Fix warning: leaked hash_seq_search scan for hash table.

…r 9 and 10.

…r 9.5.

+ use uniform algorithm for status lists
+ rename struct tagHtabToStat to tagVariableStatEntry
+ use context instead of bunch of args in list workaround routines

At the moment pg_variables can't handle cursors in a "proper" way.
It's possible to remove pg_variables while have active cursor open,
which seem to be erratic and couse sanitizer to be unpappy about it.

+ Compatibility check for Postgres Professional Enterprise.
+ Fix for cursors: switch off hash leak message.
+ Switch off cursor test, because cursors not supported completely.

Sanitizer did failed on query like "SELECT pgv_select(...) LIMIT 1"

Use pg_compatibility_check_no_error function to precheck compatibility in order to free memory.