Merge main into feat/pict-open-place

docs/file-format-parity.md

@@ -271,6 +271,7 @@ Names in the editing data that aren't UTF-8 (layers, objects, artboards, colors,
 | PDF | open, place, save | yes (authored ActualText in editable import, #6165; layers, masks, editable text, security, overprint, File Info from the XMP packet and the document information dictionary; Crop To › Trim takes the bleed box as the document's bleed) | yes (presets, PDF/X-1a/3/4, layers, marks and bleed, ICC, fast web view, 1.3+; objects' URLs as link annotations, not in PDF/X) | fuzzed import; shared guarded writer and windowed reader ([pdf-backend.md](pdf-backend.md)), metadata, incomplete-content and dependency-source regressions; PDF text opens where the file draws it, glyph by glyph (#722: kerning or none, spacing per character, text rise as baseline shift, vertical writing, justified text boxes); alpha soft masks drawn as images import with the mask (#1551); a clip only counts as covering the page when its shape does (#1542); a PDF another app copied pastes as editable vector art (#1499); letter-spaced type on a path keeps its length and its glyph-by-glyph stroke comes in as one text object (#1586, #1585); a crafted name tree with shared nodes no longer takes exponential time on open (#1781); fonts named with hyphens (`Noto-Sans-Bold`) or with a composite font's CMap (`Name-Identity-H`) find the installed family; invisible text (rendering mode 3: a scanned page's recognized text) opens as type with no fill or stroke and saves back as invisible, searchable text (#2675); PDF/X-1a and X-3 flattening in a CMYK document keeps CMYK inks where it rasterizes (#2160); an image mask (`/ImageMask true`) painted in a solid color opens as a grayscale image colorized by that color, a spot color keeping its swatch and tint, so Fill and Recolor Artwork reach it, and the same mask painted again in another color keeps that color (#6313); an image in a Separation space opens as a grayscale image of its tints colorized by its ink at 100 % (a spot swatch, a process plate or Registration), its soft mask as its opacity mask (#6314); a DeviceN image of one colorant opens the same way, and one of process plates only (the ink `None` may be among them) opens in CMYK with each colorant's tint in its plate (#6348)  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. |
 | EPS | open, place, save | yes (own PostScript L3 interpreter, Illustrator EPS editing data structure) | yes (L2/L3, TIFF previews) | CIDFonts the file doesn't carry not drawn, but for those over a predefined CMap (`UniJIS-UCS2-H`, `90ms-RKSJ-H`, `GBK-EUC-H`, `KSCms-UHC-H`, `ETen-B5-H`), set as type in the font of the CIDFont's name, vertical over a `-V` CMap; a name that ends in the row-and-cell CMap `H` or `V` splits off the CMap only when the file defines the CIDFont of that name, or the name has a hyphen and an installed face that covers Japanese has it; name-keyed fonts in a FontSet set as type by name; #505; images encoded with `CCITTFaxDecode` (ITU-T T.4 Group 3 and T.6 Group 4, with `K`, `Columns`, `Rows`, `EndOfBlock`, `EndOfLine`, `EncodedByteAlign` and `BlackIs1`) draw, and damaged data keeps the rows decoded before its first damaged row; Type 42 (TrueType) fonts set type at its size (#1553); an embedded Type 1 font's encrypted part is read for its glyphs, which space its type, outline it and draw glyphs without a character, in plain and composite fonts; a Type 42 font that carries its TrueType font gives each code its character by its glyph name or, for names that say nothing (`/g12`), by the glyph's entry in the font's own character map, spaces its type by its own widths and draws glyphs that stand for no character from its outlines; the show operators' spacing along the baseline (`ashow`, `widthshow`, `xshow`, `yshow`, `xyshow`, and the moves of `kshow`'s procedure, run between each two characters) is kept on live type as tracking, so type in a font that isn't installed keeps the file's line lengths (spacing that moves characters off the baseline, as `yshow`, `xyshow` or a `kshow` procedure can, places each character as its own text object); composite (Type 0) fonts with the 8/8 mapping, as PDF-to-PostScript converters write two-byte fonts, draw a Type 42 descendant's glyphs from the TrueType outlines it carries and an embedded Type 1 descendant's from its charstrings, and set other descendants' characters as type at the composite font's size (other mappings are still set as type in the font of their name); CID-keyed fonts (`composefont`, `FMapType` 9 over a CMap built with the CIDInit operators or `Identity-H`/`Identity-V`; `bfchar` and `bfrange` mappings give a base-font descendant its codes or glyph names) draw the glyphs of the