8.3 support broken; AbstractJdbc2DatabaseMetaData.getColumnPrivileges uses 8.4-only attribute about pgjdbc

[davecramer]

Craig,

The only one remarkable is the attacl problem. The others are all rounding
errors

Dave Cramer

On Fri, Sep 21, 2012 at 9:23 AM, Craig Ringer [email protected]:

Tests of master (9d558509d55850)
against 8.3 show that it's broken with 8.3. Unsure what broke it or what
the oldest working server is yet, just noting this so far.

[junit] Testcase: testColumnPrivileges(org.postgresql.test.jdbc2.DatabaseMetaDataTest):    Caused an ERROR
[junit] ERROR: column a.attacl does not exist
[junit]   Position: 38
[junit] org.postgresql.util.PSQLException: ERROR: column a.attacl does not exist
[junit]   Position: 38
[junit]     at org.postgresql.core.v3.QueryExecutorImpl.receiveErrorResponse(QueryExecutorImpl.java:2157)
[junit]     at org.postgresql.core.v3.QueryExecutorImpl.processResults(QueryExecutorImpl.java:1886)
[junit]     at org.postgresql.core.v3.QueryExecutorImpl.execute(QueryExecutorImpl.java:255)
[junit]     at org.postgresql.jdbc2.AbstractJdbc2Statement.execute(AbstractJdbc2Statement.java:546)
[junit]     at org.postgresql.jdbc2.AbstractJdbc2Statement.executeWithFlags(AbstractJdbc2Statement.java:417)
[junit]     at org.postgresql.jdbc2.AbstractJdbc2Statement.executeQuery(AbstractJdbc2Statement.java:302)
[junit]     at org.postgresql.jdbc2.AbstractJdbc2Statement.executeQuery(AbstractJdbc2Statement.java:276)
[junit]     at org.postgresql.jdbc2.AbstractJdbc2DatabaseMetaData.getColumnPrivileges(AbstractJdbc2DatabaseMetaData.java:2671)
[junit]     at org.postgresql.test.jdbc2.DatabaseMetaDataTest.testColumnPrivileges(DatabaseMetaDataTest.java:404)
[junit]
[junit]
[junit] Testcase: testGetTimestampWTZ(org.postgresql.test.jdbc2.TimestampTest):     FAILED
[junit] expected:<2000-07-07 23:00:00.123> but was:<2000-07-07 23:00:00.122999>
[junit] junit.framework.AssertionFailedError: expected:<2000-07-07 23:00:00.123> but was:<2000-07-07 23:00:00.122999>
[junit]     at org.postgresql.test.jdbc2.TimestampTest.timestampTestWTZ(TimestampTest.java:434)
[junit]     at org.postgresql.test.jdbc2.TimestampTest.testGetTimestampWTZ(TimestampTest.java:186)
[junit]
[junit]
[junit] Testcase: testSetTimestampWTZ(org.postgresql.test.jdbc2.TimestampTest):     FAILED
[junit] expected:<2000-07-07 23:00:00.123> but was:<2000-07-07 23:00:00.122999>
[junit] junit.framework.AssertionFailedError: expected:<2000-07-07 23:00:00.123> but was:<2000-07-07 23:00:00.122999>
[junit]     at org.postgresql.test.jdbc2.TimestampTest.timestampTestWTZ(TimestampTest.java:434)
[junit]     at org.postgresql.test.jdbc2.TimestampTest.testSetTimestampWTZ(TimestampTest.java:256)
[junit]
[junit]
[junit] Testcase: testGetTimestampWOTZ(org.postgresql.test.jdbc2.TimestampTest):    FAILED
[junit] null
[junit] junit.framework.AssertionFailedError
[junit]     at org.postgresql.test.jdbc2.TimestampTest.timestampTestWOTZ(TimestampTest.java:517)
[junit]     at org.postgresql.test.jdbc2.TimestampTest.testGetTimestampWOTZ(TimestampTest.java:307)
[junit]
[junit]
[junit] Testcase: testSetTimestampWOTZ(org.postgresql.test.jdbc2.TimestampTest):    FAILED
[junit] null
[junit] junit.framework.AssertionFailedError
[junit]     at org.postgresql.test.jdbc2.TimestampTest.timestampTestWOTZ(TimestampTest.java:517)
[junit]     at org.postgresql.test.jdbc2.TimestampTest.testSetTimestampWOTZ(TimestampTest.java:399)
[junit]
[junit]
[junit] Test org.postgresql.test.jdbc2.Jdbc2TestSuite FAILED

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/10.

