Case-insensitive filesystem lookups

Posted May 23, 2018 21:27 UTC (Wed) by saffroy (guest, #43999)
In reply to: Case-insensitive filesystem lookups by Sesse
Parent article: Case-insensitive filesystem lookups

Indeed, locale matters, and this means the definition of case also depends on the character encoding.

Unicode case folding does simplify things, but AFAIK the Turkish "i" still requires some special care (see this page for a detailed explanation).

to post comments

Case-insensitive filesystem lookups

Posted May 25, 2018 18:58 UTC (Fri) by wahern (subscriber, #37304) [Link]

There's really only a single choice for doing case-insensitive names: rely on Unicode case-folding rules. It's the only choice because it's the only one, AFAIK, with an explicit guarantee regarding forward stability:

For each string S containing characters only from a given Unicode version, toCasefold(toNFKC(S)) under that version is identical to toCasefold(toNFKC(S)) under any later version of Unicode.

Backward compatibility--mounting a volume created on a system supporting Unicode X-1 that was created on a system supporting Unicode X-- is problematic but I don't think that's entirely avoidable.