When a .ai file, or an EPS with editing data, opened from its editing
data, the reader converted each embedded 8-bit CMYK image to RGB and
stored it as a PNG, so the document lost the image's ink amounts. When
the same .ai opens from its PDF part, the PDF importer keeps them in a
CMYK TIFF.
The reader now stores those samples with ImageBlob::cmyk_tiff, as the
PDF importer does (crates/pdf/src/import_image.rs). An alpha channel
that hides part of the image becomes the image's opacity mask, a
grayscale image over it, as in PDF import. Grayscale, RGB and 1-bit
images still open as RGBA PNGs.
Docs: docs/file-format-parity.md, docs/target-app-parity.md.
hayro-syntax's Dict::entries still unwraps each value it reads, so a
dictionary cut short (or with a stray token) aborts the import. The
craft-libs ArrayIter now stops at the first unreadable item instead of
panicking, which also means a hostile item after a self-referencing
function or colour space is silently dropped from nests_without_end.
- import_color: `readable(array)` checks every item parses; `items()`
yields nothing for an unreadable array; `entries()` reads a dictionary
key by key through get_raw, skipping values that don't parse.
- pages/editing/import_scan: use the helpers instead of raw_iter() and
Dict::entries().
- nests_without_end refuses a file with an unreadable array.
- Tests: two refusal cases in pdf tests and an .ai cut inside an array
or dictionary in engine/tests/import_fuzz.rs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KcHisW1vt6pkVPYbzL6Cpg
In a dev build, opening a native file whose area type has 2^60 rows
or 2^60 columns failed with an internal error ("capacity overflow"),
and opening one with 2^32 rows and 2^32 columns failed with "attempt
to multiply with overflow". Text layout reserved room for rows times
columns frame cells. `text.areaOptions` keeps both counts within 1 to
100, but reading a document kept the counts the file gave.
Reading a document now clamps each count to 1 to 100
(`AreaOptions::MAX_ROWS_COLUMNS`): a larger count reads as 100, and 0
reads as 1, as layout already treated it. Text layout clamps both
counts the same way. Area type with 1 to 100 rows and columns lays out
as before.
The cells of a frame now share one copy of the frame's flattened
polygons and one copy of the flattened text wrap shapes. Each cell held
its own copy: in a dev build, a test process that laid out text in 100
rows and 100 columns of a circle frame of radius 1,000,000 pt peaked at
1.33 GB of memory, and now peaks at 100 MB.
Type is written as version 7 text objects (§15): point type, area type and type on a path, with
fonts by PostScript name re-encoded to Windows-1252, and each run's paint, size, scaling,
tracking, leading, alignment and indents; type 8-bit text can't hold (or all, with outlineText)
is outlines. The format hides only layers, so hidden objects go to a hidden layer above their
own (hiddenObjects "layer", the default), are left out ("omit") or show ("visible"). The .ai
reader reads legacy type (To … TO, TE, %AI3_BeginEncoding) and ignores Ar; the import fuzzer
gets legacy type tokens.
A PDF is refused when a stream in it would decode to more than 256 MB
or has decode parameters the reader can't apply, but the check looked
only at the file's objects. A stream written directly inside an array
or a dictionary, such as the /Function of a shading, was not checked.
On main at b12a52f7, a calculator function holding 257 MiB of deflated
spaces opened after 3.0 to 4.3 s with a peak of 550 to 818 MB when it
was written in an array, in a dictionary or as a shading's function.
The same stream as an object of its own was refused in 81 ms.
hostile in pages.rs now runs unsafe_stream on every stream that the
walk of the file's objects meets, at any depth in arrays and
dictionaries, before it decodes the stream as a calculator function.
It refuses such a file with the same message. The walk is the one the
nesting check made before (64 deep at most).
A shortcut set another app saved as slash-name text (a .kys file whose
text starts with a slash name and "{", such as "/Menus {") opens in
Import Shortcut Set from File › Open, from Edit › Keyboard Shortcuts ›
Import… in the desktop app, or from a drop on the window, whether or
not a document is open. The dialog proposes a target for an entry when
an earlier import kept one for that name, or else when one tool,
command or panel matches the name clearly by its id or by one of its
labels. It decodes the file's key and modifier codes from the codes
earlier imports kept, from the example in the AiCommandPalette README
(/Key 48 for the 0 key, with /Modifiers 0 taken as no modifiers), and
from VectorCraft's default shortcuts where at least two entries vote and
four votes in five agree. The user picks the remaining targets, presses
the key for each unknown key code or marks it No equivalent, and chooses
the modifiers for each unknown modifier code. Apply assigns the ready
rows with the Keyboard Shortcuts rules: a tool takes a single key,
optionally with Shift. It goes through the rows from the last one back
and leaves a row out when a later row it applies has the same shortcut
or the same tool, command or panel. Keyboard Shortcuts then shows the
result as unsaved changes until the user clicks OK. Copy Report copies
the set's name, the rows Apply leaves out and the parts of the file that
weren't read.
- shortcutSet.read is a query command: it reads a set and does not
modify the document or the preferences. It returns each entry with
its section, codes and line, and up to 1,000 skipped tokens, each cut
to 80 characters. It reads at most 4 MiB of text and 20,000 entries,
skips blocks nested more than 16 deep and cuts names at 1,024 bytes.
With a path it reads at most 4 MiB and 4 bytes of the file, and with
--automation-read-root (#832) the path must lie inside the read root.
A file whose text starts with "<", such as an XML shortcut file saved
as UTF-8, fails with "This shortcut file is in a form VectorCraft
can't read."
