v1.4.0

Added

• Add parent attribute/property (#71 via #206).

  COMPATIBILITY REMARKS:

  • As the children attribute changed to property, existing code needs to be changed like this:

    -hasattr(token, 'children')
    +token.children is not None
    
    # ...
    
    -'children' in vars(token)
    +token.children is not None
    
    # ...
    
    -getattr(token, 'children', [])
    +token.children or []

Fixed

• Incomplete sdist content (#219).

Changed

• MarkdownRenderer: Keep the original content spacing before the list marker (#213 via #215).
• Bump build/CI scripts.

v1.3.0

Added

• Add line numbers (line_number attribute) on all block tokens during parsing (#188).
• More thorough documentation and explanation of AST and tokens in the Developer's Guide (within #188).
• HtmlRenderer: Option to skip HTML tokens parsing (#74 via #204). Just pass process_html_tokens=False to the renderer's constructor.
• LaTeXRenderer: Add AMS packages for Math (#207).

Fixed

• MarkdownRenderer: Penultimate lines of multiline fragments being ignored (#201).
• MathJaxRenderer: Output inline math (in: $...$) correctly (out: \(...\)) (#195).
• Make local installation work again with pip 23.2 and above (fb231b1).

Changed

• MarkdownRenderer: Keep the original content spacing after the list marker (#196 via #197).

  COMPATIBILITY REMARKS:

  • The old behavior can be enforced by passing normalize_whitespace=True to the renderer's constructor (#202).
  • Provided you create ListItem's tokens directly via its constructor, you need to pass it a new parameter called indentation (number of spaces before the item marker):

    -def __init__(self, parse_buffer, prepend, leader):
    +def __init__(self, parse_buffer, indentation, prepend, leader):

• Allow just 1 dash in Table delimiter cells (#131 via #203).

For contributors:

• Improvements in the Python styling checks, setup and documentation - see diff.

v1.2.1

Fixed

• In v1.2.0, due to #182, using HTMLRenderer (the old class name) directly from the mistletoe package stopped working.

v1.2.0

Fixed

• FileWrapper's methods anchor/reset are DEPRECATED, use a more versatile get_pos/set_pos approach instead (#186).
• LaTeXRenderer: Escape special characters in URLs of generated links (#114 via #190).

Changed

• Do not escape double or single quotes by default (#181).
• As in GFM, tables are allowed to interrupt paragraphs by default (#166 via #187). Set mistletoe.block_token.Table.interrupt_paragraph to false if you need the old behavior.
• PygmentsRenderer no longer throws a ClassNotFound error if there is a code language specified which is not supported by the Pygments highlighter (#183). If you still need to get the error, simply pass fail_on_unsupported_language=True to the renderer's constructor.
• Use strict PascalCase for class names (#182). The old names like HTMLRenderer are marked as deprecated.

For contributors:

• Unify assertEqual() calls in all the tests (#185).
• Starting to use requirements.txt.
• Scheme renderer's __main__ section turned into a proper unit test (#189).
• The possibility of a block token to interrupt a paragraph was generalized (#187). A token class just needs to define the check_interrupts_paragraph(cls, lines) -> bool method.

v1.1.0

Added

• MarkdownRenderer - the long awaited renderer for outputting parsed AST back to the Markdown syntax (#4, #162). 🎉
  For this, parsing was extended to store additional information about the input markdown.
• New options to control whether HTMLRenderer should escape double and/or single quotes (#176).

Fixed

• Token.__repr__() no longer outputs class attributes (#172).
• Make Scheme (scheme.py) work again (7883f58).
• Make CI work again (#121). Good news for both the maintainers and the contributors.

Changed

• "Simpler is better!": Revert back to HTMLRenderer not escaping single quotes by default (#176).
  Note that we plan to extend this to not escaping double quotes in the next version.
• As a side effect of #162, the AST token classes and parsing process was changed for more uniform behavior and API
  (#163).
• ASTRenderer no longer outputs all token attributes, but only those listed in Token.repr_attributes (#172).

Special thanks go to @anderskaplan as the most active contributor of this release. 💪

v1.0.1

Fixed:

• Do include package "mistletoe.contrib" in the built artifacts (#177).

Updated:

• Make CONTRIBUTING.md more up-to-date.

v1.0.0

WARNING - Backwards compatibility changes:

• #167: For practical reasons and for following common packaging practices, contrib folder got moved under the mistletoe folder / package. So if you reference a renderer from that folder, you need to reference it as mistletoe.contrib.<renderer> now.
• See "change handling of (white)space characters in code spans" below. We keep processing extra whitespace characters at the parsing level, so that all renderers can benefit from it out-of-the-box. Provided that a custom renderer, for whatever reason, relied on all the spaces being collapsed, it needs to do that collapsing itself now (e.g. ' '.join(re.split('[ \n]+', content.strip())), or re.sub('[ \n]+', ' ', content.strip())).

Added:

• JIRARenderer: Support link title notation, i.e. [label](url "title") gets transformed to [label|url|title] (#161)

Fixed:

• Make the traverse() function actually work with various input parameters:
  • filtering by class (#157, #158)
  • filtering by depth (#159)
• Compatibility with the latest CommonMark specification v0.30 (#108):
  • Change handling of (white)space characters in code spans (#156; in line with commonmark/commonmark-spec#532 and commonmark/commonmark-spec#569)
  • Fix handling of tabs and parsing continuation lines within list items (#89 via #164)
  • Fix the other examples, mostly edge cases, from the spec (see #165 and #168 for details; #173 seems to be fixed as well)
• Make parsing of link reference definitions (a.k.a. Footnotes) more strict - spec compliant (#132)
• Fix handling of special punctuation characters when parsing emphasis (#96, #175)

Updated:

• Smaller inner working refactorings, like #171.

v0.9.0

WARNING - Backwards compatibility changes:

• Python versions below 3.5 are no longer supported (Python 3.6 end-of-life: December 2021)
  • html module (available since Python 3.4) is no longer included
• As unescaping of HTML character references (entities) is now correctly done in parsing phase already, prospective custom renderers should be altered accordingly provided they do the unescaping themselves now.
• HTMLRenderer: single quote is no longer rendered as ', but as ' (see #115; let us know if you would need the old behavior)
• BaseRenderer.__getattr__() is removed and replaced by explicit render_*() methods definitions for clearer API (#133)

Added:

• Add __repr__() methods to all token classes (#140)
• Add type hints for HTMLRenderer methods (#133; supported since Python 3.0)

Fixed:

• Correctly unescape HTML character references (entities) for LaTeXRenderer - refactored globally (#135)
• Ensure LaTeX renderer uses valid \verb delimiter - not always just | (#149)
• GithubWiki unit test failing when run via pytest (#142)

Others:

• Simplify implementation of escaping special HTML characters (#135)
• Remove unused imports and variables (#146)
• Document (for maintainers) how to create a new mistletoe release
• Fix and extend docstring documentation of various token types (#154)
• This version is about 0,5% faster according to the benchmark test. :)

v0.8.2

Fixed:

• Support emphasized inline code (#70)
• Failure to parse paragraph containing just "[" (#130) (a side-effect of the fix of #124 in v0.8.1)
• IndexError when parsing unfinished-link-like text (#98)

Others:

• Small documentation improvements.

v0.8.1

Added:

• Documentation (#122 - covering #56, #99 and some other basic topics)

Fixed:

• Avoid infinite loop when parsing specific Footnotes (#124)
• Read and write to files / console in UTF-8 (in all remaining locations)

Testing:

• Benchmark tests made up-to-date (#119)