Fix etree deprecation warnings

[nickmoreton]

I have changed the import for etree to avoid the Django deprecation warnings

[zerolab]

@nickmoreton thanks for this. Can you point me to the deprecation warnings?
Ideally in a changelog somewhere, but a traceback should do too

[nickmoreton]

Hey @zerolab this is one of our support tickets:

/usr/local/lib/python3.8/site-packages/wagtailmarkdown/utils.py:18: DeprecationWarning: 'etree' is deprecated. Use 'xml.etree.ElementTree' instead.
from .mdx import linker, tables

Is this info enough? I can get trace tomorrow if need be.

[zerolab]

That is plenty, thanks

[tbrlpld]

Issue: It looks like that import is only deprecated as of Markdown >= 3.2.

https://python-markdown.github.io/change_log/release-3.2/#markdownutiletree-deprecated

This project currently allows the version range Markdown>=3,<4.

We either need to add a version dependent import or limit the supported version range. I would say the former would be preferred as the impact should be smaller.

import markdown

if markdown.VERSION >= (3, 2):
    import xml.etree.ElementTree as etree
else:
    from markdown.util import etree

[zerolab]

🌟@tbrlpld
This was on my list to review later today, but you beat me to it, and did a great job at it too!

[nickmoreton]

@tbrlpld Thanks for the suggestion. As @zerolab says good job!

[zerolab]

@nickmoreton you probably want to use __version__info (https://github.com/Python-Markdown/markdown/blob/master/markdown/__meta__.py#L29) to avoid TypeError: '>=' not supported between instances of 'str' and 'tuple'

[nickmoreton]

@nickmoreton you probably want to use __version__info (https://github.com/Python-Markdown/markdown/blob/master/markdown/__meta__.py#L29) to avoid TypeError: '>=' not supported between instances of 'str' and 'tuple'

Thanks @zerolab I was busy working it out before I saw your message but used version instead. Is that OK?

[zerolab]

Wasn't sure how this would play with non-pip solutions (e.g. Conda) but it looks like Conda projects can use pip when a package is not available in the Conda repos.

From my research it seem it should all be fine. Happy to go with this.