v0.11.1

Fixed

• number_chars/2 and number_codes/2 read a list that holds the whole text the way the term
  reader reads a number and unify the result with the first argument, so
  number_chars(1.0e9, "1.0E9") holds, as ISO requires; before, a bound number was written out and
  compared as text. The list may start with layout and comments and a - name token, quoted or
  not, so " 1", "- 1", "'-'1", and "/**/1" read as numbers, while "1 ", "+1", and
  "0X1" are syntax errors. With a bound first argument, an element that is not a character or a
  list that is not one raises the ISO error instead of failing. A float beyond the finite range is
  representation_error(max_float). Every case of the ISO conformity table for number_chars/2
  now gives an allowed answer, and a program runs them the same with or without a compiler attached.
• A raw control or layout character after 0', such as a newline or a tab, is a syntax error, as
  it is between quotes, in both modes; write 0'\n or 0'\t instead. So is 0' followed by a
  line continuation, which denotes no character and was read as 0.