Reminds me of the day I spent debugging a SWIG-generated segfault in a ruby binding I was writing. I solved the problem by using an older version of SWIG. The XKCD that day contained both the problem and the solution. --Ken Bloom