Hello, right now sort(1) in stable/10 sorts UTF-8 text files incorrectly if I have LC_CTYPE=ru_RU.UTF-8 set. This is probably due to the problems described in recent "UTF-8 Sorting" thread on -stable, though it is a clear regression when moving from GNU sort from stable/9. I would rather ignore LC_CTYPE completly (until is it broken) and produce sort results as if LC_CTYPE=C. Or at least make a big warning in sort(1) man-page. Right now it looks rather disappointing. Thanks.