Not sure if its worth adding locale stuff to the tests right now. It's not like anything is looking at it or am I missing something?

man comm specifies that comm respects LC_COLLATE in regard to determining whether a line is later in sort order from the previous line or not. respect_lc_collate (a test added in the second commit) is an example of this, where when LC_COLLATE=C we expect that uppercase alphabetical characters are considered as before lowercase in sort order, and so a file with a row that line that starts with uppercase that follows a line that starts with lowercase should throw an error during a call like LC_COLLATE=C comm --check-order

LC_COLLATE is overridden by LC_ALL, a commonly set locale variable. I defaulted to explicitly providing the locale of en_US.UTF-8 out of (possibly reflecting my unworldliness / limited experience in these matters) concern for the possibility that someone may run these tests in a locale that uses ascii characters but considers them as having a different sort order than the sort order that was considered when writing the tests.

Nothing in uutils or the rust std lib looks at these variables, so it has no effect right now. If we ever add support for locales that's most likely through an external crate, but i don't expect that to happen anytime soon. So until we know how we want to support locales (or at all), i think it makes little sense to add locale specific tests to the utils.

cc @Heather @Arcterus for opinions

@ebfe well locale is important sure but I'm also agreed that it should be added globally, that's all I can say

@ebfe @Heather references to locale rm'd.