w:tblStylePr — header row, banded rows, first/last column — was never read, so a
table styled in Word arrived as a bare grid. Import bakes each area into the cells
now (fill, borders, run marks): the file's own style has no registry entry to
re-derive regions from, and the bake is what a re-export carries.
A band's box is the whole banded region, so its insideH is the line between two
band rows; banding counts from the first body row, as in Word and LibreOffice.
Corpus 803 -> 795, page counts unchanged.
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.
text:display="none" now drops the paragraph its style hides and the stored
value of a field that carries the attribute — LibreOffice lays out neither.
A hidden heading is the exception and still renders: it is the outline the
running head's chapter field reads, and dropping it blanked every head that
quotes one. The editor has no way to hold a block it does not draw.
ODF carries western, asian and complex-script font properties side by side and
a word processor picks by the character's script; the importer read only the
western pair, so a complex-script run came out at the document default.
Decided per run — the finest split the model has — so a document with no
complex script is untouched, and our own export (all three written alike)
round-trips. Measured: a two-column page's line pitch 7.5mm -> 5.9 against
LibreOffice's 6.0.
Two compounding bugs on a title/back-page background image:
- its host paragraph picks up position:relative from the line-height
rule, becoming the frame's containing block instead of the page and
offsetting it by a full page margin — now forced back to static.
- the frame always painted behind text (z-index -1); a graphic that
declares ODF's style:run-through="foreground" now paints in front,
as LibreOffice draws it, via a new inFront attr round-tripped through
ODF export/import.
Confirmed against LibreOffice's own render: the duplicated, ~2cm
displaced cover graphic now matches exactly.
A sweep of the whole tree: 27 blocks ran to four lines or more. Trimmed
to what the code can't say, with two of them split back onto the
function each was really about.
Word paints a link through its Hyperlink character style, never for
being one, so the editor's own blue overwrote both the file's #0000FF
and the plain black of a link carrying no style at all. The DOCX
importer marks those `plain` and the CSS leaves them alone; LibreOffice
paints every ODF text:a, so those keep the visual.
Chromium colour-manages a CMYK JPEG through its embedded ICC profile
where LibreOffice and Word convert naively, so a fixture's
logo arrived dull teal on grey instead of vivid blue on black. The
importer drops the profile — measured back onto LibreOffice's own
values, and 655 KB smaller for that one logo.
Both exports wrote one geometry for the whole document, so a title page's
margins survived import and were lost on the next save. Each ODF section master
now names a page layout of its own — the document's, shifted by the section's
difference, so what the header/footer pass folded into it survives — and every
w:sectPr carries its section's w:pgMar. DOCX import takes the document pair from
the first section, the one .tiptap's padding draws.
The parity harness spread a 458-page document's word list into Math.max and blew
the stack; the longest fixture was never measured.
Page geometry was document-wide, so a title page declaring 1cm started 1cm too
low. w:pgMar / the section's ODF page layout now ride its HfSet and feed the
per-page content area; an ODF master handing over (style:next-style-name) gives
its own layout to the section's first page and the successor's to the rest.
A spacer can only push down, so a smaller margin than the document's needs the
opposite: the section's first block gets a lifting spacer where nothing was
pushed onto its page ahead of it.
The zone is one paragraph, so its mark sizes every line of the band. Runs
that agree on a font now set that strut (applyUniformRunFont, as the body
already does), the zone schema carries blockFontSize so it renders, and
hfReachPx measures a line by the family's natural height instead of a flat
1.15. A 10pt footer reserved 12pt lines and sat 1.6mm high on every page.
`wrapDist` carries Word's distL/distR and ODF's fo:margin-left/-right onto
the frame; a file that declares none wraps flush, which is what LibreOffice
draws (probed: text runs to the frame edge, and 0.5cm of declared margin
stops it exactly 0.5cm short). The hardcoded 14px was 2px more than Word's
own 0.32cm and broke a line one word early.
Whole pixels cost a page-wide picture 0.3mm of height, and the column cap
then scaled that loss into the width too. Sizes stay fractional (snapped
back to a whole pixel inside the export's own 0.001cm quantization), and a
sub-pixel overhang trims the width without touching the height — which is
what LibreOffice draws.
Repeating the current master was read as "no switch, no break", so a book
whose chapter heading style names its own page style ran the chapters
together — a fixture lost the break before every chapter. An explicitly empty
name now ends the parent walk, so nothing inherits a break it never asked
for; the section itself still starts only where the master changes.
