Problem: "Data truncated" warning was incorrectly generated
when storing a Japanese character encoded in utf8
into a cp932 column.
Reason: Incorrect wrong warning condition
compared the original length of the character in bytes
(which is 3 in utf8) to the converted length of the
character in bytes (which is 2 in cp932).
Fix: use "how many bytes were scanned from input" instead
of "how many bytes were put to the column" in the condition.


mysql-test/r/ctype_cp932.result:
  Adding test case
mysql-test/t/ctype_cp932.test:
  Adding test case
sql/field.cc:
  Change warning condition from:
  "if number of bytes written to destination is less than full source length"
  to
  "if number of bytes read from source is less than full source length"