Type 0 CIDFonts (Type 1 charstrings in their `StartData` binary data or `GlyphDirectory`) and Type 2 CIDFonts (TrueType glyphs through `CIDMap`) they carry, and the CID-keyed fonts of a FontSet's Compact Font Format data (Type 2 charstrings); a clip only counts as covering the page when its shape does (#1542); an EPS another app copied pastes as editable vector art (`clipboard.importEps`, #1499); a stitched shading keeps its hard colour changes on import (#1654); sampled functions (type 0) interpolate between their samples, multilinearly or by cubics with `Order 3`, and a sampled axial or radial gradient has a stop at each sample; a gradient the page paints as a row of abutting one-colour strips (as Level 2 writers draw smooth shades) imports as one linear gradient, without the seams the strips leave when each is drawn alone; CMYK images, CMYK JPEGs included, keep their ink amounts on import, as from PDF, and the pixels a `MaskColor` or `ImageType 3` mask hides become a grayscale opacity mask; ruzstd 0.9.1 for compressed editing data (#1807, RUSTSEC-2026-0344) |
 | PostScript `.ps` | open, place | yes (the EPS reader reads its first page; `.epsf` files open and place as EPS) | via Print to PostScript | `tests_epsimport.rs` |
+| DCS 1.0 / 2.0 | place linked | place linked only (one image of the plates; spot plates as spot swatches; the composite's clipping path; each plate on its own ink in Separations Preview and print separations) | no | `crates/eps/src/dcs/tests.rs`, `crates/engine/src/tests/dcs.rs` |
 | SVG / SVGZ | open, place, save | yes | yes (SVG Options, Preserve Editing, symbols, filters, rich text) | strongest format; tests in `crates/svg/tests`; text on a path follows basic shapes as their SVG 2 equivalent paths (#1527); CSS transforms and the `font` shorthand (#1565), percentage lengths in symbols (#1596), Inkscape flowed text as area type (#1568) and Inkscape pages as artboards (#1601) import; an unsupported filter is named in its warning (#1598); an artboard exported on its own (Export As with Use Artboards, Export for Screens, Asset Export) has the artboard's name as `<title>`, Save the document's (#3544); attributes in other applications' XML namespaces (`plotter:pen="2"`) kept on their objects from import to export, as `data-*` attributes are (`crates/svg/tests/it/foreign_attributes.rs`); simple shapes export as `<rect>`, `<circle>`, `<ellipse>`, `<line>`, `<polygon>` and `<polyline>` where they draw the same as the path (#4652, `crates/svg/tests/it/basic_shapes.rs`); the regions of exported masks and filters (opacity masks, inverted masks, knockout groups) cover the art they apply to, and over type keep a region that covers any artwork (#6401, `crates/svg/tests/it/blend_export.rs`); SVG Options › Include Variable Data writes the text variables, data sets and bound type objects, and opening the SVG reads them back (`tests::variables::an_svg_with_variable_data_opens_with_its_variables_data_sets_and_bindings`) |
 | DXF | open, export | yes | yes (R12–2018) | R13–2018 closed SPLINE flags and control-point closure on import verified (#1299); circular cubic paths export as native CIRCLE/ARC where stroke semantics permit; R12 weighted strokes remain POLYLINE while non-circular curves keep the existing fallback (`crates/cad/src/tests.rs`); [Autodesk group 70](https://help.autodesk.com/cloudhelp/2024/ENU/AutoCAD-DXF/files/GUID-E1F884F8-AA90-4864-A215-3182D47A9C74.htm). DXF 2000–2018 LAYOUT objects write the paper image origin as 148/149 (was a stray 158), the shade plot codes (2004+) and the model type flag, and the plot style placeholder carries no subclass marker (#3085; structure test `r2000_and_later_files_have_the_structure_cad_apps_require`, checked with ezdxf's audit; not yet confirmed in Autodesk Fusion). DXF Options' Preset menu and `document.exportDxf`'s `dxfPreset: "cnc"` set the CNC / CAM options (#5500); VCarve QA remains open |
 | 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` |

docs/mcp.md

@@ -786,7 +786,8 @@ art's colours (by default as global swatches the art links to).
 them in the Swatches panel does); give colour group names instead to reorder the groups. Dropping a swatch on art
 in the app runs `paint.setFill` (or `paint.setStroke`, the active proxy) with `ids: [the object under the pointer]`.
 The Swatches panel menu's commands: `swatch.addUsedColors {selection?, global?}`, `swatch.unused` (a query: the
-names Select All Unused selects), `swatch.merge {names}` (the first is kept), `swatch.ungroup {name}` and
+names Select All Unused selects; a spot swatch counts as used while an image placed from a DCS file has a plate of
+its ink), `swatch.merge {names}` (the first is kept), `swatch.ungroup {name}` and
 `swatch.sortByKind`.