- importMatches.list, importMatches.set and importMatches.reset read
and change the matches kept with the preferences; resetting the
preferences keeps them. With Remember My Choices on (the default),
Apply keeps the targets and codes the user chose, and for the rows it
assigns, it also keeps the matched targets and the codes decoded by
agreement. A later import proposes them.
- The name matcher compares a name with ids by exact match and shared
words. It compares the name with labels, in English and in the
interface language, by exact match, shared words and character
pairs. It proposes a match only when one tool, command or panel
scores at least 0.6 and at least 0.1 more than the next one.
- When Import Shortcut Set is opened from Keyboard Shortcuts, Cancel,
Escape and ui.dialog.cancel close it and return to Keyboard Shortcuts
as it was (dialogs::PARENT). A document's queued Missing Fonts dialog
(#792) opens after Keyboard Shortcuts closes.
- File › Open has a "Shortcut sets" filter, and in the desktop app
Keyboard Shortcuts' Import… shows .json and .kys files.
Docs: docs/mcp.md, docs/control-protocol.md, docs/development.md,
docs/target-app-parity.md, docs/roadmap.md.
hayro (0.7.2 and 0.8.0 alike) inflates Flate and LZW data with no ceiling, so a few megabytes
of zeros ask for gigabytes, and applies a predictor before checking its parameters: Columns 0
divides by zero (a panic, an abort on the web), a huge Columns allocates a row that large (an
abort everywhere). Both happen inside the reader, before any of our limits.
A file is refused ("a stream in it would decode to more than 256 MB, or its decode parameters
can't be applied") when a stream's decode parameters ask for a predictor the reader can't apply,
or when it decodes past 256 MB through Flate, LZW, run-length and the ASCII encodings before
them, decoded here with a ceiling (only streams large enough to get there are decoded). Every
object's stream is looked at once the file is open (decrypted, references followed), with the
inline images of anything that can be content; object and cross-reference streams, which the
reader decodes while it opens the file, are looked at first in the file's text.
A scatter brush whose art is an instance of a symbol stroked with that brush, or any brush whose
art is painted with a pattern whose tile is stroked with it, drew itself without end: a stack
overflow in a hand-written native file. Brush pieces already strip the brushes of their own
strokes; now what they draw by name that would draw a brushed stroke (a symbol instance, or a
pattern paint, also through the symbols and patterns those use) is left out too, in one place
(`stroke_pieces`, which every renderer and exporter calls with the document).
Probed the rest of item 2 of the depth follow-up with hand-written hostile inputs, kept as tests:
graph designs of deep art and of a symbol holding itself, Affinity groups nested to and past
both of its reader's caps, DXF blocks inserting themselves, EMF/WMF saving states without end,
and plug-in manifests and output nested 30,000 to 100,000 deep. None of them overflowed.
hayro-syntax 0.8.0 reads arrays and dictionaries 64 deep and no deeper wherever they are (the
file's text, object streams, the content of pages, forms, patterns, glyphs and masks), walks the
page tree without recursion and follows a chain of object streams once: what 0.7.2 overflowed
the stack on. hayro-interpret 0.8.0 comes with it (a new device interface: the importer's device
takes it as before). Our raw pre-scan of the file's text and page content goes: the reader's own
cap covers more, and the scan miscounted the bytes of unfiltered binary streams.
What 0.8.0 still builds by recursion with no cap is refused when the file opens: functions and
colour spaces built on themselves or on one another more than 16 deep, and calculator functions
whose procedures nest more than 64 deep. The import runs on a 16 MB thread, as the SVG importer
does.
hayro-syntax 0.8 turns on flate2's zlib-rs backend, which the whole app then deflates with: a
compressed native file is the same gzip stream as before (header and data), but its deflate
blocks now depend on how the data is fed, so the streaming test checks the header and the
unpacked data instead of the bytes.
VectorCraft does not run scripts written for the reference app, and
File › Scripts held only the disabled Other Script…. File › Scripts ›
Check Script… picks a .jsx, .jsxinc or .jsxbin file, reads it without
running any of it, and shows a report: what the script needs (the
script language's own objects, the document model, files, menu
commands, dialogs, or features only the reference app has), what it
would do outside the document, the menu-command strings it passes, the
document-model member chains and type names it uses, its directives,
the files it includes, and the problems that would stop it loading.
Copy Report copies the report as JSON. File › Open and drag-and-drop
show the report for those files too.
- vectorcraft-scriptcheck, a new Standalone crate with no
dependencies, holds the checker: a lexer for ECMAScript 3 with E4X
literals and the #name and //@name directive lines, and an analysis
of the tokens that follows variables, function scopes and the files
a script includes. It never recurses: open brackets, the levels of
an XML literal and the function bodies being read are kept on stacks
capped at 4,096. It reads at most 8 MiB of a file and 32 MiB with the
included files, follows 64 included files 16 deep, looks up at most
1,024 paths for them, keeps 1,000 entries in a list, tracks 1,000
variables holding member chains and 1,000 holding File objects, and
cuts echoed text to 120 characters. A compiled script is reported as
compiled and isn't read.
- script.check (engine, a query) takes path, data or dataBase64. With a
path and includes: true, it reads the files the #include lines name
from inside the script's folder, through the automation read root,
and when the script or one of those files assigns a member chain or a
File object to a variable, the check treats the variable as holding
it in every file. It doesn't follow absolute names, names, folders or
links that lead out of the folder, names or folders longer than 4,096
bytes, or JSINCLUDE, and it reads only regular files.
- ui.checkScript shows the report (dialog scriptCheck). File › Open
lists Scripts (.jsx, .jsxinc, .jsxbin) and sends a script's path to
script.check, which reads at most 8 MiB of it.
