* fix(redact): size image clearing with checked arithmetic (#2177)
An ICCBased image's /N comes from the file and was cast with `as usize`, then multiplied
unchecked into the row size. /N 2^60 overflowed it (debug: overflow panic; release: a zero row
emptied the buffer and the write loop indexed out of bounds), and a negative /N wrapped.
/N is now limited to 1, 3 or 4 (ISO 32000-2 §8.6.5.5), every size product is checked, and the
pixel writes use `get_mut`. An image that fails any of these is refused, and the caller already
removes it whole (fail-closed). Regression test covers /N 2^60, -1 and 2.
* redact tests: pass the pixel array by value (clippy needless_borrows_for_generic_args)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Flattening the fields of a laid-out XFA form used to leave the XFA in
the file: Adobe's viewers then laid the form out again from the
template over the baked pages, so the "flattened" PDF was still a live
form there and a plain one everywhere else.
pdfcraft_xfa::strip takes the XFA out of a document: the packets
(template, data and all), /NeedsRendering on the catalog and the
AcroForm, the layout mark, the overrides, and the private keys on
every field down the tree. The engine calls it after Flatten form
fields on a document with XFA packets, so the result is a plain PDF
every viewer shows the same, and for the new Edit::XfaToStatic
("Convert to a static form"; tool xfa_to_static), which keeps the
laid-out pages and the fields with their values as an ordinary
AcroForm that no viewer lays out again and whose scripts no longer
run. Both undoable: the engine's post-edit refresh re-reads the
layout mark to drop or restore its XFA state and the cached template.
What the packets held beyond the fields is reported (packets by
name, data no field showed, buttons left inert, the layout's own
warnings, Reader usage rights) as document warnings, and a dynamic
form that was never laid out is refused rather than left with
nothing. Applying redactions shares the same strip. Closes the last
planned XFA parity entry (form.xfa-flatten shipped).
The phone pattern tried every position, so the last groups of a French
social security number, an IBAN or a SIRET matched as a trunk-0 or NANP
number once the groups before them were skipped, and the start of the
identifier stayed visible. A phone number is now a whole run of digit groups:
one reached through separators from a digit before it, or going on past the
28-character cap, is not one.
Claude-Session: https://claude.ai/code/session_01RUNQSPVXx3dDBabRZHAnjz
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Consolidate the reviewed change and its follow-ups on current upstream.
Original change notes:
M8: Bound and reuse repeated redaction overlays
Prepare appearances within a shared 8 MiB budget before changing the document. Reuse one measured and encoded row, reject excessive geometry, and preserve document and undo state on refusal.
Fixes#1326
M8: Preserve upstream font metrics in bounded redaction overlays
M8: reuse ordered redaction mark groups (#1326)
Co-authored-by: LabDennis <341377186+LabDennis@users.noreply.github.com>
55 files hand-rolled the same serialiser -- number the objects, record each one's offset, emit a
classic cross-reference table and a trailer -- 111 times over. The object bodies are properly the
business of the crate under test; the envelope around them is not, and every copy is a place for a
fixture to be quietly wrong in a way that makes a test prove less than it appears to.
`pdfcraft-testkit` holds that envelope. It has no workspace dependencies on purpose, so the crates
at the bottom of the layering can use it without a cycle, and it returns bytes rather than a parsed
document for the same reason.
The builder is byte-identical to the serialiser it replaces, asserted against a verbatim copy of
the old code across versions and trailer variations, because a migration that changes a fixture's
bytes silently changes what the test using it proves. Watched failing by breaking the free-entry
generation: three tests caught it. The one intended difference is that an empty stream dictionary
gives `<< /Length N >>` rather than a doubled space.
Migrates cos, measure and redact: 111 copies in 55 files becomes 108 in 53. The rest are
follow-ups rather than one sweep -- ui-egui alone is 27 of those files and is the most contested
area in the repository. Some copies should stay: a test about a malformed cross-reference table
needs to write its own bytes.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Flattening deleted a cloudy comment it could not draw, and a refused text-box restyle or resize left the new properties in the file. Arrow defaults were stored on plain lines. A custom match ending on a separator included the next glyph, and the vertical-text redaction error claimed an overlap it had not checked.
