>I'm still pissed about the LSB tests mandating multibyte support when >POSIX requires multibyte support only if you claim support for multibyte >locales. (I.e., we should be able to avoid this problem by saying that >we're not supporting multibyte locales for LSB purposes in sarge.) Last >I heard someone was supposed to issue a clarification on that issue.
Hmmm, I don't think this request made it to the right place. This is not an official statement on the situation; but in general, we certainly don't want the tests to be more stringent than what the spec requires, in all such disputes "the spec wins" and an interpretation (waiver) would be granted. It looks to me like the spec is not calling out explicitly that multibyte locales must be supported, although it's a desire that an LSB conforming system be "ready to go" anywhere, which would imply multibyte support. The lack of upstream support in coreutils for this would be another factor that I expect would tend to weigh in favor of an interpretation being granted.
