A custom shape whose type WordCraft has no kind for (hexagon, octagon,
smiley, can, moon, lightning, block arc, cross, pentagon, flowchart
shapes…) opened as a rectangle. It is now drawn from its own geometry
(ODF 1.3 §10.6.2) as a freeform:
- the enhanced path (§19.145): moves, lines, curves, ellipses (T, U),
arcs (A, B, W, V) and quarter ellipses (X, Y), arcs followed in
short straight steps;
- its values worked out from the shape's equations (§10.6.3) and
adjustment values: numbers, other equations, modifiers, the frame's
keywords, + - * /, abs, sqrt, sin, cos, tan, atan, atan2, min, max
and if.
Path length, equation count and nesting, and points are all bounded.
LibreOffice fills no part of a `draw:path` that has an open part. So a
freeform with both closed and open paths is now saved as a custom shape
whose path leaves the open ones unfilled (`F`).
LibreOffice's preset shapes render as LibreOffice draws them. The
saved file validates against the schema.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
ODT citations (`text:bibliography-mark`) were read as their text, and
the sources they carry were lost. So an ODT bibliography had nothing
to update from.
- Reading: each mark is now a `CITATION` field of its source, which
becomes one of the document's bibliography sources (the first mark
of an identifier gives it). The identifier is the tag, the type the
kind, plus author, title, year, publisher, address, journal, volume,
pages, URL, editor, edition, number and month.
- Writing: a `CITATION` of a known source is written back as a mark
carrying the source.
References › Update rebuilds an ODT bibliography from them. A
LibreOffice citation and bibliography round-trip pixel-identical and
validate against the schema.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Linked text boxes (`draw:chain-next-name` on a `draw:text-box`, naming
the frame its text continues in) lost their link on open. The
following box's text was gone, since only the first box holds the
story.
- Reading: each box keeps its frame name and the name of the next box.
Once all boxes are made, names are resolved to WordCraft's link
(`TextBody::next`). A link to a missing box, to the box itself or
round in a loop is dropped.
- Writing: each box is named after its story, and a linked box names
the next one.
Flow model: `BoxLook::chain` (name, next).
A LibreOffice chain round-trips pixel-identical, and WordCraft flows
the text from the first box into the second. A linked box keeps its
own height where LibreOffice grows it to the rest of the text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Two picture properties were dropped.
- Crop: `fo:clip` (top, right, bottom, left, in lengths of the
picture's own size, its pixels at 96 per inch) is now WordCraft's
crop fractions. A picture's crop is written back as `fo:clip`, with a
graphic style even for an in-line picture.
- Contour wrapping: `style:wrap-contour` with
`style:wrap-contour-mode` `outside` is Word's Tight wrapping, and
`full` (the default) is Through. The frame's `draw:contour-polygon`
becomes the wrap polygon, scaled from its view box. A polygon whose
points lie well outside its view box is in another coordinate system
(LibreOffice writes one for some pictures), so the outline is
derived from the picture instead. All of this is written back, the
polygon last in the frame.
A cropped LibreOffice picture round-trips pixel-identical, and a
contour-wrapped one round-trips with WordCraft wrapping text as
LibreOffice does. Both validate against the schema. `Body::start`'s
frame parts moved to a helper (clippy's complexity limit).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
This round added flow breaks, table placement, ruby, index marks, text
borders, connectors and custom properties. A 20,000-paragraph document
round-trips in about a second with nothing lost, and 400 fuzzed
documents opened without a crash.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Review of the custom property commit: an ODF date property may be a
bare date, while Word's `filetime` (which a DOCX save writes) wants a
UTC date and time. A bare date is now midnight UTC, a date and time
without a zone is taken as UTC, and a value that isn't a date is kept
as text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
ODT custom document properties (`meta:user-defined` in `meta.xml`) were
dropped, and the fields showing them or a user field were read as
plain text.
- Properties: `meta:user-defined` values (string, number, true/false,
date) are now WordCraft's custom document properties, and so are
user field declarations (`text:user-field-decl`).
- Fields: `text:user-defined` and `text:user-field-get` are Word's
`DOCPROPERTY` fields, which WordCraft updates from those properties.
- Writing: custom properties go back into `meta.xml`, and a
`DOCPROPERTY` naming one is written as `text:user-defined`.
