Skip to content

Latest commit

 

History

History
482 lines (442 loc) · 23.7 KB

CHANGELOG.md

File metadata and controls

482 lines (442 loc) · 23.7 KB

Changelog

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

The format is based on Keep a Changelog. This project adheres to Semantic Versioning, with the exception that 0.x versions can break between minor versions.

0.23.0 - 2024-09-16

Added

  • New extension for footnotes!
    • Syntax:
      Main text[^1]
      
      [^1]: Additional text in a footnote
      
    • Inline footnotes like ^[inline footnote] are also supported when enabled via an option in FootnotesExtension.Builder
    • Use class FootnotesExtension in artifact commonmark-ext-footnotes (#332)
  • New option omitSingleParagraphP in HtmlRenderer.Builder for not using <p> tags for when a document only has one paragraph (#150)
  • Support for custom link processing during inline parsing (e.g. [foo]), see Parser.Builder#linkProcessor
  • Support for extending inline parsing with custom inline content parsers. See Parser.Builder#customInlineContentParserFactory. This allows users/extensions to hook into inline parsing on a deeper level than before (e.g. with delimiter processors). It can be used to add support for math/latex formulas or other inline syntax. (#321)

Changed

  • The default DefaultUrlSanitizer now also allows data as a protocol. Use the constructor with a list to customize this. (#329)
  • LinkReferenceDefinition now extends Block (it was extending Node directly before)
  • MarkdownRenderer: Don't escape = text if it's the first node in a block (#335)

Fixed

  • Fix parsing of link reference definitions with incorrect title syntax (followed by characters other than space/tab). In that case, the title was set to the partially-parsed title and the source spans were wrong. (#315)
  • Fix source spans of blocks with lazy continuation lines (#337)
  • MarkdownRenderer: Preserve thematic break literals (#331)

0.22.0 - 2024-03-15

Added

  • New MarkdownRenderer for rendering nodes to Markdown (CommonMark) (#306)! Note that while care is taken to produce equivalent Markdown, some differences in the original Markdown (if parsed) are not preserved, such as:
    • The type of heading used
    • The type of link used (reference links will be rendered as inline links)
    • Whether special characters are escaped or not
    • Leading and trailing whitespace
  • Modular JAR (JPMS): All artifacts now include module descriptors (module-info) so jlink can be used; the old Automatic-Module-Name manifest entries were removed
  • New package org.commonmark.parser.beta containing classes that are not part of the stable API but are exported from the module because they might be useful for extension parsers
  • New package org.commonmark.text for text related utilities that are useful for both parsing and rendering
  • TableCell now has getWidth returning the number of dash and colon characters in the delimiter row, useful for rendering proportional width tables (#296)
  • ThematicBreak now has getLiteral containing the string that was used in the source when parsing (#309)
  • ListItem now has getMarkerIndent and getContentIndent for retrieving the space between the start of the line and the marker/content
  • Deprecated a some properties of BulletList, OrderedList, FencedCodeBlock and replaced with nullable ones because they might not be set when constructing these nodes manually instead of via parsing

Changed

  • Java 11 or later is now required (dropping support for Java 8)
  • Update to CommonMark spec 0.31.2

Fixed

  • Fix LinkReferenceDefinition having null SourceSpan when title is present and parsing with source spans option enabled (#310)

0.21.0 - 2022-11-17

Added

  • GitHub strikethrough: With the previous version we adjusted the extension to also accept the single tilde syntax. But if you use another extension that uses the single tilde syntax, you will get a conflict. To avoid that, StrikethroughExtension can now be configured to require two tildes like before, see Javadoc.

0.20.0 - 2022-10-20

Fixed

  • GitHub tables: A single pipe (optional whitespace) now ends a table instead of crashing or being treated as an empty row, for consistency with GitHub (#255).
  • GitHub strikethrough: A single tilde now also works, and more than two tildes are not accepted anymore. This brings us in line with what GitHub actually does, which is a bit underspecified (#267)
  • The autolink extension now handles source spans correctly (#209)

0.19.0 - 2022-06-02

Added

  • YAML front matter extension: Limited support for single and double quoted string values (#260)

Changed

  • Check argument of enabledBlockTypes when building parser instead of NPEing later

0.18.2 - 2022-02-24

Changed

  • Test against Java 17
  • Bundle LICENSE.txt with artifacts (in addition to Maven metadata)

0.18.1 - 2021-11-29

Fixed

  • Fix tables with leading/trailing header pipes and trailing spaces (#244). This was a regression in 0.16.1 which is now fixed.