Co-authored-by: Brandon Thomas <echelon@gmail.com>
* M8: redaction unlinks removed images and forms from inherited and shared resources
Apply Redactions and Remove Hidden Information give a rewritten page its own
pruned /Resources, but the dictionary the page drew through before could still
name the image or form that was removed:
- a /Resources inherited from a /Pages node;
- one resource dictionary object shared by several pages;
- the page's own dictionary, when a form without /Resources drew the image
through it.
A full-rewrite save keeps whatever is still referenced, so the removed image
samples and form text stayed in the saved file while the operation reported
success. Per-page resources (the case fixed in #1099) were not affected.
The interpreter now records the original object behind every XObject it drops
or replaces. After the page loop, each page-tree resource dictionary that still
names one loses that entry unless a page drawing through that dictionary still
draws the name (the same image on another page, or again outside a mark). A
dictionary that loses an entry is replaced by a private copy on its node, so a
dictionary object shared with other nodes is left alone for them. If a page's
use of a name cannot be told (unreadable content, a form that will not decode),
the operation fails closed.
Tests: redacted_xobjects_do_not_survive_through_inherited_or_shared_resources,
an_image_another_page_still_draws_through_the_same_resources_is_kept,
an_image_a_form_drew_through_the_pages_resources_does_not_survive,
a_hidden_layers_image_does_not_survive_through_inherited_resources.
Not covered here: originals named by a form XObject's own /Resources, by an
annotation appearance or by a pattern.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* M9: a cross-reference entry the reader sets aside is never an unchanged signed object
The reader that judges post-signing changes and the renderer that draws the
page resolve some cross-reference entries differently. A revision appended
after signing could redefine a page's content stream and give the object an
entry the reader sets aside while the renderer follows it:
- an in-use entry at offset 0 (the reader skips it; the renderer fails to read
the object there, rebuilds its index and takes the last definition);
- a second entry for the same object in one section (the reader keeps the
first, the renderer the last).
The classifier compared the signed object with itself, found nothing changed
and reported the signature Valid while the page showed the new content.
pdfcraft-cos now records the entries it does not take at face value
(Document::xref_oddities: offset-0 in-use entries other than object 0,
duplicate entries that differ, reserved cross-reference stream types, and a
hybrid /XRefStm entry that disagrees with the table), and can say whether an
entry locates the object itself (Document::xref_entry_is_exact). What it
resolves is unchanged.
classify_changes treats any such entry added after the signed revision as a
change that is never permitted (page content when the object was a signed page
or its content, other changes otherwise). An object whose content compares
equal at a new place is unchanged only when its entry is exact: where the
reader had to search for the object, another reader may arrive elsewhere.
Tests: a_cross_reference_entry_the_reader_sets_aside_does_not_hide_new_page_content,
a_cross_reference_entry_that_misses_its_object_is_not_an_unchanged_object,
cross_reference_entries_that_are_set_aside_are_recorded,
an_entry_that_misses_its_object_is_found_by_searching_but_is_not_exact.
This closes the two demonstrated triggers and their near variants. It does not
prove that the two parsers agree on every file: other differences in what each
tolerates may remain.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* M8: Apply Redactions is refused when the page still shows text under a mark
Redaction decides what lies under a mark from the positions its own
interpreter computes, and its verify pass reuses them. Where those differ from
the renderer's, a mark drawn over the text as displayed lies over glyphs the
interpreter places elsewhere, or does not see at all: they stayed in the file
and the operation reported success.
The case the test pins: a font and size set through an ExtGState /Font entry
(`gs`), which the renderer honours and the interpreter does not model. The same
kind of gap was observed for a font a form XObject takes from its caller, for
some predefined CMaps and for a rotated Type 3 font matrix. The standard-14
width error that first showed it is fixed at its root by d0d1ad81.
The engine now asks the renderer, which is what the user saw when placing the
marks. After redaction runs, the result is written as a save would write it
(without the marks' own overlay), its text is extracted, and a glyph still at
least half inside a mark fails the edit with the existing "could not be
verified" error. The edit runs on a copy, so the document is unchanged. A page
whose text cannot be read back fails the same way.
This makes such mismatches fail closed; it does not make those documents
redactable. Teaching the interpreter each of them is separate work.
