Re: auto_explain log_verbose causes regression failure

On Wed, Aug 19, 2009 at 10:07 PM, Robert Haas<robertmhaas(at)gmail(dot)com> wrote:
> On Wed, Aug 19, 2009 at 9:39 PM, Andrew Dunstan<andrew(at)dunslane(dot)net> wrote:
>> Robert Haas wrote:
>>>
>>> On Wed, Aug 19, 2009 at 7:57 PM, Andrew
>>> Dunstan<andrew(dot)dunstan(at)pgexperts(dot)com> wrot>
>>>
>>>>
>>>> I am getting a repeatable failure  on the HEAD regression tests when
>>>> auto_explain's log_verbose is set. If auto_explain.log_verbose is turned
>>>> off
>>>> the failure disappears. Data below.
>>>>
>>>> cheers
>>>>
>>>> andrew
>>>>
>>>> config settings:
>>>>
>>>> custom_variable_classes = 'auto_explain'
>>>> auto_explain.log_min_duration = 0
>>>> auto_explain.log_format = 'xml'
>>>> auto_explain.log_analyze = on
>>>> auto_explain.log_verbose = on
>>>> shared_preload_libraries = 'auto_explain'
>>>>
>>>
>>> I can't figure out how to make this config work.  I dumped these
>>> settings into a file called "t" and then did, from src/test/regress,
>>> TEMP_CONFIG=t make check.
>>>
>>
>>
>> I did make install, initdb, put settings in postgresql.conf, then make
>> installcheck
>
> It looks like this is enough to reproduce the cache lookup failure:
>
> begin;
> create table abc (a int);
> insert into abc values (5);
> insert into abc values (10);
> insert into abc values (15);
> declare foo cursor for select * from abc;
> fetch from foo;
> abort;

...and it also looks like the same 3 tests fail in REL8_4_STABLE,
which is the oldest branch that contains auto_explain. So I suspect
this has been broken all along. I still don't know how to fix it, but
at least now I don't feel guilty for breaking it...

...Robert