-
Use a checkbox input element to render checkboxes (#102, David Campbell). Remove checked, unchecked class from li; this is no longer needed. We still need custom CSS on ul.task-list to remove the redundant list markers.
-
Fix positions for list_item and section (#83, #82, black-desk).
-
Test: fix and test number comments (#100, Martin Fischer).
-
Update djot.md to fix a tiny typo (#97, Yuri).
-
Move package.json exports "default" condition to last (#92, Fukuda Yuma).
-
Add prepare and pretest scripts in package.json (#81, black-desk).
- Update man page and schema.
- Djot renderer:
- Handle spans (#79).
- Adjust
task_list
indent to 6 (black-desk). - Fix bug that introduced an extra newline before some blocks (black-desk).
- Omit auto generated attributes and references (#69, #75, black-desk).
- Types [API change]: The interface HasAttributes now contains an
optional
autoAttributes
field. This contains automatically generated attributes (e.g., on section elements). The interface Doc now contains anautoReferences
field. This is likereferences
but is meant only for automatically generated references to section headings. Now that these are distinguished in the AST, the Djot writer can choose to omit attributes and references that were auto-generated, enabling round-tripping. - Re-export types from ast submodule (Alex Kladov).
- Add types to 'exports' field as well (#78, Sondre Aasemoen).
- parse: add source position to reference (black-desk).
- attributes: Change regex for identifiers so that non-ASCII will work (black-desk).
- Fix task lists in pandoc renderer (#71).
- Remove most of the benchmark files. These were hastily converted from commonmark.js benchmarks, and in many cases didn't even test the same thing due to escaping.
- Add note on unpkg CDN (see #67).
- HTML renderer: use a single variation selector in backlink (#65). Previously we had a doubled one.
- Revert change in parsing code for links/images. (commit 1b5425c7ac175127288dc8412e06df23db20b89a). (See #46, #64.)
- Attributes: put events on markers and spaces (new fix for #46).
Otherwise this is lost in
strMatches
, when we have to treat this material as part of a reference or destination. - Normalize reference + footnote labels. Trim and collapse adjacent interior whitespace, including newlines, to a single space. This will allow implicit heading references to work even when headings span multiple lines (#63).
getUniqueIdentifier
: improve regex for symbol replacement (#63).
- Fix bug in parsing code for links/images (#46).
- Fix buggy table caption parsing (#57).
- Fix smart punctuation in pandoc output (#62).
- Fix filters to not run twice (#61, kibigo!).
- Add package-lock.json and yarn.lock to repository (#60).
- Allow comment end with attribute end (gemmaro, #59).
- Support pandoc-types 1.23 JSON API (#54). Note that this will require use of pandoc >= 3 with djot.
- Lists: prioritize roman numerals over alpha when ambiguous (#38, Noah Hellman).
- Fix processing of backslash escapes in attributes (#49).
- Make sure we reparse unclosed block attributes, even whene not followed by blank line (#47).
- Ensure that table cell 'align' is always non-null (#41).
- Migrate all attributes from header to section (#43).
- Don't remove heading attributes completely when promoting the identifier to the enclosing section (#43).
- djot renderer:
- Throw error if no renderer defined for a node (#48).
- Render url and email nodes instead of ignoring silently (#48).
- HTML renderer:
- Don't generate "text-align: undefined".
- Use unicode instead of entities for smart punctuation in HTML (#50).
- Playground:
- Disable "Source positions" by default, but always enable it for preview (Noah Hellman).
- Respect sourcepos checkbox (Noah Hellman).
- Make debounce dynamic. Make debounce ms proportional to length of text, so short documents update faster.
- Update README on event parser function (gemmaro).
- Fuzz tests: report string that caused failure.
- HTML renderer: insert blank dummy notes when there is no note defined for a reference (#36).
- HTML renderer: render footnotes that are only referenced from other notes (#37).
- Pandoc parser: correctly handle DefaultStyle, DefaultDelim (#34).
- Optimize parsing and HTML rendering.
- Initial public release.