 `brush.move {name | names, to?, type?}` reorders brushes among the brushes of their type, as dragging one onto
 another of its type in the Brushes panel does. `to` is an index among the brushes of that type in `brush.list`
@@ -4698,13 +4699,15 @@ an `anchor` or `segment` in the same way. The document does not change and no un
 after the new anchor on its subpath are numbered one higher, and a selected segment that the new anchor splits stays
 selected as its two halves, with the new anchor as an end of both that moves with them. Undo restores the path and
 the selection in one step.
-`path.close {id, subpath?, in?, out?, independent?}` closes an open subpath (by default the last open one). `in` places
+`path.close {id, subpath?, at?, in?, out?, independent?}` closes an open subpath (by default the last open one). `at`
+moves the first anchor to `[x, y]`, its handles with it, before `in` and `out` are placed. `in` places
 the first anchor's incoming handle as `path.setHandle` places a handle: a smooth anchor turns its outgoing handle to
 stay in line and keeps that handle's length, and a corner keeps its outgoing handle where it is. With
 `independent: true` the anchor becomes a corner first, so its outgoing handle does not move. `out`, placed after `in`,
 is the first anchor's outgoing handle, and makes the anchor a corner. A Pen press on the first anchor sends `in` at
 the anchor with `independent: true`. The Pen's and the Smart Pen's closing drags send the mirror of the pointer as
-`in`, and once Alt is held they also send `out` at the outgoing handle's position when Alt went down.
+`in`, and once Alt is held they also send `out` at the outgoing handle's position when Alt went down. Once Space held
+during the Pen's closing drag has moved the first anchor, the drag also sends `at` at the anchor's new position.
 
 `path.reshape {id, x, y, dx, dy, tol?, focal?, radius?}` (a Reshape tool drag) grabs the path at `x`, `y`, adding an
 anchor there unless one lies within `tol` (3 by default), and moves that point by `dx`, `dy`.
@@ -4753,7 +4756,7 @@ an `x` without a `y`, or a `y` without an `x`, with "x and y must be given toget
 `path.blob` refuse a point in `points` whose x or y is not a number, such as `["10", 10]` or a point written as a
 string, with "x and y must both be numbers". All of them except `path.curvature` refuse a pen pressure, tilt or
 rotation that is neither a number nor null with `` `pressure` must be a number, not "0.5" ``, naming the key
-(`pressure`, `tiltX`, `tiltY` or `rotation`). `path.close` refuses an `in` or `out` that is not `[x, y]`
+(`pressure`, `tiltX`, `tiltY` or `rotation`). `path.close` refuses an `at`, `in` or `out` that is not `[x, y]`
 with `` `in` must be [x, y] ``, naming the key.
 
 These commands refuse a string, a number or another value that is not true or false with
@@ -5943,6 +5946,40 @@ of its art, as one clipped group. A PDF, `.ai` or `.ait` page read from `path` (
 `textAs: "outlines"`) is a placed document the same way, showing page `page` clipped to its `crop` box (or its
 art's bounds); with `link: false` it is placed as one clipped group. EPS and PostScript `.ai` files are embedded. `at` centres it, `rect` fits it, `replace`
 swaps the selected object (keeping its place and transform), `template` puts it on a new template layer.
+A DCS file (Desktop Color Separation, version 1.0 or 2.0: an EPS file whose header names one EPS program per printing
+plate, <http://www.telegraphics.net/doc/DCS.txt>) is placed only from its `path` with link, as one linked image
+([Import DCS files](https://helpx.adobe.com/illustrator/desktop/add-and-import-files/import-other-file-types/import-dcs-files.html)).
+`file.place` with `link: false` or from `dataBase64`, `document.open` and `links.embed` fail with "DCS files can only
+be placed linked" and change nothing. The plates are read from the file (`%%PlateFile: (name) EPS #offset size`) or
+from the plate files it names (`%%PlateFile: (name) EPS Local file`, and the `%%CyanPlate:`, `%%MagentaPlate:`,
+`%%YellowPlate:` and `%%BlackPlate:` comments of DCS 1.0), which are looked for by file name in the DCS file's folder.