A LibreOffice document with a custom property and a user field
round-trips pixel-identical, valid against the schema. Its user field
comes back as a document property.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Review of the connector commit: a connector inside a drawing group
wasn't converted. Its group's shapes weren't recorded, so its ends
lost their glue. Group members are now recorded for their kind (never
for placement, since they sit in the group's coordinates), and
connectors in groups have their glue points made their shapes' sites
too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
ODT connectors (`draw:connector`) were left out. They now open as
WordCraft connectors: straight (`line`), curved (`curve`), or with
bends (`standard`, `lines`).
- Gluing: each end is glued to the shape it names (`draw:start-shape`,
`draw:end-shape`). The shape's four default glue points (§9.2.19:
top, right, bottom, left) become its connection sites. An end glued
to a glue point of the shape's own, or to a shape not read, is left
unglued.
- Placement: a connector's own coordinates are in LibreOffice's drawing
layer, not its anchor's. So a connector glued at both ends to shapes
anchored in its paragraph is placed between their sites.
- Writing: connectors are written back as `draw:connector` with their
type and glue, and the shapes they glue to get ids
(`xml:id`/`draw:id`). LibreOffice re-routes them from the glue.
LibreOffice's connector document round-trips with its connectors
drawn between the shapes, and WordCraft draws them between the right
sites. Other formats still get connectors as straight lines.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
LibreOffice writes a border around text as `loext:border` (or its
sides) with `loext:padding`; ODF 1.3 has no attribute for it. It was
dropped. It is now read as Word's text border (one line all round, the
first side with a line, the padding its distance) and written back
the same way, which LibreOffice draws. A document with a text border
is then extended ODF rather than strict.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Gap 19's row now says what ODT reads and writes and what is left, and
its estimate drops from 18–27 h to 10–16 h. The roadmap's progress log
covers the ODT work since text boxes: shapes through password
protection, forms, ruby and flow breaks.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
ODT index marks (`text:alphabetical-index-mark`) were dropped. So
References › Update Index on an opened ODT index found no entries and
left the old list in place. Now:
- a mark at a point is Word's `XE` field, its `text:key1`/`text:key2`
the levels above its entry;
- a mark over text (`-mark-start` … `-mark-end`, in one paragraph) is
an `XE` field whose entry is the marked text, which stays text;
- `XE` fields are written back as point marks.
Engine: Update rebuilt only lists in WordCraft's own form, a paragraph
holding the field. A list held as a field around its entries, as ODT
and Word files hold an index, bibliography or table of figures, is now
first made the field's own paragraph, then rebuilt. That happens only
when the field starts its paragraph, so no document text is replaced.
A LibreOffice document with an index mark round-trips pixel-identical,
the mark kept.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Ruby (`text:ruby`, ODF 1.3 §6.4: a base text with ruby text over it,
furigana or pinyin) was read as its base and ruby text run together,
and the writer wrote any ruby as plain text (its comment wrongly said
ODF has none).
Ruby is now read as WordCraft's Phonetic Guide. It is written back as
ODF ruby, above its base and centred, with an automatic ruby style.
A ruby inside a tracked change is still written as its base text,
which the change marks, so the tracking isn't lost. LibreOffice shows
the round-tripped ruby over its base, and the content validates.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Six features of LibreOffice documents were lost on open and save. A
two-page document came back as one page, because the page break after
a paragraph was dropped. Now read and written:
- a page break after a paragraph (`fo:break-after="page"`), as a page
break before the next paragraph, as Word has it;
- lines kept together (`fo:keep-together`);
- widow and orphan control (`fo:widows`/`fo:orphans`): on with any line
count, off with none;
- hyphenation (`fo:hyphenate`): a paragraph's in ODF, the document's in
WordCraft. A document with a hyphenated paragraph hyphenates, and its
other paragraphs opt out;
- a table's alignment and width (`table:align`, `style:rel-width`);
- its rows' heights (`style:row-height` exact, `style:min-row-height` at
least).
Flow model: `FTable::extra` (alignment, width, row heights), filled by
formats that keep them.
A LibreOffice document with all of these round-trips at 573 px
difference instead of 4,407, keeps its two pages, and validates
against the ODF 1.3 schema. No other test document changed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
A turned text box (free rotation) drew its frame turned and its text
upright. Its text now turns with it about the box's centre, as Word
draws it, on screen and in PNG and PDF export: upside down when the
box is flipped across, never mirrored by a side-to-side flip.
PDF export draws item by item, to tag each paragraph, so it turns each
of the box's lines the same way, and tags the turned lines as their
paragraph's text rather than as decoration.