The language rules come from ECMA-262 3rd edition and ECMA-357 2nd
edition. The directive, platform and document-model root names come
from the public JavaScript Tools Guide and scripting guide pages that
catalog.rs cites by URL, and executeMenuCommand, doScript, loadAction,
unloadAction and selectTool come from the public forum posts cited
there; only names are used. Every script in the tests is written from
scratch with invented names.
Docs: docs/mcp.md, docs/target-app-parity.md.
Paste read PDF only from the Windows clipboard and EPS nowhere, so on macOS a copied
equation or drawing pasted nothing new (the last thing copied in the app came back).
- macOS: the pasteboard's PDF, EPS and SVG are read, ahead of text and bitmaps.
- Windows: EPS is read beside PDF.
- Everywhere: an .eps file copied in a file manager pastes.
- clipboard.importEps loads an EPS into the clipboard as document.open reads one.
Opening an EPS whose threaded area type frame has a control point at
x = 8e115 failed with an internal error ("capacity overflow").
`eps_text_frames_never_panic` in `import_fuzz` found it with
PROPTEST_CASES=20000. Text layout flattens a frame to within 0.1 pt,
and for such a curve kurbo 0.13.1's `flatten` calls `reserve` for
`usize::MAX` quadratic pieces. Laying out text around a wrap shape with
such a curve failed the same way. Laying out type on a path along such
a curve panicked in kurbo's `solve_itp` ("attempt to shift left with
overflow") in builds with overflow checks, and in a release build it
did not finish within 20 s. Laying out area type in a triangle frame
that comes to a point at its top, at y = 1e20, did not finish either:
adding a lead to a baseline that large leaves the baseline where it is.
Text layout now leaves out a frame, a text wrap shape or the path of
type on a path that has a coordinate beyond 1,000,000,000 pt either
way, control points included, or a NaN coordinate in an anchor point.
Such a frame or path holds no text, and such a wrap shape is ignored.
PDF export uses the same bound (`MAX_COORD` in
`crates/pdf/src/export.rs`).
Text layout also steps a line down a frame cell at most 100,000 times
to find room for it. In a cell taller than 100,000 leads, a step is the
cell's height divided by 100,000; in a shorter cell, a step is one
lead, as before. Before this change, in a dev build, with triangle
frames that come to a point at their top, laying out text in one
2,000,000,000 pt tall took 9.8 s, and laying out a paragraph whose
Space Before is -1e300 pt in one 1,000 pt tall did not finish within
20 s.
- The type in a text object's drawn look (1 /FreeUndo) is a story of the
undo-free text document, which counts its stories on its own. It was read
as the story of that number in the text document, so a label could show
another object's text in another object's frame. The object itself, which
follows the look, is the type now.
- The page's type of a text object on a hidden layer stays with that object
(hidden) instead of going to the nearest text object that shows.
- The type on the page of a later artboard is matched where the layers have
that artboard, not by the first page's place.
The editing data places a gradient with matrices after Bg: Xm for a linear
gradient's whole ramp and one Bm for each pair of stops next to each other.
The reader kept the last Bm, so a gradient of three stops drew only its last
part. It now uses Xm (or, without one, the span of the Bm's from the first
stop to the last), and a radial gradient's one Bm as before. The angle and
length are in the matrices, so Bg's own numbers are used only without them.
An object's own note of its gradient after it (Bb, Bg, BB with no path) no
longer becomes the fill of the next path.
A .ai file with placed files opened from its PDF part (no guides, no sublayers,
no hidden objects), because the editing-data reader stopped at them.
- Lexer: a placed file's preview section is one token (it was skipped).
- Reader: %AI5_BeginPlace ... ~ becomes a linked image with the file's path,
showing the preview the data keeps of it (premultiplied ARGB, capped at
MAX_PIXELS, its rows checked against the bytes there), in its place: the
matrix takes the file's pixels from the canvas, where they start at the point
of art space's origin. A placed file without a preview, a path or a matrix
still makes the file open from its PDF part.
- The open result lists one warning saying linked pictures show the
low-resolution preview until their files are found (rows in the catalogs).
- Engine: a link nothing is known of whose image shows a preview is read where
it is found as the document opens, into the image's bounds, and isn't
reported as modified.
- Tests: lexer, reader (hand-written data), engine open with the link missing
and found, fuzz tokens.
- Docs: file-format-parity, target-app-parity, gaps, ROADMAP.
In builds with overflow checks (`cargo run`, `cargo test` and
`cargo xtask ci`), PDF export aborted the process when a number it
wrote rounded to -2,147,483,648 as an f32. pdf-writer 0.15 writes
-2,147,483,648.0 as an integer, and `Limits::register_int` overflows
when it takes the absolute value. krilla 0.8.2's `Surface` then fails
an assertion in its `Drop` while that panic unwinds.
`svg_hostile_markup_never_panics` in `import_fuzz` hit it with type
shifted 2,147,483,649 pt down. Paths, gradients, dash offsets, images
and artboards that far out hit it too, and so does a group around a
stroke wide enough to reach that far. Release builds have no overflow
checks.
PDF export now leaves out art whose coordinates or transform entries
lie beyond 1,000,000,000 in its own space or in the content stream it
is drawn into, and glyph lines whose spacing lies beyond 1,000,000,000
thousandths of an em. That covers paths and clipping paths, glyph
lines, images, gradients, strokes whose width, caps or miter spikes
reach that far, and the art and marks of a page whose artboard lies
that far, which is written blank. A dash offset beyond the bound is
written as its remainder after whole periods of the pattern.
