On Wed, Mar 15, 2006 at 05:54:39PM +0900, Stephen J. Turnbull wrote: > Good luck. Big5 and Shift JIS are not ASCII compatible, they use > octets in the 0-127 range for multibyte characters. Big5 even uses > control characters.
I beg to differ. Both do stomp on ASCII in the second byte, but were sane enough to avoid the control characters, and Darcs only cares about control characters, so that's okay. > UTF-16 standards don't even require a signature > for on-the-wire use, so you can't count on presence of a BOM. I'm aware of that. It's just an additional heuristic that might make Darcs work in more situations. > Because it places the responsibility for breakage squarely on > somebody, and not incidentally breaks things for a significant share > of users if done in isolation, thus providing incentive to find ways > to avoid that. So you want to deliberately break things for lots of people (e.g. all ISO8859-1 users), so that some people will not find that anything new now works, but just that the documentation now says it's their own fault when it breaks? Why not just change the documentation and leave Darcs as-is? > *You* see few problems. Are you in a position to see the problems I > see, or are they unimportant merely because you don't live with them? You just have only very vaguely explained what your problems actually are, and you certainly haven't explained how forcing everyone to use UTF-8 helps matters. (Or rather you haven't explained how /Darcs/ forcing everyone to UTF-8 helps matters; clearly if everyone already used it, there would be no problem.) -- Jamie Webb _______________________________________________ darcs-users mailing list [email protected] http://www.abridgegame.org/mailman/listinfo/darcs-users