Clicks and the caret inside a turned box still go by its upright text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Three causes of the largest remaining LibreOffice round-trip
differences:
- In-line frames and shapes centred on or hanging from the baseline
(`style:vertical-pos` `middle`, `bottom`, or `from-top` with
`svg:y`, relative to the baseline) were read as standing on it. They
are now read as their run's lowering, which the layout honours, and
written back as `from-top`.
- Leading: LibreOffice also leaves the leading after a paragraph's last
line. It is now added to the paragraph's space after.
- Text box frames didn't derive from LibreOffice's Frame style, without
which LibreOffice ignores a frame's `fo:border` and outlines it in
its default shape colour. They now do, and an empty Frame style is
declared.
LibreOffice round trips (pixel difference): shapes 13,864 → 1,273,
spacing 2,973 → 706, text box 736 → 0. No other test document changed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
An in-line picture, shape or chart always stood on the baseline,
whatever its run's position. Word moves an in-line picture with its
run's position (`w:position`, Font › Advanced › Position):
- lowered, it reaches below the baseline (at most until its top is on
the baseline) and the line's descent grows to hold it;
- raised, it stands clear of the baseline at its own height.
A drawing's cluster now reaches `obj_h` above the baseline and `obj_d`
below it (negative when raised). It is painted, hit-tested and wrapped
from both.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
ODT form controls (ODF 1.3 §13: controls in `office:forms`, each
placed in the text by a `draw:control`) were dropped on open, values
and all. They now open as WordCraft content controls, each with its
value, entries and selection; its name becomes the tag and its label
the title:
- text fields and text areas: plain text controls;
- check boxes: check boxes;
- radio buttons: check boxes with round symbols;
- list boxes: drop-down lists;
- combo boxes: combo boxes;
- date fields: date controls.
Saved as DOCX they become Word content controls.
Content controls of those kinds in the body are written back as form
controls with a `draw:control` anchor; in headers and footers they stay
their text. LibreOffice shows the round-tripped controls with their
values, and the content validates against the ODF 1.3 schema.
Not read: buttons, images, hidden and other controls. Controls' sizes
and looks aren't kept.
Flow model: new `Inline::Control`, kept only by flows that ask for
boxes (ODT); other formats get its text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
A password-protected ODT (ODF 1.3 Part 2 §3.4) failed to open. The
package is now decrypted with the password into a plain one, which the
reader opens as before.
- Supported: AES-256-CBC with a SHA-256 or SHA-1 start key, PBKDF2
with HMAC-SHA-1, and SHA-256 or SHA-1 1K checksums (LibreOffice's
encryption).
- Blowfish (the old default) is refused with a message naming it.
- A missing or wrong password gives the same errors as Word files, so
the app asks for the password.
- Hostile parameters are bounded: iteration count, IV size, ragged
data, inflated size.
Saving with a password (and saving a document opened with one) writes
the ODT encrypted as LibreOffice does. Every file but the `mimetype`
and manifest is deflated and encrypted, each with its own salt and IV,
and the manifest validates against the ODF manifest schema.
LibreOffice opens WordCraft's encrypted files with the password and
refuses a wrong one; WordCraft opens LibreOffice's.
No new dependencies: the AES, SHA, HMAC, miniz and getrandom crates
the workspace already uses for Word encryption.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Review of the objects commit found two gaps.
- The header and footer reader wasn't given the package's entry list,
so embedded objects there were still lost on open. It now shares the
body reader's list and manifest, borrowed rather than copied.
- Headers and footers were flattened for writing without drawings, so
shapes, text boxes, charts and kept objects in them were dropped on
save. They now go out as in the body.
The full LibreOffice round-trip harness shows no change on the other
test documents.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
LibreOffice writes an embedded object's replacement picture in its own
metafile format, which WordCraft can't decode. So an embedded
spreadsheet (`draw:object`, ODF 1.3 §10.4.6) or OLE object
(`draw:object-ole`, §10.4.7) vanished on open, and saving lost it.
Such objects are now kept whole: their files, media type (from the
manifest or the sub-document's body) and replacement picture.
- Display: the replacement picture when WordCraft can show it, else a
placeholder of the object's size drawn in code (ruled for a
spreadsheet).
- ODT save: written back as the same object, with its manifest entries
and original replacement. LibreOffice renders a round-tripped
spreadsheet pixel-identically.
