Word represents "restart numbering" on a style-based list by pointing
only the first item of the restarted list at a new `numId` that shares the
original list's abstract numbering definition but carries a
`w:startOverride`; the remaining items keep using the original `numId`.
Pandoc keyed list continuation and grouping on the `numId`, so the
restarted items continued the stale count from the earlier list (and
were split into a separate ordered list with the wrong start).
Key continuation and grouping off the `abstractNumId` instead (the real
running counter in Word), and treat `startOverride` as a restart that
resets the count.
Closes#8367.
Co-Authored-By: Claude <noreply@anthropic.com>
Previously, when a w:p paragraph contained runs with textboxes,
the entire paragraph was replaced by just the textbox content,
discarding all other runs (including image-bearing runs).
Now we walk the paragraph's children in order, grouping
non-textbox content into copies of the original w:p and splicing
unwrapped textbox content in place, preserving the original order.
We also treat text inside a textbox containing an image as a
figure caption. (One often finds captioned images of this
kind in docx files.)
Closes#11510.
Closes#6893.
Closes#11412.
Closes#5394.
Closes#9633.
Co-Authored-By: Claude <noreply@anthropic.com>
Table rows in Docx files can use `w:gridBefore` property to specify
a number of extra empty table cells to be inserted at the beginning
of a row. Without this, table columns in tables using the element
will become misaligned.
https://learn.microsoft.com/en-us/dotnet/api/documentformat.openxml.wordprocessing.gridbefore?view=openxml-3.0.1
In the worst case, this can lead to cells at the end of some rows
to be dropped by the parser.
The test file was produced as follows based on
the “3GPP TS 24.282 V19.5.0 (2025-12)” document obtained from
<https://www.3gpp.org/ftp/Specs/archive/24_series/24.282/24282-j50.zip>
where I encountered the issue:
1. Extracted the `document.xml`.
2. Formatted it with `xmllint --format`.
3. Found the raw XML for “Table 15.2.13-2: Payload content type”.
4. Removed `w:rsid*` and `w14:*` attributes, and `w:pStyle` elements.
5. Replaced the context of `test/docx/raw-blocks.native`.
6. Used pandoc to generate the docx.
Each worksheet turns into a section containing a table.
The common file `nativeDiff` has been extract from
the Docx and Pptx text files and put in Tests.Helpers.
When the last parPart before a header was a bookmark, no span with an
anchor was added for a bookmark in the header. But the function that
adds header anchors to the anchor map, needs a span with an anchor.
So this commit adds that span.
Some fields only have an instrText and no content, Pandoc didn't
understand these, causing other fields to be misunderstood because it
seemed like a field was still open when it wasn't.
Fields delimited by fldChar elements can contain other fields. Before,
the nested fields would be ignored, except for the end, which would be
considered the end of the parent field.
To fix this issue, fields needed to be considered containing ParParts
instead of Runs, since a Run can't represent complex enough structures.
This also impacted Hyperlinks since they can originate from a field.
When a paragraph has an indentation different from the parent (named)
style, it used to be considered a blockquote. But this only makes sense
when the paragraph has more indentation. So this commit adds a check
for the indentation of the parent style.
* Column spans
* Row spans
- The spec says that if the `val` attribute is ommitted, its value
should be assumed to be `continue`, and that its values are
restricted to {`restart`, `continue`}. If the value has any other
value, I think it seems reasonable to default it to `continue`. It
might cause problems if the spec is extended in the future by adding
a third possible value, in which case this would probably give
incorrect behaviour, and wouldn't error.
* Allow multiple header rows
* Include table description in simple caption
- The table description element is like alt text for a table (along
with the table caption element). It seems like we should include
this somewhere, but I’m not 100% sure how – I’m pairing it with the
simple caption for the moment. (Should it maybe go in the block
caption instead?)
* Detect table captions
- Check for caption paragraph style /and/ either the simple or
complex table field. This means the caption detection fails for
captions which don’t contain a field, as in an example doc I added
as a test. However, I think it’s better to be too conservative: a
missed table caption will still show up as a paragraph next to the
table, whereas if I incorrectly classify something else as a table
caption it could cause havoc by pairing it up with a table it’s
not at all related to, or dropping it entirely.
* Update tests and add new ones
Partially fixes: #6316
In the current dev version, we will sometimes add
a version of an image with a hashed name, keeping
the original version with the original name, which
would leave to undesirable duplication.