+Each plate is the one image its program draws, and a dark pixel is ink; a plate program that draws more than one
+image, art besides its image, or its image through a clipping path is an error naming the plate. Each spot plate adds a spot swatch (when
+the file is placed, and when `links.update` or `links.relink` reads it) in the
+CMYK values of the file's `%%CMYKCustomColor:` comment (else `%%RGBCustomColor:`; black, with a warning, when the file
+gives neither). A swatch of the same name that the document already has is kept, with a warning (in `warnings` of
+`file.place`, `links.update` and `links.relink`) when it is not a spot color of the same values; Separations Preview
+and `file.print` separations then show and print the plate in that swatch's color, and a plate whose name is a
+process swatch prints as process inks in that swatch's color. `swatch.unused` does not list a spot or process swatch
+while such an image that the art shows has a plate of its name; deleting the swatch makes the plate print as its CMYK values. The clipping path the
+file's composite is drawn through clips the image (a clipping group, its path under the nonzero fill rule, an even-odd
+clipping path kept as the same area), and `file.place` answers the image's id as
+`image`, since the Links commands take the image and not the group. `links.update` and `links.relink` clip the image
+by the file's clipping path again (in a new clipping group when none holds it, and inside a clipping mask made in the
+document, whose path is kept), unless Placement Options clip the image, the same way for the images in symbols, graph designs, patterns and opacity masks; when the new file has no
+clipping path, a clipping group holding just the image whose path is the old file's clipping path loses that path and
+stays as a plain group of the image, with its id, name, visibility and transparency (one whose path was given the
+even-odd fill rule stays as it is). The link shows as modified when a plate file of its own changes
+or is gone, as it does when the DCS file changes, and the app's link watch acts on a changed plate file as on a
+changed DCS file. `links.relink {copy: true}` fails for a DCS file that reads plate files of its own, and copies
+nothing. These rules are VectorCraft's own: the screen, composite print and exports show the image as its plates
+mixed into CMYK (each ink amount times the ink's CMYK values, plates combined as `1 − (1 − a)(1 − b)`); Separations
+Preview and `file.print` separations show each plate on its own ink, and a spot plate prints as the document's swatch
+of its ink (else its CMYK values) with `spotsToProcess`. Separations Preview shows a reduced image's plates as the
+mean of the plate pixels each shown pixel covers. A file may name 64 plates, a plate file may be 1 GB, and the plates together may hold 512 MB of
+ink amounts; a plate that can't be read is an error naming the plate, and nothing is placed. The plates go on the first
+plate's pixel grid, grown to hold every plate where it prints; a plate of another size or position is resampled onto
+it, and a file naming more than 64 plates is an error. `file.place.info` answers `plates` (the ink names) for a DCS file.
 The art's bounds are the bounds of the art that shows, with its strokes and effects, as an export measures them:
 hidden objects, guides and template layers do not count, and symbol instances are measured by their art. A page or artboard holds the objects
 whose shown art lies on it; a hidden object on it is still placed and stays hidden, as do the hidden members of the groups placed.
@@ -6662,6 +6699,14 @@ allowance has bytes, drawn as a crossed gray box on the canvas and left out of e
 counter-clockwise), `placement` and, for a linked file, `fileName`, `location`, `fileSize`, `created`, `modified`.
 With automation read roots, a file outside them has no `path` or `found` in `links.list` and `links.check`, and no
 `link`, `path`, `location`, `fileSize` or `modified` in `links.info`; its row keeps the file's name.
+An image placed from a DCS file has the `format` `DCS`. `links.list {dcsTransparency: true}` (the panel menu's Show
+DCS Transparency Interactions, off by default) adds `dcsTransparency: true` to the rows of such images that
+[are transparent or overlap transparent objects](https://helpx.adobe.com/illustrator/desktop/add-and-import-files/import-other-file-types/eps-and-dcs-file-import-in-illustrator.html),
+and the panel shows a warning mark on them. VectorCraft's rule for it: the image, or a group or layer holding it, has
+an opacity below 100%, a blend mode other than Normal, an opacity mask or an effect, or its bounds overlap the
+bounds of another visible object that has one of those or a gradient with a transparent stop (in a fill, a stroke or
+a character's fill or stroke); a DCS image with transparency of its own counts as such an object. The check looks at
+every object in the layers each time the list is read. A `dcsTransparency` that is not true or false is an error.
 `links.goTo {id}` selects an image; `links.embed {ids?}` keeps linked files' pixels in the document (an image whose
 file is missing stays linked: relink it first); `links.unembed {id, path}` writes an embedded image to a file
 (another extension converts it) and links to it, and without `path` returns `{name, dataBase64}`.
@@ -6910,7 +6955,11 @@ visible type that fits its frame or path, on the layers `printLayers` picks, on
 that print, and the type in the symbols, patterns, scatter brushes, opacity masks and placed documents of the art that
 prints. The warnings also give the ids of the area type and type on a path that print and whose text doesn't fit
 (objects of the document only, not type in the art of symbols, patterns, brushes, opacity masks or placed documents;
-see [Fitting area type](#fitting-area-type-auto-size-shrink-text-to-fit-overflow)).
