This is an error i stumbled upon converting my system to unicode following the gentoo unicode guide. While using the et_EE.utf8 locale with the unicode USE flag set, different packages fail to emerge. The packages - ranging from mozilla to jpeg - are always the same, and every one of them always dies with its own error message. The problem disappeared when i added LC_ALL="C" to make.conf (been running 4 servers and multiple desk/laptops without problems for some 2-3 months for now. I have not tried to duplicate this problem using different locales and/or different architectures. It seems to me that this is an issue with the toolchain and some other apps (e.g. some packages' emerges died with sed errors) not fully supporting unicode. I am, however, not experienced in software development and my suggestion is to modify the gentoo unicode guide to reflect the possible need to set locales in make.conf until unicode works out-of-box.
I haven't got any problems for fi_FI.utf8. Are you sure the problem is caused by UTF-8 and not for example the Estonian locale? Because there's a (little)known problem that in Estonian collation the character range [a-z] doesn't yield all alphabets like most programmers expect, which tends to cause borkage in scripts. Anyhow, I really wouldn't like to add notes to few documentations if this is just an issue of stupid programming from some packagers or upstreams, they should be fixed instead. Should someone correct all documentations related, I think best workaround would be to suggest preferring user-based locale setting over global environment, just to be sure.
i'm 99% sure this has nothing to do with unicode and everything to do with the weird EE alphabet a lot of packages use [A-Z] in grep/sed/etc... which just plain doesnt work in EE
