😳 o-okay
as seen on cctv monitors, now cherry-flavored
remember, there's a person behind everything
| Pronouns | he/they |
| Site | https://eclair.cafe |
| Alt | @vyivel |
as seen on cctv monitors, now cherry-flavored
remember, there's a person behind everything
| Pronouns | he/they |
| Site | https://eclair.cafe |
| Alt | @vyivel |
never really thought about password normalisation before but it came up in a recent audit and now I'm kinda fascinated by it.
TL;DR if you treat passwords as byte strings in some particular encoding (say, UTF-8), and derive keys or hashes from those bytes, it breaks in a bunch of common use-cases with CJK scripts, because different input methods (e.g. desktop vs. mobile keyboard) may generate different encodings or forms/widths of the same visual characters.