The type's transform is now pushed inside the form that marks a fill
or stroke of type to overprint. The exporter checks what that form
holds through the type's transform, as it checks the same fill or
stroke without overprint.
A PDF saved with the reference app's editing data keeps that data in
streams named AIPDFPrivateData1, AIPDFPrivateData2, ..., where a .ai
file uses AIPrivateData1, AIPrivateData2, .... illustrator_data read
only the .ai names, so such a PDF opened like any other PDF: plain
groups opened ungrouped, and hidden objects and guides were missing. The
import gave no warning that the PDF's editing data went unread.
illustrator_data now reads the AIPDFPrivateData streams when the file
has no AIPrivateData1, and the PDF opens from its editing data as a .ai
file does. editingData: false still opens it like any other PDF.
has_private_data still looks for AIPrivateData keys only.
layered_ai also returns whether it read the layers. When it did, art
reaching past a PDF's page opens whole, and through_editing_data drops
the PDF importer's note that such art was clipped (PAST_PAGE_NOTE, now
public) as it drops the off-artboard note. When it didn't, both notes
stay. On main, a .ai whose editing data can't be decoded or has no
layers opened from its PDF part without the off-artboard note.
testkit::ai::pdf writes the .ai file of testkit::ai::ai with the PDF
key names. The editing-data tests open it as a PDF, and the import fuzz
test opens the damaged sample as a PDF too.
Docs: the document.open params, docs/mcp.md,
docs/file-format-parity.md, docs/gaps.md (G9: the export direction of
#868 remains), one evidence line in docs/target-app-parity.md, and #868
moved from layer structure to editing data in the File formats row of
docs/target-app-parity.md and in docs/roadmap.md.
The engine's guard turns a panic inside a command into
EngineError::Internal and rolls the document back. The panic checks in
command_sweep.rs counted that error like any other error. They include
probe, which the empty-params, junk-params and structured-junk sweeps
call. The tests in import_fuzz.rs that open, place or paste a file
through a command counted it the same way. None of them reported the
panic.
Each of these checks now passes the command's result through no_panic,
which the library file tests in import_fuzz.rs already use, and the
change adds a copy of it to command_sweep.rs. A case whose command
panics now fails. The panic from no_panic reads "<command> panicked:
<message>", and the checks that print a panic's message show it. Each
check handles any other error as before.
A .ai in PostScript form went only to the PostScript interpreter, which
stopped on the first operator its missing prolog would define
(`Adobe_packedarray`, `Lb`, `Np`): files from Rhino and other CAD apps name
the procsets without including them.
- Such a program is its own editing data: when it has layers
(`%AI5_BeginLayer`) and no private data, the editing-data reader reads it,
compared with the page when the program runs, from the layers alone when
it can't be run. If the layers can't be read either, the error says why.
- The reader takes the legacy ten-operand `Lb`, ignores flatness `i`, and
uses the bounding box as the page when nothing else says where it is.
- Tests on synthetic files; fuzz fixture and hostile-program layers added.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Conflicts:
# ROADMAP.md
# docs/file-format-parity.md
# docs/gaps.md
The editing-data reader stopped on `Xk`/`XK` (global process colours) and
on `p`/`P` (pattern fills), so the file fell back to its PDF part, which has
only the top-level layers.
- `Xk`/`XK` read like `Xx`: a global swatch (not a spot ink), linked at its tint.
- `%AI3_BeginPattern` sections are read by a reader of their own into
pattern swatches (tile from `(name) llx lly urx ury`, art in pattern space);
`p`/`P` fill or stroke with them, placed by the operator's matrix. A pattern
it can't read whole still makes a fill with it unreadable, as before.
- Tests on synthetic editing data; fuzz tokens and fixture extended.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Search: a search field at the top of the Libraries panel filters the
current library's colours, character and paragraph styles and graphics
by name, ignoring case; groups without a match are hidden, groups with
one open, and a search nothing matches says so ("No results").
Groups: items can be put in user-named groups, kept in the library file
(`groups: [{name, items: [{kind, item}]}]`, left out when there are
none, so files written before groups load as they are). New commands:
library.createGroup {library?, name?, items?}, library.renameGroup
{library?, group, name}, library.deleteGroup {library?, group} (its
items stay, ungrouped) and library.moveItem {library?, kind, item,
group?}. An item is in one group at most; deleting an item takes it out
of its group. The panel shows each group as a collapsible section with
its item count after the ungrouped items (rename and delete on its
context menu), each item's context menu has Move to Group (No Group,
the groups, New Group...), and a New Group button sits in the footer.
library.get lists the groups, library.list counts them. The generic
command dialog takes fixed params in `__params` for this.
Export / import: library.export {library?, path?} writes a library
(colours, styles, graphics with art and thumbnails, groups) as a
.vclibrary file, the JSON the library folder keeps (without a path:
{data}); library.import {path | data | dataBase64, name?} adds one as a
new library, made current, numbered on a name clash ("Brand (2)").
The panel menu has Import Library... and Export Library..., through the
app's file dialogs (the web downloads, and picks through File > Open,
which now lists Libraries files). Imported and folder files are made
safe: at most 10,000 items of a kind and 1,000 groups, names clipped and
made unique, sizes finite, thumbnails kept only when their PNG header
says at most 1024 px a side, group entries naming missing items or
unknown kinds dropped; a file over 512 MB or that isn't a library is an
error. Paths go through the engine's file helpers (automation roots).
The importer is fuzzed (import_fuzz: garbage and mutated libraries).
Also fixed: the third graphic added to a library got the second's id
(`graphic-2` twice).
