Comment 1 for bug 585648

Hello,

1) I cannot reproduce this on current Debian.
2) The respective code is not located in find, re_match() is part of libc.
3) The fact that regex are locale dependent is expected behavior, e.g. in the Estonian alphabet Z is not the last letter and therefore e.g Y is not in 'A-Z'.
4) To matching upper case letters you should use the respect collation sequence ([[:upper:]] instead of [A-Z]) or reset LC_COLLATE to C.

Given all that, afaict from Google it looks like some in some versions of libc '[A-Z]' includes lower case letters in en_US.UTF-8 locale while in others it does not. See also #120687