On Wed, Jan 29, 2014 at 2:21 PM, Andreas Karlsson <andreas@proxel.se> wrote:
> On 01/28/2014 09:39 PM, Tom Lane wrote:
>>
>> I'm for doing the measurement in ExplainOneQuery() and not printing
>> anything in code paths that don't go through there.
>
> Reading the discussion here and realizing that my last patch was wrong I am
> now in agreement with this. I have attached a patch which no longer tries to
> measure planning time for prepared statements.
Cool. I propose adding one parameter rather than two to
ExplainOnePlan() and making it of type instr_time * rather than
passing an instr_time and a bool. If you don't want to include the
planning time, pass NULL; if you do, pass a pointer to the instr_time
you want to print. I think that would come out slightly neater than
what you have here.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company