Docs: the commands' params docs, docs/mcp.md, ROADMAP.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
File › Place and File › Open read PSD and PSB files: the merged image
every such document carries after its layers, which is what a linked PSD
shows. A placed PSD links like any image, so the Links panel updates and
relinks it, and it is sized by its ResolutionInfo resource.
The reader is written from the public file format specification. It
reads Bitmap, Grayscale, Duotone, Indexed, RGB, CMYK (through the colour
settings' CMYK) and Lab, at 1, 8, 16 and 32 bits per channel (32-bit
values are linear), raw, PackBits or zipped with or without prediction.
A document with transparency stores its merged image on white with the
alpha in the first extra channel. It comes back with straight colours,
also for 16- and 32-bit files, which say so in their Lr16/Lr32 block.
Multichannel files are refused with a message. The file is untrusted:
every read is bounds-checked and the image's size is checked against
the decode budget before anything is allocated.
PSB is a read-only format of its own, so exports never write a PSD under
a .psb name. The desktop packaging still doesn't associate either
extension with VectorCraft (UNASSOCIATED_EXTS): they belong to raster
editors. Layers (import options) are left for later.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Illustrator EPS and .ai type is made whole from the file's text
document, shown or hidden: point type, area type (a story in several
frames as threaded type that flows when the file opens) and type on a
path, with its fonts, size, leading, tracking, scaling, baseline shift,
fill and stroke colours (grey, RGB or CMYK), alignment, indents and
paragraph spacing. The page's drawing of that type, often in pieces, is
left out; where the text document can't say, type that shows still
comes from the page.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
When a document opens with fonts that aren't available, fonts used
only in its symbols included, VectorCraft lists them in a Missing
Fonts dialog after the missing linked file questions, one document at
a time, with Find Fonts and Close. Find Fonts opens Type › Find Font
on the first missing font of the layers' type, to replace it. After
Find Fonts or Close the dialog doesn't open again for that open; when
another dialog replaces it, it opens again once that dialog closes.
Search Fonts Folder (the Fonts folder File › Package writes next to
the document) and Find in Folder… search a folder and its subfolders
for the fonts' files, and Add Fonts copies the files the user chooses
into a Fonts folder next to the preferences. The desktop app and
vectorcraft-cli read that folder with the system's font folders. Type
› Find Font has a Find in Folder… button too when fonts the document
uses are missing. The web app shows no dialog; a status message says
how many fonts aren't available.
- text.missingFonts lists each missing family and style once,
whatever version the document's type names, and leaves out CSS
generic families such as sans-serif. text.findFontFiles searches a
folder on up to 8 threads and returns at once; text.findFontFiles {}
reports the search, and {stop: true} stops it. A search reads only
what the font scan reads of font files (their table directories and
name, fvar and OS/2 tables) and, on macOS, of suitcase fonts, and
lists a file for a font only when the file makes the font resolve
exactly. It does not enter app packages, the Program Files and
ProgramData folders, the system's folders at the root of a volume,
the Shared and Public folders in Users, a home folder's Library,
AppData, Applications, snap and dot folders, a Library folder's app
data folders or the folders the environment names for apps, and it
skips other hidden folders and does not follow links below the
picked folder. It ends after 60 s by default, after 10,000,000
entries listed, with more than 1,000,000 folders waiting, after
10,000 font files read or 1,000 kept, or once every font has a file;
Stop and Close end it too.
- text.addFontFiles copies only files the last search found. Of
those, it copies only regular .ttf, .otf, .ttc, .otc, .woff and
.woff2 files and suitcase fonts with their resource fork, at most
256 MB a file and 1 GB in all. It never replaces a file: a file with
the same contents is kept, and when another file has that name, the
copy gets a number (Name 2.otf). The fonts are then scanned again,
and type set in them redraws.
- ui.missingFontsDialog and ui.findFontsInFolder open the dialog for
agents; docs/control-protocol.md lists its fields.
- File › Package gives a font file whose name has no font extension
the extension that matches its data (.ttf, .otf or .ttc).
- vectorcraft-text adds set_app_font_dir, which system_font_dirs
reads, WantedFonts, which reads a font file's names as the font scan
does, is_font_file and is_suitcase. #762 split its face reader, which
now takes read budgets: file_face_names_within tries the suitcase
reader first, and reader_face_names_within applies the budgets to
each font. vectorcraft-format adds write_new_with.
Docs: docs/mcp.md, docs/control-protocol.md, ROADMAP.
- Fit to Artwork Bounds fits the art exports draw: hidden objects and
layers, guides and template layers don't count, and the bounds stop at
a layer's clip.
- The comparison with the page catches an object one draws that the
other doesn't (pixels whose neighbours all differ, away from type), a
light colour against white, and, for an EPS whose page is its art's
box, art the layers print outside it.
- textAs: outlines opens a .ai whose type shows from its PDF part, which
has the outlines.
- document.open takes editingData: false to open an Illustrator EPS or
.ai as its page or PDF part only, for print pipelines.
Part of #637.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Save Swatch Library dialog lists Swatch Exchange (.ase) as a
format, and swatch.library.save takes format: "ase" or a path ending
in .ase. Each solid color keeps its model (RGB, CMYK, Lab or Gray) and
its type (spot, global or process), and color groups stay groups.
Gradients and patterns are left out, and a tint swatch is written as
the color it shows. Without a path or user: true, the command returns
the file as dataBase64.
A swatch exchange library opens under the name of its file. When Save
Swatch Library or Save Graphic Style Library saves to a file, the save
panel suggests the Name from the dialog as the file name, as a save to
User Defined does, and the web app downloads the file under that name.
