I never knew releasing a program was so much work. Between checking all documentation, exporting CVS, uploading tarballs, filling in forms on a multitude of websites, sending email to various lists, argh!
Well, it's done now. Synopsis 0.2 is released. I can go and relax.
It's kind of fun parsing different languages at different levels -- the C++ parser represented the parse-tree as a singly-linked list with lispisms (Car, Cdr, Cons?), the Python parser I've just started uses tuples, and the HTML formatter only uses the Abstract Syntax Tree. For a while I was considering converting the C++ parse-tree into python lists but decided it would slow things down too much. Have you ever seen just how many lines, functions and classes gcc has to deal with just because your code includes an STL header? I'm thinking of implementing some kind of precompilation to speed this up, but I'll have to grep the web for useful papers on the subject before I dive in.