0.18.0 - 2021-06-30

Changed

  • Update to CommonMark spec 0.30:
    • Add textarea to list of literal HTML block tags. Like script, style, and pre, textarea blocks can contain blank lines without the contents being interpreted as commonmark.
    • Fix case folding for link reference labels in some cases (e.g. and SS should match)
    • Allow lowercase ASCII in HTML declaration
    • Don't let type 7 HTML blocks interrupt lazy paragraphs either
  • Preserve the original case for the label of LinkReferenceDefinition. Before, we used to store the normalized version (lowercase, collapsed whitespace).

0.17.2 - 2021-05-14

Changed

  • Pass original instead of normalized label to InlineParserContext for lookup (#204). This allows custom contexts to change the lookup logic and have access to the original label content. In case you have a custom implementation of InlineParserContext, you might need to adjust it to do normalization.

0.17.1 - 2021-02-03

Fixed

  • Fix emphasis surrounded by non-BMP punctuation/whitespace characters (characters that are longer than one UTF-16 "char"). Note that this is an edge case with rarely used Unicode characters, which a lot of other implementations don't handle correctly.
  • Fix tables where the row starts with spaces and then the first | - rows that didn't have spaces before were not affected (#199). This bug is present in 0.16.1 and 0.17.0.

0.17.0 - 2021-01-15

Changed

  • ACTION REQUIRED: Maven groupId has changed from com.atlassian.commonmark to org.commonmark
    • To continue getting new versions of commonmark-java, change the Maven coordinates in your dependencies:
    • Old: <groupId>com.atlassian.commonmark</groupId>
    • New: <groupId>org.commonmark</groupId>

0.16.1 - 2020-12-11

Added

  • Support for including source spans on block and inline nodes (#1):
    • Answer for "Where in the source input (line/column position and length) does this node come from?"
    • Useful for things like editors that want to keep the input and rendered output scrolled to the same lines, or start editing on the node that was selected.
    • Use includeSourceSpans on Parser.Builder to enable, either with IncludeSourceSpans.BLOCKS or IncludeSourceSpans.BLOCKS_AND_INLINES
    • Read data with Node.getSourceSpans
    • Note that enabling this has a small performance impact on parsing (about 10%)

Changed

  • In order to support source spans (see above), a few of the extension APIs changed. It should only affect users implementing their own extensions. See the Javadoc to see what changed.
  • YAML front matter extension: Support dots in key names

0.15.2 - 2020-07-20

Fixed

  • image-attributes extension: Fix unexpected altering of text in case parsing of attributes fails, e.g. {NN} text -> {NN text}, thanks @jk1

0.15.1 - 2020-05-29

Added

  • Add text content rendering support for InsExtension

0.15.0 - 2020-05-21

Added

  • Extension for width/height attributes for images, thanks @dohertyfjatl
    • Syntax: ![text](/url.png){width=640 height=480}
    • Use class ImageAttributesExtension in artifact commonmark-ext-image-attributes
  • Extension for task lists (GitHub-style), thanks @dohertyfjatl
    • Syntax:
      - [x] task #1
      - [ ] task #2
      
    • Use class TaskListItemsExtension in artifact commonmark-ext-task-list-items

0.14.0 - 2020-01-22

Added

  • Add sanitizeUrls to HtmlRenderer.Builder to enable sanitizing URLs of <a> and <img> tags. Sanitizing logic can be customized via urlSanitizer. Thanks @VandorpeDavid

0.13.1 - 2019-11-25

Fixed

  • Fix potential StackOverflowError for regular expressions used in the inline parser (e.g. when parsing long HTML), thanks @lehvolk

0.13.0 - 2019-07-15

Added

  • LinkReferenceDefinition nodes are now part of the document (not rendered by default).
  • InlineParserContext.getLinkReferenceDefinition was added to allow custom inline parsers to look up definitions for reference links.

Changed

  • Performance improvements compared to previous version:
    • Parsing 7-10% faster
    • HTML rendering 105% faster - or in other words, twice as fast!
  • Update to CommonMark spec 0.29 (#156):
    • Change how newlines/spaces are handled in inline code
    • Info strings for tilde code blocks can contain backticks and tildes
    • Allow spaces inside link destinations in pointy brackets
    • Disallow link destination beginning with < unless it is inside <..>
    • Disallow unescaped '(' in link title in parens
    • Disallow indenting list item marker by more than 3 spaces
    • No longer treat <meta> as a block tag
    • Link reference definitions can now be in setext headings too
  • Tables extension: Changes to match GitHub implementation:
    • Escaping now only considers pipe characters when parsing tables: \| results in a literal | instead of a column, everything else is passed through to inline parsing.
    • Table body can now contain lazy continuation lines (without |). An empty line or another block is needed to interrupt the table.
    • For tables without a body, <tbody> is no longer rendered in HTML
    • See https://github.github.com/gfm/#tables-extension- for details
  • Check non-null arguments early and provide a nicer message

Fixed

  • Fix incorrectly preserving HTML entities when rendering attributes
  • Fix pathological case with input [\\\\... (a lot of backslashes)
  • Fix pathological case with input []([]([](...

