At Fri, 31 Jul 2009 18:42:37 -0700, Andreas Raab wrote:
BTW, one thing I was wondering about is how line breaking is assumed to work with CJK-scripts (or more generally with Unicode). Do you have any information on that? It seems that we're treating all of the CJK characters as non-breakable at this point which leads to odd effects when there are a few spaces included in the text.
Unicode has a standard algorithm to find possible positions, but the entire thing can be really complex. For Japanese, #isBreakableAt:in: defines a very similar rule to the one used in Emacs to more or less conform the standard writing rules in Japanese. And, the rule cannot be totally language independent.
To make JapaneseEnvironment version of isBreakableAt:in: take effects, the system has to know to that the characters around it is supposed to be in Japanese. If the language tag is set, it looks reasonable like the attachment.
-- Yoshiki