We used 'IDENT' for labels as a temporary fix for the  parser conflicts
  introduced if the proper rule 'ident' was used. Now a specially tailored
  'label_ident' rule is used for labels instead.


mysql-test/r/sp.result:
  New test cases for BUG#7088.
mysql-test/t/sp.test:
  New test cases for BUG#7088.
sql/sql_yacc.yy:
  Make labels in stored procedures work with other character sets than latin1.
  Using a separate 'label_ident' rule (instead of 'ident') and splitting
  the 'keyword' rule into two got rid of the temporary fix of using 'IDENT' for
  labels (which didn't heed character sets).