The picker lists .odt and .docx, and the chosen extension decides which
exporter runs and which format the document keeps — no silent switch to
.odt for a file that came in as .docx. Both save paths now build their
bytes through one helper instead of repeating the argument list.
The heading constants move to styles/headings.ts and the filename logic to
storage/documentName.ts, which were all the editor needed of the exporter
at startup. Main chunk 2.1 MB → 1.37 MB (gzip 615 KB → 411 KB).
The corpus legs pass seven of the exporters' nineteen arguments, so the other
twelve — note configuration, page decoration, line and page numbering, the
header/footer zones, document properties — were never round-tripped from a
file. A new unit test sets all of them and reads them back out of both formats.
The kitchen sink had only ever been written, never imported; it now makes both
round trips and comes back identical. Two fixture repairs it needed: notes in
anchor order, and a section break with sections to open.
Caught: a document holding no note yet lost its note configuration on .docx.
The corpus leg out of one format and into the other compared block types and
text, so anything the trip dropped below that passed. It now compares every
attribute each node carries, with the values a format states in its own units
excluded: proportional vs absolute column widths, and a formula's inline flag,
which ODF has no place for.
Two the stricter leg caught: a note lost the paragraph style it names and fell
to the stock 10pt, and a link stored an attribute equal to its default.
A header row is w:tblHeader on the row, the same thing ODF spells with
<table:table-header-rows>. An index's leader and tab stop ride its first row,
the level styles are a family per kind, and the stop is written even with no
page number running to it. A box carries its own ring (w:bodyPr/@lIns) and
where it sits across the column when behind the text.
A section past the first now names its zones even when blank: a sectPr
without a headerReference is Word's Link to Previous, which put a chapter's
running head on the pages a blank section was meant for.
Chapter numbering rides the heading styles' w:numPr, so re-import read every
chapter as a list item — which swallowed the page break in front of it, since
the break is read for body paragraphs only. Headings now also write their own
w:outlineLvl, so a document setting its chapters in a style of its own
("Appendix 1") keeps them headings. SVG goes out rasterized: Word reads none,
and the importer makes one of every EMF too.
The corpus cross-format leg now carries the file's own stylesheet, where both
the numbering and a heading's style live; measured on one fixture, the leg
ODT → .docx → back loses 0 of 34 headings and 0 of 7 page breaks, was 7 and 7.
Five legs in tests/lo-password.test.ts, driven through the repo's first UNO
helper because --convert-to cannot pass a password; self-skips without soffice
or python3-uno and runs with npm run test:lo. The probed format details live in
docs/architecture/encryption.md.
One button per kind shows or hides it and its caret says where its cards go —
balloons or the pane, which is why the two pane toggles are gone. Display for
review carries Word's four modes beside them, and the accept / reject / walk
pairs sit under each kind. The PDF path takes the mode as page attributes, so
a printout of a page in no markup reads as the page does.
The list quoted its text but named no place, so on a five-page printout the
entry was a needle. It now leads with the page, the format LibreOffice prints
its own end-of-document comments in. Only the raster paths can say it — they
print the editor's own pagination, the vector one re-paginates in the browser.
commentsHeading gives way to commentLabels, since the page needs a word too.
The Review tab's "Print markup", on by default and stored only when off
(edentext-print-markup). Off, the raster paths drop the bar layer from the
clone and skip the comment pages, the vector path emits neither the markup CSS
nor the list.
printPdf rebuilds its options field by field, so it never saw commentsHeading
either — the printed list was headed in English whatever the UI language.
The bar in the margin beside every changed or commented block, and the comment
bodies as a list after the document — a comment's text is nowhere on the page,
so a bar alone could only say that one exists. Resolved comments print nowhere,
and a document with no markup prints exactly as before.
The print engine clips to the page area (measured), so the vector path buys the
bar's strip out of the left @page margin and gives it back as padding. The
raster paths clone the editor's own layer and append the list as whole pages.
