use the interface sense of their English: Desired is वांछित and the range message keeps both comparisons, rulers are रूलर, text characters कैरेक्टर, vector points पॉइंट, fills फ़िल्स and High Contrast उच्च कंट्रास्ट; the glossary test rejects the unrelated senses in every row; native review remains; scores and estimates unchanged.
- Hindi terms (2026-11-13): every row that names a variable (data merge, variable fonts and widths, environment variables) says वेरिएबल, and joining paths, anchor points, point type and lines says जोड़ें instead of शामिल हों, which stays for joining the community; a test checks both terms in every row; scores and estimates unchanged.
- Hindi English words (2026-11-13): 264 rows no longer keep English words in their sentences, repeat a parameter's name in Hindi or garble them, among them "unknown align", which now names the alignment; lists of accepted values such as `(left|right|top|bottom)` stay as their English writes them; two tests check every row; native review remains; scores and estimates unchanged.
- Message field names and values (2026-11-13): 49 status and error messages in the Hindi, Bulgarian, Danish, Indonesian, Dutch, Brazilian Portuguese, Swedish and Turkish catalogs keep the field names in brackets, such as `[subpath, anchor]`, and the values an `unknown` message lists, such as `(points, lines)`, as their English writes them; a test checks every catalog; scores and estimates unchanged.
@@ -950,7 +954,8 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- Align New Objects to Pixel Grid (`document.setup {alignNewObjectsToPixelGrid}`, a checkbox in New Document's Advanced Options and a checked item in the Transform panel menu) gives every path and compound path drawn in the document, the paths of a new grid or flare included, the pixel-aligned property, snapped in the step that draws it; it is on in the Mobile and Web presets and off in the others, saved presets and Recent keep it, and the document saves it. Type, images and groups do not take it, and art already in the document, pasted, duplicated or placed keeps its own property (VectorCraft's own rules) ([Pixel-aligned property](https://helpx.adobe.com/illustrator/desktop/get-started/learn-the-basics/pixel-preview-mode-for-web-graphics.html); `tests::pixelalign::align_new_objects_makes_new_paths_pixel_aligned`, `align_new_objects_reaches_grids_flares_and_add_on_shapes`, `align_new_objects_is_saved_and_read_back`, `tests::newdoc::align_new_objects_to_pixel_grid_follows_the_preset`, `panels::transform::tests::align_new_objects_menu_item_toggles_the_active_document`, `dialogs::new_document::tests::align_new_objects_checkbox_follows_the_preset_and_reaches_the_document`).
- Break Link to Symbol and Object › Expand give a symbol instance's art the instance's own bounding box: after Reset Bounding Box the art stands square to the page, an object turned inside the symbol keeps its angle relative to the instance's box, a group added for the instance's opacity takes the instance's box, and the clip groups of 9-slice art take the angle of their cell. Compound shapes and Pathfinder receive an instance's art with the same angles (VectorCraft's own behavior; `bboxrotate::reset_bounding_box_then_expand_symbol_instance`, `reset_bounding_box_then_expand_keeps_the_angles_inside_the_symbol`, `expanding_a_faded_symbol_instance_gives_its_group_the_instance_box`, `expanding_a_reflected_symbol_instance_keeps_its_box`, `symbol_nine_slice::break_link_turns_the_cells_of_a_sliced_instance_with_its_box`, effects `outline_object_turns_an_instances_art_with_its_box`).
- Object › Transform › Reset Bounding Box squares the box of every object inside the selected ones as well: the children of groups, compound paths, clip groups, blends, envelopes, repeats and intertwines, and the art of linked opacity masks. Ungroup and the release commands then bring back boxes square to the page (VectorCraft's own behavior; `bboxrotate::reset_bounding_box_squares_the_objects_inside_a_group`, `reset_bounding_box_squares_the_turned_members_of_a_square_group`, `reset_bounding_box_then_release_clipping_mask` and the other `reset_bounding_box_then_*` tests, `reset_bounding_box_leaves_the_art_of_an_unlinked_opacity_mask`).
- After a rotation the Free Transform tool shows and drags the same turned bounding box as the Selection tool: its handles, their cursors and its rotate zone sit on the selection's own sides, a handle scales along those sides from the opposite handle ([the help page](https://helpx.adobe.com/illustrator/desktop/manage-objects/reshape-transform-objects/scale-objects.html)), Cmd/Ctrl on a side handle shears along that side, and Perspective Distort and Free Distort start from the turned box's corners (`object.distort` takes `angle` with `from`). The Scale and Reflect tools, their dialogs and commands, and the Transform panel's flips keep the page's axes (VectorCraft's own rule; `bboxrotate::free_transform_handles_turn_with_the_object_and_scale_along_its_axes`, `free_transform_shears_along_the_turned_side`, `free_transform_distort_starts_from_the_turned_box`, `xform::distort_reads_from_in_a_turned_frame_and_rejects_a_bad_angle`, tools `xform::free::tests::the_box_and_its_handles_turn_with_the_selection`).
- After a rotation the Free Transform tool shows and drags the same turned bounding box as the Selection tool: its handles, their cursors and its rotate zone sit on the selection's own sides, a handle scales along those sides from the opposite handle ([the help page](https://helpx.adobe.com/illustrator/desktop/manage-objects/reshape-transform-objects/scale-objects.html)), Cmd/Ctrl on a side handle shears along that side, and Perspective Distort and Free Distort start from the turned box's corners (`object.distort` takes `angle` with `from`) (`bboxrotate::free_transform_handles_turn_with_the_object_and_scale_along_its_axes`, `free_transform_shears_along_the_turned_side`, `free_transform_distort_starts_from_the_turned_box`, `xform::distort_reads_from_in_a_turned_frame_and_rejects_a_bad_angle`, tools `xform::free::tests::the_box_and_its_handles_turn_with_the_selection`).
- After a rotation the Scale and Reflect tools, the Scale and Reflect dialogs, `object.scale`, `object.reflect` with a named axis and the Transform and Properties panels' flips act along the selection's turned bounding box until Reset Bounding Box: the horizontal and vertical scale stretch along its own width and height, and the vertical and horizontal reflection axes are its own. The Reflect tool's default axis and its Shift steps start from the box's axes, and an axis drawn with the tool is the line drawn on the page. `basis: "page"` keeps the page's axes (VectorCraft's own rules; `bboxrotate::scale_and_reflect_follow_the_turned_box`, tools `xform::transform::tests::scale_and_reflect_drags_follow_a_turned_bounding_box`).
- Perspective Distort and Free Distort (`object.distort`) turn each warped object's bounding box as the warp turns the object at its own center, the objects inside groups, compound paths and clip groups and the art of linked opacity masks included, so the next Selection or Free Transform drag uses the tilted box; Object › Transform › Reset Bounding Box squares it again, and corners left where they are add no undo step (VectorCraft's own rule; `bboxrotate::distort_turns_an_upright_box_with_the_warp_at_its_center`, `distort_turns_each_member_and_linked_mask_by_its_own_warp`, `mirrored_and_collapsing_distorts_keep_valid_angles`, `distort_angles_undo_redo_and_reopen_together`, `xform::unchanged_distort_records_no_history`).
- The Free Transform widget's Constrain works in Free Distort as well as in Free Transform ([the help page](https://helpx.adobe.com/illustrator/using/tool-techniques/free-transform-tool.html)). With Constrain on, or Shift held, a corner dragged in Free Distort moves only along the line through the opposite corner, in the box's own proportions, and the other three corners stay in place; on a turned box the line runs through the box's own corners. The line is VectorCraft's own rule. Cmd/Ctrl+Alt+Shift on a corner still drags in Perspective Distort, where Constrain stays unavailable (tools `xform::free::tests::free_distort_corner_keeps_to_its_diagonal_with_shift_or_constrain`, `shift_pressed_during_a_free_distort_drag_constrains_it`, `a_turned_box_is_constrained_along_its_own_diagonal`, `along_diagonal_of_a_degenerate_box`; engine `xform::free_transform_constrained_distort_keeps_the_corner_on_its_diagonal`; ui-egui `free_transform::tests::the_widget_sets_the_mode_and_a_corner_drag_then_distorts`).
- Native macOS menu bar, vector tool cursors, a ⌘K command palette, middle-button panning with any tool. Native menus explicitly end tracking on command selection, outside window presses and Escape; menu updates wait until tracking ends, and a dismissing press never starts a canvas gesture (M14.8).
@@ -986,7 +991,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- **Delete Symbol asks first** ([Symbols panel](https://helpx.adobe.com/illustrator/desktop/manage-objects/traces-mockups-symbols/symbols-panel-options.html), [symbol libraries](https://helpx.adobe.com/illustrator/desktop/manage-objects/traces-mockups-symbols/create-or-import-symbol-libraries.html), [Alt-click Delete](https://helpx.adobe.com/illustrator/using/default-keyboard-shortcuts.html)): the Symbols panel's Delete Symbol button and menu item ask "Delete the selected symbol?" with Yes and No for a symbol no instance uses. For a symbol with placed instances they ask with Expand Instances (the instances become ordinary art), Delete Instances and Cancel; either Instances button deletes the symbol in one undo step. Alt-click on the button deletes an unused symbol without asking and still asks about one with instances. A question left open while another document became active deletes nothing. `symbol.delete` itself does not ask. Tests: `panels::symbols::tests::delete_symbol_asks_first_and_alt_click_deletes_an_unused_symbol_at_once`, `the_in_use_question_expands_or_deletes_the_instances`, `a_stale_delete_symbol_question_deletes_nothing`, `the_delete_symbol_questions_show_their_buttons`.
