> Vlad,
>
> Also why was the syntax changed to "SET PER_STATEMENT", whereas the original
> patch had "SET STATEMENT"?
>
> SET STATEMENT follows the current logic of possible SET modifiers (i.e.
> "GLOBAL", "SESSION"), i.e. you specify the scope of the variable being set.
> But PER_STATEMENT is not in line with that naming.
That's because of some tests were failed due to intersection the "STATEMENT" keyword with one of the possible values of binlog_format variable (set binlog_format=STATEMENT).
> Vlad,
>
> Also why was the syntax changed to "SET PER_STATEMENT", whereas the original
> patch had "SET STATEMENT"?
>
> SET STATEMENT follows the current logic of possible SET modifiers (i.e.
> "GLOBAL", "SESSION"), i.e. you specify the scope of the variable being set.
> But PER_STATEMENT is not in line with that naming.
That's because of some tests were failed due to intersection the "STATEMENT" keyword with one of the possible values of binlog_format variable (set binlog_ format= STATEMENT) .