Fix 'character maps to <undefined>' issue on windows

[alx00x]

Fixes #21

[raethlein]

Hey @koaleksa, thanks a lot for the PR and fixing the issue! I was just wondering whether this could break the behavior in certain cases / for certain users where the markdown file was not saved as 'utf-8'? I think this is still much better than leaving it out as in the original, though. But maybe we should make the default 'utf-8' overridable via a CLI argument? Do you have any thoughts about this?

[alx00x]

Hi @raethlein, there are probably very few reasons why one would want to write a text file (markdown in particular) with any other encoding these days. An argument could be made, of course... If there are such users, the code will still run for them as function simply opens the file in write mode and replaces the content, but the encoding will indeed change for them under their feet.

I think, for a lightweight tool, I'd advocate for keeping things simple and adding options only when there's specific expressed interest, rather than going by unlikely scenarios. However, a CLI override would be trivial to implement so if you'd still prefer to expose the option, I'll be happy to amend this PR.

Thank you guys for the awesome tool, btw. 👍

For some interesting read on this topic, I suggest having a look at PEP597

[raethlein]

Thanks a lot for your detailed thoughts and I totally agree with your assessment! I am going to merge it then :)