Fix bug with character enumeration literal

Fix bug with character enumeration literal

gnat encodes character enumeration literals using a few different
schemes.  The gnat compiler documented the "QU" and "QW" encodings,
but failed to document that a simpler encoding was used for certain
characters.

This patch updates gdb to handle this simple Q encoding.  Note that
wide character literals are still not handled.

gdb/ChangeLog
2019-08-15  Tom Tromey  <tromey@adacore.com>

	* ada-exp.y (convert_char_literal): Handle "Q%c" encoding.
	* ada-lang.c (ada_enum_name): Likewise.

gdb/testsuite/ChangeLog
2019-08-15  Tom Tromey  <tromey@adacore.com>

	* gdb.ada/char_enum.exp: Add regression tests.
	* gdb.ada/char_enum/foo.adb (Char_Enum_Type): Use '_'
	and '0'.
	(Char, Gchar): Update.
	* gdb.ada/char_enum/pck.ads (Global_Enum_Type): Use '+'.