A _very_ underfeatured NuParser is now (or will be in a few minutes) in CVS. The fronend's grammar support should be done; its integrated debugger and tester are not even started (nor will they be until the backend is done). The backend only supports 'normal' nodes (i.e., strings). It does not actually reduce at the moment, either. Here is an example grammar: # Comments are denoted by a hash. Blank lines are # ignored. The rule seperator is the newline character. # The comment end character is also the newline # character. However, it often looks best to add extra # newlines where possible to prevent wrapping as in this # comment. "http://" "www." "erols.com" ":80" "/derobert/" "http://" "www." "sourceforge.net" ":80" "/" [Obviously, this grammar is not optimal; it is a test] And the input: http://www.sourceforge.net:80/ And the debugging output: Got new data: http://www.sourceforge.net:80/ Attempting match for: http:// ... Match found. String http:// matched. Attempting match for: www. ... Match found. String www. matched. Attempting match for: erols.com ... Could not match Attempting match for: http:// ... Match found. String http:// matched. Attempting match for: www. ... Match found. String www. matched. Attempting match for: sourceforge.net ... Match found. String sourceforge.net matched. Attempting match for: :80 ... Match found. String :80 matched. Attempting match for: / ... Match found. String / matched. Could not gather new data. If you can come up with a grammar consisting solely of quoted literals that does not parse right, file a bug report. Otherwise, don't bother; it's not supposed to work.