- **Symbols panel context menu** ([Symbols panel](https://helpx.adobe.com/illustrator/desktop/manage-objects/traces-mockups-symbols/symbols-panel-options.html)): right-clicking a symbol in thumbnail view or list view displays a context menu with Place Symbol Instance (passes view center or artboard center), Edit Symbol, Duplicate Symbol, Delete Symbol, Break Link to Symbol (enabled when canvas selection includes instances of this symbol), and Select All Instances. Right-clicking an unselected symbol selects it first (`symbol.setCurrent`). Tests: `panels::symbols::tests::right_click_context_menu_in_symbols_panel_runs_actions`.
- **Symbols › Select All Unused** ([Symbols panel](https://helpx.adobe.com/illustrator/desktop/manage-objects/traces-mockups-symbols/symbols-panel-options.html), #4918): the Symbols panel menu's Select All Unused shows the symbols nothing places selected (`symbol.unused`): instances in the art (opacity masks included), graph designs, pattern tiles and brush art count, and so do instances in a used symbol's art. Delete Symbol, the button or the menu item, then asks "Delete the selected symbols?" once and deletes those still unused in one undo step (`symbol.delete {names}`); Alt-click on the button deletes them without asking, and a click on a symbol picks just it again. Tests: `tests::brushsym::symbol_unused_lists_the_symbols_nothing_places_and_delete_takes_several`, `panels::symbols::tests::select_all_unused_picks_the_unused_symbols_and_delete_deletes_them`, `dialogs::symbol_delete::tests::several_symbols_get_the_plural_question`.
- **Symbols › Sort by Name and the list views** ([the help page](https://helpx.adobe.com/illustrator/desktop/manage-objects/traces-mockups-symbols/symbols-panel-options.html)): the Symbols panel menu's Sort by Name puts the symbols in alphabetical order (`symbol.sortByName`), and its Thumbnail View, Small List View and Large List View show them as thumbnails, in a compact list or in a list with larger thumbnails. VectorCraft's own rules: the sort ignores case and keeps names that differ only in case in their order; it changes the document's order, which `symbol.list` reports, as one undo step, keeps the current symbol and adds no undo step when the symbols are already in order; Thumbnail View is the default, Small List View has 18-point rows with 14-point thumbnails and Large List View 26-point rows with 22-point thumbnails, and a click or double-click on a symbol does the same in every view. Tests: `tests::brushsym::symbol_sort_by_name_orders_the_symbols_in_one_undo_step`, `panels::symbols::tests::sort_by_name_and_the_small_and_large_list_views`.
- **Symbols › Sort by Name and the list views** ([the help page](https://helpx.adobe.com/illustrator/desktop/manage-objects/traces-mockups-symbols/symbols-panel-options.html)): the Symbols panel menu's Sort by Name puts the symbols in alphabetical order (`symbol.sortByName`), and its Thumbnail View, Small List View and Large List View show them as thumbnails, in a compact list or in a list with larger thumbnails. VectorCraft's own rules: the sort ignores case and keeps names that differ only in case in their order; it changes the document's order, which `symbol.list` reports, as one undo step, keeps the current symbol and adds no undo step when the symbols are already in order; Thumbnail View is the default, Small List View has 18-point rows with 14-point thumbnails and Large List View 26-point rows with 22-point thumbnails, and a click or double-click on a symbol does the same in every view. `ui.panelView` sets the view of the Symbols panel, and of the Swatches and Graphic Styles panels, as their panel menus do. Tests: `tests::brushsym::symbol_sort_by_name_orders_the_symbols_in_one_undo_step`, `panels::symbols::tests::sort_by_name_and_the_small_and_large_list_views`, `panels::symbols::tests::panel_view_sets_the_view_of_each_panel`.
- **Expand of symbol instances:** Object › Expand turns selected symbol instances (without effects, not in perspective, in groups or placed documents too) into their art as Break Link to Symbol places it, one level deep, as one undo step; the Expand dialog's Fill and Stroke reach that art (`object.expand`, `crates/engine/src/tests/expand_symbols.rs`; estimated, score unchanged).
- **Reset Transformation** ([Transform symbols](https://helpx.adobe.com/illustrator/desktop/manage-objects/traces-mockups-symbols/transform-symbols.html), [Symbols panel options](https://helpx.adobe.com/illustrator/desktop/manage-objects/traces-mockups-symbols/symbols-panel-options.html)): the Symbols panel menu's Reset Transformation and the Control bar's Reset, shown while symbol instances are selected, give the selected instances their symbol's own scale again, without rotation or shear, as one undo step (`symbol.instance.resetTransform`); an instance in perspective is left as it is. VectorCraft's own rules: each instance keeps the center of its bounding box in place, reflection is removed too, and the menu item and the button are enabled only while a selected instance would change. Tests: `tests::brushsym::reset_transformation_restores_the_symbols_scale_and_keeps_the_center`, `reset_transformation_keeps_the_center_of_art_away_from_the_origin`, `reset_transformation_ignores_a_stored_size_that_is_not_positive`, `reset_transformation_leaves_unchanged_instances_and_refuses_bad_ids`, `panels::symbols::tests::reset_transformation_is_enabled_for_a_transformed_instance_and_resets_it`.
- **Expand of dimensions:** Object › Expand, Expand Appearance, Create Outlines and Outline Stroke turn a dimension whose art they change into a plain group of that art, which stays as it is through later moves, scales and edits of what it measured; undo gives the live dimension back (`crates/engine/tests/it/dimensions.rs`; VectorCraft's own rule; estimated, score unchanged).
@@ -1137,6 +1142,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
With Smart Guides on, a straight Knife cut snaps to anchor points and paths: [Cut objects using the Knife and Scissors tools](https://helpx.adobe.com/illustrator/desktop/manage-objects/edit-objects/cut-objects.html) describes the Knife snapping to anchor points and paths while Smart Guides are on. Its ends snap as the ends of the Pencil's straight line do, to anchor points and path segments of art that is shown: the press, and the release of an Alt-drag. Shift+Alt steps 45° from the snapped press, and its release stays on that step. With Smart Guides off, or with Snap to Pixel or Snap to Grid on, the cut is not snapped. The dashed line and `path.knife` use the snapped points.
Eraser Tool Options has Angle and Roundness besides Size, as [Erase paths using the Eraser tool](https://helpx.adobe.com/illustrator/desktop/draw-shapes-and-paths/modify-paths/erase-paths-using-eraser-tool.html) describes: the Eraser's tip is an ellipse Size across, its long axis at Angle and its short axis Roundness percent of the long one, and a drag with a tip that is not round shows the outline of the area it erases (`path.eraseRegion` `angle` and `roundness`).
The outline a Blob Brush or Eraser drag shows is built a bounded amount per frame, so a long stroke does not slow the canvas: each frame takes at most 256 new points of the drag and outlines again only the pieces of up to 32 segments they change, and the parts of a piece's outline inside the rest of the area are removed a limited amount at a time. While work is left, the canvas draws again with the pointer still, the points not outlined yet show as a thin line, and the tip shows at the pointer. With Size set to Pressure the outline also shows only the edge of the whole area. The release sends the whole stroke as before (`path.blob`, `path.eraseRegion`).
The settled part of that outline is kept in paths of up to 64 pieces each. A path is shared from one frame to the next until a piece in it changes, and the canvas keeps the triangles it drew the path with while the view stays the same, so a frame does not copy, flatten or tessellate a settled path again; a pan, zoom or rotation draws each one once more. This is VectorCraft's own rule. The outline of a drag with a round Eraser tip is drawn whole each frame.
- **Transform:** Rotate, Reflect, Scale, Shear (click or Alt-click snaps the reference point to anchors and centres; Alt-click opens the tool's dialog, and a drag with Alt down at the release, from the press or pressed mid-drag, transforms a copy), Reshape, Free Transform (free and perspective distort and shear, switched by the keys held during a handle drag).
A Reshape drag stretches part of a path from the dragged point without losing the path's general form, as [Adjust path segments](https://helpx.adobe.com/illustrator/desktop/draw-shapes-and-paths/modify-paths/adjust-path-segments.html) has it. On an open path the stretch reaches both ends: the end anchors stay in place, each anchor between moves less the nearer it lies to an end, and the handles of a smooth anchor stay in line (`path.reshape`; the reach and the shares are VectorCraft's own rule, in `docs/mcp.md`). A click sets a focal point and a click on a segment adds a highlighted anchor there; Shift-click adds more, and a drag from any of them moves them all. The anchors that are not direct-selected stay in place and a closed path selected whole moves whole (VectorCraft's own rule, in `docs/mcp.md`).
Align, Distribute, Distribute Spacing, Reflect and Shear return an error for any `horizontal`, `vertical`, `to`, `axis`, `bounds` or `spacing` value they don't take, listing the choices, and Align returns one when a call gives neither `horizontal` nor `vertical`; nothing moves and no undo step is recorded. Align and Distribute act on the anchor points selected with Direct Selection: the points of one path align to each other, not to the artboard, and Distribute spaces them (#1641, fixes #1636; aligning to a key anchor is still missing).
