I think that downloading the whole history applies to things like counting commits and attributing them to developers. ohcount should always look at the latest snapshot, and probably is not incremental.

The code in GNU Smalltalk is being recognized mistakenly as C/C++. To recognize it, look for a lot of keywords matching "[a-z][a-zA-Z0-9_]+:". The rules to parse it are extremely easy: the extension ... [More] is "*.st", the comments are delimited by double quotes, strings are delimited by single quotes and do not support escaping (i.e. '\' is a one-character string), and a $ eats the following character whatever it is (e.g. $\ is a two-character sequence, and $' does not require a matching single-quote). [Less]