A browser always kerns; ODF kerns unless told not to and Word kerns only
above the size w:kern names — nothing, in most files. Measured 5% on a
kerning-heavy string, enough to break a line a word early throughout every
Word document in the corpus. Carried on the style, since that is where
w:docDefaults and the default paragraph style put it.
fo:padding defaults to 0 in ODF, but an undeclared side was left open and
fell through to Word's implicit 0.19cm, indenting every cell of a table
whose file names no cell style — a fixture's figure captions by 1.9mm.
ODF style:may-break-between-rows="false" — probed: LibreOffice leaves four
lines of a page empty rather than split such a table, and splits the identical
one when the attribute says "true". pageBreaks now emits a single atomic leaf
for the wrapper unless the table is taller than a page, where the rule cannot
be met. Imported, exported and round-tripped; Word has no equivalent.
fo:min-height is a floor, not the band: LibreOffice grows it for the zone's
own lines and, in a footer, for the first paragraph's space above (probed —
a header's space below only hangs into the body). The zone paragraph now
carries those margins, so the reach counts them at its real font size and a
page with no zone of its own keeps the floor.
Word writes 1.08 into w:docDefaults, so a style declaring single spacing was
dropped as "the default anyway" and inherited 1.08 instead — every Title and
every heading that asks for 1 rendered a line too tall. The parent style is
the yardstick, which ownProps already applies.
ODF style:vertical-pos/-rel on an inline frame, which every LibreOffice
formula image carries: baseline, character area, or a plain offset from the
baseline. The default moves from the line-box bottom to the baseline, so a
line with a picture in it keeps the descent below it, as LibreOffice and
Word both do.
Mirrors the ODF fix: a block that cannot name its style is measured against the
default paragraph style but renders at its own, so a run matching the default was
suppressed and came out at the wrong size.
A list item is measured against the default paragraph style, but it renders at
its own style's size — so a 12pt run under an 11pt ListParagraph was suppressed
as "the style supplies it" and came out 11pt. Runs are now measured against the
size the block actually gives them.
The guidance document: 15 of 16 pages start on the same line, from 1.
The body margin reserved the zone's height *plus* its fo:margin-top/-bottom, so
every page with a footer lost a line — a fixture ran a page over. Probed against
soffice: moving a footer's fo:margin-top leaves the last body line exactly where
it was and only shifts the footer text down its band; style:dynamic-spacing
changes neither side. The band is the declared height, the gap sits in it.
One fixture paginates to LibreOffice's 63 pages again.
An as-char figure frame sits in a paragraph of its own; the importer lifts the
box to top level and dropped that paragraph whole, so its space above and below
went with it — every figure in a fixture sat 2.1mm too tight. The box now carries
them (ParagraphSpacing covers textBox), space above as padding so it adds to the
block above, and the export mints an anchor style to write both back.
One fixture: 142 → 100 parity issues.
ODF puts fo:margin-top/-bottom on the table style and LibreOffice honours
them like a paragraph's; we dropped both, so every callout table in a fixture sat
half a line too tight and the document ran a page short. With them the Math
Guide paginates to LibreOffice's 63 pages exactly.
The TOC rendered every level with our 0.5cm-per-level indent, an editor-only
1.5 line height and no space above — so a fixture's contents drifted 7.6mm off
LibreOffice down one page and pushed entries onto the wrong one. The per-level
entry template names a paragraph style; the rows now carry it as data-style
and the document stylesheet does the rest. Our own indents stay the fallback
for an index whose file names none.
A manual's footer carries text:chapter, which we imported as the string the
producer happened to cache — so every page of a fixture claimed the last chapter
of the book. It is now a live field: the layer resolves the last heading at
or above its outline level to have started by that page. Round-trips to ODF
text:chapter and Word STYLEREF; PDF keeps the cached name (Chrome's print
engine has no named strings).
Import read style:vertical-align but neither exporter wrote it, so a
mid-cell callout label jumped to the row's top on every save. odf-kit's
CellOptions and docx's TableCell both take it natively; DOCX import maps
w:vAlign in as well.
A table-of-content entry template names the stop its page numbers are set
against — the deputy-standards guidance puts it 45mm inside the column,
and every number of its index hung out at the margin instead.