0.12.1 - 2018-11-13

Changed

  • Speed up parsing significantly: Compared to the previous version, the benchmarks show up to 55% faster parsing for both small and large documents! (#137, #140)
  • Parse backslash followed by unescapable character the same way as the reference implementations.
  • Build and test on Java 11 as well.
  • autolink: Stop URLs at " and ` as well

Fixed

  • Fix tab handling in ATX and Setext headings.

0.11.0 - 2018-01-17

Added

  • The extension for tables now also renders to plain text (when using a TextContentRenderer), thanks @ahjaworski

Changed

  • Add Automatic-Module-Name manifest entries so that library can be used nicely in Java 9 modules. The module names correspond to the root package name: org.commonmark, org.commonmark.ext.autolink, etc.
  • Java 7 is now only supported on a best-effort basis (but it has been EOL for quite some time, so yeah)

0.10.0 - 2017-09-14

Added

  • Support multiple DelimiterProcessor with the same delimiter char as long as they have different length, thanks @szeiger
  • Add tests for thread-safety and a section to the readme (#83)

Changed

  • Update to CommonMark spec 0.28 (#94):
    • Adapt to changed emphasis parsing rule
    • Allow nested parentheses in inline link destinations

Fixed

  • Fixes for text content rendering, thanks @JinneeJ:
    • Support for mixed lists
    • Fixed that whitespaces between text elements are removed in "stripped" mode. For example **text** and text had rendered as textand text
    • Improved rendering for auto links
  • Fix [\] being parsed as link label
  • Fix [foo](<\>) resulting in \ in href
  • Fix multiple of 3 rule for emphasis parsing (see commonmark/cmark#177)
  • Fix text node merging when opening/closing delimiters are adjacent (#96)
  • autolink: Fix linking of URLs without host, e.g. http://. (#99)

0.9.0 - 2017-03-03

Added

  • Support restricting which block types are parsed, see enabledBlockTypes method on Parser.Builder (#43), thanks @marksliva, @pivotal-graham-bell and @lalunamel. This allows you to disable parsing of e.g. headings, they will just be parsed as paragraphs instead.
  • Allow customizing the inline parser, see inlineParserFactory method on Parser.Builder (#68), thanks @vreynolds and @lalunamel. Note that this is experimental and currently requires using internal classes.

Changed

  • Wrap escaped HTML blocks in a <p> tag (#78)
  • Add missing ext-heading-anchor to dependencyManagement in parent pom, thanks @drobert

0.8.0 - 2016-12-09

Changed

  • Update to CommonMark spec 0.27 (#73):
    • Treat h2..h6 as HTML blocks well
    • Allow shortcut reference link before open parenthesis (if parenthesis is not part of a valid inline link)
  • AttributeProvider.setAttributes now has an additional tagName argument and is called for all HTML tags of a block. This allows users to add attributes for the pre tag of a code block in addition to code. Also added attribute provider support for additional HTML tags, namely em, strong, code and br. (#74)

Fixed

  • ext-heading-anchor: Fix IllegalArgumentException on Android (#71)

0.7.1 - 2016-10-05

Added

  • Allow to configure prefix/suffix for ID on HeadingAnchorExtension (#66), thanks @paulthom12345

0.7.0 - 2016-09-23

Added

  • Plain text content renderer (#58), thanks to @JinneeJ!
    • Renders a plain text representation of a document instead of HTML, see TextContentRenderer in core.
    • Extensible in the same way as HTML rendering.
  • Heading anchor extension (#26), thanks to @paulthom12345!
    • Adds "id" attribute to heading tags (e.g. <h1 id="heading">Heading</h1>), useful for linking to sections of a document.
    • ID generation logic can also be used by itself via the IdGenerator class.
    • Use class HeadingAnchorExtension in artifact commonmark-ext-heading-anchor
  • Ins (underline) extension (#54), thanks to @pabranch!
    • Enables underlining of text by enclosing it in ++. It's rendered as an ins tag in HTML.
    • Use class InsExtension in artifact commonmark-ext-ins.

Changed

  • HtmlRenderer and related classes moved from org.commonmark.html to org.commonmark.renderer.html
  • HtmlRenderer.Builder no longer takes an AttributeProvider, but uses a AttributeProviderFactory to instantiate a new provider for each rendering. Code needs to be changed to create a factory and then return the existing provider from its create method, similar to node renderers.
  • NodeRendererFactory was renamed to HtmlNodeRendererFactory, same for related classes (there's a corresponsing interface for text content rendering)

0.6.0 - 2016-07-25

Added

  • Add coverage data to build. Currently at 97 %.

