markdown-it - v15.0.2
    Preparing search index...

    Changelog

    All notable changes to this project will be documented in this file.

    The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

    • Fixed quadratic complexity in smartquotes when quote types don't match, #1209. Also limited the smartquotes stack to 1000 unmatched openers.
    • doc: replace oxide theme with custom one.
    • Fixed code span parsing after lookaheads for unclosed link and image labels, #1201.
    • Preserve spaces in code spans whose content consists only of spaces, #1180.
    • Preserve brackets around IPv6 address literals when normalizing links, #1204.
    • Fixed quadratic complexity when replacing fuzzy links.
    • Fixed quadratic complexity in scheme backscan (inline linkify rule).
    • Exposed parser internals classes as static properties on markdownit.
    • Bundled TypeScript declarations. Remove @types/markdown-it if you used it.
    • Added the markdown-it/browser export with bundled ESM and UMD builds.
    • Added colored CLI help on supported terminals via argparse 3.
    • Added reference labels to link/image tokens, #938.
    • Added reference_definition tokens. They remain stripped by default for backwards compatibility, #1055. Also adjusted the line-break heuristic to keep output exactly the same.
    • [breaking] linkify-it => v6
      • No fuzzy links by default.
      • No auth part check by default.
      • Unicode punctuation terminates the link by default (should help with CJK).
      • See linkify-it changelog for other changes.
    • Package root now resolves to prebuilt ESM and CJS files instead of raw sources. Distribution files were reorganized under dist/ and dist/browser/.
    • Migrated to Typescript.
    • entities => v8. Can be rolled back to v7 if compatibility issues happen.
    • Moved validateLink, normalizeLink and normalizeLinkText from properties to prototype methods.
    • Reworked issue templates and contribution guidelines.
    • [breaking] Removed package-internal subpath exports (markdown-it/lib/* and similar). Use the static classes exposed on markdownit instead.
    • [breaking] Removed obsolete StateBlock#ddIndent, #1139. Update markdown-it-deflist to keep it working.
    • Preserve a literal backslash before a terminating space in link destinations, matching CommonMark 6.3, #1188.
    • Recognize lowercase declarations as HTML blocks, matching CommonMark 4.6, #1189.
    • Require semicolons for named entities in all decoding paths, completing the fix for #1096.
    • Enhanced the text_join rule to process image alt text.
    • Include inline code content in image alt text, #1142.
    • Reworked build pipeline & tools.
    • Added source maps.
    • Bumped linkify-it to 5.0.2.
    • Preserve backslash-space hard line breaks, matching CommonMark 6.7, #1185.
    • isPunctCharCode to utilities.
    • Don't end HTML comment blocks on a blank line, #1155.
    • Properly recognize astral chars (surrogates) in delimiter scans for emphasis-like markers, #1072. Big thanks to @tats-u for his global efforts with improving CJK support.
    • Preserve unicode whitespaces when trimm headings/paragraphs, #1074.
    • More strict entities decode to avoid false positives ;, #1096.
    • Restore block parser state on fail in lheading rule, #1131.
    • Fixed poor smartquotes perfomance on > 70k quotes in single block
    • Bumped linkify-it to 5.0.1 with fixed potential perfomance issues.
    • Fixed regression from v13 in linkify inline rule. Specific patterns could cause high CPU use. Thanks to @ltduc147 for report.
    • Updated CM spec compatibility to 0.31.2, #1009.
    • Fixed quadratic complexity when parsing references, #996.
    • Fixed quadratic output size with pathological user input in tables, #1000.
    • Drop ancient browsers support (use .fromCodePoint and other features).
    • Rewrite to ESM (including all plugins/deps). CJS fallback still available. No signatures changed, except markdown-it-emoji plugin.
    • Dropped dist/ folder from repo, build on package publish.
    • Set punicode.js as external dependency.
    • Html tokens inside img alt are now rendered as their original text, #896.
    • Hardbreaks inside img alt are now rendered as newlines.
    • Fixed crash/infinite loop caused by linkify inline rule, #957.
    • Throw an error if 3rd party plugin doesn't increment line or pos counters (previously, markdown-it would likely go into infinite loop instead), #847.
    • Bumped linkify-it to 4.0.1. That should fix some hangs, caused by wrong data, returned from linkify-it.