A table-of-content entry template that declares its tab stop without a
style:leader-char means the gap is empty — the deputy-standards guidance
has one, and got a row of dots it never asked for. Round-trips.
An as-char figure frame is centred by the paragraph it sits in, and the
importer lifts it out of that paragraph — so a frame whose own style says
"from-left" landed at the margin instead of the middle of the column.
ODF style:vertical-align (middle/bottom) on a table cell, which every
LibreOffice callout table uses for its Tip/Note label. Import and render
only: odf-kit dedupes cell styles by property, so writing it back needs a
pass of its own — noted in docs/architecture/tables.md.
The inset ring was a fixed 0.15cm the editor imposed on every box, so a
figure frame — which declares 0 — sat 1.5mm right of where LibreOffice
draws it and stood 3mm taller, on every figure of the document.
An as-char text box takes its centring from the paragraph it is anchored
in, and the importer lifts it out of that paragraph — so every figure in
a fixture sat at the left margin. Carry style:horizontal-pos on the
box (an image needs none: with no side to float to it centres anyway).
A hyperlink in a fixture's copyright footer threw generateHTML and
blanked the whole zone. Drop what the one-paragraph schema lacks, keep
its text. A master handing over to a successor is also the different-first-
page flag, so a title master with no zones leaves page one blank.
The default master was the one literally named Standard, which a document
need never reference: a fixture's body runs on OOoPageStyle, so its
footer band went unreserved and every page held ~15mm too much text.
An index longer than a page has no inner document positions for a
pagination spacer, so it now breaks itself between entries.
style:master-page-name on a paragraph style is LibreOffice's "page break
with page style": the block always opens a page, except for the master
the document itself opens with.
The LibreOffice "page break" paragraph style carries the break on its
own end, not on the block that follows; unread, a fixture's title
page ran straight into the copyright page.
ODF text:anchor-type="page" is a cover graphic or a watermark: out of
flow, placed from its page's corner, behind the text. In flow, the two
full-page frames of a fixture's covers cost three pages.
The style-registry path divided an ODF fo:line-height by 1.15 where the
block path, both DOCX paths and the export all treat it as the plain
percentage, so every paragraph whose spacing came from its style
rendered 15% too tight. Its suppression yardstick is the block's own
style too, so a 100% under a 115% style stays as direct formatting.
LibreOffice keeps a paragraph's space above and below in one item, so a
style that declares only one of them does not inherit the other from its
parent chain — the missing half falls back to the default paragraph
style. Probed against soffice with the OPG fixture's own styles, where
it is the difference between a heading sitting flush and 4mm too low.
LibreOffice adds the upper block's space-below to the lower one's
space-above, and drops the space-above where a page break put the block
at a page top — but for a Word document it takes only the larger of the
two (all probed against soffice). CSS margins can only collapse, so
space-before rides --space-before and editor.css renders it as padding
('add') or margin ('max'); the model is read from settings.xml, assumed
for DOCX, and written back on ODF export.
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.
The TOC node regenerated over every heading level and minted a title of its
own. A file asking for level 1 only (text:outline-level, Word's TOC \o range)
got a 43-entry index instead of 12 — a full page where LibreOffice needs a
quarter, plus a second "Table of Contents" under the file's own heading.
maxLevel and an empty title now round-trip through both formats.
An ODF master page whose style:next-style-name names another master applies
to a single page. Read as the section's running header, this file's full-page
title background repeated on every page and its height pushed each block onto
its own — 276 pages instead of 16.
Its zones now become the first-page variants, the successor's the running
ones, and a positioned frame in a zone is dropped as in DOCX.
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.
The DOCX importer read w:tblBorders off the table but never off its table
style, so a table whose style declares none (Word's Normal Table, and the
thesis' Formel-tab around every formula) fell back to the editor's 0.5pt
black default. Resolved per side along the w:basedOn chain, and an
undeclared side is not drawn — LibreOffice writes fo:border="none" for
exactly those tables when it re-saves the file. The ODF importer already
treated an undeclared side that way.
A top-and-bottom frame carrying positionH align / style:horizontal-pos
keeps it only where a nearby frame is set against the other end
(pairAlignedFrames, both importers); such a pair floats to its own side
instead of filling the band, so the two sit side by side as they do in
LibreOffice. A pair slightly wider than the column is scaled to it —
Word lets the two overlap, two floats cannot.
Steffen Becker