sort is not conformant with POSIX (collation, i18n environment variables)

https://pubs.opengroup.org/onlinepubs/9799919799/utilities/sort.html#tag_20_112_08

The following environment variables shall affect the execution of sort:

LANG

Provide a default value for the internationalization variables that are unset or null. (See XBD 8.2 Internationalization Variables for the precedence of internationalization variables used to determine the values of locale categories.)

LC_ALL

If set to a non-empty string value, override the values of all the other internationalization variables.

LC_COLLATE

Determine the locale for ordering rules.

Under Hungarian locale, sort result is expected as follows:

$ printf '%s\n' g{x..z} | sort
gx
gz
gy

I tested GNU, BSD, and uutils sort’s under product of LANG, LC_ALL, and LC_COLLATE where

• unset
• set to empty
• set to en_US.UTF-8
• set to hu_HU.UTF-8

GNU and BSD behave as the same, in conformity to POSIX, but uu-sort e939eda7bdea0391161eb99687e8a8a7054671bd does not give correct ordering under the following conditions:

e.g.

$ unset LANG
$ export LC_ALL=
$ export LC_COLLATE=hu_HU.UTF-8
$ printf '%s\n' g{x..z} | gnusort
gx
gz
gy
$ printf '%s\n' g{x..z} | ./target/debug/sort
gx
gy
gz

Start by reproducing the Hungarian-locale examples with the uu-sort binary and compare them with GNU sort across the listed LANG, LC_ALL, and LC_COLLATE combinations. Trace sort's locale and collation handling, then verify that the expected gx, gz, gy ordering is produced for every failing environment combination.