A TOC field regenerates its rows from the styles named `toc 1`…`toc 9`, and its
cached rows are skipped on import — so those styles were never seen and the index
fell back to the editor's 0.5cm per level, dropping the file's space-after with it.
Measured on a 60-row index: a 5.1mm row pitch against LibreOffice's 6.9mm.
They now land as Contents 1…10, and the DOCX export writes the rows back under
them instead of the fallback indent.
textBox becomes an inline node carrying block content, like the frame both
formats write, so "in line with text" really is in the line and a box reaches
a table cell or a list item through their paragraphs. Both exports write it
where it sits (an as-char draw:frame / a wp:inline drawing) and both imports
leave it there, which drops the hoisting machinery on either side.
Two guards the schema cannot express: no box inside a box, and a selection
reaching into a box grows to the box's bounds rather than dissolving it.
wrapAlign only ever came from a file, so a box inserted here stayed at
the left in both modes that leave it in flow. The floating toolbar now
offers left/centre/right for `inline` and `topBottom` (a side wrap names
the side itself) and drops a coordinate placement when one is picked.
DOCX carried neither: `topBottom` now writes wp:align and an as-char box
the w:jc of its rebuilt marker paragraph, both read back on import. ODF
already wrote both. Measured centred at 297px of a 76…718px column, and
rendered through LibreOffice from .odt and .docx alike.
Without a parent chain to the Frame style LibreOffice read the box as a
drawing object, whose text loses lists, nested images and borders. The
frame takes fo:background-color/fo:border and its writing mode in the
graphic properties; an inline box's centring rides the anchor paragraph.
The import reads the frame spellings back and drops LO's from-top/0
respelling of the default alignment.
The three index kinds exported as fields with an empty result; Word for
Mac never updates fields on open, so every index rendered as nothing.
They are now complex fields whose result is the editor's cached entries
(indent per level, leader tab, pages), the form Word itself writes; the
importer recognizes the complex INDEX/BIBLIOGRAPHY form and keeps the
attrs the fldSimple path set.
Three legs against the closed-loop blind spot the duplicate-style bug
exposed: package-lint.test.ts checks invariants the schemas cannot
express (unique style ids/names, dangling style/num/rel references,
balanced ranges, manifest completeness); lo-roundtrip gains a
soffice→fodt leg asserting the styles LibreOffice resolves (Heading
chain Arial/bold/sizes, Word defaults); lo-render-consistency.test.ts
renders the kitchen sink from both formats through LibreOffice and
compares pagination and pixels (ratchet thresholds). CI's LO job
gains poppler-utils; test:lo runs both LO files.
The docx package always emits its own Title and Heading1-6 styles; our
same-id copies in paragraphStyles made styles.xml define them twice, and
both Word and LibreOffice then dropped the basedOn chain — headings
rendered serif, regular, at the raw size. buildStyles now feeds those
seven through styles.default.title/headingN, leaving a single
definition with the chain intact. Regression test asserts unique
styleIds and Heading1's basedOn.
ODF: every part declares 1.3 (the version LibreOffice writes; header-first
is a 1.3 element), draw:image gains the mandatory xlink:type trio,
text:time-value becomes an xsd dateTime (import still reads the legacy
duration), index-entry-link-start/-end stay in TOC templates only, and
dont-balance-text-columns moves onto style:section-properties (probed: LO
reads and re-writes it there). DOCX: a final pass re-sorts w:settings into
the CT_Settings sequence, and comments.xml declares w14 mc:Ignorable.
effectiveListLevel resolved a level only for its own node species, so a
Tab-nested <ul> under a number level kept cycle bullets while the named
ODF definition (and LibreOffice) numbered it. The level's kind now wins:
the walk numbers a <ul>/bullets an <ol> (CSS keys generalized), both
exporters resolve clones and forks through eff.kind (new ODT pass
applyListLevelKinds switches L# level species), and setListStyle retypes
the subtree. Node marker attrs still override with their own species.