    • Added a new token type text_special to store escaped characters, same as text but unaffected by replacement plugins (smartquotes, typographer, linkifier, etc.).
    • Added a new rule text_join in core ruler. Text replacement plugins may choose to insert themselves before it.
    • (p) is no longer replaced with § by typographer (conflicts with ℗), #763.
    • text_collapse rule is renamed to fragments_join.
    • Smartquotes, typographic replacements and plain text links can now be escaped with backslash (e.g. \(c) or google\.com are no longer replaced).
    • Fixed collision of emphasis and linkifier (so http://example.org/foo._bar_-_baz is now a single link, not emphasized). Emails and fuzzy links are not affected by this.
    • Fix possible ReDOS in newline rule. Thanks to @MakeNowJust.
    • Fix corner case when tab prevents paragraph continuation in lists, #830.
    • StateInline.delimiters[].jump is removed.
    • Fixed quadratic complexity in pathological ***<10k stars>***a***<10k stars>*** case.
    • Ordered lists: add order value to token info.
    • Always suffix indented code block with a newline, #799.
    • Updated CM spec compatibility to 0.30.
    • Newline in alt should be rendered, #775.
    • HTML block tags with === inside are no longer incorrectly interpreted as headers, #772.
    • Fix table/list parsing ambiguity, #767.
    • Fix crash introduced in 12.0.3 when processing strikethrough (~~) and similar plugins, #742.
    • Avoid fenced token mutation, #745.
    • [](<foo<bar>) is no longer a valid link.
    • [](url (xxx()) is no longer a valid link.
    • [](url\ xxx) is no longer a valid link.
    • Fix performance issues when parsing links (#732, #734), backticks, (#733, #736), emphases (#735), and autolinks (#737).
    • Allow newline in <? ... ?> in an inline context.
    • Allow <meta> html tag to appear in an inline context.
    • Three pipes (|\n|\n|) are no longer rendered as a table with no columns, #724.
    • Fix tables inside lists indented with tabs, #721.
    • .gitattributes, force unix eol under windows, for development.
    • Added 3rd argument to highlight(code, lang, attrs), #626.
    • Rewrite tables according to latest GFM spec, #697.
    • Use rollup.js to browserify sources.
    • Drop bower.json (bower reached EOL).
    • Deps bump.
    • Tune specsplit.js options.
    • Drop Makefile in favour of npm scrips.
    • Fix mappings for table rows (amended fix made in 11.0.1), #705.
    • %25 is no longer decoded in beautified urls, #720.
    • Fix blockquote lazy newlines, #696.
    • Fix missed mappings for table rows, #705.
    • Bumped linkify-it to 3.0.0, #661 + allow unlimited . inside links.
    • Dev deps bump.
    • Switch to nyc for coverage reports.
    • Partially moved tasks from Makefile to npm scripts.
    • Automate web update on npm publish.
    • Fix em- and en-dashes not being typographed when separated by 1 char, #624.
    • Allow opening quote after another punctuation char in typographer, #641.
    • Assorted wording & typo fixes.
    • Fix quadratic parse time for some combinations of pairs, #583. Algorithm is now similar to one in reference implementation.
    • Minor internal structs change, to make pairs parse more effective (cost is linear now). If you use external "pairs" extensions, you need sync those with "official ones". Without update, old code will work, but can cause invalid result in rare case. This is the only reason of major version bump. With high probability you don't need to change your code, only update version dependency.
    • Updated changelog format.
    • Deps bump.
    • Remove extra characters from line break check. Leave only 0x0A & 0x0D, as in CommonMark spec, #581.
    • Fix possible corruption of open/close tag levels, #466
    • Updated CM spec compatibility to 0.29.
    • Update Travis-CI node version to actual (8 & latest).
    • Deps bump.
    • Fix --no-html CLI option, #476.
    • Fix smartquotes around softbreaks, #430.
    • Updated CM spec compatibility to 0.28.
    • Fix blockquote termination inside lists, #386.
    • Fix blockquote termination by list item, #338.
    • Remove tabs at the beginning of the line in paragraphs.
    • Better error message for bad input type, #324.
    • Fix table indentation issues, #325, #224.
    • Fix blockquote termination inside indented lists, #329.