@@ -1381,6 +1387,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
The Glyphs panel's Alternates for Current Selection (#5514) applies the OpenType feature that draws the glyph to the selected character alone, as [Replace characters with alternate glyphs](https://helpx.adobe.com/illustrator/desktop/design-with-text/special-characters-glyphs/replace-characters-with-alternate-glyphs.html) describes (Alternates For Current Selection in the Show menu; double-click the glyph to replace the selected character). One deliberate difference from a literal reading: the text keeps the Unicode character and gains a character-style feature with a value (`salt=2`, `ss03`; `text.setRangeStyle {featureChanges}`) instead of taking an unencoded glyph in its place. The grid lists the alternates of the glyph the shaper makes for the character in its word (a medial letter shows its medial variants; the default glyph comes first and the current glyph is highlighted); a variant no supported feature draws there stays enabled as a replacement glyph (`text.setGlyphAlternate`, including positional alternates), and its tooltip says it may not join its neighbors; the default glyph takes the character's alternate features off again. Runs that differ only in their OpenType features are shaped as one piece of text, each run's features over its own characters, so the letters of an Arabic word keep their joining forms. Show › Entire Font also lists the glyphs no character maps, with the font's name for them (or `gid N`) and the feature that gives them. It applies the selected letter's contextual alternates the same way, and applies a glyph no feature reaches whose name does not belong to another letter as a glyph choice; other glyphs are dimmed with the reason. An Arabic presentation form picked there inserts its base letter. These rules are VectorCraft's own. Tests: `tests::glyph_alternates` in the text crate (a fixture font; Kufi Murabba PIIIX when it is installed), `panels::glyphs::tests` (including positional glyph choices, unlisted glyph name ownership, apply and Undo), `tests::textedit::feature_values_are_set_replaced_and_saved`.
With a Type tool and exactly one character selected, the canvas shows that character's alternate glyphs in a row beside it, and a click on one replaces the character's glyph, as [Replace characters with alternate glyphs](https://helpx.adobe.com/illustrator/desktop/design-with-text/special-characters-glyphs/replace-characters-with-alternate-glyphs.html) describes ("Replace a character with on-canvas alternate glyphs"). The row shows at most five alternates, and an arrow at its end opens the Glyphs panel on Alternates for Current Selection when the character has more, as the [reference manual](https://web.archive.org/web/20210601162814/https://helpx.adobe.com/pdf/illustrator_reference.pdf#page=479) describes. A click runs the command the panel's double-click runs, as one undo step, so a letter of an Arabic word keeps the word joined. These rules are VectorCraft's own: the row leaves out the glyph the character is drawn with and puts the default glyph first for a character drawn with an alternate; it stands under a character of horizontal type and on the right of a character of vertical type, or on the other side when the canvas ends there or the Contextual Task Bar is there; Preferences › Type › Show Character Alternates (`showCharacterAlternates`, on by default) turns it off. Tests: `canvas::alternates_bar::tests`, `tests::prefs::show_character_alternates_preference_is_on_by_default_and_saved`.
The OpenType panel's Stylistic Sets menu offers `ss01`–`ss20` sets present in the selected text's font, using the font's custom labels when available and numbered labels otherwise. Sets can be combined and deselected individually. With mixed fonts, the menu shows the first selected character's font's sets and applies the feature tag to all selected text; a font without that set draws unchanged. The mixed-font rule is a VectorCraft design choice. Behavior reference: [Add stylistic sets to selected text](https://helpx.adobe.com/illustrator/desktop/design-with-text/special-characters-glyphs/add-stylistic-sets-to-selected-text.html) and the [OpenType specification](https://learn.microsoft.com/en-us/typography/opentype/spec/features_pt).
With Show East Asian Options on, the OpenType panel menu has H or V Style and Roman Italics, as [Format East Asian characters](https://helpx.adobe.com/illustrator/using/formatting-asian-characters.html) describes. H or V Style sets the feature tag `hkna`, which draws kana with the forms the font makes for horizontal type, and with `vkna` for vertical type in vertical type; Roman Italics sets `ital`, the font's italic forms of half-width alphanumerics ([OpenType feature list](https://learn.microsoft.com/en-us/typography/opentype/spec/featurelist)). Each is checked when every selected character has it and turns it on or off for all of them in one undo step; character styles carry both in OpenType Features. Tests: `panels::opentype::tests::kana_style_and_roman_italics_turn_on_and_off`, `features::tests::kana_style_and_roman_italics_are_kept_and_shaped`.
`text.setStyle {stroke, strokeWidth}` sets the character stroke's color and weight of the selected type or of a range of its characters (the weight in document points, as `stroke.set` sets it), in the same undo step as the command's other changes. Tests: `tests::textedit::set_style_sets_the_character_stroke`.
- **Paragraph alignment:** the shared Paragraph/Properties alignment row leaves every button unselected when the paragraphs it acts on differ. Auto resolves per paragraph; choosing an alignment uses the existing command. Headless paragraph_mixed_alignment_* regressions exercise mixed → Center → Undo, caret/uniform states, Unicode range boundaries, Auto/direction, split runs and grouped text. These regressions and full integration CI pass. Original captures confirm mixed → Center → Undo in the existing alignment-button layout. The [public Paragraph panel guide](https://helpx.adobe.com/illustrator/desktop/design-with-text/edit-format-text/paragraph-panel-overview.html) lists the seven controls, but does not specify mixed-state rendering. Other paragraph fields remain outside this change; scores and hours unchanged.
For Auto Size area type in columns, the bisection for the frame's height runs on the lines of the one-column layout, and its result is checked with at most two layouts of the columns. When the text wraps around objects or does not fit the one-column layout, or when the check fails, the columns are laid out at every step of the bisection, as before. Typing at the start of a story of 50,000 characters in 2 columns took 1,670 ms a keystroke and now takes 292 ms (release build, `auto_height_in_columns_typing_timing`).
@@ -1525,7 +1532,9 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
Object › Live Paint › Merge adds the other selected paths and Live Paint groups to the bottom selected Live Paint group ([Live Paint groups](https://web.archive.org/web/20241230205225/https://helpx.adobe.com/illustrator/using/live-paint-groups.html)), which keeps its id, the paint of its faces and edges, and its name, lock, visibility, transparency, appearance, opacity mask and graphic style link. Object › Live Paint › Make with a Live Paint group selected, and a Live Paint Bucket click on a selected path beside a selected Live Paint group, add the paths to that group the same way. Make, Merge and the Live Paint Bucket take paths, compound paths and groups of them; with type, an image or another object among the selection they report that it has to be converted to paths first and change nothing, as type and raster images can't go into a Live Paint group directly ([Create Live Paint groups](https://helpx.adobe.com/illustrator/desktop/paint-and-fill/learn-painting-basics/create-live-paint-groups.html)); before, such an object went among the group's hidden sources and vanished from the canvas. Evidence: `tests_build::live_paint_refuses_type_and_images_and_changes_nothing`.
With the Live Paint Bucket selected, Alt-click on vector art samples its fill, stroke and stroke weight into the paint defaults ([Paint with the Live Paint Bucket tool](https://helpx.adobe.com/illustrator/desktop/paint-and-fill/learn-painting-basics/paint-with-the-live-paint-bucket-tool.html)). On a placed image it samples the image pixel as a solid fill with no stroke. Alt+Shift-click samples only the color under the pointer, including an intermediate gradient color or image pixels, into the active fill or stroke ([Default keyboard shortcuts](https://helpx.adobe.com/illustrator/using/default-keyboard-shortcuts.html)). Neither click paints or changes the selection or selected art, and the next Bucket click paints with the sampled paint. Evidence: `tests_build::live_paint_bucket_alt_sample_paints_next_face_without_changing_selection`.
Double-clicking the Live Paint Bucket opens Live Paint Bucket Options ([Paint with the Live Paint Bucket tool](https://helpx.adobe.com/illustrator/desktop/paint-and-fill/learn-painting-basics/paint-with-the-live-paint-bucket-tool.html)): Paint Fills (on) and Paint Strokes (off) set whether a click paints faces, edges or both, Highlight turns the outline of the face or edge under the pointer on or off, and Color and Width set it (red and 4 px by default). Faces are outlined at the Width and edges at half of it, as the page says faces are highlighted with a thick line and edges with a thin line. With both on, a click paints the edge within 4 px of the pointer, else the face, and Shift switches to Paint Fills only; with one on, Shift switches to the other. The options last across tool switches and are saved with the preferences (`tool.setOption {tool: "livePaintBucket"}`). Evidence: `builder::tests::{bucket_options_choose_what_a_click_paints, bucket_options_set_the_highlight, bucket_highlights_an_edge_thinner_than_a_face}`, `dialogs::live_paint_bucket::tests::double_clicking_the_tool_opens_its_options_and_ok_sets_them`.
Live Paint Bucket double-click and triple-click, as [the help page](https://helpx.adobe.com/illustrator/desktop/paint-and-fill/learn-painting-basics/paint-with-the-live-paint-bucket-tool.html) describes them: a double-click on a face fills it and the faces reached across edges that draw no stroke, and on an edge it paints the edges joined to it with the same stroke; a triple-click fills every face of the group with the face's fill, or paints every edge with the edge's stroke (`livePaint.fill` and `livePaint.strokeEdge` `{scope: "connected" | "samePaint", match?}`; `pointer_gesture` and `ui.pointer` take `tripleclick`). Each click paints as it arrives, as its own undo step, and the paint compared is the one the face or edge had before the first of its clicks that painted it; a first click that made the Live Paint group counts when the face it filled is the clicked face. With Space held for the view tools, a click, double-click or triple-click paints nothing and no face or edge is outlined, and with Cmd/Ctrl held a double-click or triple-click paints nothing. Which faces and edges join and which paints match are VectorCraft's own rules. Evidence: `tests_build::{live_paint_bucket_double_click_floods_across_unstroked_edges, live_paint_bucket_double_click_floods_curved_and_narrow_faces, live_paint_bucket_double_click_floods_across_hidden_edges_and_past_gaps, live_paint_bucket_triple_click_fills_every_face_of_the_same_fill, live_paint_bucket_triple_click_after_making_the_group_matches_the_made_fill, live_paint_bucket_double_and_triple_click_paint_joined_and_same_stroke_edges, live_paint_scope_and_match_are_checked}`, `builder::tests::{bucket_double_and_triple_click_widen_what_a_click_paints, bucket_triple_click_after_making_the_group_on_another_face_matches_its_own_face, bucket_triple_click_with_a_click_on_another_face_between_matches_the_first_fill, bucket_highlights_nothing_with_space_held}`, `tests::build::{live_paint_scoped_fill_and_stroke_skip_parts_with_the_paint, live_paint_bucket_cmd_double_and_triple_click_paint_nothing}`, `tests::livepaint::{a_triple_click_fills_every_face_of_the_same_fill, clicks_with_space_held_paint_nothing}` (MCP).
