Parse -1.
IND00 as the floating point number returned by nan(), -1.INF00 as -inf(), and 1.INF00 as inf(). Note that the C99 standard specifies that a variety of formats like "NaN" and "nan" are to be used; these are easier to parse yourself and not currently supported by readNumber.
An alternative to specifying msvcSpecials is to read numbers as:
Token x = t.read();
Token y = t.peek();
if ((x.string() == "-1.") &&
(y.string() == "#INF00") &&
(y.character() == x.character() + 3) &&
(y.line() == x.line()) {
t.read();
return nan();
}
... similar cases for inf
If the single-comment character was #, the floating point special format overrides the comment and will be parsed instead.
If signedNumbers is false msvcSpecials will not be parsed.
Default is true.