Robustness to control chars

[bittlingmayer]

Certain input causes the error not well-formed (invalid token).

(That's the value of e in parser.on('error', e...).)

If we use bash's built-in xmllint, the error message is more revealing:

parser error : PCDATA invalid Char value 8

ASCII char 8 is of course a control char, Backspace.

Is this expected? Or is there an option to let the parser handle or ignore such segments?

Right now a large file can fail cryptically just because 1 or 2 segments in a million have this character which is like any ASCII not especially exotic.