Found while testing crates/textimport/src/docx.rs:183 (run_attrs, w:sz) and :206-214 (para_attrs, w:spacing/w:ind via the `tw` closure)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Found while testing crates/textimport/src/tagged.rs:173
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Found while testing crates/textimport/src/xlsx.rs:62 and :112 (shared strings are read with El::text, which includes <rPh> children; inline strings use the same call, crates/textimport/src/docx.rs:37)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
crates/engine/src/dtd.rs: keep the branch's version; #319 already bounds DTD group nesting (with the same limit of 64) along with the entity budget and matching fixes, so this PR's DTD part is dropped. The XML element limits are taken.
crates/textimport/src/docx.rs: keep the branch's page_break_ends_its_paragraph test (already present further down); add the PR's inline content control test.
Found while testing crates/textimport/src/docx.rs:354
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
ST_OnOff (ECMA-376 17.17.4) is a boolean plus the words on and off, but
the toggle reader only counted 0/false/none as off — so <w:b w:val="off"/>
(a legal spelling written by non-Word producers) imported as bold, and
the same for <w:i>, <w:strike>, <w:caps>, <w:isLgl> and <w:u>.
Add "off" to both off-lists. Regression test:
off_valued_toggles_stay_off — on main the off run imports bold; with the
patch it stays regular and an on run stays bold.
The DOCX and IDML element-tree builders accepted any depth, and the
tree is then walked and dropped recursively; a DTD content model's
groups recursed without a budget. Each parser now refuses nesting past
a fixed budget with an error, before any recursive consumer runs.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011K1peyrh5MdZ6ttHkVq8Gh
Table cell paragraphs go through the same run and inline path as body paragraphs, so deleted and moved-from text and field instructions stay out of cells, and field results, tabs and breaks come in.
Word's Normal imports as a paragraph style named Normal whose definition is w:docDefaults with Normal's own properties on top; paragraphs without a style use Word's default paragraph style. A styleMap entry can still map Normal to [Basic Paragraph].
The .docx/.xlsx text import iterates attributes with quick-xml's default duplicate check, so one crafted
start tag with N attributes cost O(N^2) (RUSTSEC-2026-0194, fixed in quick-xml 0.41). Bumping also clears
RUSTSEC-2026-0195 (NsReader namespace allocation), which this workspace does not reach (no NsReader).
unescape_value() is deprecated in 0.41 (clippy -D warnings); normalized_value(Implicit1_0) is what the
deprecated method now does. Regression test: an element with 100,000 attributes parses in < 5 s.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Round cached doubles to 15 significant digits, so =0.1+0.2 reads 0.3,
and write very large or small values with an exponent (1E+300).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Keep small and high-precision numbers intact in table placement and data merge. Honor cached formula strings as text, including leading zeros and whitespace.
Co-Authored-By: OpenAI Codex <noreply@openai.com>
Preserve paragraph breaks after a ParaStyle tag starts the story so leading empty paragraphs and their styles survive Tagged Text import.
Validated individually and after #638 with all six cargo xtask ci gates (1,278 tests passed, four existing tests ignored, 15 WASM compilation checks), six independent paragraph/header/Unicode probes, and an actual export-to-placement command smoke test.
Refs #641
Paragraphs with w:numPr, directly or through their paragraph style, become
bullet or numbered paragraphs: Word's number formats, label text, start
numbers and overrides, level restarts, the text after the label and the
level's indents. Parent levels in a label are written as text. Unreadable or
hostile list definitions leave plain paragraphs.
A break character is stored as the last character of the paragraph it
ends, followed by the usual '\n' separator. The story model, paragraph
ranges, styles, editing and every exporter keep splitting paragraphs on
'\n' only, so the change stays in the places that create, lay out or
write break characters:
- IDML import writes the break and ends the paragraph, so the range that
follows keeps its own paragraph style. NextOddPage and NextEvenPage map
to new odd and even page break characters; IDML export writes a
paragraph that ends in a break as that one Br with its
ParagraphBreakType, so export and re-import give the same story.
- Word and RTF import, Type > Insert Break Character (with odd and even
page breaks added) and typing end the paragraph after a break. Typing
between a break and its return goes to the next paragraph; deleting
either deletes both.
- Layout: a break that ends a paragraph leaves no empty line; the next
paragraph starts in the next column, frame or page, as a paragraph start
option would. Page breaks and Next Page start options move to the next
frame on a later page; odd and even ones to the next page with that
page number parity (no such frame: overset). A break inside a paragraph
from an older document still moves the rest of that paragraph on.
Storing the break as a paragraph separator of its own would have changed
every place that counts or splits paragraphs on '\n'.
Co-authored-by: Brandon Thomas <echelon@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Text inside a tracked move destination was dropped on .docx import, so
moved passages vanished from both places. Treat moveTo like ins (both
kept as accepted); moveFrom stays skipped like del. Regression test
covers ins, del, moveFrom and moveTo together. Fixes#256.
