dagrevis / mdx_linkify Goto Github PK

When using markdown I sometimes specify the output_format as "xhtml". But, if I use LinkifyExtension then this is ignored and any tags that would be XHTML-style are HTML5 style.

For example, the \n\n----\n\n in text here should come out as <hr />, but it doesn't in the second case:

from markdown import markdown
from mdx_linkify.mdx_linkify import LinkifyExtension

text = "Hello\n\n----\n\nBye"

markdown(text=text, extensions=[], output_format="xhtml")

markdown(text=text, extensions=[LinkifyExtension()], output_format="xhtml")

Output:

'<p>Hello</p>\n<hr />\n<p>Bye</p>'
'<p>Hello</p>\n<hr>\n<p>Bye</p>'

I guess this is to do with Bleach?

Some of the new TLDs are not supported, and bleach doesn't plan to fix it.

mozilla/bleach#519

Can you add ability to override tlds/url_re to this project? If it's considered good idea, i can make pull-request.

When mdx_linkify sees a URL in a string surrounded by backticks like this:
<img src="http://dontescapethis.com/img.png" />

It linkifies the URL when it should leave it alone.

https://pypip.in/d/mdx_linkify/badge.png

https://pypip.in/v/mdx_linkify/badge.png

After update to 2.0 i encountered a strange bug, sometimes, very rarely, linkify fails with following traceback:

...
  File "markdown/core.py", line 292, in convert
    output = pp.run(output)
  File "mdx_linkify/mdx_linkify.py", line 34, in run
    return self._linker.linkify(text)
  File "bleach/linkifier.py", line 170, in linkify
    dom = self.parser.parseFragment(text)
  File "bleach/_vendor/html5lib/html5parser.py", line 317, in parseFragment
    self._parse(stream, True, *args, **kwargs)
  File "bleach/html5lib_shim.py", line 399, in _parse
    self.mainLoop()
  File "bleach/_vendor/html5lib/html5parser.py", line 243, in mainLoop
    new_token = phase.processEndTag(new_token)
  File "bleach/_vendor/html5lib/html5parser.py", line 485, in processEndTag
    return self.endTagHandler[token["name"]](token)
  File "bleach/_vendor/html5lib/html5parser.py", line 1372, in endTagP
    node = self.tree.openElements.pop()

And i cannot reproduce it in test case.

But i presume, that's becouse HTMLParser.parseFragment function is not pure.

Before 2.0 bleach.linkify created new Linker instance for each call, here we creating single instance and reusing it.

I moved Linker from __init__ to run in my private project, i'll test it for few days, if it works, i'll create PR with fix.

First of all, many thanks for providing a maintained package!

I've come by an apparently less frequently experienced behaviour that may not be intentional, the partial linkification of Posix paths. (It may not be an issue.) A (real life :) ) example:

from markdown import markdown

markdown("tests/test_markdown.py", extensions=["mdx_linkify"])

gives

'<p>tests/<a href="http://test_markdown.py" rel="nofollow">test_markdown.py</a></p>'

Albeit .py is a valid TLD, I suspect this may not be as intended, at the very least due to the conversion being partial - I guess it should cover the "tests/" thing then as well ... is there a dedicated workaround and/or a setting that I missed?

I personally wouldn't like this linkified without a protocol part, or at least a "www." at the front - I'll look into accomplishing this via the "Ignoring TLDs" example. I am partial to imagining this to be a more common default preference (I can be biased of course! for now it is just a guess).

(I would also suggest/bring up the ability of configuring https versus the now-less-desirable http protocol, maybe even making it the default. Let's spread good security practice?)

I'll try to give it a look over the weekend, but if it's like something quite trivial (so far as priorities allow ... almost never the case :) even if I'm otherwise quite interested), please do feel free to let me know

• what I should have noticed, or
• beat me to investigating the issue, if any

Thanks in advance!

I get this error when using mdx_linkify 0.5 with Markdown 2.5.1.

Exception Type: TypeError
Exception Value:
'NoneType' object is not iterable
Exception Location: /home/nnyby/src/dmt/ve/local/lib/python2.7/site-packages/markdown/extensions/init.py in setConfigs, line 75
Python Executable: /home/nnyby/src/dmt/ve/bin/python
Python Version: 2.7.8
Server time: Fri, 7 Nov 2014 13:59:44 -0500

Here's how I'm calling mdx_linkify:

{{object.description|markdown:"linkify"}}

Hi, thanks for the work on this package!

Minor issue, but after upgrading the markdown package to version 3.0, I'm seeing deprecation warnings for the md_globals parameter, as well as the add call:

Using the add method to register a processor or pattern is deprecated.
'Use the `register` method instead.', DeprecationWarning
/home/circleci/repo/venv/lib/python3.6/site-packages/markdown/core.py:128: 
DeprecationWarning: The 'md_globals' parameter of 
'mdx_linkify.mdx_linkify.LinkifyExtension.extendMarkdown' is deprecated.

Are there plans to upgrade the markdown dependency any time soon? I think the fix should only need to touch the LinkifyExtension class: