(All the markdown file extensions it knows, plus rst and asciidoc. Reference: https://github.com/github/linguist/blob/master/lib/linguist/languages.yml )
(the list will still be truncated after the 9 billionth item)
When the string is read (from Mojo UA), it's not flagged as utf8. To then encode_utf8($str) (convert a Str to a Buf, basically), and then write that buf to a :encoding<UTF-8> handle, causes double-encoding. Instead, the raw string should be decoded from utf8, which turns on the utf8 flag. Then writing to the :encoding<utf-8> filehandle will do the right thing.
…eeded This reverts commit e5af4fc.