A name made only of dots and spaces, such as "...", gives the file name
Library (Library.vcswatches) in place of a hidden file.
palette_io::write returns the file's bytes. A property test writes
generated libraries and reads them back, and an import fuzz test loads
mutated copies of a file the writer wrote.
The layers reader of the editing copy (operators translated into
PostScript for the interpreter, any file with RGB colours, gradients,
transparency or images falling back to its page) is replaced by a reader
of the whole structure (import/ai): layers and sublayers with their
options, groups, compound paths, clipping groups and layer clipping,
names, hidden and locked objects, RGB, CMYK, grey, named and spot
colours, gradients, transparency, fill rule, overprint, embedded images
with their alpha channel, guides and every artboard, and objects drawn
as their look when they have several fills or strokes, effects or a
brush. Containers: Zstandard, zlib or plain data in EPS and .ai.
The text placement, the comparison with the page and the overwrite guard
stay around it. The page's type takes its text object's place without a
group; the comparison is by lightness, on nothing, without layers that
don't print. Symbols, pattern fills, placed files and unknown operators
make the file come in as its page where they show, and are left out of
layers that don't.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An Illustrator EPS is a flattened page plus, after %%EOF, an editing copy of
the document; a PDF-compatible .ai carries the same in its private data.
Importing only the page lost layer names, order, visibility, hidden layers
and hidden objects and hidden text.
- eps: read the editing copy (Zstandard %AI24_DataStream, older zlib
%AI9_DataStream) into layers, sublayers, groups, compound paths, clip
groups and hidden objects, run by the existing PostScript interpreter.
Falls back to the page, with a warning, when a shown layer uses an
operator that isn't read or the art differs from the page by over 5%.
- eps: read the text document (ate.rs) to place type on hidden layers and
objects; the page's type is matched to the file's text objects by
position (pieces of a line included) and kept with the nearest one when
none claims it. Area type and type on a path are left out with a note.
- pdf/engine: .ai files go through the same reader.
- engine/cli: a file whose import left things out isn't overwritten unless
asked (acknowledgeLoss, --replace-lossy).
- eps: font names re-encoded with a *<n> suffix resolve to their family.
- i18n: es/it/ru for the new import messages; ROADMAP updated.
Container and text-document formats were inferred from user-owned files; the
Adobe file format specification was not read.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Conflicts: README and ROADMAP (both sides kept: the French and Russian
interface with the Affinity sentence; the placed-documents bullet with
the Affinity one); place and swatch libraries (upstream's byte sniffing
and comments, with the preview-only refusal kept).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GQPh8UDgMyARR5Hat9UchC
File › Open, Other Library… and the User Defined library folder read
.ase files. Colors keep their model, global and spot colors stay global
and spot, and color groups stay groups. The library readers take a
file's bytes. On the web, File › Open passes a swatch library file to
swatch.library.load as dataBase64 instead of as lossy UTF-8 text.
Swatch and group names in a library are now unique, in .gpl files too.
Add to Swatches puts each color group of a library into one document
group when the document already has a swatch with the group's name.
The library fuzz helpers fail when a command panics.
Open and Place read Affinity documents natively (.af from Affinity 3,
.afdesign/.afphoto/.afpub from Affinity 1 and 2): layers and sublayers,
groups, artboards with their names and backgrounds, Publisher pages side
by side, curves with live corners, parametric shapes, compound shapes,
clipping by nesting, vector and pixel masks, solid and gradient fills,
strokes, fill layers, artistic and frame text, placed images (their
original JPEG/PNG), pixel layers, embedded documents as their cached
picture, opacity, visibility, locking and blend modes, at the document's
resolution in points. Unsupported content (effects, adjustments, brushes,
master pages, special shapes, Affinity-only blend modes...) is listed in
the import warning, one line per kind. A file whose native data can't be
read opens as its embedded preview with the reason, and that preview
cannot be placed or used as a template or library. Save never writes back
over the Affinity file and there is no Affinity export.
The document semantics were fitted to the thumbnails Affinity embeds in
176 public documents. `cargo xtask corpus --affinity` fetches 22 public
CC0/MIT/Apache-2.0 documents at pinned commits, sha256-verified; the new
engine test renders each and compares it with Affinity's own thumbnail
(0-4.4 of 255 mean difference, per-file ceilings). import_fuzz mutates a
synthetic native document's stream and container through open, render
and export; the fuzz crate gains a stream target.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GQPh8UDgMyARR5Hat9UchC
The PostScript Language Reference (3rd ed., §3.7.3 "Save and Restore")
has `restore` reset composite objects in local VM to their values at the
`save`. The reader only restored the graphics state, so dictionary
entries changed between `save` and `restore` stayed changed: a program
that counts its open saves in a dictionary and refuses to go past a depth
stopped after that many saves in sequence, and the file fell back to its
preview. EPS files with many images, each drawn between `save` and
`restore`, are a likely case of #505; its reporter hasn't sent a file, so
this is part of it.
- While a `save` is open, `def`, `store`, `put`, `undef` and dictionary
`copy` keep each entry's old value in a journal (counted against the
memory budget); `restore` undoes it, newest change first, down to its
save. A save restored already, itself or through an outer one, is no
longer valid (`invalidrestore`, §3.7.3).
- `cachestatus` reports a device's cache sizes (chapter 8) instead of
zeros: its last value, the most bytes one cached glyph may take, is what
programs divide by.