[ringerc]

On 09/21/2012 09:28 PM, Dave Cramer wrote:

Craig,

The only one remarkable is the attacl problem. The others are all
rounding
errors

Looks like that broke in:

commit 367663d
Author: Dave Cramer [email protected]
Date: Mon Feb 13 16:43:57 2012 -0500

 resolve ACL getTablePriveledges for later servers

which added a.attacl to the query on
org/postgresql/jdbc2/AbstractJdbc2DatabaseMetaData.java:2642.

Any idea what server version that got added in? Ideas for how to avoid
adding another version-specific query to maintain there, which will
become another thing that has to be tested with a particular narrow
range of server versions or it'll break?

As for the rest: Gotcha, the rounding issues aren't breakage. It'd be
nice to make the tests more tolerant by knowing about server version
specific results, but that's something I can TODO for later, not
something important.

Craig Ringer

[davecramer]

Craig,

8.4 is where it first appeared.

http://www.postgresql.org/docs/8.3/static/catalog-pg-attribute.html

If you go to the top of the page you can click on 8.4 and see it show up.

I can fix this if you like or if you are in the code and motivated.

Dave Cramer

On Fri, Sep 21, 2012 at 9:47 AM, Craig Ringer [email protected]:

On 09/21/2012 09:28 PM, Dave Cramer wrote:

Craig,

The only one remarkable is the attacl problem. The others are all
rounding
errors

Looks like that broke in:

commit 367663d
Author: Dave Cramer [email protected]
Date: Mon Feb 13 16:43:57 2012 -0500

resolve ACL getTablePriveledges for later servers

which added a.attacl to the query on
org/postgresql/jdbc2/AbstractJdbc2DatabaseMetaData.java:2642.

Any idea what server version that got added in? Ideas for how to avoid
adding another version-specific query to maintain there, which will
become another thing that has to be tested with a particular narrow
range of server versions or it'll break?

As for the rest: Gotcha, the rounding issues aren't breakage. It'd be
nice to make the tests more tolerant by knowing about server version
specific results, but that's something I can TODO for later, not
something important.

Craig Ringer

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/10#issuecomment-8765118.

[ringerc]

Sorry, I should've just checked. I forget the catalogs are so well documented.

I'm going to have to go to sleep as it's 10pm here, so if you get the chance to sort it out that'd be fantastic.

BTW, I'm going to be out of action most of next week due to family stuff. Sorry. I've tried to get what I can done before that.

[davecramer]

Ok, I'll have a look at it

Have a great weekend.

Thanks,

Dave Cramer

On Fri, Sep 21, 2012 at 9:53 AM, Craig Ringer [email protected]:

Sorry, I should've just checked. I forget the catalogs are so well
documented.

I'm going to have to go to sleep as it's 10pm here, so if you get the
chance to sort it out that'd be fantastic.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/10#issuecomment-8765300.

[davecramer]

Well it's broken much worse than I thought

Dave Cramer

On Fri, Sep 21, 2012 at 9:54 AM, Dave Cramer [email protected] wrote:

Ok, I'll have a look at it

Have a great weekend.

Thanks,

Dave Cramer

On Fri, Sep 21, 2012 at 9:53 AM, Craig Ringer [email protected]:

Sorry, I should've just checked. I forget the catalogs are so well
documented.

I'm going to have to go to sleep as it's 10pm here, so if you get the
chance to sort it out that'd be fantastic.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/10#issuecomment-8765300.

[ringerc]

Looks like you sorted that out. Thanks very much. I'll look at fixing the timestamp tests so other breakage doesn't hide in the noise there.