+see [Fitting area type](#fitting-area-type-auto-size-shrink-text-to-fit-overflow)). The warnings of `file.print`
+and `print.preview` list what the job leaves out of the document or prints differently first (art outside the
+imageable area, type in a fallback font, transparency flattened for PostScript). The notes on the options follow: an
+option not applied or applied another way, such as halftone screens, a fixed flatness or copies per sheet that don't
+fit. Each kind keeps the order it is given in.
 
 ```json
 {"name":"run_command","arguments":{"command":"print.setup","params":{"settings":{"media":"a4","scaling":"fit","marks":{"trim":true}}}}}

docs/target-app-parity.md

@@ -591,7 +591,7 @@ since 2026-10-01.
 - **Clipped bottom-bar text (#6622):** A Help Bar hint wider than the window, such as the Dimension tool's hint in an 800-point window, ends in an ellipsis, as the status message does. In a narrow status bar the tool name ends in an ellipsis and leaves the status message its room. Resting the pointer on a clipped hint, tool name or message shows the full text in a tooltip that wraps within the window. Text that fits shows no tooltip, and screen readers receive the full text. Scores and estimates are unchanged.
 - **Notes of an open in the status bar:** After File › Open the status message names the first note that says reading the file left something out (the notes that make Save ask before writing over the file), else the first note, and counts them all. With several notes, resting the pointer on the message lists them in that order, as many as fit in the window and at most 30, then how many more. Screen readers receive the list after the message. Scores and estimates are unchanged.
 - **Notes of a save or an export in the status bar:** After Save, Save As, Save a Copy, Save as Template, Save a Timestamped Copy, Save as PDF, Export As, Export Selection, Export for Screens, Asset Export, Save for Web, Preview in Browser, Save for Office Documents, Text Export Options, a print to a file or the CSS Properties panel's export, the status message names the first note and counts them all. With several notes, resting the pointer on the message lists them in the order the reply gives them (a save's notes start with what the format loses), as it lists an open's notes. Scores and estimates are unchanged.
-- **What a save or an export loses, named first:** Each writer marks its notes as a loss (something of the document the file leaves out or writes differently) or a note on the options (an option it didn't apply or applied another way while the art reaches the file as it looks, or something an option turns off). The replies of the save and export commands list the losses first, each kind in the order given, across every file an export writes. The status message after a save or an export then names a loss when there is one: an export to lossy WebP of type in a font that isn't installed names the fallback-font note instead of the lossy WebP note. A print to a file keeps the order its notes are given in. Scores and estimates are unchanged.
+- **What a save or an export loses, named first:** Each writer marks its notes as a loss (something of the document the file leaves out or writes differently) or a note on the options (an option it didn't apply or applied another way while the art reaches the file as it looks, or something an option turns off). The replies of the save and export commands list the losses first, each kind in the order given, across every file an export writes. The status message after a save or an export then names a loss when there is one: an export to lossy WebP of type in a font that isn't installed names the fallback-font note instead of the lossy WebP note. A print to a file and the print preview list theirs the same way: art outside the imageable area, type in a fallback font and transparency flattened for PostScript come before notes such as a fixed flatness that isn't written or copies per sheet that don't fit. Scores and estimates are unchanged.
 
 - **Canvas review follow-up:** Cached canvas dimensions match the renderer limit before cache-key creation, view construction and worker submission. Controlled-worker regression checks oversized width and height, high DPI and the render transform without allocating oversized images. Scores and estimates are unchanged.
 - **Cached canvas frames during viewport changes:** Cached screen textures use their own raster dimensions and pixels-per-point when reprojected. `canvas::tests::cached_frames_keep_their_scale_during_resize_rotation_and_dpi_changes` holds the worker and checks every UV corner against the cached render transform, including isolated viewport-only resize at zero/30-degree rotation, fractional raster dimensions, zoom/pan, changed rotation, monitor-DPI changes in both directions, Pixel Preview and replacement-frame delivery. The contributor rebuilt and confirmed the original resize issue is fixed with their own artwork on 2026-10-19; the subsequent upstream rebase is checked by regressions and full CI. The artwork stays local. Scores and estimates are unchanged.
@@ -1025,7 +1025,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
     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).