One shared abstract per style (each list its own w:num instance, so numbering
restarts); a post-pack pass injects w:styleLink and the numbering style's real
numId, which the docx library cannot express. Import follows w:styleLink and
w:numStyleLink — a linked abstract used to resolve to {} and silently import
as a bullet list — maps the style onto listStyleName + sheet.list, and derives
no per-level attrs. Probed: LO ignores DOCX numbering styles both ways (mints
WWNum# on import, flattens to the look on export); ODF carries the identity.
Fourth style family (LibreOffice's Listenformatvorlagen): per-level model in
styles/listStyles.ts, five probed LO built-ins, listStyleName on the outermost
list, one effective-level resolver shared by the decoration walk and the export.
ODF writes the full named text:list-style into styles.xml and repoints
unoverridden lists at it (overrides keep the resolved automatic clone);
import maps named list styles back to the attr + registry, attr-free.
Probed: LO keeps the named reference and display-name across a re-save.
An inline atom drawn gray as ‹label›: a click selects the whole field,
typing replaces it, Tab/Shift-Tab jump between fields. Exports to ODF
<text:placeholder> (PLH sentinel U+E022) and a DOCX w:sdt content
control tagged edentext-placeholder; both importers read it back, the
export-minted gray stripped. Covered in roundtrip Leg 31 + fuzzDoc.
The fallback download shows a one-time hint about the browser's
ask-where-to-save setting, keeps its anchor in the DOM and the blob
URL alive past the click (Gecko). Failed saves and exports now report
the underlying error, with a clear message when a script blocker
prevented a lazy chunk from loading. A full or blocked localStorage
no longer fails the save when recording the recent-files entry.
Word's in-front-of / behind-text (wp:wrapNone + behindDoc) and ODF's
style:wrap="run-through" both read as a side float, so a near-column-wide
picture reserved its whole height and pushed the text 23-52mm down. They
are a wrap mode of their own now: out of flow at the static position it
was anchored at, z-order from the file (inFront). Both exports carry it
back; tests/unit/wrap-through.test.ts covers the round trip.
The image fixture 72 -> 38 over both formats and 7 pages -> 6, matching
the reference; corpus 1108 -> 1074, no other fixture moved.
Each of the ~30 post-processing passes re-zipped every entry (images
included) at level 6 just for the next pass's unzipSync. Intermediate
archives are now stored; zipFinal compresses once at the end, keeping
mimetype first/stored and storing already-compressed picture formats.
Measured ~10x on an image-heavy document (853ms -> 87ms).
A new document starts with it off and an import adopts the file's own
flag, so an editor left recording no longer marks the next document it
opens. Both formats carry it, probed against soffice: ODF on the change
registry (text:track-changes, written for the flag alone where nothing
is recorded yet), Word as w:trackRevisions in settings.xml — LibreOffice
writes and reads exactly those two, and reading w:trackChanges instead
would have found nothing.
Both come from the document's language, read off Intl: the symbol's side
and spacing, the short date's order and padding. Each minted data style
names that language — without it LibreOffice prints its own locale's
separators (probed: a German document's 1.234,00 € came out $1,234.00).
A date is a serial day from LibreOffice's day 0, 1899-12-30. Word's
formula dialog offers no date, so the DOCX leg uses the field's \@
picture. Also drops a document's name from three older comments.
Word's Text Direction and LibreOffice's, on the frame's own toolbar: the
browser lays the vertical flow out, so the box needs nothing but the mode.
ODF keeps it in the frame style's *paragraph* properties — probed: in the
graphic properties LibreOffice drops it, there it keeps it and lays the box
out — and Word in the shape body's `vert`, the VML fallback's
`layout-flow:vertical` read too. Both survive a LibreOffice re-save.
The Formula dialog gains the number format Word's has: whole numbers, two
decimals, grouped thousands and the two percentages beside the general
one.
ODF keeps it on the cell *style* as a reference to a data style — odf-kit
shares one cell style across every cell, so a formatted cell gets a copy
of it — and LibreOffice writes both back unchanged. Word keeps it on the
field as its \# switch; LibreOffice's DOCX filter neither reads nor writes
that switch, so that leg holds through our own round trip.
