It is a bit weird to let . actually match \n, but this was documented behaviour so i would keep it.

Also added a bunch of tests now. IMO this is ready to be merged.

I've pushed 2 commits on your branch. One with a testcase that breaks on your code, but worked before. I guess there's more incompatibility between regex implementations, also in specifying character classes.
See also https://www.regular-expressions.info/charclass.html

thx for the typo fixes :D

One with a testcase that breaks on your code

Yes, currently the mini spec I wrote intentionally excluded the special case of "if [ or ] is at the first position of the charset they are fine without escape". And tests test_regex2_data.err15 and test_regex2_data.err16 checked for that.

Do you think this is an important case that should be included? Personally I would say this is one of the cases where its good to be more restrictive...

I updated the tests (again disallowing [[] and []]) and added documentation about the supported regex syntax.