Hrvoje> wmperry(a)aventail.com (William M. Perry) writes: Hrvoje> Right. Worse yet, for compose characters and representations of, say German or Japanese, using just ASCII or another character set that can only approximate the original, d-c-r and e-c-r may actually create ("split") or merge characters. Rounding may not really work in those cases, although I guess it should be OK. The correct pronunciation is "Oh SHIFT, that's SHIT-jis". :-) Hrvoje> In that case I would round it to a character. Hrvoje> No, I don't think so. I think that if you somehow manage to create an extent that ends up with endpoints inside a *-coding-region, you've screwed up somehow anyway so an approximation should be fine. The real problem (and what I was really asking, now that I've slept on it) is "what do you do when the text properties you have applied don't make sense?" For example, Japanese has no italic or bold faces (systematically, in the typography sense); emphasis in text is indicated by underlining (actually, "overdotting") and use of the "katakana" set of phonetic characters. It seems to me that you probably want to split such extents, preserve the meaning outside of the *-coding-region, and do something (probably at user option) inside the *-coding-region. How far do you want to preserve that "all"?, is what I want to know. And which is more important, endpoints (ie, not splitting) or preserving the user's intended semantics (yes, I've slanted that, sorry I don't have better wording). -- University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091 __________________________________________________________________________ __________________________________________________________________________ What are those two straight lines for? "Free software rules."