>>>>> "Lars" == Lars Gullik Bjønnes <[EMAIL PROTECTED]> writes:
Lars> f.ex. a construct that would make us more robust is: Lars> vector<string> a = getVectorFromString(l, " \n\t\r"); Lars> Where l is f.ex. a line read from a file, and I just want to Lars> split on all "whitespace". (or could be a whole file for that Lars> matter...) We have lyxlex for that :) Real parsing is better than approximations... And I know you don't want to use lyxlex for parsing because of the keyword table issue. Let me say this is bogus since separating the class into LyxParse and LyxLex would be so easy. I know this is not really on-topic, but I never understood your excuses for using regexps for parsing files (cdef files, for ex). Regexp will never catch corner cases as well as a parser. JMarc