- Tests (programs written from the reference): 100 saves counted in a
dictionary under a depth limit, each dictionary change undone, saves
restored twice, `cachestatus`'s glyph limit. Mutation-checked: without
the journal or with a zero `cachestatus` the test fails. The import fuzz
vocabulary has saves and restores around dictionary changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An image can take one data source per colour component
(MultipleDataSources, or `colorimage` with multi true). The PostScript
Language Reference (3rd ed., §4.10 "Images" and `image`/`colorimage` in
chapter 8) has the operator read the sources in turn, the same amount
from each. The reader drained each source on its own instead: when the
sources are procedures reading one file a row at a time, the first plane
took the rows of every component and the image came out in sheared
horizontal bands, as in #498.
- Data sources are read in turn by one reader, `read_sources`: a procedure
once a round, a file or string a row a round, until each plane is full or
its source ends. Single sources, procedure filters and masked images use
it too; a string source resumes where it left off.
- Tests (`tests_images.rs`, programs written from the reference):
procedures reading one file in the dictionary and `colorimage` operand
forms (fails without the fix), separate string sources, rows padded to a
byte at 1, 2, 4, 12 and 16 bits with an odd width (one and several
sources), Flate PNG predictors with every row filter at an odd width.
The import fuzz vocabulary has several sources reading one file and
sources of uneven lengths.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Two EPS import tests become programs written from the PostScript Language
Reference (3rd ed.), with their own names and structure, covering the
interpreter features the importer relies on:
- `level_2_and_3_probes_read_as_vectors`: a resource category made from
Generic, resources defined and found in it (§3.9), local and global VM
(`currentglobal`, `setglobal`, `gcheck`, §3.7.2), `internaldict` probed
through `stopped`, a halftone dictionary set and restored, `cachestatus`,
`rootfont` and `currentcolorrendering`, drawing with its own procedures.
- `executable_strings_and_integer_keys`: shading dictionaries closed by an
executable string (`(>>) cvx`, or a Level 1 builder with `counttomark`),
procedures kept under integer keys run with `load` through `stopped`,
`cvx`/`cvlit`/`xcheck` on strings.
Each was mutation-checked: making any of findresource, defineresource,
currentglobal, setglobal, gcheck, internaldict, sethalftone,
currenthalftone, currentcolorrendering, cachestatus, rootfont, stopped,
cvx, load, counttomark, cvlit or xcheck unknown fails a test, and so do
strings that `cvx` leaves literal, `load` refusing integer keys and
`gcheck` answering true.
The ROADMAP's EPS paragraph, file-formats row and last-updated line state
the PostScript features the reader supports; the legacy `.ai` comments in
the import module and the interpreter cite that format's published
specification (PRONOM fmt/423) for the `%AI` header comments and `u`/`U`
group operators the reader keys on.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Type left unfinished when a layer's marked content ends, and a clip that closes in a later layer,
stay in their own layer (no stray "Page N" layer, no art moved to the next layer).
- Top-level layers stack as the pages mark them, so a layer whose art starts on a later page keeps its
place; layers marked with no art come in empty.
- A stroke written as each glyph's outline stroked, right after (or before) the type, becomes the type
object's stroke above (or below) its characters: the type stays one editable object.
- Spaces read from gaps wider than a space get their own tracking where even tracking would put the
glyphs after them more than a tenth of an em off (barcode digit groups); lines are measured to their
last visible glyph, so a wrapped line's trailing space no longer spreads into the tracking.
- Lines wrapped in a frame (each wrapped line ending in a space, one left edge, font, size and leading,
with an optional paragraph spacing) come back as one area type object, left aligned or justified,
only where laying it out breaks the lines exactly where the file does.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Read only the indexed PNG thumbnail observed in version-12 Affinity containers. Bound and validate the PNG before decoding, warn about omitted native content, clear the source save target, and reject auxiliary imports that cannot display that warning. Add malformed-input, lifecycle, property and optional native-oracle tests plus a fuzz target. Editable native layers and Affinity export remain unsupported.
Stock EPS files saved in the Illustrator 8 format failed with "typecheck in `a
dictionary`": its gradient procset builds shading dictionaries with `bd` …
`ed`, where `ed` is `(>>) cvx` (or a Level 1 dictionary builder as a
string), and `cvx` left a string literal, so `shfill` got a string. Its
`discard` procedure also `load`s procedures by integer keys, which `load`
refused.
- Executable strings: `cvx` makes a string executable (`xcheck` true,
`cvlit` back, still a `stringtype` for `length`/`get`/`forall`/
`getinterval`); run by name or met in a procedure, it runs its text as a
program, every time (an executable file shares the same runner).
- `load` takes any key (integers, booleans), as `where` already did.
- Test: an Illustrator 8-style gradient procset with both `ed` forms and
`discard`; the import fuzz vocabulary has executable strings and integer
keys. An Illustrator 2026 file saved in the Illustrator 8 format (made
here, not committed) now opens as vectors with its radial gradient.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Discord feedback: "EPS files are always rasterized: they open as bitmaps, in
other programs as vectors". Every EPS saved by Illustrator 2026 fell back to
its TIFF preview (`gcheck` was unknown), and cairo/Inkscape files stopped at
their first image (`status`), dropped masked images (ImageType 3) and failed
on mesh gradients; the warning didn't say where the program failed.
Fallback never silent:
- The warning names the PostScript error, the operator that raised it and
the named procedures it ran in, innermost first ("a PostScript error
(typecheck in `add`), in `inner` in `outer`"); errors caught by `stopped`
aren't reported. It reaches the status bar, document.open, app.open and
file.open (which now return document.open's result, so MCP's open_file over
the control channel gets it), and `vectorcraft-cli info` ("warnings").
- A file whose PostScript can't be read opens as its TIFF preview, else its
Windows metafile preview (as vectors, scaled to the bounding box), else its
EPSI bitmap. es @msg rows for the new messages.
