(From Pekka via email)
Hi
I use your product, and apart for a few web sites that spend enormous effort to break simple user interfaces find it a delight.
Until today, when I unintentionally had a Japanese IME active (Windows 10) and then tries to paste a username and password into a site.
The keystrokes took on the behaviour of the IME and resulted in a mix of katakana and kanji and the plaintext password in the username/email field. If I had walked away, a (very) knowledgeable observer could have worked out the problem.
While I realise this is rather an edge case, perhaps there is merit in recording the locale used when recording the credentials and selecting the same one when they are pasted. While this may not affect western European languages, it probably applies even to Cyrillic layouts and will likely impact other scripts.
I'd be prepared to test any changes you make in support of your development.