Live Paint Bucket Options has Cursor Swatch Preview, as the [help page](https://helpx.adobe.com/illustrator/desktop/paint-and-fill/learn-painting-basics/paint-with-the-live-paint-bucket-tool.html) lists: with the Bucket selected, the active fill or stroke shows above the pointer between the color swatches before and after it in the Swatches panel, and the Left and Right arrow keys pick the swatch before or after it, as the page describes. The selected art keeps its paint. As VectorCraft's own rules, the option is on at first, the arrows step the stroke when the Stroke proxy is in front, they wrap around at the ends and still step with the preview off, a color that is no swatch shows alone until Right picks the first swatch or Left the last, and the arrows do nothing during a fill drag or with no color swatch. Evidence: `tests::build::{live_paint_bucket_arrow_keys_step_the_active_paint_through_the_swatches, live_paint_bucket_arrow_keys_handle_unmatched_paint_few_swatches_and_drags, live_paint_bucket_arrow_keys_pass_swatches_of_the_same_color_for_fill_and_stroke, swatch_arrow_keys_leave_modified_arrows_to_the_shortcuts}`.
Object › Live Paint › Make turns the selected paths into a Live Paint group of faces and edges ([the help page](https://helpx.adobe.com/illustrator/desktop/paint-and-fill/learn-painting-basics/create-live-paint-groups.html)). As VectorCraft's own rule, each face takes every fill row that draws (shown, with paint and an opacity above 0) of its front-most covering source that has one and an opacity above 0, and each edge every stroke row that draws (also with a weight above 0) of its source, in order with their opacity, blending mode, effects and stroke settings, with gradients kept at their fit on the source and with the source's opacity and blending mode. Where a path is cut into several edges, each edge's dashes go on from where it starts along the path with the lengths set (dashes aligned to corners and path ends are not fitted to the cut edges), the arrowheads stay at the path's own ends, and a width profile or brush stays only on an edge that runs the whole path; each edge runs the way its path does, a gradient applied along a stroke starts again on each cut edge, and an inside or outside stroke is drawn centered on an open piece of a closed path. A painted lower row under an unpainted or hidden top row is kept, a hidden top row does not become visible, and a source in front whose fill does not draw does not hide the fill behind it. Rows that do not draw stay with the stored sources, which Release returns unchanged; the source's object effects are not carried. Evidence: `tests::build::{live_paint_make_keeps_lower_drawn_stroke_rows, live_paint_make_fits_dashes_and_arrowheads_to_each_edge, live_paint_make_starts_a_closed_edge_where_its_path_starts, live_paint_make_keeps_the_direction_of_gradient_strokes, live_paint_make_keeps_all_drawn_fill_rows_and_attributes, live_paint_make_uses_frontmost_source_with_drawn_fill}`.
Live Paint Make preserves the front-most covering source's fill when compound paths share a curved boundary, and Expand keeps the face fills. Regression evidence: `planar::tests::coincident_curves_keep_their_source_faces` checks source order, winding, fill rules and nearby boundaries; `tests::build::live_paint_make_and_expand_keep_abutting_compound_colors` checks native compound paths, Make, Expand and Undo/Redo. This preserves VectorCraft's source-fill rule without changing arrangement precision.
Object › Live Paint › Expand keeps the faces and edges that draw on the canvas, with all their fill and stroke rows, so the expanded paths look as the painted group did ([Create Live Paint groups](https://helpx.adobe.com/illustrator/desktop/paint-and-fill/learn-painting-basics/create-live-paint-groups.html)). As VectorCraft's own rule, a face or edge draws when it is shown, its opacity is above 0, and any of its fill or stroke rows is shown, has paint and an opacity above 0 (a stroke also needs a weight above 0), whichever row is on top: a painted lower row under an unpainted top row is kept, and a part whose only painted row is hidden is removed. Merge copies the fill and stroke rows, visibility, opacity and blending mode of each face that draws by this rule, and of each edge, to the rebuilt parts, with their gradients kept at their fit on the old outline and without the strokes' pen pressure, tilt, bearing and rotation. Evidence: `tests::build::{live_paint_expand_keeps_the_faces_and_edges_that_draw, live_paint_merge_carries_the_faces_that_draw, live_paint_merge_leaves_pen_pressure_behind, live_paint_merge_keeps_the_fit_of_carried_gradients, live_paint_merge_keeps_the_visibility_opacity_and_blending_of_carried_parts}`.
`livePaint.info` lists every fill and stroke row of each face and edge, bottom first, with its paint, opacity, visibility, weight for a stroke, and whether it draws by the rule above, beside the top fill, stroke and weight it already reported. Evidence: `tests::build::live_paint_info_lists_every_appearance_row`.
@@ -1762,7 +1771,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
A `.ai` opened from its editing data opens its Objects on Path groups live, with their path, their objects in order along it, Pivot and Rotate, and their art where the file draws it; a group whose spacing has been dragged, or with options an Objects on Path group doesn't hold, opens as a plain group (`crates/eps/src/import/ai/tests.rs`).
`.ai` symbol definitions and placed instances open as live symbols, in the Symbols panel's order; an instance of a symbol that isn't read, or whose appearance differs from its symbol's, comes in as the page draws it (`crates/eps/src/import/ai/tests.rs`).
In a `.ai` opened from its editing data, an object in a pattern's art or a brush's art with an art style VectorCraft reads opens as itself with its effects live, as an object on a layer does, and an object in a pattern's art keeps the id of a graphic style in the file's list (`crates/eps/src/import/ai/tests.rs`).
A graph in a `.ai` whose records have the shape and values the published version 7 specification gives ([the specification](https://13thmonkey.org/documentation/fonts/AI7FileFormat.pdf), printed pp. 109–112 and 146–158) opens as a live grouped column, stacked column, line or area graph with its saved art, table, axes and settings, and Graph Data edits it ([the help page](https://helpx.adobe.com/illustrator/desktop/automate-visualize-data/visualize-data/add-graph-data.html)); other graphs keep their art as plain objects with a note, and saving a graph to `.ai` warns that its data and settings aren't written (`crates/eps/src/import/ai/tests.rs`, `crates/engine/src/cmd/graph/tests.rs`).
A graph in a `.ai` whose records have the shape and values the published version 7 specification gives ([the specification](https://13thmonkey.org/documentation/fonts/AI7FileFormat.pdf), printed pp. 109–112 and 146–158) opens as a live grouped column, stacked column, line or area graph with its saved art, table, axes and settings, and Graph Data edits it ([the help page](https://helpx.adobe.com/illustrator/desktop/automate-visualize-data/visualize-data/add-graph-data.html)); other graphs keep their art as plain objects with a note. Saving to `.ai` writes these records for a graph that reads back as the same graph, and warns that any other graph's data and settings aren't written (`crates/eps/src/import/ai/tests.rs`, `crates/engine/src/cmd/graph/tests.rs`).
Blends and envelopes in the definitions of a `.ai`'s symbols, patterns and brushes' art count against the same per-file limits on blend layout and envelope checks as the layers' art, so a file with many definitions does not get those limits again for each one (`crates/eps/src/import/ai/read/directives.rs`).
A `.ai` or EPS whose editing data has brushes opens with the file's Calligraphic, Scatter, Art and Bristle brushes in the Brushes panel in place of the default ones, in the file's order: the Calligraphic size, roundness and angle with their variation, the Scatter size, spacing, scatter and rotation, the Bristle settings, and the Scatter and Art brushes' art (Art brushes in the older and newer forms of their objects). A brush library `.ai` (brushes and no layers) opens from its PDF part with its brushes, so Other Library lists them. An Art brush in None or Tints keeps that colorization, and one in the newer form also keeps its fixed width, Flip Along and Flip Across, and Scale Proportionately, Stretch to Fit Stroke Length or Stretch Between Guides with its guides ([Art brush options](https://web.archive.org/web/20220911143322/https://helpx.adobe.com/pdf/cs6/illustrator_reference.pdf), printed page 193). Pattern brushes and brushes that can't be read are left out with a note. Paths with a Calligraphic brush, and paths with an Art brush in the newer form that bends art of solid-filled paths with those options, open as the path with the brush on its stroke, except paths whose brush scales its art proportionately. Paths with an Art brush in the older form, and other brushed paths, still come in as their drawn look (part of #372 and #637; `crates/eps/src/import/ai/tests.rs`, `crates/eps/src/import/tests_native.rs`).
In a `.ai` opened from its editing data, a path or type object with Offset Path, Outline Object or Outline Stroke on the whole object opens as itself with those effects live, in their order and shown or hidden; on a fill or stroke of type they stay on that fill or stroke. Offset Path keeps its Offset, Joins and Miter Limit, and Outline Stroke takes its width from the stroke. An Offset Path with a setting missing, out of range or not read keeps the drawn art, as does a path with a Path effect on one fill or stroke (`crates/eps/src/import/ai/style/tests.rs`, `crates/eps/src/import/ai/tests.rs`, `crates/eps/src/import/tests_native.rs`).