Test: redaction_is_refused_when_the_page_still_shows_text_under_a_mark (the
line whose font is set by `gs` is refused and left unchanged; the same line set
with `Tf` is redacted).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* M8: Retire inherited redaction resources
* Redaction: pin inherited resources only on redacted pages, shared by reference
localize_resources rewrote every page in the document on every apply and copied a direct
inherited /Resources dictionary into each one, so a large document grew by N x the dict.
Now only redacted pages that inherit resources are pinned; a direct inherited dictionary is
made one indirect object and shared by reference. The parent's /Resources is retired when no
other page still inherits it (the leak fix), and kept (through the shared reference) when an
unredacted page still draws with it. A /Parent chain that can't be followed or disagrees with
the page tree retires nothing. Tests: an unredacted page is left untouched and the rewrite is
bounded; redacting every inheriting page retires the parent's resources. ROADMAP entry moved
into the revision-history table.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
pdfcraft_fonts::helvetica_width was a fifteen-line guess by character class,
and seven crates generated appearance streams through it: annot, edit, forms,
ocr, redact, sign and xfa. Times was Helvetica x 0.9, bold was x 1.05, Courier
a flat 0.6. Against the published metrics a Times capital A is 722, so that
estimate said 600 -- a 17% error on a letter, compounding across a line.
Std14 carries the real advance widths for all fourteen faces, indexed by
character code: WinAnsiEncoding for the twelve text fonts, which is what we
write into the appearance streams we generate, and their own built-in
encodings for Symbol and ZapfDingbats. The tables are generated by joining the
published encodings with the per-glyph widths.
helvetica_width now delegates, so every call site gained exact Helvetica with
no change. The three places that faked another face by scaling it -- Family
in edit::added, the paragraph measure in edit::text, and the redaction overlay
fonts -- now select the right face instead. So does Metrics::width in
fonts::pdf, which measures existing document text when a font carries no
/Widths; that path feeds redaction's glyph removal, where a wrong width means
cutting the wrong glyphs.
Tests spot-check the table against values typed in from the specification by
hand, so it is verified against the spec rather than against whatever produced
it; plus Courier being fixed pitch everywhere, full printable-ASCII and Latin-1
coverage, Symbol and ZapfDingbats not being WinAnsi, /BaseFont aliases and
subset prefixes resolving, and hostile text staying finite.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Remove Hidden Information counted and removed as a comment every
annotation that was not a link, widget, pop-up, file attachment or
redaction mark, so a Comments-only cleanup also deleted Watermark,
PrinterMark, TrapNet, Screen, Movie, RichMedia and 3D annotations.
Use pdfcraft_annot::is_comment_subtype, which the comments list already
uses, and keep excluding file attachments (removed with Attachments)
and redaction marks as before.
Adds tests::removing_comments_keeps_annotations_that_are_not_comments,
which fails before this change.
* M14: text the fonts can't draw is kept and reported, never written as ? silently
Form values and captions, text boxes, callouts, Fill & Sign text, measurement labels and stamp by-lines keep text the standard fonts can't draw, draw ? for it, mark the annotation (PCUndrawn) and report the edit (interface notice, tools' warnings); forms set /NeedAppearances (not in PDF/A). Flattening, redaction overlay text, typed and visible signatures and create-from-text refuse instead. Reports escape such characters as [U+XXXX].