- Document model: the files live in `passthrough` under `odf-object/`,
linked from the picture's `ole` by a marker. The DOCX writer skips
the marker and writes the picture alone.
Flat ODT, a fix to the chart commit: inline sub-documents lacked the
required `office:mimetype`, so the schema rejected flat files holding
charts. Sub-documents now go inline as full `office:document`s (media
type, styles, settings) and are split back into their files on
opening. Replacement pictures and OLE files go inline as base64. A
picture's data is matched to the element that holds it, no longer to
the last `draw:image` before it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Review of the chart commit: the engine drew opened charts only at the
paragraph level, so a chart that was a member of a `draw:g` group
opened blank. The pass now goes into groups (bounded depth).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
A chart object (`draw:object` whose sub-document is an `office:chart`,
ODF 1.3 §10.4.3, Part 3 §11) used to open as its replacement picture.
It is now read into WordCraft's chart model: column, bar (stacked or
not), line, area, pie, doughnut and scatter charts, with categories,
series names and values (gaps included) taken from the chart's own
table through the series' cell range addresses, plus title, legend
position and data labels. The engine draws opened charts in the
document's theme colours, and Chart Design's commands edit them.
Charts WordCraft can't edit as they are (3D, percent-stacked, mixed
plot types, a second value axis, radar and the like) keep their
picture.
WordCraft charts are written as ODF chart objects shaped like
LibreOffice's: a data table, each series and slice in its accent
colour, pies clockwise from the top. They validate against the ODF 1.3
schema, and LibreOffice draws them like WordCraft does. In flat
documents they are written inline as an `office:document`.
Bug fixed: flat ODT reading ended the body at the first nested
`</office:body>`, so an inline sub-document dropped the text after it.
Element lookup now follows nesting, with a bounded depth.
Flow model: new `Inline::Chart`, kept only by flows that ask for
boxes (ODT); other formats still get the alt text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
LibreOffice's leading (`style:line-spacing`, extra space between lines)
has no WordCraft line spacing mode. It was read as a multiple of a
12 pt line, which spreads the extra space differently from LibreOffice,
so the paragraph started too high. It is now read as at least a 12 pt
line and the leading: the extra space goes above each line, as
LibreOffice draws leading (it also leaves some after the last line,
which isn't followed).
A LibreOffice spacing document went from 6,190 px apart after a round
trip to 2,973, and WordCraft's drawing of it from 8,570 to 5,521.
Regression test: `leading_is_at_least_a_line_and_the_leading`.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Tracked formatting changes (`text:format-change` regions, ODF 1.3
§5.5.7, between `text:change-start`/`-end` marks) were dropped. The
text between the marks now gets a formatting revision with the change's
author and date. ODF records no old formatting, so the old formatting
is the text's own: the revision shows and can be accepted, and
rejecting it changes nothing (LibreOffice can't reject one after
reopening either). WordCraft's formatting revisions are written as
format changes around their text.
Fixed: changed regions were written with `text:id` only, but ODF 1.3
requires `xml:id` (which the change marks refer to), so any document
with tracked changes failed schema validation. Both are written now,
and a region named by `xml:id` alone is read.
A LibreOffice document with a tracked formatting change round-trips
pixel for pixel; documents with tracked insertions, deletions and
format changes validate.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Drop caps (`style:drop-cap`, ODF 1.3 §17.9) are read into WordCraft's
drop cap (lines and distance; one line is none) and written back as
the paragraph style's `style:drop-cap`. WordCraft drops one letter, so
a drop cap of several letters or a word opens as its first letter.
Fixed: the paragraph style key kept its child elements after a
separator, and the background, borders, contextual spacing and
right-to-left direction were added after the tab stops, so a paragraph
with its own tab stops wrote them into the tab stops' place instead of
onto the paragraph properties, and lost them. Attributes and child
elements are now gathered apart
(`tab_stops_and_other_paragraph_properties_together`).
A LibreOffice document with a drop cap round-trips pixel for pixel.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
A flat ODT holds the whole document in one `office:document` element
(ODF 1.3 Part 3 §2.1.2), its pictures inline as base64
`office:binary-data` and its formulas inline in their `draw:object`.
It is opened by taking it apart into a package's files (content,
styles, meta, settings; pictures and formulas as entries, each part's
root keeping the document's namespaces) and reading that package, and
saved by putting a package's files together: one set of automatic
styles, every part's namespaces declared once on the root, pictures and
formulas inline without the link attributes.
`.fodt` opens and saves from Open, Save As, the CLI and the web app.
LibreOffice's own flat documents round-trip as their packages do, and
the written files validate against the ODF 1.3 schema.
The parity doc's ODT fidelity is re-estimated at ~80%, and its status
line, which hadn't been updated since the Asian fonts change (a text
replacement that matched nothing), now lists every change since.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
- Page number format: the page layout's `style:num-format`, and a page
number field's own (`PAGE \* Arabic`, `roman`, `ROMAN`,
`alphabetic`, `ALPHABETIC`), both ways.
- Page colour (the first page style's `fo:background-color`).
- Page borders: ODF's sit at the margin with the text inside their
padding; WordCraft measures a page border from the page edge and the
text margin from the edge, so the border's space is ODF's margin and
the margin takes in the border and padding, and back on save.
- Footnote and endnote numbering (`text:notes-configuration`: format,
first number, counted from 0 in ODF, restart, page foot or document
end); only what the document sets is written.
- Line numbering (`text:linenumbering-configuration`, on unless it says
otherwise, ODF 1.3 §19.845: interval, distance, restart on each page).
A LibreOffice document with all of these round-trips pixel for pixel;
WordCraft draws it 7,213 px from LibreOffice (24,025 with the border at
the page edge). The written styles validate against the ODF schema.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Headers and footers were put into the document under the ids their
sections named (1, 2…), assuming the body had made no stories of its
own; but notes, comments and text boxes now are stories made first, so
in a document with a footnote and a footer the footer took the note's
id and the note showed the footer's text (a page number field).
Headers and footers now take free ids and the sections are pointed at
them.
Regression test: `a_footnote_keeps_its_text_beside_a_footer`.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Fields that were read as plain text now open as Word's, so Update Field
recomputes them, and are written back as ODF's:
- dates and times in their date or time style (`number:date-style`,
`number:time-style`) ↔ `DATE`/`TIME \@ "picture"`: day, month (by
number or name), year, weekday, hours (12-hour with AM/PM), minutes,
seconds and the text between, both ways; data styles are written as
automatic styles and the `number` namespace declared;
- sequences numbering figures and tables (`text:sequence`) ↔ `SEQ` with
its number format; on save the sequences are declared
(`text:sequence-decls`); on open a bookmark named by the sequence's
reference name covers the caption's label and number ("Figure 1"),
what references to it show;
- chapter fields (`text:chapter`) ↔ `STYLEREF "Heading n"`, by name or
number (`\n`);
- reference marks as bookmarks, and references to them and to sequences
(`text:reference-ref`, `text:sequence-ref`) as `REF`/`PAGEREF`.
Captions renumber, references to them and chapter fields refresh after
Update Field (engine test `odt_fields.rs`). A LibreOffice document with
all of them round-trips 13 px apart with the same text; the written
content and styles validate against the ODF 1.3 schema.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Word's date-time pictures (`DATE \@ "…"`) put literal text in single
quotes, two quotes in a row standing for one: `H' o''clock'` shows
"8 o'clock". The formatter read the quotes as text and the letters
inside as instructions (an `a` became AM or PM). Quoted text is now
shown as it is; an unclosed quote runs to the end.
Regression test: `date_pictures_keep_quoted_text`.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
A bookmark over text (`text:bookmark-start`…`text:bookmark-end`, in one
paragraph or across several) was read as a point at its start, and a
Word bookmark over text was written as a point: a reference to it then
updated to nothing. The flow model gets the end of a bookmark
(`Inline::AnchorEnd`, in flows asked for fields, as ODT's is), the
document gets the bookmark's start and end where they are, and the
writer writes a start and an end (a point, `text:bookmark`, when it
ends where it starts). A bookmark started twice, ended twice or ended
before it starts is kept once, as a point when it has no end after it.
References to bookmarks, LibreOffice's references to headings included,
now update to the bookmarked text (engine test `odt_fields.rs`, through
a save and reopen).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Tables of contents, tables of figures (`text:illustration-index`,
`text:table-index`), alphabetical indexes and bibliographies open as
WordCraft's: Word's `TOC` (levels from the source's
`text:outline-level`; `\c` and the caption name for figures), `INDEX`
and `BIBLIOGRAPHY` fields over their entries as last generated. With a
title the field goes in the title paragraph, as References › Table of
Contents inserts one; without, it spans the entries (new flow inlines
`FieldStart`/`FieldEnd`, mirroring the document's range fields). Update
Table then regenerates them from the headings (engine test
`odt_toc.rs`, titled and untitled, through a save and reopen).