@@ -1812,6 +1821,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- Targa files (`.tga`, `.vda`, `.icb`, `.vst`) open, place linked or embedded, relink and update at their pixel size: color-mapped, true-color and black-and-white, raw or run-length encoded, in any of the four row orders. The image descriptor's attribute bits and the extension area's Attributes Type decide the transparency: a file with no attribute bits, or with attribute data the Attributes Type calls undefined, opens opaque; with one bit, a 16-bit true-color pixel takes that bit as its alpha, and a color-mapped pixel takes the bit of the 16-bit color map entry it indexes; premultiplied colors are divided by their alpha. Open and Place recognize a Targa file by its extension, and `clipboard.importImage` and Glass's Load Texture… by its header; a Targa file copied in a file manager pastes as a bitmap (`crates/engine/src/cmd/fileio/tests_tiffbmp.rs`, `crates/engine/src/tests/links.rs`).
- BMP files named `.dib` or `.rle` (run-length encoded bitmaps included), Photoshop files named `.pdd` and PNG files named `.pns` open and place, linked or embedded, as BMP, PSD and PNG files, in any letter case: the Open and Place dialogs list them with their format, the Links panel names their format, and Edit Original opens them. Exports write only `.bmp`, `.psd` and `.png` (`crates/engine/src/cmd/fileio/tests_tiffbmp.rs`, `crates/engine/src/cmd/fileio/tests_psdread.rs`, `crates/engine/src/tests/links.rs`).
- File › Open converts a PSD or PSB file's layers to objects, keeping the layer structure, merging clipped layers into one layer and leaving out hidden layers that would need a merge, as [PSD import options](https://helpx.adobe.com/illustrator/desktop/add-and-import-files/import-from-other-apps/photoshop-import-options.html) describes for the option that converts layers to objects; hidden layers come in with `hiddenLayers`, as Import Hidden Layers brings them in, and `layers: false` opens the merged image. VectorCraft's own rules: each layer becomes a layer holding its pixels with its name, visibility, opacity and blend mode, groups become layers of sublayers, layers in Dissolve, Linear Burn, Linear Dodge, Vivid Light, Linear Light, Pin Light, Hard Mix, Subtract, Divide, Darker Color and Lighter Color are merged with the layers below, layer masks become opacity masks, fill opacity multiplies the opacity (in a layer that others are clipped to, it fades only that layer's own pixels), and adjustment layers, layer effects, knockouts and vector masks aren't applied and are counted in the warnings. 8-, 16- and 32-bit files, PSB, raw, PackBits and zipped channels. There is no PSD import options dialog yet (`crates/engine/src/cmd/fileio/tests_psdlayers.rs`).
- A PSD or PSB file's clipping path, the saved path that image resource 2999 names in the [PSD file format specification](https://www.adobe.com/devnet-apps/photoshop/fileformatashtml/), clips its merged image on File › Place and on File › Open of the merged image, as [PSD import options](https://helpx.adobe.com/illustrator/desktop/add-and-import-files/import-from-other-apps/photoshop-import-options.html) describes for a file imported as one flattened image. The image comes in a clip group whose clipping path has no paint and uses the even-odd rule, and the path can be edited or released. `links.update` and `links.relink` keep the clip group, and files that name no clipping path come in as before (`crates/engine/src/cmd/fileio/tests_psdread.rs`).
- PCX files (`.pcx`) open and place, from the ZSoft PCX Technical Reference Manual: 1 bit in the header palette's first two colors (black and white when the version has no palette or both colors are the same), 1 bit in two to four planes and 4 bits with up to 16 colors from the header's palette (the EGA colors for versions 0 and 3), 2 bits in the CGA colors that the header's CGA palette settings pick (background, color burst, palette and intensity), 8 bits with the 256-color palette after the image data (gray levels without one) and 24-bit color in three planes, run-length encoded. The header's HDpi and VDpi size the image. A file of another depth or plane count, one cut short, and one whose pixels as RGBA pass the pixel allowance are refused with a message before its pixels are decoded. Open and Place recognize a PCX file by its extension, and `clipboard.importImage` and Glass's Load Texture… by its header; a PCX file copied in a file manager pastes as a bitmap (`crates/engine/src/cmd/fileio/tests_pcxread.rs`).
- Pixar picture files (`.pxr`) open, place and paste, from "The Format of Stored Pictures" in the Pixar Image Computer manuals: any set of the R, G, B and A channels in 8 or 12 bits, dumped or run-length encoded with all four packet kinds and block ends, in tiles of any size (pixels of border tiles past the picture are dropped). R alone is gray, a missing color channel is 0 and a missing A opaque; 12-bit samples are clamped to 0.0 to 1.0; color matted to black becomes straight color; a null tile is transparent when the picture has A and black otherwise. The offsets aren't read, and the image takes the default resolution. Version 2 headers read like version 0. A file cut short, one with a tile whose writing wasn't finished, one with header values or packets the format doesn't define, and one whose pixels as RGBA pass the pixel allowance are refused with a message. Every reader recognizes a Pixar file by its magic number. Tested on synthetic files only (`crates/engine/src/cmd/fileio/tests_pxrread.rs`).
- JPEG 2000 JP2 and raw codestream images open and place from `.jpf`, `.jpx`, `.jp2`, `.j2k`, `.j2c` and `.jpc`. Each axis uses the valid display resolution (`resd`), else the valid capture resolution (`resc`), whichever box comes first ([ITU-T T.800](https://www.itu.int/rec/T-REC-T.800/en), I.5.3.7.2); an axis with none uses 72 ppi. Multiple codestreams, JPX composition or layers, and unsupported color spaces are refused with a message. The decoder checks image size before allocating (`crates/engine/src/cmd/fileio/tests_jp2read.rs`).
@@ -1824,6 +1834,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- EPS and DXF in and out, EMF/WMF in and out; PNG (with PNG-8), JPEG, WebP, GIF, TIFF, BMP, Targa and layered PSD export; grayscale TIFFs with an alpha channel or other extra channels place, relink and open, the alpha as the image's transparency (#1466 in part); PSD and PSB files open and place, linked or embedded, as their merged image in every colour mode and depth (#918). EPS import runs a PostScript Level 3 interpreter written from the PostScript Language Reference: resource categories and `resourceforall`, local and global VM (`currentglobal`/`setglobal`/`gcheck`), `clipsave`/`cliprestore`, subarrays and substrings that share storage, the standard dictionaries in `systemdict`, `dictstack`, `internaldict`, executable filters and strings, `load` by any key, `//name` immediate names, `token`, `execform`, user paths, `cshow`, `rootfont`, `cachestatus`, colour rendering and halftone queries, file `status` and data read with `flushfile`, and `pdfmark` marks dropped (Discord feedback, #466, #469, #474). Masked images (`ImageType 3` in its three interleavings) and reusable streams are read, and an image's several data sources are read in turn, as the reference defines (procedures that read one file get their own rows: such images came in banded, #498). `restore` puts dictionaries back as they were at the `save`, so programs that count their saves in a dictionary read to their end (part of #505). Tiling patterns come in as pattern swatches, uncoloured ones too; Type 3 fonts' glyphs as their outlines, spaced by `ashow`/`widthshow`/`xshow`/`kshow`; mesh and function-based shadings (types 1, 4–7) as gradient meshes; sampled functions (type 0) and Flate/LZW predictors are read. It is tested with programs written from the specification in the ways common generators use these features; Inkscape/cairo EPS files open as vectors. When a file's PostScript still can't be read, it opens as its TIFF preview, its Windows metafile preview (as vectors) or its EPSI preview, and the warning (in the status bar, `document.open`/`app.open`/MCP `open_file` replies and `vectorcraft-cli info`) names the error, the operator that raised it and the procedures it ran in. Legacy PostScript `.ai` files (Illustrator 3–8, written with the prolog that defines their operators) keep their groups and nested groups (`u` … `U`), between and inside clipping groups; EPS and EMF/WMF import nest clips (and groups) at most 128 deep, so a hostile file can't exhaust the stack. DXF TEXT justified Fit or Aligned spans its two points (stretched, or scaled as a whole). PDF export writes type as real, searchable text in embedded subset fonts by default, as Illustrator's PDF presets do (Convert Text to Outlines is an option; printer's marks stay shapes), so text-heavy PDFs stay small (#733). DXF export writes each fill as its outlines followed by its hatch, so laser and cutter software, which skips hatches, gets filled art and type too (#546).
- TIFFs that store each channel as its own plane (PlanarConfiguration 2) place, relink and open as the same image stored pixel by pixel does, 8 and 16 bits: grayscale with an alpha channel or other extra channels, and CMYK with or without them. A planar CMYK TIFF keeps its ink amounts and is stored as a CMYK TIFF with each pixel's inks together, at the file's resolution. A TIFF whose header claims more strips or tiles than can be counted is refused as too large to draw (#1466 in part; `crates/engine/src/cmd/fileio/tests.rs`).
- A 16-bit CMYK or grayscale TIFF with an alpha channel is converted from its 16-bit samples, and only the resulting color and alpha are rounded to 8 bits; associated alpha is divided out at 16 bits, and a pixel whose associated alpha is zero has no color. [TIFF 6.0](https://printtechnologies.org/standards/files/tiff-v6.pdf) gives extra samples the depth in BitsPerSample, and its Associated Alpha Handling section states the arithmetic and the zero-alpha rule for RGB data; applying them to CMYK and gray samples is VectorCraft's own rule. Each sample was cut to its high byte first, so a faint ink or a low alpha could be lost (`crates/doc/src/cmyk.rs`, `crates/engine/src/cmd/fileio/tests.rs`).
