Apologies if this just an issue on my side. But I found that the follow two lines would not be joined into the same quote, when passed to my custom renderer:

> first quote line
second line underneath

I assumed they would both be passed together to the block_quote function.

Hi there !

Are you on a strict 'no backward compat' (as the version number is < 1.x) ?

I guess we miht try to reintroduce default_features as a class property that print a warning and maps to default_rules ?

Just saying that as all our CI test are failing since 0,5 release, but not a big deal we can update and.or pin mistune version.

Thanks !

Normally with escape=False, nested HTML block is corectly not escaped:

>>> print markdown('<div id="special-part"><div class="subsection">text</div></div>', escape=False)
<div id="special-part"><div class="subsection">text</div></div>

But when I add parse_block_html=True, only out-most element is not escaped and the rest is escaped:

>>> print markdown('<div id="special-part"><div class="subsection">text</div></div>', escape=False, parse_block_html=True)
<div id="special-part">&lt;div class="subsection"&gt;text&lt;/div&gt;</div>

For rendering of markdown documents in text format.

The concrete use case is the ability to render templated markdown documents directly within a terminal environment.

I just looked at how the grammar is extended and in it's current form it's not really usable. There are a few problems with the way the inline lexer is supposed to modified alone:

• right now the example modifies default_rules which is a class attribute and as such shared across all other lexers as well.
• because the text rule is greedy it needs to be modified whenever another rule is inserted but it is impossible for a plugin to do so because it cannot now which sequences indicate the start of a rule. The current example only works by chance.

I think the current design does not really suit itself to extension well but I'm not sure how to fix it without breaking what's already there.

Congrats @lepture for this project! (I'm doing a benchmark of many Markdown to HTML libraries, and mistune seems to be the best!)

I used timeit with the same Markdown document, and:

• version 0.8.4: mistune.markdown(s) took 4ms on average
• version 2.0.0a4: mistune.markdown(s) took 13ms on average (for the same document)

What could be the reason of 2.0.0a4 be 3x slower than version 0.8.4?

Can I use 0.8.4 for my current project, is it still stable?

s = 'my_sample_document'
import timeit
print(timeit.timeit("markdown(s)", setup="from mistune import markdown;from __main__ import s", number=100)/100)

PS: I also found this article, it's linked to this topic: https://getnikola.com/blog/markdown-can-affect-performance.html

Is it possible to retrieve information about the line number of block-level elements in the AST? We are looking into using Mistune for a project and this would be quite helpful.

E.g., running:

markdown = mistune.create_markdown(renderer=mistune.AstRenderer())
markdown("# Heading 1\n\n## Heading2")

to return something like

[{'type': 'heading',
  'children': [{'type': 'text', 'text': 'Heading 1'}],
  'level': 1,
  'lineno': 0},
 {'type': 'heading',
  'children': [{'type': 'text', 'text': 'Heading2'}],
  'level': 2,
  'lineno': 2}]

If it's not possible currently, is it something that would be doable w/ a PR?

Great work on this; it is indeed fast!

Would you consider a change something like the prototype in this pull request? It provides the inline_lexer to the renderer method and sets a couple extra values that can be used for more advanced rendering.

Of course, a full implementation would change all the renderer methods to accept the lexer as an argument and would be breaking backwards compatibility, but it is a very new project so maybe it is in time? If so, I'm willing to make the change and submit it.

In the exact example usage I'm thinking of, I detect if an image is all alone or surrounded by other elements. Depending on its surroundings or lack thereof, I change the rendered CSS class. So for example, an image followed by text all in one paragraph might float the image left and flow the text around it.

Maybe an image would better explain:

Hi,

I have been testing markdown rendering in our archiver for GNU Mailman project, Hyperkitty for some time now using the alpha/rc version of mistune 2.0. While it works okay in our test environments, i was hoping that released stable versions rely on the non-pre-release dependencies.

I was wondering if you had any plans for the 2.0 release for mistune? There seems to be no issues or PRs marked with the 2.0 milestone at the moment.

I have a blog with markdown files generated by Jekyll. They use Markdown files that has a metadata block on top that contains yaml. Is this something that mistune would be willing to support? Jekyll is a very common blog engine, so I think it would be useful to many!

---
id: 29
title: How this website was built
date: 2005-12-20T00:00:36
categories:
  - CSS
---
Hi there early readers!