LibreOffice's Contents 1–10, Contents Heading and Figure Index 1 styles
are WordCraft's TOC 1–9, TOC Heading and Table of Figures, both ways,
also when a file names them without defining them. WordCraft's tables
are written as ODF indexes with their source and entry templates, so
LibreOffice updates them, the entries' links in the Index Link style.
User and object indexes are read as their paragraphs; index marks are
not read yet.
Found on the way, both shown by the table of contents' round trip:
- a style based on LibreOffice's generic Heading style (as its headings
and index headings are) lost that parent on save, and with it its
font and spacing;
- a paragraph's tab stops were measured from its own indent only, not
from an indent its style gives (LibreOffice's `TabsRelativeToIndent`),
and a named style's stops not from its indent at all.
A LibreOffice document with a table of contents, a table of figures and
an alphabetical index round-trips pixel for pixel (it was 29,717 px
apart and a page short). The parity doc's ODT fidelity is re-estimated
at ~72%.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
A pure move, no behaviour changed: `odt/body.rs` (1,720 lines) and
`odt/writer.rs` (1,716) are split into child modules by topic, the
methods of `Body` and `Writer` moved in `impl` blocks of their own.
- body.rs (930): paragraphs, text, lists, element dispatch;
body/drawing.rs frames, shapes, groups, formula objects;
body/notes.rs notes, comments, fields; body/tracked.rs tracked
changes; body/tables.rs tables.
- writer.rs (618): the writer, paragraphs, inlines, blocks, styles;
writer/props.rs text and paragraph style properties;
writer/drawing.rs pictures, text boxes, shapes, groups, formulas;
writer/tracked.rs tracked changes; writer/notes.rs notes and
comments; writer/tables.rs tables.
Items the rest of `odt` uses are re-exported where they were. Every
harness document (21) exports content.xml, styles.xml and settings.xml
byte for byte as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Polygons and polylines (`draw:points`), paths (`svg:d`) and regular
polygons and stars (`draw:corners`, `draw:concave`, `draw:sharpness`)
open as WordCraft freeforms in their view box (none: the box around
their points) instead of being skipped; circles (`draw:circle`, by size
or by centre and radius) open as ellipses.
The path data reader follows SVG path syntax: moves, lines, horizontal
and vertical lines, cubic and quadratic curves and their smooth forms,
and elliptical arcs (converted from endpoint to centre form), absolute
and relative, with repeated commands; curves are followed in 16 straight
steps, a subpath closed by `Z` is a closed path, and input is bounded
(numbers, points, paths, coordinates).
WordCraft's freeforms (ink included) are written as `draw:path` in a
view box of hundredths of a point (whole numbers, as ODF's view box is).
A LibreOffice document with a polygon, a polyline and a closed Bézier
curve (relative `c` and smooth `s` commands) round-trips 333 px apart,
and WordCraft draws it 529 px from LibreOffice; the written content
validates against the ODF 1.3 schema.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Graphic styles' arrowheads (`draw:marker-start`/`-end`), line dashes
(`draw:stroke-dash`) and gradient fills (`draw:gradient`), defined by
name in the styles, open as WordCraft's arrowheads, dash patterns and
gradients:
- dashes: up to two kinds of dash and one space, in line widths or
absolute lengths, matched to WordCraft's named dashes when equal;
- gradients: linear and axial along their angle (ODF's 0° runs top to
bottom, anticlockwise; WordCraft's left to right, clockwise; an angle
without a unit is tenths of a degree, as ODF 1.2 wrote it), radial,
ellipsoid, square and rectangular out from their centre (end colour in
the middle), with border and intensity.
They are written back as definitions in `styles.xml` (WordCraft's own
triangle arrowhead marker; equal definitions share a name; the header
writer's names are prefixed apart). Frames keep gradient fills too.
A line now keeps the direction it is drawn in (`svg:x1`/`y1` is its
start): one drawn leftwards or upwards is flipped that way, so its
arrowheads stay at their ends, and it is written from its start.
A LibreOffice document with arrowheads, dashes and gradients round-trips
215 px apart; WordCraft draws it 518 px from LibreOffice. Exported
styles.xml and content.xml validate against the ODF 1.3 schema.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
The renderers draw a line from its frame's top left to its bottom right
(DrawingML's `line`; flips turn the frame), but the arrowhead code and
`end_direction` assumed a line from the bottom left to the top right, so
a sloped line's arrowheads were drawn at the other diagonal's corners,
apart from the line. Horizontal and vertical lines hid it. The
arrowheads now start from the line's own ends, pointing along it.