    • Add -o / --output option to CLI, #312.
    • Add missed h2..h6 to whitelisted block tags.
    • Updated CM spec compatibility to 0.27 (no significant changes).
    • Fix backticks handle inside tables, #303.
    • Fix edge case for fenced blocks with ~~~ in info, #301.
    • Fix fallback to reference if link is not valid, #302.
    • Make link parse helpers (md.helpers) pluggable, #299.
    • Tables: allow tab characters in markup
    • Benchmarks src cleanup.
    • Remove testing in old nodes (but still use es5).
    • Updated CM spec compatibility to 0.26 (see list below):
    • Two consecutive newlines no longer terminate a list.
    • Ordered list terminating a paragraph can now only start with 1.
    • Adjust emphasis algorithm (*foo**bar**baz* is now parsed as <strong> inside <em>).
    • Fix tab width calculation inside lists and blockquotes.
    • Fence renderer: fix concat of class array, #276.
    • Code renderer: do not render double space before attrs, #275.
    • Replacer: disable replacements inside autolinks, #272.
    • Bump linkify-it dependency to 2.0.0. --- no longer terminates autodetected links by default. md.linkifier.set('---', true) will return old behaviour.
    • Major version bumped, because internals or linkify-it was changed. You will not be affected anyhow, if not used direct access to require('linkify-it/re') for customizations.
    • Render code_inline & code_block attributes if exist.
    • Updated fence renderer to not mutate token. Token stream should be immutable after renderer call.
    • Process \r the same way as \n and \r\n, #252.
    • Added Token.attrGet() method for convenience, #251.
    • Security fix: possible ReDOS in linkify-it (forced bump of linkify-it & uc-micro dependencies). New installs will use fixed packages automatically, but we bumped markdown-it version for sure & for web builds.
    • Fix: should not escape twice content of image alt attribute, #246.
    • Improve support of missing values in tables, #224.
    • Maintenance release. Version bump caused by notable changes in CM spec (multiline setext headers, no spaces inside links, ...). API was not changed.
    • Fit CM 0.24 spec requirements.
    • Fixed nesting limit check in inline blocks, #197.
    • Fixed posible tail loss in CLI ouput.
    • Token: added .attrSet() & .attrJoin() methods.
    • Highlighter: allow wrapper override (if result starts with "<pre").
    • Allow single column and mismatched columns count in tables.
    • Smartquotes: take into account adjacent tokens.
    • Fill content property in image token with alt source.
    • Fix meta information (token.markup and token.info) for autolink tokens.
    • Improved tables compatibility with github, #120.
    • Internal API change. Due to new CM spec requirements, we had to update internals. That should not touch ordinary users, but can affect some external plugins. If you are plugin developper - see migration guide: https://github.com/markdown-it/markdown-it/blob/master/docs/migration/migration_v5.md.
    • Updated CM spec compatibility to 0.22 (see list below).
    • Keep tabs (don't replace with spaces).
    • Don't wrap iframes with paragraphs.
    • Rewritten emphasis algorithm.
    • Fix closure compiler collisions (don't use reserved words), #159.
    • Updated HTML blocks logic to CM 0.21 spec.
    • Minor fixes.
    • Fix class name injection in fence renderer.
    • Allow numbered lists starting from zero.
    • linkify-it dependency update (1.2.0). Now accepts dash at the end of links.
    • CM spec 0.20.
    • Added support for multichar substituition in smartquites, #115.
    • Fixed code block render inside blockquites, #116.
    • Doc fixes.
    • Minor emphasis update to match CM spec 0.19.
    • Bumped linkify-it version to 1.1.0. Now links with IP hosts and without protocols are not linkified by default, due possible collisions with some version numbers patterns (0.5.0.0). You still can return back old behaviour by md.linkify.set({ fuzzyIP: true }).
    • Bumped linkifier version. More strict 2-chars tald support for links without schema. Should not linkify things like io.js and node.js.
    • Improved pipe chars support in table cells, #86 (thanks to @jbt).
    • Disabled data: URLs by default (except some image mimes), to avoid possible XSS. Version bumped, because features changed (formally). If you did not used data: URLs, consider this version as 4.0.4 (no API changes).
    • Simplified link validator code. Now more easy to understand and to copy into your projects for customization.