- A CMYK or grayscale TIFF whose alpha channel comes after another extra channel places and opens with that alpha, 8 and 16 bits, chunky or planar, and the other extra channels are left out. A channel that the file does not mark as alpha is not read as transparency, even when its ExtraSamples tag has fewer values than the file has extra channels. A CMYK TIFF with two or more extra channels past its ExtraSamples values, or with no ExtraSamples tag, opens opaque when none of them is marked as alpha, and a white-is-zero grayscale TIFF with an alpha channel opens too. A document saved with such a CMYK TIFF as a CMYK image keeps it as one, with its ink amounts. Such a TIFF is read when its samples take more than 256 MiB, as long as its samples, the copy a planar file needs, the row the decoder reads each strip or tile into and the RGBA pixels together fit the decoding allowance; a TIFF whose SamplesPerPixel is more than 64 is not read, which is VectorCraft's own limit, and both checks run on the file's tags before any decoding. [TIFF 6.0](https://printtechnologies.org/standards/files/tiff-v6.pdf) gives each extra channel its own ExtraSamples value, in order (pp. 31–32), and tells readers to skip extra channels they do not use (p. 27); reading the first channel marked as alpha is VectorCraft's own rule (`crates/doc/src/cmyk.rs`, `crates/engine/src/cmd/fileio/tests.rs`).
- 1-bit TIFFs compressed with CCITT Group 3 (ITU-T T.4) place, relink and open: Compression 2 (modified Huffman) and Compression 3 with one-dimensional rows or, with T4Options bit 0, two-dimensional ones, with or without fill bits, in strips or tiles, with either FillOrder and either polarity. Rows in T.4's uncompressed mode are not read. A file whose pixels would take more than the decoding allowance as RGBA is refused as too large to draw (#1466 in part; `crates/engine/src/cmd/fileio/tests.rs`).
- EPS and PostScript import keeps a CMYK image's ink amounts, as PDF import does: the samples of `image` and `colorimage` in a CMYK color space, and of a CMYK JPEG read through `DCTDecode`, come in as a CMYK TIFF, read through the image's Decode array, and the pixels a `MaskColor` or `ImageType 3` mask hides become a grayscale opacity mask (`crates/eps/src/import/tests_images.rs`). Such images came in as RGB PNGs.
- EPS and PostScript import reads image data through `CCITTFaxDecode`: Group 4 (ITU-T T.6) and Group 3 one-dimensional and mixed (ITU-T T.4) data, with the filter's `K`, `Columns`, `Rows`, `EndOfBlock`, `EndOfLine`, `EncodedByteAlign` and `BlackIs1` parameters, decodes to 1-bit rows for `image` and `imagemask`. Damaged data keeps the rows decoded before it (damage in the first row is a PostScript `ioerror`), and parameters out of range or past the 256 MiB limit on decoded data are refused (`crates/eps/src/import/tests_images.rs`). Reading such a file stopped at an error naming the filter, and the file opened from a fallback, such as its preview or the art drawn before the image, or did not open.
@@ -1865,7 +1876,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- SVG import shows a raster `<image>` whose EXIF orientation or TIFF Orientation tag turns or flips it upright, linked or as a data URL, and inside an SVG shown as an image: the pixels are stored turned, the width and height swap, and the `<image>` box fits the turned image under `preserveAspectRatio`, as CSS `image-orientation: from-image` describes (<https://www.w3.org/TR/css-images-3/#the-image-orientation>). A turned CMYK JPEG or TIFF keeps its inks as a CMYK TIFF, as does a CMYK TIFF that is not turned. TIFF data URLs (`image/tiff`, or untyped TIFF bytes) are read, and TIFFs with an alpha or other extra channel, CCITT Group 3 TIFFs and CMYK TIFFs stored one plane per ink are read as Open and Place read them. The pixels are decoded within the decode allowance, and a TIFF claiming more strips or tiles than the decoder can count is refused before it is decoded (`crates/svg/tests/it/image_orientation.rs`).
- An image that SVG import stores again, as PNG or as a CMYK TIFF, keeps the resolution its PNG, JPEG, TIFF, WebP or BMP file declares, as Open and Place keep it: in a PNG `pHYs` chunk (<https://www.w3.org/TR/png-3/#11pHYs>) or the TIFF's resolution tags, with the horizontal and vertical values swapped when the image turns a quarter. An image that declares no resolution is stored without one (`a_turned_raster_keeps_its_resolution`, `an_upright_tiff_keeps_its_resolution` and `a_cmyk_tiff_keeps_its_resolution_turned_or_not` in `crates/svg/tests/it/image_orientation.rs`).
- SVG import leaves out an SVG shown as an image that it can't read (text that isn't XML, an SVG without a size) and goes on with the images after it, as SVG 2's error processing describes (<https://www.w3.org/TR/SVG2/conform.html>). Only images that decode count toward the 256 nested SVG images an SVG may show, so an image after many broken ones is imported. Every nested SVG image read, decoded or not, counts toward VectorCraft's own limits of 1,024 reads and 512 MB of text per import; past them, nested SVG images are left out with a warning (`crates/svg/tests/it/import_details.rs`).
- SVG import measures what an SVG expands to before the SVG reader builds it: every copy a `<use>` shows (<https://www.w3.org/TR/SVG2/struct.html#UseElement>), chains of `<use>`s included, and the style values each copy gets from its `style` attribute and from the style sheet rules that match it. An SVG that expands to more than 999,999 elements (500,000 on the web) or to more than 64 MB of copied style values is refused with an error, and so is a loop of `<use>` references that would be copied without end; both limits are VectorCraft's own. An SVG shown as an image counts toward the same 64 MB as the SVG showing it, and one past either limit refuses the whole import (`crates/svg/src/import/budget.rs`, `crates/svg/tests/it/import_details.rs`).
- SVG import measures what an SVG expands to before the SVG reader builds it: every copy a `<use>` shows (<https://www.w3.org/TR/SVG2/struct.html#UseElement>), chains of `<use>`s included, and the style values each copy gets from its `style` attribute and from the style sheet rules that match it. An SVG that expands to more than 999,999 elements (500,000 on the web) or to more than 64 MB of copied style values is refused with an error, and so is a loop of `<use>` references that would be copied without end. The declarations the import's own style cascade keeps count toward the 64 MB too, once for each element a rule matches, including those of rules the reader drops, such as ones with `:not()`. Matching the style sheets' selectors against those elements may take at most 50,000,000 tests; past that the SVG is refused, so a selector that never matches can't keep the import busy. These limits are VectorCraft's own. An SVG shown as an image, from a data URL or a linked SVG file, counts toward the same limits as the SVG showing it, and one past any of them refuses the whole import (`crates/svg/src/import/budget.rs`, `crates/svg/tests/it/import_details.rs`).
- SVG import reads the files that a linked SVG file's own `<image>` elements link to, relative paths from that file's folder (<https://www.rfc-editor.org/rfc/rfc3986#section-5.1.3>), as it reads the imported SVG's links: a raster stays linked, an SVG file imports as vector art, and one that can't be read gets the placeholder, its link and a warning. SVG 2 tells a viewer not to load these links (<https://www.w3.org/TR/SVG2/conform.html#secure-static-mode>); reading them is VectorCraft's own rule. The same href in SVG files of two folders names two files. A link that leads back to an SVG file showing it gets the placeholder and a warning, and reading the links of linked SVG files stops at the nesting depth cap and at 1,024 reads and 512 MB per import, the same limits decoded nested SVG images have, counted separately from them. The text of every linked SVG file read counts, and so does every file read for the links of a linked SVG file, raster or SVG; rasters the imported SVG links to don't. Once either limit is passed, no file linked from a linked SVG file is read and its image is left out with a warning, and an SVG file the imported SVG links to is still shown, without the files its own links name, with a warning. Another href to an SVG file already read, spelled differently or in another SVG file, does not read it again, unless a link back or the depth cap changes what it shows there (`crates/svg/tests/it/import_details.rs`, `crates/svg/src/import/files.rs`).
- File › Place and a file dropped on the canvas refuse a PNG, JPEG, GIF or WebP file whose pixels can't be read, such as one cut short by a download or a sync still running, and the status bar shows "Couldn't place <name>: the image's pixels can't be read". Paste and Relink refuse such an image too, and Update Links, or Open when it reads the file again, counts it as missing. A JPEG cut short still places when its pixels can be decoded (`file.place`, `crates/engine/src/tests_place.rs`).
- On the desktop, Copy and Cut make the system clipboard's formats (text, SVG, PDF, PNG) from a copy of the clipboard on the thread that writes background saves and exports, and write them when they are ready; while that thread has a save, an export, a print or a recovery copy to finish, they are made and written at once. Until they are written, Paste in VectorCraft takes the internal clipboard, and while a text field or the Type tool has the keyboard the app waits for them. Another Copy or Cut, or a copy of text in VectorCraft, drops them.
@@ -1877,6 +1888,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- With Enable Background Export on, an Export for Screens with no folder of its own also writes in the background into the folder its `exportPreset` keeps, with the same progress dialog and Cancel, and the status bar names that folder. When that folder is outside an MCP client's read roots, the reply, its errors and the status bar name it by its own name, as for an export made at once (`dialogs::tests_screens::an_export_into_a_presets_folder_writes_in_the_background`, `fileio::tests_screens::a_prepared_export_into_a_presets_folder_outside_the_read_root_leaves_it_out`). Behavior reference: the help page, https://helpx.adobe.com/illustrator/using/exporting-artwork.html (exports run in the background by default).
