Currently, if click-and-release is enabled, the OCR functionality is pretty much useless.
If I hit the “Extract Text”-Button, it just takes a screenshot of the whole screen.
I think the “cleanest” solution is to implement “Extract Text” in the same way as crop.
You switch to OCR and select the area you want to extract text from afterwards.
That way it should work whether click-and-release is enabled or not.
Also, this works in a way everyone is already used to for taking their screen captures.
Sadly, I do not possess the skills to make any such changes myself, so I can only suggest and ask if I am overlooking anything in how this may or may not work for everyone.
Are you using Meta+Shift+PrintScreen or just pressing PrintScreen? With accept on click and release, you should get a viewer window where you can do OCR on the cropped image if you arrived at the rectangle UI from PrintScreen. Meta+Shift+PrintScreen is a shortcut to quickly save a rectangle screenshot, so you probably aren’t getting the viewer window if you’re using that one.