     Transform Each opens on the scale, move, rotate, reflect, Random, reference point, Transform Objects and Preview it was last confirmed with (kept with the UI state; Copy never is), and its Copy button, left of Cancel, transforms a copy of each object as the [Scale multiple objects](https://helpx.adobe.com/illustrator/desktop/manage-objects/reshape-transform-objects/scale-multiple-objects.html) help page has it; the Copy checkbox is gone (#1732; the remembered values follow a maintainer comment on the issue, not the help page). Its Scale section has the Relative and Absolute radio buttons the same page describes: percentages of each object's size, or the width and height it becomes (`object.transformEach` `relative`, which Transform Again repeats); Absolute measures each object's own box, and the scales act along its axes about that box's reference point, so a turned shape keeps its shape and objects of different sizes come out alike; the fields show the width and height the selected objects share and read "Mixed" when they differ or a side is 0 (a line's), each object measured by its own box, rotated with it, as the Transform panel and the Control bar show it (sizes they show alike count as the same), and a size typed into them sets every object to it (a side left Mixed keeps its size), our choices where the page leaves them open; a live Repeat scales its spacing by the geometric mean of the two factors, so its Absolute size is approximate. The link beside the two scales keeps them in proportion, as the shape dialogs' Width and Height link does (`__constrain`, kept with the dialog's values; where it sits is our choice, the public pages don't establish it). The panel lists Scale, Move and Rotate, then the Options in two columns (Transform Objects, Transform Patterns, Scale Strokes & Effects, Scale Corners; Reflect X, Reflect Y, Random), the Reference Point, and Preview at the left of the button row, level with Copy, Cancel and OK.
     Transform Again uses the Copy, Transform Patterns, Scale Strokes & Effects and Scale Corners choices of the transform it repeats; after a bounding-box handle drag on area type, it resizes the type area again and the type keeps its size (`object.transformAgain`). With anchors selected with Direct Selection, Move, Rotate, Scale, Reflect, Shear and Transform Again (the commands, their dialogs and the four tools' drags) transform only those anchors, their handles with them, about the centre of the box around them unless a point was clicked; a copy still copies the whole object (#1408; behaviour reference: [adjust path segments](https://helpx.adobe.com/illustrator/using/adjust-path-segments.html); the default reference point and the copy are our choices). The Move, Rotate, Scale, Reflect and Shear dialogs have a Preview checkbox, off when they open: while it is on the canvas follows the values, Cancel rolls it back and OK is still one undo step; Up/Down step the Rotate and Shear angles and the Scale percentages by 1, Shift by 10 (#1412). The five dialogs and Transform Each have Transform Objects beside Transform Patterns: off, the art stays and only its pattern tiles transform (`objects` on the transform commands), and Transform Again repeats the choice (#1448; behaviour reference: [scale objects](https://helpx.adobe.com/illustrator/desktop/manage-objects/reshape-transform-objects/scale-objects.html); the checkboxes are always enabled and the Scale page's pair is applied to all of them, our choices). Reflect's axis is Horizontal, Vertical or an angle (#1460; [reflect objects along an axis](https://helpx.adobe.com/illustrator/desktop/manage-objects/arrange-objects/reflect-objects-along-an-axis.html)).
-  - **Pen:** Alt pressed in the middle of dragging out an anchor splits its handles and keeps the curve already shaped into it (only the outgoing handle follows the pointer); Space held while dragging out an anchor moves it, handles and all. A press on the first anchor of the path being drawn closes it whatever the drag then does: the drag shapes the closing curve and the release closes the path wherever it ends (#501). The first anchor's incoming handle mirrors the pointer, and a click leaves it retracted and the anchor a corner. Its outgoing handle stays where it is on a corner, so the first segment keeps its shape, and on a smooth anchor it turns to stay in line at its own length; Alt makes the anchor a corner and leaves the outgoing handle where it was when Alt went down. The Smart Pen's closing drag does the same (fixes #6634; `path.close`'s `in` and `out`; behavior reference: [draw with the Pen, Curvature, or Pencil tool](https://web.archive.org/web/20220614224656/https://helpx.adobe.com/illustrator/using/drawing-pen-curvature-or-pencil.html), where a click or drag on the first anchor closes the path and Alt breaks the pairing of the closing anchor's handles; which handles the drag sets are our choices). Moving the closing anchor with Space held, which that page also describes, is still missing. Clicking the last anchor of the path being drawn retracts its outgoing handle (convert cursor), and dragging from it pulls out a new one on its own. Auto Add/Delete: over a selected path, a click on a segment adds an anchor and a click on an anchor deletes it (add and delete cursors); Shift held, or Preferences › General › Disable Auto Add/Delete, starts a new path instead. Alt held over a handle end or an anchor of a selected path (the one being drawn too) works as the Anchor Point tool, with its convert cursor: dragging a handle moves it alone, clicking a smooth anchor makes it a corner, dragging an anchor pulls out new symmetric handles, dragging a segment reshapes it; drawing then goes on. Cmd (Ctrl) held drags with the selection tool chosen last, Direct Selection until one is (handles, anchors and segments of the path being drawn or another), and the Pen then goes on drawing its path while that stays selected; the pointer shows that tool's while Cmd is held. The same holds for agents (`pointer_gesture` with `mods.cmd`), as the borrowing now lives in the engine (#494). A drag's preview (the Pen's, the Anchor Point tool's, Direct Selection's handles and anchors) follows every position the pointer is seen at (#834): the canvas sends the tool a drag whenever the pointer's document position differs from the last one it sent, not only when egui's frame delta is non-zero (zero when a compositor coalesces motion and in the frame the pointer comes back into the window after leaving it), and a move that arrives in the release's own frame shapes the preview before the release commits it, once the press has dragged. With the Pen, a press that moves under 3 screen pixels is a click (a corner without handles), and a Precise Cursors command toggles the existing preference and can take a shortcut (#1663).