* Catalogs: the two new can't-draw messages in Czech, Slovene, Kazakh, Turkish and Greek; refresh hashes
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
* fix(create): cap decoded size of imported PNGs
A PNG's declared dimensions sized an infallible allocation, so a few-byte file claiming huge dimensions aborted the process when the allocation failed. Cap the decoded size at the TIFF ceiling and reserve every input-sized buffer fallibly, returning CreateError::Image instead. (Security audit run-2: v3-024.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(xfa): bound total picture bytes kept during layout
Each laid-out picture kept its own copy, so one picture in a repeating subform multiplied without bound and could exhaust memory on open. Add a per-layout aggregate cap (256 MiB) that stops with XfaError::TooLarge. (Security audit run-2: v3-103.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(xfdf): bound XML nesting and protect redaction marks on import
Deeply nested XFDF/XML overflowed roxmltree's recursive parser and aborted the process; a non-recursive depth scan now rejects it before parsing. Imported comments could also replace a locked or redaction mark by /NM or add a hidden one; place() no longer replaces locked/Redact annotations and clears the hidden/invisible flags on imported redaction marks. (Security audit run-2: v3-105, v3-111.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(sign): classify post-signing changes by the signed revision's roles
A signed object's modification was classified from labels in the attacker-written new version, so relabelling a signed page /Type /Sig (or giving it /ModDate), or relabelling a drawn stream /ObjStm or /XRef, let a forged page validate as Valid even under DocMDP P=1. Decide an object's role (page content, signature, metadata, file plumbing) from what it was in the signed revision, not its new dictionary. (Security audit run-2: v3-082, v3-083.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(redact): remove redacted content from the saved file, fail closed otherwise
Apply Redactions and Remove Hidden Information left the originals behind: dropped or cleaned image/form XObjects kept their resource names (so the full rewrite saved the originals), an XFA form kept the redacted value in its datasets packet, and vertical-writing text under a mark was never hit by the horizontal removal geometry yet reported success. Prune replaced XObject names the page no longer draws, drop /XFA and /NeedsRendering when redacting an XFA form, and fail closed (RedactError::VerticalText) rather than silently leave vertical text. (Security audit run-2: v3-064, v3-067, v3-062.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(ui-egui): trust the displayed signer, not the document's embedded root
The Signatures panel's one-click 'Add to trusted certificates' trusted the top of the chain the document supplied, installing the author's self-signed root as an anchor instead of the signer shown. Trust the displayed signer certificate. (Security audit run-2: v3-098.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* Use main's decoded-size ceiling for PNGs too; keep the depth tests on main's checker
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
The Email pattern of Search & Redact accepted only ASCII letters and
digits, so addresses such as "ayşe.yılmaz@örnek.com.tr", "müller@bücher.de"
or "иван@пример.рф" (RFC 6531 local parts, IDN domains) were not found
and stayed visible after a redaction by pattern.
Letters and digits of any script written with spaces now count in the
local part, the domain and the top-level domain. Scripts written without
spaces (Chinese, Japanese, Korean, Thai, …) still don't, so text that
runs straight into an address, as in "メールはtaro@example.jpまで", is not
swallowed into it. A domain still needs a dot and an alphabetic top-level
domain of two or more letters.
The Dates pattern of Search & Redact knew only English month names, so
"10 Ekim 2026", "3 Mayıs 2025" or "5. Oktober 2026" stayed visible after a
redaction by pattern. It also lowercased with str::to_lowercase, under
which "MAYIS" never equals "mayıs" and "EKİM" grows a combining dot.
Month names are now listed for English, Turkish, German, French and
Italian, with English and Turkish abbreviations ("Şub.", "Ağu"), and are
compared one character at a time with I, i, İ and ı folded to i, so any
case matches. A day may end in a full stop before the month name
("5. Oktober 2026"). Words that only begin like a month ("Martı",
"Ekimde", "Oktoberfest") and month names without a day and year still
don't match.
* M8: redact lists of words or phrases
Find text and redact gains "Multiple words or phrases": a list with one word
or phrase per line, typed or imported from a .txt file (Import list…), whose
every match is marked in one undo step. Lists are trimmed, deduplicated and
bounded (1,000 entries of 256 characters; files up to 1 MB are read).
redact_mark takes words and reports matched_words, the marks each entry made.
protect.redact-word-lists is shipped; whole-word and case-sensitive matching
and exporting the list are still missing.
* M8: translate word-list labels into Arabic
The Arabic catalog arrived on main after this branch, so the new Find
text and redact strings were still English and failed
arabic_covers_ui_literals.
* Tests: drop sharp grid renders when the thumbnail will cover the cell
Zooming the organize grid back out left the sharp page raster in place
until a thumbnail texture existed, so the zoom test failed when that
texture had not arrived yet.
* M8: Czech rows for the word-list redaction labels
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* M8: redaction code sets (U.S. FOIA, Privacy Act) as overlay text
Redaction properties offers "Custom text" or "Redaction code" under "Use overlay
text". Picking the U.S. FOIA (5 U.S.C. 552(b)) or U.S. Privacy Act (552a) set and
ticking codes makes new marks carry them as overlay text, in the set's order.
redact_mark takes code_set and codes and rejects unknown sets or codes with the
allowed list.