    • Updated linkifier.
    • Smartquotes rule cleanup (#76).
    • Fixed replacements rule bug in PhantomJS (#77).
    • Fixed emphasis marker fields in tokens (#69).
    • Fixed html block tokens with numbers in name (#74).
    • Added custom container plugin demo.
    • Updated linkify-it version.
    • Breaking internal API changes. See v4 migration notes. In usual case you will need to update plugins.
    • Token internals changed
    • Unified the most of renderer methods.
    • Changed tokens creation - use state.push(...) (see sources)
    • Moved normalizeUrl() to root class as .normalizeLink() & added normalizeLinkText() method.
    • Moved .validateUrl() to root class and simplified logic - no more need to replace entities.
    • Joined md unescape & replace entities logic to utils.unescapeAll().
    • Removed replaceEntities() in utils.
    • md.utils.lib now exposes useful libs for plugins.
    • Use entities data from external package.
    • Fixed emphasis regression, caused by CM v0.18 spec (#65).
    • Spec conformance update to 0.18.
    • Significantly improved autolinking quality (use linkify-it package), #2.
    • Rewritten links normalizer to solve different edge cases (use mdurl package), #29.
    • Moved link title entities replace out of renderer.
    • Fixed escaped entities in links (foo\&amp;/bar).
    • Improved smartquotes logic, #61.
    • Added basic CLI support.
    • Use external package for unicode data (punctuation).
    • Added \v \f to valid whitespaces.
    • Match table columns count by header.
    • Sync scroll result => source in demo.
    • Moved normalizeReference() to utils.
    • Fixed hang on long vertical list of links. Appeared in 3.0.5. See #54 for details. Thanks to @fengmk2 for report!
    • Table lines now can have escaped pipe char \| (#5).
    • Significantly improved tests coverage (with dead code removal and other related things).
    • Fixed link validator - could skip some kind of javascript links with uppercase digital entities (thanks to @opennota)
    • Improved errors processing in url normalizer (for broken sequences).
    • Improved nesting limit processing in inline parser.
    • Reorganized tests & improved coverage.
    • Show inline diffs for failed tests.
    • Fixed punctuation check in emphasis.
    • Allow dashes in HTML tag names (needed for custom HTML tags).
    • Added # to terminator chars.
    • Improved link encoder - fix invalid surrogates to avoid errors.
    • Big split. All "rare" rules moved to external plugins (deflist, abbr, footnote, sub, sup, ins, mark).
    • Updated CM spec conformance to v0.15 (better unicode support).
    • Added md (parser instance) link to all state objects (instead of former options/parser).
    • References/Footnotes/Abbrs moved to block chain.
    • Input normalization moved to core chain.
    • Splitted links and images to separate rules.
    • Renamed some rules.
    • Removed full preset. Not needed anymore.
    • enable/disable methods now throw by default on invalid rules (exceptions can be disabled).
    • Replace NULL characters with 0xFFFD instead of strip.
    • Removed custom fences sugar (overcomplication).
    • Rewritten link components parse helpers.
    • More functions in md.utils.
    • Fixed inline html comments & cdata parse.
    • Added development info.
    • .use() now pass any number of params to plugins.
    • Fixed line breaks in definitions lists.
    • API docs.
    • Added 'zero' preset.
    • Updated CM spec conformance to v0.13.
    • Fixed several crashes, when some basic rules are disabled (block termination check, references check).
    • Added curring to set/configure/enable/disable methods.
    • Demo rework - now can include plugins.
    • Docs update.
    • Exposed helpers into parser instances (for plugins).
    • Removed utils from global export - been in instances seems enougth.
    • Refactored demo & added markdown-it-emoji to it.
    • Refreshed browser builds, missed in prev release.
    • Minor changes.
    • Separated method to enable rules by whitelist (enableOnly).
    • Changed second param of enable/disable ruler methods.
    • Shortcuts in main class for bulk enable/disable rules.
    • ASCII-friendly browserified files.
    • Separate package for spec tests.
    • New project name & home! Now it's markdown-it,
    • Sugar for constructor call - new is not mandatory now.
    • Renamed presets folder (configs -> presets).

    Settings
    Member Visibility

    On This Page
    Changelog