- Missing linked files can be looked for in a folder and its subfolders: Find in Folder… in the Linked File Not Found dialog and in the Links panel menu lists the files with each one's name, those with the size and contents saved with the link first, and relinks the files chosen, copying an image's file into the document's folder when asked, without replacing a file there (`links.findFiles`, `links.relink {copy}`). In the Links panel it looks for the selected linked files when any are selected.
- The Links panel's Relink… opens a Relink dialog whose Replace menu picks which links take the new file: Selected Instances (the selected images, as before), All Instances (every image linked to the same file), Missing Links from This Folder (then every other missing link whose file name is in the new file's folder) and Each Link Separately (a file for each selected link in turn). Each relinks as one undo step (`links.relink {scope}`, `ui.relinkDialog`). Relink to Folder… stays as it is.
- The Links panel menu's Copy Info ([Links panel](https://helpx.adobe.com/illustrator/desktop/add-and-import-files/import-from-other-apps/links-panel-in-illustrator.html)) copies the selected link's file name to the clipboard (`links.copyInfo`). VectorCraft's own rules: with several rows selected it copies one name per line in the order the panel shows them, an embedded image gives the name its row shows, and the item is off with no row selected.
- Linked files inside symbols, graph designs, patterns and opacity masks are saved with their paths relative to the document, as linked files in the layers are, so they are found after the document and its files move to another folder together; on open each of those links takes the path where its file was found. Update Links, and opening with Update Links: Automatically, read a modified file again for that art too, and saving a document refreshes its placed copies inside symbols and patterns of the other open documents. Evidence: `links_outside_the_layers_find_their_files_after_the_folder_moves`, `update_links_reads_files_again_inside_symbols_patterns_and_masks`, `saving_a_document_refreshes_its_placed_copies_inside_symbols_and_patterns`.
- A linked file that is missing when a moved document is saved keeps the folder its relative path names, instead of a relative path worked out from the out-of-date absolute path, also for a `.ai` file another app wrote (such as `src/`), so the file is found once it is back in that folder. Evidence: `a_missing_link_keeps_its_relative_path_when_the_moved_document_is_saved`, `links_of_an_imported_file_are_found_by_their_relative_folder`, `a_missing_link_of_a_moved_document_keeps_its_relative_place`.
- On Windows, a document opened from a mapped network drive finds its links stored with the share's own path (`\\server\share\…`), as the same document opened from the share does; links on other computers stay missing.
@@ -1998,6 +2010,8 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- **Smart Blur modes:** Effect › Blur › Smart Blur… has a Mode menu with Normal, Edge Only and Overlay Edge (`blurMode`: `normal`, `edgeOnly` or `overlayEdge`; an effect saved without it draws as Normal). Edge Only draws the edges in white on black, and Overlay Edge draws them in white over the Normal blur. A pixel is on an edge when one of the pixels the search reads within Radius of it differs from it by more than Threshold in a channel and its channels, alpha included, add up to at least that pixel's. The search reads neighbors as the Normal blur does, kept within Radius, and always the four nearest pixels, so a color step gives a line about as wide as Radius, and at least one pixel wide, on the side with the larger sum (lines on both sides when the sums are equal). The white and the black keep each pixel's coverage. Behavior reference: [Summary of effects](https://helpx.adobe.com/illustrator/using/summary-effects.html).
- **Radial Blur center:** Effect › Blur › Radial Blur… has a Blur Center box. A press or a drag in the box sets the point the spin or zoom starts from, and the arrow keys move it 1% (10% with Shift). Horizontal and Vertical fields beside the box take the same position in percent. The effect stores the point as `centerX` and `centerY`, from 0 to 1 across and down the object's bounds (0.5 each by default, VectorCraft's own form), so it moves and scales with the object. An effect saved without them draws about the object's center as before, and `effect.apply` and `effect.setParams` refuse a value outside 0 to 1. Tests: `radial_blur_turns_about_its_blur_center` in `crates/effects/src/pixel/tests.rs` and `radial_blur_center_is_stored_drawn_and_checked` in `crates/engine/src/cmd/effectcmd/tests.rs`. Behavior reference: [Summary of effects](https://helpx.adobe.com/illustrator/using/summary-effects.html).
- **Dragging objects with raster effects when zoomed in (#2558):** during a drag with a tool on the canvas, the renderer draws filtered content (Gaussian Blur, Feather and the Photoshop-style filters) within 1,000,000 pixels, at a lower resolution where it would need more, unless it can draw the content from an earlier frame's raster. In the desktop app it does the same for drop shadows, outer glows and inner glows. No raster made that way is kept. The frame after the drag renders them as a still frame does, and a dialog's preview, exports and printing render as before. `zoomed_drag_perf` times a drag's frames at 100% to 6400%.
- **Dragging other art over a wide blur (#2558):** an effect raster too large to keep whole (a wide blur, a shadow or glow, or any blur zoomed in) is cropped to what can show in the view, or made at a lower resolution and resampled to the view's pixels, and kept while the view stays the same, so dragging other art over the object draws its effects from the still frame instead of re-rendering them at the drag's lower resolution. A drag stays on the render worker once the worker drew the frame shown, since only that renderer holds the rasters. `tests_viewraster` covers each effect, and `a_drag_over_a_wide_blur_draws_the_blur_from_the_still_frame` the canvas on and off the worker. Not yet: a worker frame shows one display frame after the overlays, and the first frame of an interaction can land on the renderer that has not made the rasters yet (follow-up: render every canvas frame on the worker, waiting briefly for fast ones).
- **Opening a heavy document (#4155):** the canvas renders a document's first frame (when it opens, or its tab is chosen) on the render worker, as it renders the frames of a slow document, instead of on the UI thread, where a heavy one froze the window until it was done. The UI thread waits for it up to 16 ms (`FIRST_FRAME_GRACE`), so a light document's art shows in the same frame as before, without a frame of bare paper; otherwise the document's paper and artboards show, the window stays live, and the art shows when it lands. The web build, which has no worker, renders it as before. Measured on an M3 Pro, one run each, from launch with the file: the file of #4155 kept the UI thread busy for 8.7 s at a stretch, now at most 64 ms (its art lands as before, after about 9 s of rendering); a 50,000-path SVG 1.4 s, now 62 ms. `a_documents_first_frame_renders_on_the_worker` and `the_render_worker_waits_for_a_frame_only_as_long_as_asked` (`crates/ui-egui/src/canvas/tests.rs`).
@@ -2009,6 +2023,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
- **Moving objects with raster effects (#1257):** during a drag or a dialog's preview, an object whose content only moved draws its cached drop shadows, outer glows and filtered content (Gaussian Blur, Feather and the Photoshop-style filters) from the rasters of an earlier frame at the nearest whole pixel, without running the effects again. The renderer first draws the moved content where the old content lies in its raster and compares the pixels. De-Interlace runs again. The frame after the drag runs the effects again and draws what a fresh renderer draws; on the desktop, the canvas renders that frame on the render worker. `texture_drag_perf` times a drag's frames and the frame after it.
- **Glass's Load Texture:** Effect › Distort › Glass… has a Load Texture… button that reads a PNG, JPEG, GIF, WebP, TIFF, BMP, Targa, PSD, PSB, PCX, Pixar, JPEG 2000, HEIC/HEIF or JPEG XL file and adds Loaded Texture to the Texture menu (`textureImage`: `{path}` or `{name?, dataBase64}`, stored in the effect and saved with the document as a grayscale PNG in base64). The texture lies one image pixel per point at Scaling 100%, centered on the object and repeated; each pixel's luma is the surface's height, from 0 for black to 1 pt for white, and a transparent pixel counts as middle gray. Scaling and Invert act on it as on the built-in textures. A Glass saved without `textureImage` draws and saves as before.
- **Rough Pastels' Load Texture:** Effect › Artistic › Rough Pastels… has the same Load Texture… button, and its Texture and Light Direction controls are menus. The loaded image's gray levels are the paper surface (Loaded Texture in the Texture menu, `textureImage` stored with the effect), and Scaling, Relief, Light Direction and Invert act on it as on Brick, Burlap, Canvas and Sandstone (`pixel::tests::rough_pastels_draws_a_loaded_texture_as_its_surface`, `rough_pastels_loads_a_texture_saves_it_and_draws_it`).
- **Bounded effect length inputs (#4638):** Effect dialog fields with defined length ranges (such as Crystallize Cell Size 3..300 pt) clamp immediately on commit and reject invalid inputs visibly with an accent highlight without reverting to sub-minimum values.
- **Color Halftone in raster pixels (#3260):** Max Radius is 4–127 px of the Document Raster Effects resolution, as public documentation gives it, shown in px whatever the document units; documents from before open with their radii in points converted (document version 3), and a file saved for an older app version holds them in points again.
