> I think the heart of the matter is that JS_EvaluateScript cannot cope > with non-ASCII things, even if encoded in UTF8.
Yes it can, as shown in my previous example. utf8 chars can be in strings, or even regular expressions. This works. var x = "a b"; alert(x.replace(/ /, "-")); It prints out a-b. It doesn't seem to handle breakspace as whitespace however, and yet it should, and apparently does in other browsers. I don't think converting everything to unicode would change anything. It would still accept my code fragment above, and likely barf on breakspace elsewhere. As a really imperfect solution, I'm thinking about changing every breakspace to space, unless: it is in a line less 200 characters && the line does not start with // && it is not part of a string wholly contained in this line using a rather simple " " criterion. I'm trying to avoid writing a js scanner, which is really not a trivial thing, beyond what lex can handle, mostly because of those cursed regular expressions that aren't even quoted, just appear free, and we're suppose to recognize them as such, even though almost any character can appear between the slashes. I know about this nightmare because I parsed and ran my own javascript in edbrowse version 2. It's really best left to someone else! But then again, that someone else isn't handling breakspace properly. Karl Dahlke _______________________________________________ Edbrowse-dev mailing list [email protected] http://lists.the-brannons.com/mailman/listinfo/edbrowse-dev