Ruby, as both products' phonetic guide writes it: one element holding a
base and its reading, so the editor keeps it as one inline atom and the
browser sets it as a real <ruby>.
ODF gets <text:ruby> plus a minted ruby-family style (centred, above —
probed: LibreOffice writes that style back unchanged); Word a <w:ruby> run
from a post-pack pass, the package having none. Both legs verified through
a LibreOffice re-save, our .docx to .odt included.
HTML stops at h6, so 7-10 render as unknown elements editor.css makes
blocks; their built-in styles continue level 6 at 12pt, since LibreOffice
writes these styles with no properties and resolves them from its own
pool. Both importers read a two-digit level again (the style-name regexes
stopped at one digit). Verified through a LibreOffice re-save of both
formats and in the live editor.
LibreOffice's Format > Paragraph > Text Flow checkbox. Only "off"
travels (Word has no per-paragraph "on") and only under a document that
hyphenates, or it would say what is already true. ODF writes
fo:hyphenate="false" into the paragraph's text properties, DOCX gains
w:suppressAutoHyphens through a post-pack pass — the docx package
exposes no option for it. Both survive a LibreOffice re-save.
The formula rides the cell, as ODF writes it; Word's field syntax is what
is stored, since it is the wider language. `ooow:` in the ODF attribute
value is a namespace prefix — undeclared, LibreOffice reads it as part of
the formula and prints "Expression is faulty".
Sort takes one key and detects the type per cell where both dialogs make
you pick. Number recognition is off by default, as it is in LibreOffice,
and a formula parses a cell's text either way — exactly as there.
The citation is an inline atom holding the whole record, which is what ODF
writes on text:bibliography-mark — so no database travels beside the file.
The list is a new index family: one row per source, in document order and
without page numbers, as LibreOffice prints one.
DOCX cites by tag and keeps the sources in a custom-XML part the docx
package knows nothing about, so the post-pack pass mints the part, its
properties, both relationships and the content type itself.
References ▸ Index entry marks a place with the term the index should
list — under a key of its own where it belongs elsewhere — and the Index
in the same menu lists each term once, sorted case-insensitively, with
every page it was marked on.
Both formats write a place, not a range, so the entry is an inline atom:
ODF text:alphabetical-index-mark plus text:alphabetical-index, Word XE
plus INDEX. LibreOffice reads our file back unchanged. The bibliography,
the other half of that gap, is still open.
A .ott/.dotx is the built document with one label changed — ODF's
package media type, Word's content type for the document part — so the
picker offers both and the chosen extension decides which exporter runs.
Opening or saving also records the file in a recent list. The name is in
localStorage so the menu draws at once; the handle that reopens it is in
IndexedDB beside it. A file that has moved, or whose permission the
browser drops on reload, is reported and dropped from the list.
A list inside a box was flattened to literal bullet/number runs and a
picture was dropped: the box XML is a hand-built string, and neither had
the package parts it needs.
applyTextBoxesDocx already holds the whole unzipped package, so it mints
them — a media entry with its relationship, and an abstractNum/num pair
above the package's own ids, the paragraphs carrying a sentinel until
those ids exist. Verified through LibreOffice and back through our own
importer, nesting, marker symbols and start values intact.
Two marks carry a revision with its author and date: an insertion is
marked text, a deletion is text kept and struck through. Consecutive
typing joins one change; accept/reject work at the cursor or on all.
ODF splits the pair — text:change-start/-end around an insertion, a
text:change where a deletion was, its text in the tracked-changes
registry. DOCX uses w:ins/w:del with w:delText. Both round-trip, and
accepting every change gives the document an unrecorded edit would.
A caption is a paragraph in the Caption style holding a sequence field —
one counter per category, in document order, renumbered as the document is
edited and on load (a file's cached numbers are unreadable wherever the
format is not arabic). Round-trips as an ODF text:sequence with
LibreOffice's own ooow:Illustration+1 formula and a Word SEQ field.
Caption style values probed off LibreOffice: 10pt italic, 0.212cm around.
