I guess that's the way the markdown library (that I don't want to hack) works. Did you try the script Markdown-it markdown parser if that suits your needs?
Or are you proposing a fix for that behavior?

The used library is: https://github.com/hoedown/hoedown

I didn't use the script until now but just tried it. With the script all elements A B C D E F get the <p> tag when there are one ore more empty lines (and no <p> tag when there is no empty line).
So it looks like:

• A

• B

• C

• D

• E

• F

Which means there is also an inconsistency ...

I'm not sure if there is a specification for that... If you write some text between the lists it works.

When the text does not start with 3 spaces - yes.
It does not flip to code style

This is the version of the library we are using: https://github.com/pbek/QOwnNotes/tree/develop/src/libraries/hoedown

The Markdown-it is similar. Probably I must first get used to the (im-)possibilities of MD.
But at least I understand now a little how QOwnNotes works - so including other features like mathjax.org could be solved by the right transformation script ...

Won't fix
Upstream problem in QTextBrowser