No more memory leaks in the code.

Still problems with hashtable acl_roles
Need to create a copy of grant_roles, currently it uses the same buffer

The function calls delete_dynamic, after if calls a free function on every
array element.

TODO:
Memory allocated for the array is never freed

Changed ACL_USER.user from char * to LEX_STRING.
Refactored every section that made use of ACL_USER.user as a char*.

This was done so as to be able to quickly check the hash_key of the acl_user.

role_grants is no longer used as it will be added to the ACL_USER class

It was a leftover from attempting to use Hash_set

There is an issue with correct searching of keys in the HASH.

TODO: Free structures on flush

MDEV-5123 Remove duplicated conditions pushed both to join_tab->select_cond and join_tab->cache_select->cond for blocked joins.
    
BNL and BNLH joins pre-filter the records from a joined table via JOIN_TAB::cache_select->cond.
There is no need to re-evaluate the same conditions via JOIN_TAB::select_cond. This patch removes
the duplicated conditions from the top-level conjuncts of each pushed condition.

The added "Using where" in few EXPLAINs is due to taking into account tab->cache_select->cond
in addition to tab->select_cond in JOIN::save_explain_data_intern.

depends on uninitialised value(s) in JOIN::save_explain_data_intern"
- Make find_best()  /* the old join optimizer code */ also use table
  condition selectivity.

Local variable table_name_buffer went out of scope while
its content was still being used by a String instance.
Moved the variable to the function scope.

fix for SP & PS

- It turns out, there are statements that will call lex_start(thd->lex) 
  after parsing has been finished. lex_start() will set lex->explain=NULL,
  which will lose the pointer to already allocated Explain_plan object.
- To get rid of this, switch to lazy creation of lex->explain.  Now, it is 
  created only when we get a part ot query plan.

- MYSQL_MULTI_DELETE_DONE probe compile failure
- show_explain_non_select.test