News
- 09 Sep 2008: New source tarball and Windows pre-compiled executables. Checked with valgrind for memory leak. Tests are still to be fixed!
- 25 Aug 2008: Back to the NFA approach! Also, now every modifier is greedy. The expressions like ".+z" or "\d+3" will never succeed as the first term of the expression will eat all the available characters. Current version under SVN supports:
- Captures (up to 8)
- Back reference to captures (e.g. "{a+)b\1" matches aba, aabaa, etc )
- Negated expressions that match 0 characters of the input.
- 07 Feb 2008: Stuck on an issue on the determinization step. The algorithm described in the PDF and implemented in the code succeedn on (a*)(a)a but fails on ([ab]*)(a)([ab]*). I devised a different algorithm, closely following the standard subset construction, that succeeds on ([ab]*)(a)([ab]*) but, alas, fails on (a*)(a)a ! I guess I have to take a step back an rethink the whole thing. For those interested, the problem is to determine when a set of states needs to be merged again, even if they've been already, because of a different tags assignment. Any suggestion is greatly welcome!
- 04 Feb 2008: New tarball and Windows binary. This version produces C code! Use "sh mktest 'expr1' 'expr2' ..." to create a test program named x (or x.exe). Requires gcc.
- 31 Jan 2008: New tarball and Windows binary. Able to generate ASM-like code. Also performs some optimization.
- 25 Jan 2008: New tarball and Windows binary.
- 21 Jan 2008: Discussion group and SVN commits group created
- 20 Jan 2008: Posted a request for comment about the YRX approach on comp.compilers http://compilers.iecc.com/comparch/article/08-01-051
About YRX
YRX is a tool to ease the creation of lexical scanners similar to re2C. It's in its early stage, any comment or feedback is appreciated.
