Do not unnecessarily inject the '=' function into classification data.

A statement like INSERT INTO t SET f = 42 will be reported to be using the function =, but so will INSERT INTO t (f) VALUES (42). In the latter case it is clearly nonsensical and in the former case quite dubious as = in that context is assignment and not comparison. From the point of view of the firewall filter (no longer supported in 23.02) and the masking filter the use of = for comparison is meaningful, but not for assignment.

In the latter case, since = is not present in the statement, it means that extra scratch memory will have to be allocated in the classification data. In DEBUG mode it will cause a warning to be logged.