6a6c3bcd5f47cb7b2ff706134abc1eb143311cdc
like western ones, not CJK: words are indexed, not n-grams, but we use the Xapian spellcheck feature on them, not aspell.
Languages
C++
40.3%
C
20.2%
Shell
11.3%
HTML
7.8%
M4
6.8%
Other
13.3%