The block's own direction, set in the Paragraph dialog where LibreOffice
keeps it, so a quoted Hebrew or Arabic paragraph turns inside a left-to-right
document. Rides ODF style:writing-mode / Word w:bidi through cells and list
items as well as top-level blocks; the direction the page already has is
inheritance, so neither importer stamps it on every block.
LibreOffice's style pool, probed: Heading 6 is 12pt bold italic and
Heading 4 italic too. MAX_HEADING_LEVEL drives the extension registry,
the shortcut plugin, the TOC depth and both importers' yardstick, so
only the constant table and the CSS selector lists had to grow.
Bookmarks/cross-references move to Implemented — they already are.
The marker moves from a CSS ::before to a widget decoration: generated
content is no text node, so it reached neither the PDF nor a measurement.
The note also carries the paragraph style its file names, so font, size and
colour are the document's; only the hanging indent stays the editor's,
because the style model has no first-line half to pair with the margin.
Adds a unit test for the numbering and the settings, plus the architecture
doc and the two open parity findings.
ODF style:writing-mode="rl-tb" and Word's section-level w:bidi were not
read, so a Hebrew two-column page put column 1 on the left where
LibreOffice puts it on the right — checked against pdftoppm of its own
PDF, since the harness bands both columns into one line and never gets
that far. It is a document setting now, applied as `dir` on the body
host, which is all CSS multi-column needs; it also gives bidi the base
direction the file means. Round-trips through both formats. The
header/footer band keeps its own direction, as LibreOffice lays it out.
ODF style:page-usage="mirrored" and Word's w:mirrorMargins make the
declared left/right the inner/outer pair, which an even page swaps.
Read, rendered and written back in both formats; the body rides the
per-block inset a section's own margins already use, the header and
footer band moves with it. Measured on a mirrored two-column file:
its page 2 went from 12.7-184.1mm to 31.8-203.2, which is LibreOffice's
to 0.1mm, and its running head with it.
A positioned frame in a header is out of flow — the title-page background or
watermark. It keeps its wrap and its offset from the page corner, and the
layer paints it per page under the body instead of flowing it inline, where
it reserved a page-tall line of body space.
The page surface moves from .tiptap to .paper so the image has a layer
between the sheet and the text; ODF export writes it back page-anchored.
Bookmarks are a mark, not a point: both formats store a range, and only a
range lets a REF recompute its text. A cross-reference caches its display
like a Word field and resolves live on each pagination settle — the
bookmark's text, or the page it renders on — keeping the cached value when
the target is gone. Ctrl+click jumps to it, internal #name links too.
ODF splices its sentinels into the run text so cells and custom paragraphs
need no separate emitter; DOCX needs no post-pass at all.
A formula node stores only its LaTeX source; the MathML on screen, the ODF
formula object and Word's OMML are derived from it, and both importers parse
foreign markup back into that source, so an imported formula stays editable.
Word documents were losing their formulas outright: m:oMath sits beside the w:r
runs in its own namespace, past convertInline's `namespaceURI !== W` guard.
The default tab interval was fixed at 1.25cm, so a file's own
w:defaultTabStop / style:tab-stop-distance was ignored and the DOCX
export wrote 1.25 back over Word's 1.27. It is a document setting now.
A stop can also carry the character it repeats across its gap. odf-kit
has no leader option and mints one automatic style per option set, so
the leader rides style:type behind a sentinel — matching on position
and alignment alone gave every stop at 6cm the same leader, which
LibreOffice drew back plain as day.
The root file is read at every session start and had grown to 87 KB, so every
session paid for knowledge it did not use. Move the per-directory detail into
nested CLAUDE.md files, which load when a file in that directory is touched,
and the five deep-dive topics into docs/. The root keeps the commands, the hard
rules and an index of where the rest lives.
Content moved verbatim — the probed Word/LibreOffice behaviour, sentinel order
and shared constants are what the code cannot say. CI fails the root over 120
lines, since a prose budget alone is what let it grow.