Am 07.04.2011 um 12:29 schrieb Jean-Marc Lasgouttes: > Le 07/04/2011 11:11, Stephan Witt a écrit : >> What exactly looks weird? >> The test for hard hyphen altogether or the test for '-' or for space around >> it >> or some other implementation detail? > > The method that tests for the standalone hyphen seems complicated to me and > hard-codes one particular case. There is a similar problem with ' character, > for example, that can either be part of the word "don't" or not, like in > possessive forms. > > I think hyphen and commas are in the class of characters that are part of the > word only when there is stuff on their right and on their left. The > discussion I pointed to earlier from mozilla would probably be enlightening > and give solutions that feel better than special casing.
In principle you're right. But then the method "Paragraph::isWordSeparator(pos_type pos)" did it wrong since ages then. I cannot imagine a patch that provides an implementation of the "mozilla discussion" be applied in RC stage of development. > However, I do not have more intelligent things to propose right now. I'll try to make another patch addressing the ' character case like the hard hypen test. This one I'll present then and we can decide if it is save enough to apply it. Stephan