Also merges a duplicated evidence/notes pair in form.field-list that made
parity/acrobat-features.toml fail to parse on main.
* M8: say why the default redaction code set can't fail
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Tests: pin the clock in form_bulk (saves a second apart differed in /ModDate on Windows CI)
---------
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
A page's /Contents array is one content stream in pieces that writers may
split between any two tokens (ISO 32000-2 §7.8.2). Edit text parsed each
piece on its own, so an operator whose operands ended one piece and whose
keyword started the next arrived with no operands: a `[(After) -20 (wards)]`
/ `TJ` split made the line vanish from text_lines, text_paragraphs and
text_edit, and from the Edit text tool.
The joined-parse logic the redaction interpreter used (#153) moves into
pdfcraft-content as `Pieces`: join the pieces, parse them as one stream,
map positions back to pieces, and splice joined byte ranges back into
per-piece bytes (an edit replacing or removing an operator that spans
pieces rewrites every piece it touches; untouched pieces keep their
objects). text_lines now interprets the joined operators, and replace_line
and rewrite_block splice through `Pieces`. Redaction uses the same helper.
Lines and paragraphs are grouped by the piece their first operator starts
in, as before, so pages whose pieces are separate drawings don't merge.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
The Phone Numbers redaction pattern accepted only 10 digits (11 with a
country code), the North American form. For "+49 2151 123456" (12 digits)
the `+` candidate was rejected and the scan then matched "2151 123456",
leaving "+49" visible: a partial redaction in a document the user thinks
is clean.
The pattern now accepts, still grouped by at least one separator:
- E.164 numbers with a leading `+`: 8-15 digits in all;
- national numbers with a trunk 0 followed by 6-14 digits, whose first
group has at least two digits (so decimals don't match) and which are
not a day-first date (left to the date pattern);
- North American numbers as before.
A separator may now be followed by `(`, so "+49 (0) 2151 123456" is one
match instead of a partial "(0) 2151 123456".
Refs #525. The other items (page-number and date tokens, stamps) are not
addressed here.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Redact: operators split across a page's content streams are redacted whole
A page's /Contents array is one content stream in pieces, and a writer
may split it between any two tokens (ISO 32000-2 §7.8.2). Some office
and layout exporters do: a TJ array ends one piece and its TJ operator
starts the next, or a marked-content dictionary closes in the next piece.
Redaction parsed each piece on its own, so such an operator was cut off
from its operands. Glyphs shown by a split TJ were never checked against
the marks: Apply drew the black box but left the text in the file, and
Verify reported nothing under the box. When the piece holding the
operands was rewritten for another mark, they were dropped, so the text
after it vanished and the next piece began with an operator without
operands, which other readers reject.
process() now parses the pieces joined, gives each operator back to the
piece its keyword is in, and rewrites every piece a split operator spans.
Pages whose pieces split only between operators are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Use the pdfcraft crate names after the rename
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: EthanSuquet <292885762+EthanSuquet@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Brandon Thomas <echelon@gmail.com>
The app, crates, binaries, bundle id, packaging, icons, screenshots and docs are now
PdfCraft / pdfcraft (repo storytold/pdfcraft). On first launch the desktop app moves the
PrintCraft settings and crash-recovery folders to the new name, and both apps read settings
saved under the old storage key, so upgrades keep recent files, preferences and unsaved work.
The signing test fixtures keep their PrintCraft subject names (they are binary).
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
annot, compare, content, cos, export, geom, model, ocr, preflight, print and redact have no
production `unwrap`/`expect`/`panic!`/`unreachable!`/`todo!`/`unimplemented!`. The attribute
keeps it that way.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Stable clippy 1.99 denies chunks_exact/chunks_exact_mut with a constant
size (32 sites), which failed the test job on main and every open PR.
Behaviour unchanged; MSRV 1.90 has as_chunks.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Text added with Edit ▸ Add content keeps its source string in the stream dictionary's /PCAdded.
Applying a redaction rewrote the stream's glyphs but copied the dictionary unchanged, so the
redacted text survived in the saved PDF. The redacted item now stays as plain drawn content and
is no longer editable.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