I am getting a traceback with 0.3.1 (which works with 0.3.0):

Traceback (most recent call last):
  File "website/crossbario/__init__.py", line 191, in <module>
    pages = DocPages('../crossbar.wiki')
  File "website/crossbario/__init__.py", line 153, in __init__
    self._renderer = mistune.Markdown(renderer = rend, inline = inline)
  File "mistune.py", line 899, in mistune.Markdown.__init__ (mistune.c:16104)
    inline = inline(renderer, **kwargs)
  File "mistune.py", line 502, in mistune.InlineLexer.__call__ (mistune.c:8497)
    return self.output(src)
  File "mistune.py", line 510, in mistune.InlineLexer.output (mistune.c:9059)
    src = src.rstrip('\n')
AttributeError: 'DocPageRenderer' object has no attribute 'rstrip'
make: *** [test] Error 1

I was checking the math plugin and found out four incompatibilities with latex compilers (pdfLaTeX and XeLaTeX) and MathJax:

• Block math ($$) requires a newline before and after content, effectively changing the delimiters to $$\n and \n$$.
• Block math should be able to have text before and after it, in the same line.
• Block math doesn't work with empty content ($$$$), the regex is done with .+ instead of .*.
• Both math modes don't take care of escaped dollar signs (\$).

So the following markdown:

Dollar sign $\$$

Empty $$$$

No newline $$x$$

Is compiled to:

<p>Dollar sign <span class="math">\(\\)</span>$</p>
<p>Empty <span class="math">\($\)</span>$</p>
<p>No newline <span class="math">\($x\)</span>$</p>

Whereas it should probably be:

<p>Dollar sign <span class="math">\(\$\)</span></p>
<p>Empty <div class="math">$$$$</div></p>
<p>No newline <div class="math">$$x$$</div></p>

For comparison, here are the latex and MathJax versions:

\documentclass{article}

\begin{document}
    Dollar sign $\$$ \\
    Empty $$$$ \\
    No newline $$x$$
\end{document}

<html>
    <head>
        <script src="https://polyfill.io/v3/polyfill.min.js?features=es6"></script>
        <script id="MathJax-script" async src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>
    </head>
    <body>
        <p>Dollar sign \(\$\)</p>
        <p>Empty $$$$</p>
        <p>No newline $$x$$</p>
    </body>
</html>

Would you like a donation of inline type hints? I can send a PR to apply @fmigneault type hints from https://github.com/common-workflow-language/schema_salad/commit/47332275ae04540a4b71ccfa0ff160edbef01496#diff-b6fac26c3692bb3b814467d88fe4b15140151671dc85689c3d2eb7385e95afd8 to as inline type hints , if you would release a new 2.x with it

Thanks for this awesome library! :heart:

Motivation

I'm using mistune to parse documents that always have one level 1 heading. For example, the document might look like this:

# mistune

A fast yet powerful Python Markdown parser with renderers and plugins.

## Sponsors

...

## Install

...

The generated table of contents always includes the level 1 heading, which is redundant since it is used as a page title, and is the only level 1 heading.

Suggestion

I think a new keyword argument on the directive's class and/or an option on the directive could be helpful. If, for example, the minimum is 2, then only ## to ###### headings would be included in the table of contents.

Mistune users are able to write there own plugins, and when they do this locally, it is easy enough for them to import and use there plugin code, so it's available when they use Mistune. Is there a preferred way or best practice for how users should publish their plugin? What is the best way one could find community-made plugins? If there are such things, I propose adding them to the docs.

Taking some inspiration from Lektor, projects could be published to pypi with a mistune- prefix so they are discoverable via pypi's search bar, or you could maintain a docs page linking to them. Perhaps PyPA would allow a new trove classifier to list Mistune as a framework, as well (makes sense to me).

I thought of this coming from the idea that Lektor could more easily pull in arbitrary Mistune plugins. https://github.com/lektor/lektor/issues/1076

https://github.com/lepture/mistune/blob/1264a1c954396fa31304bfac39588381f015a5d8/mistune/plugins/abbr.py#L56-L63

maybe here def_abbr and should the same as abbr? I run my code local no error, but on github action error occurs

full log here https://github.com/teedoc/teedoc/actions/runs/3069280480/jobs/4957741255

We're embedding some pyscript inside out markdown, and having some trouble with code comments being rendered as headings:

It appears to be an issue within a custom tag where there is a blank line before the line prefaced with the #

To reproduce:

$ python3
Python 3.10.4 (main, Jun 29 2022, 12:14:53) [GCC 11.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import mistune
>>> mistune.__version__
'2.0.4'
>>> mistune.html(r'''<some-tag>
... 
... # a comment
... something with # a trailing comment
... </some-tag>
... 
... ''')
'<some-tag>\n<h1>a comment</h1>\n<p>something with # a trailing comment</p>\n</some-tag>\n'

What is expected:

The code/comments within the custom tags (in our case they're <py-script></py-script>) should not be interpreted as markdown (there weren't in a pre-2.x version of mistune), and produce: <some-tag>\n<p># a comment</p>\n<p>something with # a trailing comment</p>\n</some-tag>\n

When there is no blank line preceeding the code comment, this behavior is not observed:

>>> mistune.html(r'''<some-tag>
... # a comment
... something with # a trailing comment
... </some-tag>''')
'<some-tag>\n# a comment\nsomething with # a trailing comment\n</some-tag>\n'