+  - **Pen:** Alt pressed in the middle of dragging out an anchor splits its handles and keeps the curve already shaped into it (only the outgoing handle follows the pointer); Space held while dragging out an anchor moves it, handles and all. A press on the first anchor of the path being drawn closes it whatever the drag then does: the drag shapes the closing curve and the release closes the path wherever it ends (#501). The first anchor's incoming handle mirrors the pointer, and a click leaves it retracted and the anchor a corner. Its outgoing handle stays where it is on a corner, so the first segment keeps its shape, and on a smooth anchor it turns to stay in line at its own length; Alt makes the anchor a corner and leaves the outgoing handle where it was when Alt went down. The Smart Pen's closing drag does the same (fixes #6634; `path.close`'s `in` and `out`; behavior reference: [draw with the Pen, Curvature, or Pencil tool](https://web.archive.org/web/20220614224656/https://helpx.adobe.com/illustrator/using/drawing-pen-curvature-or-pencil.html), where a click or drag on the first anchor closes the path and Alt breaks the pairing of the closing anchor's handles; which handles the drag sets are our choices). Space held during the Pen's closing drag moves the first anchor, handles and all, as that page also describes (`path.close`'s `at`). Clicking the last anchor of the path being drawn retracts its outgoing handle (convert cursor), and dragging from it pulls out a new one on its own. Auto Add/Delete: over a selected path, a click on a segment adds an anchor and a click on an anchor deletes it (add and delete cursors); Shift held, or Preferences › General › Disable Auto Add/Delete, starts a new path instead. Alt held over a handle end or an anchor of a selected path (the one being drawn too) works as the Anchor Point tool, with its convert cursor: dragging a handle moves it alone, clicking a smooth anchor makes it a corner, dragging an anchor pulls out new symmetric handles, dragging a segment reshapes it; drawing then goes on. Cmd (Ctrl) held drags with the selection tool chosen last, Direct Selection until one is (handles, anchors and segments of the path being drawn or another), and the Pen then goes on drawing its path while that stays selected; the pointer shows that tool's while Cmd is held. The same holds for agents (`pointer_gesture` with `mods.cmd`), as the borrowing now lives in the engine (#494). A drag's preview (the Pen's, the Anchor Point tool's, Direct Selection's handles and anchors) follows every position the pointer is seen at (#834): the canvas sends the tool a drag whenever the pointer's document position differs from the last one it sent, not only when egui's frame delta is non-zero (zero when a compositor coalesces motion and in the frame the pointer comes back into the window after leaving it), and a move that arrives in the release's own frame shapes the preview before the release commits it, once the press has dragged. With the Pen, a press that moves under 3 screen pixels is a click (a corner without handles), and a Precise Cursors command toggles the existing preference and can take a shortcut (#1663).
   - **Direct Selection (#1278):** clicking a segment picks its connection, with hollow endpoint anchors and only that curve's direction handles. Delete leaves the neighboring curves and appearance intact, opens/splits contours in one undo step, and another Delete removes the remaining selected path. A whole-path drag and deliberate anchor selection keep their existing behavior; egui frames and native control/art checked, OS mouse/visual QA pending. Where anchors of overlapping objects coincide, a click or drag takes the anchor of the selected object, also when another object lies over it (#1509, fixes #1508). Delete removes every selected anchor of a closed path (it used to remove the wrong ones), removes objects selected whole along with the selected anchors of others, and keeps the remainder selected so a second Delete removes it (#1674). In isolation mode Direct Selection and Group Selection pick only inside the isolated object (#1666). Escape cancels a Direct Selection drag (#1803, #1743). Double-clicking type (including type nested inside a group) switches to the Type tool and puts the caret at the clicked character.
     With Direct Selection and Group Selection, a drag that starts on an object while the selection also holds direct-selected anchors or segments moves those anchors and the objects selected whole in one undo step (`path.moveAnchors`), as a drag of a selected anchor does, and the other anchors of a partly selected path stay in place. While Alt is held during the drag, it copies the selected objects instead (`object.transform`). When direct-selected anchors are moved or transformed together with objects selected whole (such a drag, an anchor drag, an arrow key, or Move, Rotate, Scale, Reflect and Shear without a copy), a path inside a selected group is transformed once, with the group.
     A path segment with a control point whose x or y is infinite or larger in size than `f64::MAX / 16` is measured by its control points, so its bounds hold the whole curve and the canvas check refuses an edit that leaves such a path selected (`path.create` included).
@@ -1572,6 +1572,7 @@ Moved from the ROADMAP's "Shipped so far" on 2026-10-10. Update it when a featur
   - JPEG XL (`.jxl`) images open and place, linked or embedded, as their first frame, decoded with jxl-oxide and stored as PNG: 16-bit and floating-point samples become 8 bits, alpha is kept, and an image past the pixel allowance is refused from its header (part of #2306; `crates/engine/src/cmd/fileio/tests_jxlread.rs`).
   - The Open and Place dialogs list PostScript `.ps` files and `.epsf` files. A `.ps` file opens and places through the EPS reader (its first page), also when its header is only `%!`, as the `ps` format, which exports don't write (Print to PostScript does); an `.epsf` file opens and places as EPS. Copied in a file manager, these files paste as EPS, and `clipboard.importEps` also reads PostScript whose header is only `%!` (`crates/engine/src/cmd/fileio/tests_epsimport.rs`).
     `document.pdfInfo` reads an EPS, a `.ps` file or a `.ai` in PostScript form as one page the size of its bounding box (a US Letter page without one), and its `thumbnail: 1` is a PNG of what that page prints (`crates/engine/src/cmd/fileio/tests_pdfimport.rs`).
+  - DCS 1.0 and 2.0 files (an EPS file with one program per printing plate) place linked as one image of their plates, and Open, Embed and Place without Link refuse them ([Import DCS files](https://helpx.adobe.com/illustrator/desktop/add-and-import-files/import-other-file-types/import-dcs-files.html)). Spot plates add spot swatches, the composite's clipping path clips the image, Separations Preview and print separations show each plate on its own ink, and the Links panel menu has Show DCS Transparency Interactions (`links.list {dcsTransparency}`). The plates are read as <http://www.telegraphics.net/doc/DCS.txt> describes them (`crates/eps/src/dcs/tests.rs`, `crates/engine/src/tests/dcs.rs`, `crates/doc/src/separated.rs`).
   - EPS export with Use Artboards writes a central EPS file of the whole document, bounded by its visible art, beside the file of each artboard when it writes two or more artboards; a document with no visible art gets only the artboard files (`crates/engine/src/cmd/fileio/tests_eps.rs`).
     EPS export writes a linear or radial gradient of tints of one spot color (and paper white) in that ink's Separation space, at Level 2 and 3, and at Level 3 one that mixes a spot color with process or gray colors, or two spot colors, in a DeviceN space of the spot inks and the process inks its stops use, so each spot ink prints on its own plate; an off-center midpoint is written as a stop with the inks halfway between its two stops. At Level 2, such a mix is written in process colors with a warning. In an RGB export, a gradient that mixes a spot color with RGB or Lab colors is written without the spot ink, with a warning; a CMYK export converts those stops to CMYK first, and the gradient follows the rules above. A Perceptual gradient of spot colors, and a mix whose DeviceN space would need more than 32 inks or a spot color named Cyan, Magenta, Yellow or Black, are also written in process colors with a warning (`crates/eps/src/tests.rs`). A spot color named All or None (in a Separation space, the names of every plate and of no plate: [ISO 32000-1:2008](https://opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandards/PDF32000_2008.pdf), 8.6.6.4), is written as the ink All Spot or None Spot (All Spot 2 and on when a spot swatch already has that name) in its fills, strokes, gradients and DSC comments; Registration still prints on every plate (`tests::spot_inks_named_all_or_none_are_written_under_names_of_their_own`).
     EPS export and EPS Save make each file's PNG thumbnail from its TIFF preview, averaging the preview's pixels down to the thumbnail's size on white, instead of drawing the document a second time; a file without a preview still draws its thumbnail (`crates/engine/src/cmd/fileio/tests_eps.rs`).

docs/three-d.md

@@ -326,7 +326,8 @@ These rules are VectorCraft's own:
 
 - A graphic is drawn on the visible face of each of the object's solids (a fill's, and a stroke's outline),
   turned, foreshortened by the perspective and twisted and tapered with the face, and clipped to it. It is shaded
-  with the face's light: each color is multiplied by the light that falls on the face, and drawn with the
+  with the face's light: each color is multiplied by the light that falls on the face, except in pattern tiles,
+  which keep their own colors. It is drawn with the
   opacity and blending mode of the fill or stroke the face is painted with. A graphic on a face turned away from
   the viewer is not drawn.
 - On the back cap, X runs to the right as the back is seen from behind, so the graphic reads the right way round