Regression test: `tests::a_lines_arrowheads_sit_at_its_ends`.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Groups (`draw:g`) of shapes, text boxes and pictures open as one
WordCraft group instead of being skipped: anchored and wrapped as the
group is, over the box around its members, each member at its place in
it (from its `svg:x`/`svg:y`, its line ends or its `draw:transform`,
turned members turned); groups inside a group are flattened into it.
WordCraft's groups are saved as `draw:g`, members at their offsets with
no anchor of their own, nested groups nested (a canvas as a plain group).
A frame, shape or group whose styles state no `style:wrap` now has text
on both sides, as LibreOffice wraps it (ODF 1.3 §20.400 states no
default), not above and below only; `biggest` is read as the larger
side. A shape whose only paragraph is LibreOffice's empty `<text:p/>`
gets no empty text story.
A LibreOffice document with turned shapes and a group was 7,464 px apart
after a round trip; now 586, with the same text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Shapes and pictures turned by `draw:transform` (rotate, translate and
scale, applied in turn to the object's own coordinates) open turned: the
file's anticlockwise turn about the object's top left corner, followed by
a move, becomes WordCraft's clockwise turn about the centre of the
unrotated box at its offset, as Word places them. They are written back
the same way, in line as well as floating. LibreOffice writes no
`svg:x`/`svg:y` for a turned shape, so before, a turned shape lost both
its turn and its position.
Custom shapes keep their mirroring (`draw:mirror-horizontal`/`-vertical`
on their geometry), and a turned line is written by its turned ends. The
flow keeps the placement of an in-line object that is turned or mirrored.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Drawing shapes (`draw:rect`, `draw:ellipse`, `draw:line`, and
`draw:custom-shape` of a predefined geometry named by its
`draw:enhanced-geometry` `draw:type`) open as WordCraft shapes instead of
being skipped: size, placement (in line or floating, as frames are),
fill and line from the graphic style chain and the default graphic style
(LibreOffice's solid fill and hairline where nothing is stated), their
text, and a line's direction from its two ends. LibreOffice's type names
and its `ooxml-` preset names map onto WordCraft's kinds; a geometry
WordCraft has no shape for opens as a rectangle.
WordCraft's shapes are saved the same way: a custom shape by its type
alone (LibreOffice rebuilds the geometry from it), a line by its ends,
connectors as lines. A shape is saved with `draw:auto-grow-*` off unless
it resizes to fit its text, so LibreOffice keeps its size.
The flow model's text box becomes `Inline::Shape` with a kind; text
boxes are the `TextBox` kind and are written as frames, as before.
Not yet: shapes in groups, polygons, paths, rotation, arrowheads,
dashes, gradients.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
A text box frame's line (colour and width), fill and padding are read
from its graphic style and that style's parents (LibreOffice's Frame
style included) instead of a fixed thin black line on white, and written
back. ODF's defaults (no line, fill or padding) apply where the styles
say nothing. A frame without a height, or with a `fo:min-height` on its
text box, opens as a box that resizes to fit its text and is saved with
a least height again. A LibreOffice text-box document was 1,117 px apart
after a round trip; now 736.
Layout: a floating text box that resizes to fit its text now takes its
text's height (drawn and wrapped around), as inline ones already did;
before, it kept its stored height (an ODT frame without one showed as a
1 pt line with the text spilling over the paragraph).
A formula sub-document's content.xml stays a bare `math:math` root, as
ODF 1.3 Part 3 §2.2.9 B requires (pinned by a test), and the reader now
skips a wrapping `office:math` some writers add instead of taking it for
the MathML root.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
The Asian and complex-script fonts (`style:font-name-asian`/`-complex`,
through their font-face declarations, or `style:font-family-*`), the
complex script's size, weight and posture, and the Asian and complex
languages (`style:language-*`, `style:country-*`) are read on runs and
from the default style, into the run's and the document's
character properties, and written back. Hebrew, Arabic and CJK text
lost its fonts on a save (an RTL test document: 2,043 px apart in
LibreOffice after a round trip, now 529).