+1
View File
@@ -34,6 +34,7 @@ against Illustrator's public documentation; **estimated** unless noted.
| Preserve font sizes while scaling a group | missing | Separate enhancement; scaling still scales text with the other artwork. |
| OpenType panel, Glyphs panel, alternates | done | Highlight Alternate Glyphs preference (#394, off by default) paints pink behind OpenType substitutions on screen when on (required shaping such as Arabic joining forms and the lam-alef ligature, including a lam-alef older fonts file under `liga`, is not highlighted; an optional feature that changes the ligature is); stylistic sets ss01–ss20, stylistic and titling alternates, proportional and lining figures, superscript and subscript are character-style features (`text.setStyle {features}`, #1607); the OpenType panel's Position menu sets superscript, subscript, numerator or denominator glyphs (`sups`, `subs`, `numr`, `dnom`). The OpenType panel's Stylistic Sets menu lists the selected font's `ss01`–`ss20` features, using a font-supplied name where available and `Set N` otherwise; several sets can be active, and selecting an active one removes it. For a mixed-font selection it uses the first selected character's font, a VectorCraft design choice ([public help](https://helpx.adobe.com/illustrator/desktop/design-with-text/special-characters-glyphs/add-stylistic-sets-to-selected-text.html), [OpenType feature specification](https://learn.microsoft.com/en-us/typography/opentype/spec/features_pt)). The Glyphs panel highlights the character selected with the Type tool (a range's first character, else the one before the caret) and scrolls it into view as the selection moves (#4908; `panels::glyphs` tests); its Show menu lists Unicode ranges and Alternates for Current Selection, the selected character's alternate glyphs from the font's `GSUB` table. A double-click applies the OpenType feature that draws a variant to that character alone, or uses a glyph choice when no supported feature draws it; see the Glyphs panel row below for Entire Font and the joining limits. The glyph-cell pop-up of alternates is not there yet |
| Glyphs panel: alternates as OpenType feature values, unencoded glyphs | done | #5514, as [Replace characters with alternate glyphs](https://helpx.adobe.com/illustrator/desktop/design-with-text/special-characters-glyphs/replace-characters-with-alternate-glyphs.html) describes (Show › Alternates For Current Selection; replace the selected character by double-clicking the glyph). A character-style feature takes a value (`"salt=2"`, the font's alternate number; `ss01`–`ss20`, `swsh`, `titl` and `aalt` too) in `features` and `featureChanges`, in SVG `font-feature-settings` (`"salt" 2`) and in native files as the plain string. Show › Entire Font applies an unencoded glyph the same way when it is an alternate of the selected character in its word, applies one no feature reaches (`A.salt` in a font with no `salt`) to the selected character as a glyph choice (it may not join), and dims the others with the reason in the tooltip (e.g. `yehfarsi.fina.alt2 belongs to ی (U+06CC) at the end of a word; your selection is ي (U+064A)`); an Arabic presentation form picked there inserts its letter so the word joins. Runs that differ only in their features shape as one piece of text, each run's features over its own characters, so a variant on one letter of an Arabic word keeps its neighbours' initial, medial and final forms and the letters still touch. The panel lists the alternates of the glyph the shaper makes for the selected character in its word (a medial letter shows medial variants), and a double-click applies the OpenType feature that draws the variant (`salt=N` or `ssNN`) to that character alone, in one undo step (`text.setRangeStyle`); a variant no feature reaches falls back to a replacement glyph (`text.setGlyphAlternate`), which its tooltip says may not join. Show › Entire Font also lists the glyphs no character maps, with their name or `gid N` and the feature that gives them. The text crate reads each face's `GSUB` once (alternate sets, single substitutions, the feature that reaches a glyph). Not done: alternates reached through contextual or ligature lookups, and features of the font that the style has no tag for (`cv01`–`cv99`, `hist`). Evidence: `tests::glyph_alternates` in the text crate (a fixture font with a medial teh at `salt=1` to `3`, the neighbours' forms, no gap; Kufi Murabba PIIIX when it is installed), `features::tests::feature_values_are_parsed_validated_and_shaped`, `tests::textedit::feature_values_are_set_replaced_and_saved`, the `panels::glyphs` tests. |
| OpenType panel menu: H or V Style, Roman Italics | done | With Show East Asian Options, as [Format East Asian characters](https://helpx.adobe.com/illustrator/using/formatting-asian-characters.html) describes: H or V Style is the character feature `hkna` (kana in the font's horizontal forms, shaped as `vkna` in vertical type) and Roman Italics is `ital` (the font's italic half-width alphanumerics), set with `text.setStyle` / `text.setRangeStyle {features, featureChanges}`, on or off only, and in Character Style Options › OpenType Features ([OpenType feature list](https://learn.microsoft.com/en-us/typography/opentype/spec/featurelist)). |
| Font menu: samples, live preview, filters, favourites, recent families; font matching by any name; Find Font | done | #3515: recently applied families appear first, once each; missing families are omitted and the existing Number of Recent Fonts setting applies; #339; script filter: Japanese, Latin, Thai; the font scan on Linux also reads the folders fontconfig's configuration lists, such as TeX fonts under `/usr/share/texmf/fonts` (#1548, fixes #1529; not verified on Linux); Select Text by Font takes a scope: the document, chosen artboards or chosen layers; Find Font's Find In offers Document, Artboard (the active one) and Layers (the selected ones, else the current one) (#5529, a VectorCraft addition); Replace With Font lists family and style rows, one click setting both, with a search over them; its list scrolls on its own, apart from the document's font list (#4972); wherever the font menu opens, its filters keep the same four rows, and it opens at least wide enough for them and for a family name up to 220 points wide with its sample (up to 60% of the window), each sample filling its row's height (a VectorCraft layout); a filter chosen in the font menu of the Character panel, the Control bar or the Properties panel is on in the others, and the filters are cleared when the app quits ([Font browser overview](https://helpx.adobe.com/illustrator/desktop/design-with-text/fonts-and-scripts/font-options-overview.html)); the font menus of the Glyphs panel and of the dialogs that pick a font share the same filters (a VectorCraft choice). |
| Missing glyphs set in an installed font made for their script (Thai, Lao, Khmer, Myanmar first; then any font that has them) | done | Thai: Leelawadee UI, Tahoma, Thonburi, Sukhumvit Set, Noto Sans Thai, TLWG |
| Full font-name matching (OpenType name ID 4) | done | #1258: loaded/cataloged faces and Missing Fonts searches; localized names and variable default faces; Windows font-service QA pending |
+2 -1
View File
@@ -207,6 +207,7 @@ tools, never checked side by side) isn't captured row by row.
| Ruler guides align to a key object: a click on an object selected with guides makes it the key (one object is enough); Align moves the selected guides to its edge or centre (vertical guides to left, centre or right, horizontal ones to top, centre or bottom) and the key stays put | done | #2143; aligning to the selection or the artboard leaves guides alone |
| Align panel menu › Align to Glyph Bounds › Point Text and Area Text (checked while on): Align and Distribute measure point type or area type by the ink of its glyphs instead of its em box or frame | done | #4231; `alignGlyphBoundsPoint`, `alignGlyphBoundsArea` ([Align text](https://helpx.adobe.com/illustrator/desktop/design-with-text/edit-format-text/align-text.html)); `tests::glyphbounds`, `panels::align::tests::align_to_glyph_bounds_toggles_point_and_area_text`; the Character and Properties panels' Align to Glyph Bounds controls are not done |
| Align and distribute artboards: with the Artboard tool chosen, the Align panel's Align Objects, Distribute Objects and Distribute Spacing buttons, and the Control bar's, move the selected artboards with the art inside them (Move Artwork with Artboard) | done | `artboard.align`, `artboard.distribute`, `artboard.distributeSpacing` ([Align and distribute artboards](https://helpx.adobe.com/illustrator/desktop/create-manage-artboards/organize-manage-artboards/align-and-distribute-artboards.html)); VectorCraft's own rules: Align to Artboard keeps the active artboard in place and sends the Distribute Spacing value, Align to Key Object is off, a locked artboard stays where it is; `cmd::artboardalign::tests`, `panels::align::tests::the_align_buttons_align_the_selected_artboards` |
| Object › Distribute, below Object › Align: Horizontal Distribute Left, Center and Right, Vertical Distribute Top, Center and Bottom, Horizontal Distribute Space and Vertical Distribute Space, each running what a click on its Align panel button runs (Align To, the key object, the Distribute Spacing value and, with the Artboard tool chosen, the selected artboards) | done | The `align.distribute*` commands; the public page names Object › Distribute beside Object › Align for the Align and Distribute commands ([Align and distribute objects](https://helpx.adobe.com/illustrator/desktop/manage-objects/arrange-objects/align-and-distribute-objects.html)); the items and their order are VectorCraft's own, taken from the Align panel; `menus::tests::object_distribute_runs_the_distribute_buttons` |
| A marquee selects the objects whose paths it touches (anchors or segments): drawn inside a filled shape, over a background, it leaves that shape (Preview and Outline) | done | #1103 |
| Click a segment independently of its endpoint anchors; only its direction handles show | done | #1278: hollow endpoint squares, selected edge emphasized; headless egui checked |
| Drag a curved segment to bend it, a straight one to move its anchors | done | |
@@ -268,7 +269,7 @@ tools, never checked side by side) isn't captured row by row.
| Bounding-box handles: Shift proportional, Alt from centre, rotate outside the corners | done | |
| Rotated objects keep a rotated bounding box and their angle; Reset Bounding Box | done | |
| Vector Halftone Preview keeps the selection box and hides selection marks on the temporary dots | done | #6369; Preview off and Cancel restore the original marks. |
| Free Transform, Scale and Reflect along a rotated box | missing | |
| Free Transform, Scale and Reflect along a rotated box | done | The Scale and Reflect tools, dialogs and commands and the Transform and Properties panels' flips act along the turned box until Reset Bounding Box; `basis: "page"` keeps the page's axes |
| Free Transform widget: constrain, free distort, perspective distort; the key-held variants | done | |
| Transform tools: click sets the reference point (snaps to anchors/centres), Alt-click opens the dialog, Alt-release copies | done | with anchors selected with Direct Selection, the tools, the dialogs and Move transform those anchors alone, about the centre of their box by default (#1408; the default point and whole-object copies are our choices) |
| Move, Rotate, Scale, Reflect and Shear dialogs: live Preview; Up/Down step angles and percentages by 1, Shift by 10 | done | #1412; Preview starts off (our choice: the public documentation gives no default), Move's as it was last confirmed; OK is one undo step and Transform Again repeats it |