If I run use POSIX qw(locale_h strftime); setlocale(LC_ALL, ""); print (strftime("%a, %d %b %Y %H:%M:%S %z", gmtime()), "\n"); under ru_RU.UTF-8, then the result is completely miscoded - the output is not in any legal encoding. The "encoding" of the output can be described as a result of applying a iso-8859-1..utf-8 conversion to a legal UTF-8 encoding of a Cyrillic time string. In other words, currently (perl-5.8.0-50, glibc-2.2.93-2) env LC_ALL=ru_RU.UTF-8 /tmp/test.pl | recode utf-8..iso8859-1 produces a utf-8 encoding of a Cyrillic time string.
Created attachment 76324 [details] Test perl script
Still exists in Phoebe.
Works correctly in Cambridge alpha1 (perl-5.8.1-0.pre19333.1)