I'm still alive 
Although I still haven't received any payments from Google (they are
doing their best to solve this), I've began working on implementing
the Knuth's layout algorithms.
I have studied a bit the TeXbook, the existing mathtext parsing code,
and I have decided to rewrite the parsing almost from scratch.
Although I don't no too much about parsing, I think it won't be that
much of a problem.
My idea is to first transform a TeX string to a Python list (tree-like
structure), which can be composed of strings, commands, and/or other
lists, and so on.
Then, I plan to write some classes to trasnform this list/tree to the
actual boxes needed for displaying.
The first part is done (although bugs are possible). Now I'm
concetrating on the remaining part.
The current module is attached. It doesn't need any third-party
libraries currently.
The following works:
Going from:
r"asdf { \horse{}\ \zztop{} \ Hello\^^a^{b_c}}"
to:
['asdf', ' ', [' ', '\\horse', [], '\\space', '\\zztop', [], ' ',
'\\space', 'Hello', '\\circumflex', '\\superscript', 'a',
'\\superscript', ['b', '\\subscript', 'c']]]
Please John, do comment (others with spare time are welcome too :).
texparser.py (6.44 KB)