A run's own complex-script weight or posture is written once, not also
copied from its Latin bold or italic.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
`style:writing-mode` `rl-tb` (or `rl`; ODF 1.3 §20.404) on paragraph
properties opens as a bidirectional paragraph, from automatic and custom
paragraph styles alike, and is written back on the paragraph's or the
custom style's properties. Right-to-left paragraphs opened left to right
and left-aligned.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
A `draw:frame` holding a `draw:text-box` (ODF 1.3 §10.4.3) opens as a
WordCraft text box: a text box shape of the frame's size, in line or
floating as the frame is anchored and placed, whose story holds the
box's paragraphs (read as a note's are, with the paragraph and frame
around them put aside meanwhile, so a picture inside the box is a frame
of its own). The text was dropped on open.
Export writes each text box as a frame with a thin border and white fill
(`draw:fill` as well as `fo:background-color`: LibreOffice fills a frame
without `draw:fill` with its default blue). The module documentation now
lists what the reader and writer handle.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
LibreOffice gives a heading style the outline's numbering unless the
style names no list style (`style:list-style-name=""`, as LibreOffice
writes its own). Our heading styles named none, so a heading whose level
the outline doesn't number still took that level's default indent: a
Word-made document's headings sat indented after a save. Heading styles
of unnumbered levels now name no list style; numbered ones stay
attached. Round-trip of that document: 3,302 px apart, now 1,121.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Contextual spacing (`style:contextual-spacing`, ODF 1.3 §20.255: no
space between paragraphs of one style) is read from paragraph, custom
and kind styles and written back, and the List Paragraph style is
written like any other, so bulleted items keep their spacing; they were
written in Standard and spread apart.
LibreOffice switches many layout behaviours by `settings.xml`, and a file
without them is laid out its own way. A file's own `settings.xml` is now
kept through a save; a document from elsewhere gets the switches
LibreOffice uses for Word files, as WordCraft lays text out as Word
does. Tab stops are written from the margin either way. A Word-made
document saved as ODT drew 38,652 px apart from LibreOffice's own ODT of
it; now 3,302, with the same line breaks.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
A common text style other than LibreOffice's link, note-mark and
list-label styles (Emphasis, a user's own) opens as a document character
style with its display name, the custom character style it is based on
and its own formatting; the text in it names it, and only what the
spans' automatic styles add is direct formatting. Before, the style was
lost and its formatting copied onto the text. An id a paragraph style
already has gets "Char" added.
Export writes each custom character style the text uses, and those it is
based on, as common text styles; a span's automatic style is based on
its style and holds only the run's direct formatting. LibreOffice keeps
the style when it saves the file. Automatic styles of styles.xml are
never taken for custom styles.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Each list level's indent and hanging (`fo:margin-left` and
`fo:text-indent` of `style:list-level-label-alignment`) and its label's
alignment (`fo:text-align` of `style:list-level-properties`, ODF 1.3
§17.19-17.20) are read into the document's list levels, for lists and
heading numbering, and written back. Every list was written 36 pt a
level with left-aligned numbers, so LibreOffice's right-aligned numbers
and closer bullets moved on a save (lists test document: 658 px apart
in LibreOffice after a round trip, now 8).
The flow label keeps the position in twentieths of a point, its prefix
and suffix now up to 8 bytes ("Chapter " still fits), so list
paragraphs stay as small as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Each paragraph's style chain was searched for its custom style by
building every custom style in it (cloning its tab stops and the rest)
just to see whether there was one. A cheap check now decides; the style
is built once, by the import that adds it to the document. A test opens
20,000 paragraphs on a 15-deep chain of styles with 60 tab stops each.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
A common paragraph style other than Standard, the paragraph kinds' and
the note texts' (LibreOffice's Text body, a user's own style) opens as a
document style with its display name, the style it is based on (a
custom parent, a kind's style or Normal) and its own formatting:
spacing, indents, line spacing, alignment, keep with next, borders,
background, tab stops and text formatting. Its paragraphs stay in it;
only what the automatic styles over it add is direct formatting. Before,
the style was lost and its formatting copied onto every paragraph.
Headers, footers, notes and comments are covered too.
Export writes each custom style the paragraphs use, and the custom
styles they are based on, as common styles named by an XML-safe form of
the style id and shown by the style's name; LibreOffice keeps them when
it saves the file. A custom style may not take a name the writer uses
for its own styles (Standard, Footnote, Endnote, the kinds'), so no
style is written twice.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GBVQLyDnrG8Tn1yDQcUabC
Claude