Some of the recent releases of Python-Markdown, along with some of the

recent fixes merged in from release-3.0.x, resulted in some regressions

in the expected Markdown rendering results in unit tests. These

regressions are, for the most part, not something users will need to

worry about.
Some of the regressions were in the XSS unit tests. These aren't

security regressions, just expected string regressions. Python Markdown

3.0 changed some of the parsing, supporting "title" in link URLs, and

these captured strings in JavaScript code, turning them into title=

attributes. Tests were updated for the new strings.
Python Markdown 3.2 added <code> elements to each line in a code

block, which new tests from release-3.0.x weren't accounting for. Since

we support 3.1 and 3.3+ (depending on the Python version), these testsn

ow check for the right string for the right version.
And finally, the one user-facing change is that nested bold/italic (in

the form of *this is **a** test* renders entirely as a series of

italic blocks on Python Markdown 3.2. The working alternative is

*this is __a__ test*. Tests were updated for this as well.

Unfortunately, there doesn't seem to be a way to restore prior

formatting, but hopefully this doesn't affect many people.
See https://github.com/Python-Markdown/markdown/issues/792 for the

bold/italic changes in Python Markdown.

Unit tests pass on all supported versions of Python, with all supported

versions of Python Markdown.