Also adds my contributor credit entry.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Placing an RTF or Tagged Text file dropped every character outside the
Basic Multilingual Plane: emoji, mathematical letters and the rarer CJK
ideographs such as 𠮷 (U+20BB7, used in Japanese names). The text
around them came in, so the loss was silent. It also broke our own round
trip: Export Text writes these characters correctly and Place lost them.
Both formats carry such a character as a UTF-16 surrogate pair, two
units in a row: \u-10179?\u-8704? in RTF, <0xD83D><0xDE00> in Tagged
Text. Both readers passed each unit to char::from_u32 on its own. Half a
pair is not a character, so each half was dropped.
A small shared Utf16Units joins the units: a first half waits for the
second, and the pair becomes one character. Half a pair with no partner
is still left out, as before.
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Conflict: crates/compose/src/lib.rs: start options, the break after the previous paragraph and a break inside a paragraph go through a local `go` that sends NextColumn to next's span-aware next_column (#276) and the rest to #285's Cursor::start; the in-paragraph break takes #283's `brk` (the line's last glyph); #283's ends_para uses is_break_char; Cursor keeps next's resume() and adds #285's start/next_page (#285's old next_baseline header dropped, next's is kept); the table branch is #285's
Conflict: crates/idml/src/tests.rs: tests only; kept both
Conflict: crates/textimport/src/docx.rs: tests only; kept both
A break character is stored as the last character of the paragraph it
ends, followed by the usual '\n' separator. The story model, paragraph
ranges, styles, editing and every exporter keep splitting paragraphs on
'\n' only, so the change stays in the places that create, lay out or
write break characters:
- IDML import writes the break and ends the paragraph, so the range that
follows keeps its own paragraph style. NextOddPage and NextEvenPage map
to new odd and even page break characters; IDML export writes a
paragraph that ends in a break as that one Br with its
ParagraphBreakType, so export and re-import give the same story.
- Word and RTF import, Type > Insert Break Character (with odd and even
page breaks added) and typing end the paragraph after a break. Typing
between a break and its return goes to the next paragraph; deleting
either deletes both.
- Layout: a break that ends a paragraph leaves no empty line; the next
paragraph starts in the next column, frame or page, as a paragraph start
option would. Page breaks and Next Page start options move to the next
frame on a later page; odd and even ones to the next page with that
page number parity (no such frame: overset). A break inside a paragraph
from an older document still moves the rest of that paragraph on.
Storing the break as a paragraph separator of its own would have changed
every place that counts or splits paragraphs on '\n'.
Text inside a tracked move destination was dropped on .docx import, so
moved passages vanished from both places. Treat moveTo like ins (both
kept as accepted); moveFrom stays skipped like del. Regression test
covers ins, del, moveFrom and moveTo together. Fixes#256.
Also adds my contributor credit entry.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Placing an RTF or Tagged Text file dropped every character outside the
Basic Multilingual Plane: emoji, mathematical letters and the rarer CJK
ideographs such as 𠮷 (U+20BB7, used in Japanese names). The text
around them came in, so the loss was silent. It also broke our own round
trip: Export Text writes these characters correctly and Place lost them.
Both formats carry such a character as a UTF-16 surrogate pair, two
units in a row: \u-10179?\u-8704? in RTF, <0xD83D><0xDE00> in Tagged
Text. Both readers passed each unit to char::from_u32 on its own. Half a
pair is not a character, so each half was dropped.
A small shared Utf16Units joins the units: a first half waits for the
second, and the pair becomes one character. Half a pair with no partner
is still left out, as before.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- Remove the docs/superpowers planning spec.
- textimport: cap decompressed zip parts at 64 MiB (protects Place and
xlsx data sources); oversized part is an error. Regression test.
- datamerge: refuse data files over 256 MiB and skip image candidates
over 256 MiB by metadata before reading.
- datamerge: assign_page uses get_mut instead of indexing spreads.
- datamerge: one/limit over u32 are errors, not truncated.
- select_records dedupes with a HashSet (was O(n^2)), order kept.
- Hyperlink cells: only http, https, ftp and mailto schemes link.
- Document data.* commands and data.merge creating a new document.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Records keep empty cells inside the header width and cap the table at
100,000 rows and 200 columns. Place as a table still uses the existing
Excel import, which drops empty cells.
An .xlsx cell reference far from the others no longer allocates the whole
range before the cell cap (billions of cells, out of memory), and long
column names can't overflow. IDML import doesn't trust zip part sizes for
up-front allocation and adds header/footer counts saturating. Tagged Text
and IDML export use let-else; the font fallback has a compiled-in last
resort. These crates deny the panicking lints.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Brandon Thomas