The "badwords" script we use in curl CI to detect words in documentation and source code that we want to avoid took 48 seconds to run just a few days ago.

Then it took 8 seconds after an optimization take by me.

Then 3.6 once the regexes were improved by @icingStefan Eissing

Then 1.1 seconds with more tweaks on the regex use.

Now @vsz works at taking it yet a little further...