Re: Auditing extension for PostgreSQL (Take 2)

Hi Sawada,

On 3/25/15 9:24 AM, David Steele wrote:
> On 3/25/15 7:46 AM, Sawada Masahiko wrote:
>> 2.
>> I got ERROR when executing function uses cursor.
>>
>> 1) create empty table (hoge table)
>> 2) create test function as follows.
>>
>> create function test() returns int as $$
>> declare
>>     cur1 cursor for select * from hoge;
>>     tmp int;
>> begin
>>     open cur1;
>>     fetch cur1 into tmp;
>>    return tmp;
>> end$$
>> language plpgsql ;
>>
>> 3) execute test function (got ERROR)
>> =# select test();
>> LOG:  AUDIT: SESSION,6,1,READ,SELECT,,,selecT test();
>> LOG:  AUDIT: SESSION,6,2,FUNCTION,EXECUTE,FUNCTION,public.test,selecT test();
>> LOG:  AUDIT: SESSION,6,3,READ,SELECT,,,select * from hoge
>> CONTEXT:  PL/pgSQL function test() line 6 at OPEN
>> ERROR:  pg_audit stack is already empty
>> STATEMENT:  selecT test();
>>
>> It seems like that the item in stack is already freed by deleting
>> pg_audit memory context (in MemoryContextDelete()),
>> before calling stack_pop in dropping of top-level Portal.

This has been fixed and I have attached a new patch.  I've seen this
with cursors before where the parent MemoryContext is freed before
control is returned to ProcessUtility.  I think that's strange behavior
but there's not a lot I can do about it.

The code I put in to deal with this situation was not quite robust
enough so I had to harden it a bit more.

Let me know if you see any other issues.

Thanks,
--
- David Steele
david@pgmasters.net