Changed

  • Update to CommonMark spec 0.26 (#55)
    • empty list items can no longer interrupt a paragraph; this resolves an ambiguity with setext headers
    • ordered lists can interrupt a paragraph only when beginning with 1
    • the two-blank-lines-breaks-out-of-lists rule has been removed
    • the spec for emphasis and strong emphasis has been refined to give more intuitive results in some cases
    • tabs can be used after the # in an ATX header and between the markers in a thematic break
  • Simplify and speed up brackets processing (links/images)
    • Improves the nested brackets pathological case (e.g. [[[[a]]]] with a lot of brackets)
    • Also contributed these changes upstream to commonmark.js
  • Simplify merging of adjacent text nodes
  • Extended DelimiterProcessor interface so that implementations get more information in getDelimiterUse and can reject delimiters by returning 0 from it. Also rename the methods:
    • getOpeningDelimiterChar -> getOpeningCharacter
    • getClosingDelimiterChar -> getClosingCharacter
    • getMinDelimiterCount -> getMinLength

Fixed

  • Fix max length for link labels (999, not 1000)
  • autolink: Stop URLs at more invalid characters, notably '<' and '>'. According to RFC 3987, angle brackets are not allowed in URLs, and other linkers don't seem to allow them either.

0.5.1 - 2016-05-25

Fixed

  • Fix StringIndexOutOfBoundsException on line after tab (#52)

0.5.0 - 2016-04-22

Added

  • Add YAML front matter extension for document metadata blocks (#24), thanks to @chiwanpark
  • Add information about delimiter character and length to delimiter nodes (#10), thanks to @pcj
  • Make HTML rendering for nodes extensible (#35)
  • Add support for asymmetric delimiters (#17): DelimiterProcessor#getDelimiterChar was split into getOpeningDelimiterChar and getClosingDelimiterChar

Changed

  • Make AttributeProvider work for image and table nodes (#31)
  • Update to CommonMark spec 0.25:
    • Changes how partially consumed tabs are handled.
  • Add Android test project to build so that we won't break Android support (#38), thanks to @JinneeJ
  • Replace CustomHtmlRenderer with NodeRenderer which also allows overriding rendering for built-in node types (#35)

Fixed

  • Fix blank line after empty list item to terminate list
  • Fix nested bullet list indented with mix of tab and spaces (#41), thanks to @derari
  • Fix package name in Javadoc, thanks to @jiakuan
  • autolink: Treat more special characters as trailing delimiters to not include ">, "/> and "); at the end of URLs
  • autolink: Fix unexpected link end with unfinished delimiter pairs in URLs
  • autolink: Fix Android incompatibility by not using java.util.Objects

0.4.1 - 2016-02-11

Fixed

  • Fix problematic regex that doesn't work on some Java versions and Android
  • Fix problems with Android (usage of java.util.Objects, StandardCharsets, ProGuard, see #30), thanks to @JinneeJ!

Changed

  • autolink extension: Update to autolink 0.3.0. This stops recognizing "abc://foo" within "1abc://foo" as a link

0.4.0 - 2016-01-18

Changed

Update to CommonMark spec 0.24 (#28):

  • No longer allow whitespace between link text and link label
  • Don't allow whitespace in link destination even with <>
  • Don't use whitelist for schemes in autolinks, recognize all 2-32 length schemes (see spec)
  • Allow multi-line content in setext headings

API breaking changes (caused by changes in spec):

  • Rename Header to Heading
  • Rename HorizontalRule to ThematicBreak
  • Rename HtmlTag to HtmlInline
  • Replace MatchedBlockParser#getParagraphStartLine with #getParagraphContent that returns the current content if the matched block is a paragraph

0.3.2 - 2016-01-07

Fixed

  • Add more bounds checks to internal Substring class (might affect extensions)

0.3.1 - 2015-12-01

Fixed

  • Fix StringIndexOutOfBoundsException with unclosed inline link (#27)

0.3.0 - 2015-10-15

Changed

  • Update to spec 0.22 (#14)
  • Allow block parsers from extensions to override core behavior (#18)
  • Fix compilation without install (#19)
  • Parent pom, build and README updates

0.2.0 - 2015-08-20

Added

  • Add method Node parseReader(java.io.Reader) to Parser (#2)
  • Extend Javadoc and publish online (#4)

Fixed

  • Fix StringIndexOutOfBoundsException on some inputs (#13)
  • ext-gfm-tables: Implement single-column tables (#7)

0.1.0 - 2015-07-22

Added

Initial release of commonmark-java, a port of commonmark.js with extensions for autolinking URLs, GitHub flavored strikethrough and tables.