From: Milan B. <albis@EUnet.yu> - 2005-11-26 09:00:50
|
Michael Hieke wrote: > Milan Babuskov wrote: >> The main problem that remains is the statement parser. As far as I >> understand, it is possible to have a table called "with space". >> However, that breaks our existing parser. Basically, we need a >> tokenizer, which would detect double-quoted identifiers, >> single-quoted strings and both types of comment (-- and /**/) and >> give us the parts of statements. I might look into that after I clean >> up the Identifier stuff. > > I can look into this if you don't have the time. Using wxString instead > of std::string would be slower, but we should maybe not change too many > things at the same time. Michael, I'd like to get the tokenizer/parser as soon as possible. Since I'm still waiting for Olivier regarding the other features, this is the only thing I can work on to bring us closer to 1.0. Once we have the tokenizer working, we can release 0.5.0. If you'd like to work on tokenizer, I'm fine with that. It's most important use is in Database::parseCommitedSql(). If you take the tokenizer, I'll probably work on MetadataSearch stuff in the meantime (which is completely isolated issue), so we don't step on each others toes. If not, then I'll try to bite the bullet and write a tokenizer myself. BTW, I've read into some parsers like Boost's Spirit and ANTRL, but I'm not sure we need that kind of overhead or complexity. -- Milan Babuskov http://fbexport.sourceforge.net http://www.flamerobin.org |