markdown2

New "tables" extra for table syntax that matches GFM https://help.github.com/articles/github-flavored-markdown#tables and PHP-Markdown Extra https://michelf.ca/projects/php-markdown/extra/#table. For example:

  | Header 1 | *Header* 2 |
  | -------- | -------- |
  | `Cell 1` | [Cell 2](http://example.com) link |
  | Cell 3   | **Cell 4** |

See https://github.com/trentm/python-markdown2/blob/master/test/tm-cases/tables.text for examples and edge cases.

If you have documents using the 'wiki-tables' syntax and want to convert to the 'tables' syntax, there is a script to help with that here: https://github.com/trentm/python-markdown2/blob/master/tools/wiki-tables-to-tables.py

["nofollow" extra, issue #74, pull #104] Add rel="nofollow" support (mostly by https://github.com/cdman):

  $ echo '[link](http://example)' | markdown2 -x nofollow
  <p><a rel="nofollow" href="http://example">link</a></p>

Limitation: This can add a duplicate 'rel' attribute to raw HTML links in the input.

["toc" extra] Unescape Markdown special chars in TOC entries. See https://github.com/trentm/restdown/issues/15.

Now 'tox' testing support (by github.com/msabramo):

  [sudo] pip install tox
  tox

confirming that markdown2 works with jython (not sure which version) and pypy! Also added pypy to travis-ci testing (http://travis-ci.org/#!/trentm/python-markdown2).

[issue #90] Add a Markdown.preprocess(text) -> text hook for subclasses. This is a match for the Markdown.postprocess(text) -> text hook added in an earlier version. (by @joestump).

[issue #90, backward incompatible change] Require a space between the '#' and a text for a title. I.e.:

  # This still works

  #This doesn't work

  ##Nor this

This keeps comments, hash tags, and ticket numbers at the beginning of the line from turning into an h1. (by @joestump)

This is a backward incompatible change, however small, hence the version change to 2.0.0.

[issue #67] Fix an sub-ul inside a ol not working with an indent less than 4 spaces.

Fix code blocks and fenced-code-blocks to work with a single leading newline at the start of the input.

[issue #86, 'fenced-code-blocks' extra] Fix fenced code blocks not being parsed out before other syntax elements, like headers.

[issue #83, 'fenced-code-blocks' and 'code-color' extras] Allow 'cssclass' code coloring option (passed to pygments) to be overridden (by https://github.com/kaishaku). E.g.:

  import markdown2
  html = markdown2.markdown(text,
      extras={'fenced-code-blocks': {'cssclass': 'mycode'}})

Deprecate code-color extra. Use the fenced-code-block extra and its cleaner mechanism for specifying the language, instead. This extra will be removed in v2.0 or so.

New fenced-code-blocks extra. It allows a code block to not have to be indented by fencing it with '```' on a line before and after. Based on [GFM](<http://github.github.com/github-flavored-markdown/).

  Some code:

  ```
  print "hi"
  ```

It includes support for code syntax highlighting as per GFM. This requires the pygments Python module to be on the pythonpath.

  ```python
  if True:
      print "hi"
  ```

[issue #76] Ensure "smarty-pants" extra doesn't destroy image links and links with title text.

[issue #72] Support reading from stdin for command line tool like any well-behaved unix tool, e.g.:

$ echo '*hi*' | markdown2
<p><em>hi</em></p>

Thanks Ryan!

Drop this "1.0.1.*" version silliness. The idea was that the first three numbers tracked the Markdown.pl on which markdown2.py was originally based. I don't believe Markdown.pl really gets releases anymore tho, so pointless.

[Issue 57] Add html5 block tags (article, section, aside, et al; see "_html5tags" variable) to Markdown literal HTML block tag handling. Thanks Tim Gray!

[Issue 56] Fix setup.py install.

[Issue 54] Fix escaping of link title attributes. Thanks FND!

Tweak list matching to NOT make a ul for something like this:

  - - - - - hi there

Before this change this would be a silly 5-deep nested li. See "not_quite_a_list" test case.

[Issue 52] Fix potential pathologically slow matching for <hr> markdown ("slow_hr" test case).

