>This seems a little weird (as a thing to do), since wouldn't the table >of "word" characters still be used outside of RE searches? > >Also, does this make life any harder should *someone* eventually make >a "plugin" interface to RE searching for, oh say, PCRE? :) The idea is that the table of word characters used by Scintilla for word moves and the table use by RE parsing for words would become the same.
There is a trial implementation at: https://sourceforge.net/tracker/?func=detail&atid=352439&aid=1437523&group_i d=2439 As Neil suggested, I have moved the character classification array into its own file and left the instantiation of it in Document. I pass a pointer to the table into the RE code, so any changes made in Document can be seen by the RE parser. Any future RE evaluator will need a table of words and spaces... It would be a lot neater (now) to include the character class code in the RE code, but if you want to expand the system to plug in RE Evaluators for different lexers/code pages then it has to be separate. Greg _______________________________________________ Scintilla-interest mailing list [email protected] http://mailman.lyra.org/mailman/listinfo/scintilla-interest