Interpreter:
- Operators programs probe or use: gcheck/scheck, rootfont,
currentcolorrendering, currenthsbcolor, status, token, read,
bytesavailable, fileposition/setfileposition, resetfile, write*, pdfmark
(marks dropped, as Ghostscript does), execform, nulldevice, strokepath,
pathforall, user paths (ufill/ueofill/ustroke/uappend/upath/setbbox, the
encoded form too), setcachedevice2, findcolorrendering, cache, device and
VM queries, revision; (%stdout) (w) file gives a sink.
- `//name` is replaced by its value when read, inside procedures too.
- Tiling patterns become pattern swatches (their PaintProc drawn in pattern
space, one cell; uncoloured ones in the colour under them; makepattern
copies share one swatch); nested patterns stop at 8 levels.
- Type 3 fonts draw their glyphs (BuildGlyph/BuildChar, widths from
setcachedevice/setcharwidth) as outlines grouped per string, spaced by
ashow/widthshow/awidthshow/xshow/yshow/xyshow/kshow; glyphshow and
stringwidth/cshow use them.
- Shadings of types 1 and 4-7 become gradient meshes (data from arrays,
strings or files, packed or not, flags sharing edges), as fills through a
shading pattern too; sampled functions (type 0) are evaluated;
GradientMesh::coons/triangle are shared with PDF import.
- ImageType 3 masked images, interleaved by sample, by row or in separate
sources; ReusableStreamDecode reads its data when made; Flate/LZW PNG and
TIFF predictors.
- Bounded: DeviceN at 32 inks, sampled functions at 2^24 samples and 32
outputs, 4096 patches a shading, strokepath ignores dashes finer than 1e5
pieces, Type 3 glyphs nest 8 deep (a glyph that shows itself overflowed the
stack before the limit), 64-bit index maths for 32-bit targets.
Tests: a synthetic program per generator (cairo/Inkscape, its Type 3 fonts,
matplotlib, CorelDRAW, Affinity, LibreOffice, Ghostscript eps2write,
Illustrator procsets), meshes and function shadings, masked images, fault
messages, probing operators, metafile/EPSI previews, `//name`, user paths,
hostile patterns/glyphs/shadings (run against main, 12 of the 14 generator
and feature tests fail; CorelDRAW- and LibreOffice-style programs already
read there). The import fuzz vocabulary covers the new operators. Checked
against real files made here with Inkscape 1.4 (cairo) and Illustrator 2026
(its own format and the CS-to-CC compatibility formats; all fell back or
stopped early on main and open as vectors now; its Illustrator 8 format
needs the next commit); those files aren't committed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Legacy PostScript .ai files (Illustrator 3-8, and their EPS) open through
the EPS interpreter, which only recorded what was drawn: every group came
in flattened. In a file in Illustrator's format (its %AI comments, or
Illustrator named as the creator) the `u` ... `U` it writes at the top
level now open and close groups, whatever its prolog defines them as, so
groups come in nested as they were, between and inside clipping groups.
clipnest's Clip carries an optional region: without one it is a plain
group, nested in the same chain as the clips (ids given in opening order,
so a chain of clips and groups sorts by id).
Clips nested without end used to nest clipping groups past the stack
(a stack overflow in EPS and EMF/WMF import): clips and groups now nest
at most 128 deep with a warning, and each object's chain counts against
the EPS reader's drawing budget.
PDF-compatible .ai (Illustrator 9+) can't get the same: its PDF content
draws each object on its own and doesn't mark plain groups (only layers,
clipping groups and groups with opacity, blending or a mask, which
already come back); the groups live only in the private data, which is
out of scope. Documented in ROADMAP and docs/mcp.md.
Addresses Discord feedback.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Laying out type (show, stringwidth, cshow) costs about 300 operations plus
64 a character, measured, and now counts that much against MAX_OPS: a
300-byte file running cshow over a 4 MB string took 60 s, and
{ (a) stringwidth pop pop } loop ran for over 5 minutes (already the case on
main); both now stop with "the program runs too long" within seconds.
Tests: executable filters run inside themselves stop at the nesting limit
(directly, mutually and under stopped), currentfile cvx exec runs nothing
twice, cshow over a long string is refused; the hostile PostScript fuzz
tokens gain self-recursive filter execution, currentfile filters, cvlit and
cshow.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
#249 made the reduced repros of #234 open, but a real file saved by the
reference app (version 30) still fell back to its screen-resolution preview
with "typecheck in `a dictionary`". Running its prolog to the end needed:
- The standard dictionaries (systemdict, userdict, globaldict, statusdict,
errordict, $error) as values in systemdict, so `/globaldict where
{ /globaldict get begin } if` gets a dictionary, not an operator.
SharedFontDirectory names the font directory too. errordict has a
handleerror for prologs that wrap it.
- `dictstack`, `internaldict`, `currentsystemparams` and
`currentuserparams` (with the cache sizes programs `get`), and `cshow`.
- Executable files: `cvx` and `cvlit` set a file's attribute, `xcheck`
reads it, and executing one runs its data as a program, as the prolog
does with `(...) 0 () /SubFileDecode filter cvx exec`.
- SubFileDecode with a count of 0 and an empty end string passes its
whole source through instead of nothing.
- `flushfile` on a filter over the program's data skips that data (the
file's XMP packet). A filter chained from a pending one leaves it
nothing to skip, so the writer's own `VCsrc flushfile` is unchanged.
The interpreter empties its standard dictionaries when it finishes, since
systemdict now holds itself. Both real files tested now import as 15
vector paths in CMYK with no warnings, and round-trip through SVG and PDF
pixel for pixel. The EPS fuzzer exercises the new operators.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
dotCooCoo