Add a Markdown.postprocess(text) -> text hook that is called near the end of markdown conversion. By default this does no transformation. It is called just before unescaping of special characters and unhashing of literal HTML blocks.

["header-ids" and "toc" extras] Add "n" argument to Markdown.header_id_from_text hook. This allows a subclass using this hook to differentiate the header id based on the hN number (e.g. h1 diff that h2). Also allow a None return value to not add an id to that header (and exclude that header from the TOC).

Note: If you used this hook, this is an incompatible change to the call signature.

Add a "markdown-in-html" extra similar to (but limited) http://michelf.com/projects/php-markdown/extra/#markdown-attr. I.e. this:

  <div markdown="1">
  Yo **yo**!
  </div>

becomes:

  <div>

  Yo <strong>yo</strong>!

  </div>

[Issue 39] Test case fix for pygments 1.3.1 from thomas.moschny.

[Issue 42] Add "smarty-pants" extra for transforming plain ASCII punctuation characters into smart typographic punctuation HTML entities. Inspiration: http://daringfireball.net/projects/smartypants/ Implementation by Nikhil Chelliah. Also add \' and \" escape sequences for forcing dumb quotes when this extra is in use.

Guard against using True instead of None as follows markdown(..., extras={'header-ids': True}). None is wanted, but True is commonly (at least I did it twice) used.

[Issue 36] Fix "cuddled-lists" extra handling for an looks-like-a-cuddled-list-but-is-indented block. See the "test/tm-cases/cuddled_list_indented.text" test case.

Experimental new "toc" extra. The returned string from conversion will have a toc_html attribute.

New "header-ids" extra that will add an id attribute to headers:

  # My First Section

will become:

  <h1 id="my-first-section">My First Section</h1>

An argument can be give for the extra, which will be used as a prefix for the ids:

  $ cat foo.txt
  # hi there
  $ python markdown2.py foo.txt
  <h1>hi there</h1>
  $ python markdown2.py foo.txt -x header-ids
  <h1 id="hi-there">hi there</h1>
  $ python markdown2.py foo.txt -x header-ids=prefix
  <h1 id="prefix-hi-there">hi there</h1>

Preliminary support for "html-classes" extra: takes a dict mapping HTML tag to the string value to use for a "class" attribute for that emitted tag. Currently just supports "pre" and "code" for code blocks.

[Issue 33] Implement a "cuddled-lists" extra that allows:

  I did these things:
  * bullet1
  * bullet2
  * bullet3

to be converted to:

  <p>I did these things:</p>

  <ul>
  <li>bullet1</li>
  <li>bullet2</li>
  <li>bullet3</li>
  </ul>

[Issue 26] Fix bug where email auto linking wouldn't work for emails with underscores. E.g. Mail me: <foo_bar@example.com> wouldn't work.

Update MANIFEST.in to ensure bin/markdown2 gets included in sdist.

[Issue 23] Add support for passing options to pygments for the "code-color" extra. For example:

  >>> markdown("...", extras={'code-color': {"noclasses": True}})

This formatter_opts dict is passed to the pygments HtmlCodeFormatter. Patch from 'svetlyak.40wt'.

[Issue 21] Escape naked '>' characters, as is already done for '&' and '<' characters. Note that other markdown implementations (both Perl and PHP) do not do this. This results in differing output with two 3rd-party tests: "php-markdown-cases/Backslash escapes.text" and "markdowntest-cases/Amps and angle encoding.tags".

"link-patterns" extra: Add support for the href replacement being a callable, e.g.:

  >>> link_patterns = [
  ...     (re.compile("PEP\s+(\d+)", re.I),
  ...      lambda m: "http://www.python.org/dev/peps/pep-%04d/" % int(m.group(1))),
  ... ]
  >>> markdown2.markdown("Here is PEP 42.", extras=["link-patterns"],
  ...     link_patterns=link_patterns)
  u'<p>Here is <a href="http://www.python.org/dev/peps/pep-0042/">PEP 42</a>.</p>\n'

Fix Python 2.6 deprecation warning about the md5 module.

Pass XML processing instructions and one-liner tags. For example:

  <?blah ...?>
  <xi:include xmlns:xi="..." />

Limitations: they must be on one line. Test: pi_and_xinclude. Suggested by Wolfgang Machert.