This logic is torture and needs some rework.

Rules:

- Always strip leading (, as it can't be part of a URL
- Short circuit to only strip leading if no trailing
- If valid email address when trailing ) stripped, we can strip trailing )
- If not even a valid URL without trailing ), short circuit to only strip leading
- If query parameters detected, strip trailing. It should have been encoded as %29.
- If there isn't a / the trailing ) can't be part of the URL, strip trailing.
- If there isn't at least one ( in the URI.path, only strip leading. Assume ) is not part of the URL.
- If we have an equal count of ( and ) chars with the leading ( already stripped, only strip leading

Release/0.4.1

See merge request !36

thx @href !

Release: Change changelog and mix version

See merge request !34

priv/tlds.txt: Bump to 2020111900

See merge request !32

Version 2020111900, Last Updated Thu Nov 19 07:07:01 2020 UTC

curl https://data.iana.org/TLD/tlds-alpha-by-domain.txt | grep -v '^#' | tr '[:upper:]' '[:lower:]' > priv/tlds.txt

Add .onion in the default TLDs, Validate IPv6, IDN compatibility in email and mentions

See merge request !29

Add handle_href option for href pre-processing

Closes #19

See merge request !31

Support linking URLs with FQDNs

See merge request !30

Add example to README showing how we link user mentions

See merge request !28

Release v0.3.0

See merge request !26

Fix incorrect mentions parsing

See merge request !27