On Tue, 2005-07-05 at 15:30 -0400, Mark E. Mallett wrote:
I do case-insensitive comparisons, because in general I prefer to ignore
differences in case unless there's a reason not to; however, our script
generation tools always generate lowercase. So changing to case-sensitive
wouldn't hurt here, and I wouldn't really object to it.
However, there has been other support for case-insensitivity in other
discussions; e.g. for variable names and in the Sieve language
definition itself. As I say, I don't particularly object to making
"require" case-sensitive, but consistency is always good (as is the "be
liberal in what you accept" tenet).
I agree, case-insensitive matching is pervasive in IETF standards. e.g.
all MIME types and parameters are case-insensitive, and everything else
in Sieve is case-insensitive. the problem is that either way we will be
making some existing implementations non-conforming. if the role of
3028bis is to document current practice, case-sensitive seems to be the
appropriate choice, even if it is a wart.
--
Kjetil T.