This change separates the media's filename from the
media's canonical name (which is the path of the link
in the document itself). Filenames are based on SHA1
hashes and assigned automatically.
In Text.Pandoc.MediaBag:
- Export MediaItem type [API change].
- Change MediaBag type to a map from Text to MediaItem [API change].
- `lookupMedia` now returns a `MediaItem` [API change].
- Change `insertMedia` so it sets the `mediaPath` to
a filename based on the SHA1 hash of the contents.
This will be used when contents are extracted.
In Text.Pandoc.Class.PandocMonad:
- Remove `fetchMediaResource` [API change].
Lua MediaBag module has been changed minimally. In the future
it would be better, probably, to give Lua access to the full
MediaItem type.
For security reasons, some legal firms delete the date from comments and
tracked changes.
* Make date optional (Maybe) in tracked changes and comments datatypes
* Add tests
* Fix hlint suggestions, update hlint.yaml
Most suggestions were redundant brackets. Some required
LambdaCase.
The .hlint.yaml file had a small typo, and didn't ignore camelCase
suggestions in certain modules.
* Simplify by collapsing a do block into a single <$>
* Remove an unnecessary variable: `all` takes any Foldable, so only blocksToInlines needs toList.
PR #5884.
+ Use pandoc-types 1.20 and texmath 0.12.
+ Text is now used instead of String, with a few exceptions.
+ In the MediaBag module, some of the types using Strings
were switched to use FilePath instead (not Text).
+ In the Parsing module, new parsers `manyChar`, `many1Char`,
`manyTillChar`, `many1TillChar`, `many1Till`, `manyUntil`,
`mantyUntilChar` have been added: these are like their
unsuffixed counterparts but pack some or all of their output.
+ `glob` in Text.Pandoc.Class still takes String since it seems
to be intended as an interface to Glob, which uses strings.
It seems to be used only once in the package, in the EPUB writer,
so that is not hard to change.
The first list item of a sublist should not resume numbering
from the number of the last sublist item of the same level,
if that sublist was a sublist of a different list item.
That is, we should not get:
```
1. one
1. sub one
2. sub two
2. two
3. sub one
```
Motivating issues: #5523, #5052, #5074
Style name comparisons are case-insensitive, since those are
case-insensitive in Word.
w:styleId will be used as style name if w:name is missing (this should
only happen for malformed docx and is kept as a fallback to avoid
failing altogether on malformed documents)
Block quote detection code moved from Docx.Parser to Readers.Docx
Code styles, i.e. "Source Code" and "Verbatim Char" now honor style
inheritance
Docx Reader now honours "Compact" style (used in Pandoc-generated docx).
The side-effect is that "Compact" style no longer shows up in
docx+styles output. Styles inherited from "Compact" will still
show up.
Removed obsolete list-item style from divsToKeep. That didn't
really do anything for a while now.
Add newtypes to differentiate between style names, ids, and
different style types (that is, paragraph and character styles)
Since docx style names can have spaces in them, and pandoc-markdown
classes can't, anywhere when style name is used as a class name,
spaces are replaced with ASCII dashes `-`.
Get rid of extraneous intermediate types, carrying styleId information.
Instead, styleId is saved with other style data.
Use RunStyle for inline style definitions only (lacking styleId and styleName);
for Character Styles use CharStyle type (which is basicaly RunStyle with styleId
and StyleName bolted onto it).
* Require recent doctemplates. It is more flexible and
supports partials.
* Changed type of writerTemplate to Maybe Template instead
of Maybe String.
* Remove code from the LaTeX, Docbook, and JATS writers that looked in
the template for strings to determine whether it is a book or an
article, or whether csquotes is used. This was always kludgy and
unreliable. To use csquotes for LaTeX, set `csquotes` in your
variables or metadata. It is no longer sufficient to put
`\usepackage{csquotes}` in your template or header includes.
To specify a book style, use the `documentclass` variable or
`--top-level-division`.
* Change template code to use new API for doctemplates.
Quite a few modules were missing copyright notices.
This commit adds copyright notices everywhere via haddock module
headers. The old license boilerplate comment is redundant with this and has
been removed.
Update copyright years to 2019.
Closes#4592.
This is difficult to recreate with a modern version of Word, so I'm
using the file submitted with the bug report. It would be preferable
to find a smaller example with Latin characters, though, so as not to
confuse the issue being tested.
