quote:
Originally posted by CookieRevised
In extremely many cases...
Like I said, the password's characters should be handled as unicode, not as ascii. It isn't some small limitation that you can fix later. It is a very serious bug which will result in wrong (un)scrambled passwords and thus in wrong encryption/decryption (no matter if the original password is ascii or not).
eg: test your program with the (pure ascii) password "éééé" (=4 times 0xE9) and you'll see. The scrambled new password is not "ÒÒÒÒ" (=4 times 0xD2), like it is in your current tool, but should be a unicode string containing 4 times the unicode character 0x01D2.
Ahhh thanks, I understand now.
I've modified it so that unicode is used now. I'll never underestimate unicode again
/ –
