Re: Bug fix for cache lookup failure for statistic_ext type

From: Tomas Vondra <tomas(dot)vondra(at)enterprisedb(dot)com>
To: Mark Dilger <mark(dot)dilger(at)enterprisedb(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Bug fix for cache lookup failure for statistic_ext type
Date: 2021-08-31 17:50:58
Message-ID: d57cdb40-6870-bca2-224f-9728074849d0@enterprisedb.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 8/5/21 12:32 AM, Tomas Vondra wrote:
> On 8/5/21 12:03 AM, Mark Dilger wrote:
>> Hackers,
>>
>> You can easily get a cache lookup failure by changing the regression
>> tests as included in this small patch.  The failure looks thus:
>>
>> +COMMENT ON STATISTICS ab1_a_b_stats IS 'new comment';
>> +CREATE ROLE temp_role;
>> +SET SESSION AUTHORIZATION temp_role;
>> +COMMENT ON STATISTICS ab1_a_b_stats IS 'changed comment';
>> +ERROR:  cache lookup failed for type 27447
>> +DROP STATISTICS ab1_a_b_stats;
>> +ERROR:  cache lookup failed for type 27447
>> +ALTER STATISTICS ab1_a_b_stats RENAME TO ab1_a_b_stats_new;
>> +ERROR:  must be owner of statistics object ab1_a_b_stats
>> +RESET SESSION AUTHORIZATION;
>> +DROP ROLE temp_role;
>>
>> I believe this case simply has not had any test coverage, as I don't
>> see any way the current code would ever work.  It treats the Oid of the
>> statistics object as a type, which it is not.
>>
>
> Yeah, you're right - this is broken since 7b504eb282c. Thanks for the
> fix, I'll get it committed.
>

I've pushed a fix for this. And then a fix for the fix :-( because I
forgot about the rule that role names in regression tests should start
with regress_ prefix, so animals enforcing this failed.

regards

--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Bossart, Nathan 2021-08-31 18:09:36 Re: archive status ".ready" files may be created too early
Previous Message Tom Lane 2021-08-31 17:40:13 Re: Patch: shouldn't timezone(text, timestamp[tz]) be STABLE?