This is expected - it is side effect of PostgreSQL implementation of function overloading and type conversions
after installation citext, you will have more instances of function strpos
strpos(citext, citext)
strpos(text, text)
Do you think it would be appropriate to simply add an strpos(citext, text) overload to the extension to make sure this behaves more as expected? If so I can try to submit a patch at some point.
It has sense in this case. But I have not a idea if there can be some more broken.