Thread: Buildfarm now building but with test failures

Buildfarm now building but with test failures

From

"Mike Fowler"

Date:

04 January 2012, 09:52:01

Hi Dave,

I've noticed that the buildfarm failures have changed from jdbc-build to
jdbc-test so your work to fix the 1.4 problem appears good. The failures
for 8.4, 9.0 and 9.1 are all the same and look like so:

[junit] Testcase:
sslhostnossl8-disableGG2(org.postgresql.test.ssl.SslTest):        FAILED
[junit] Exception thrown: Connection refused. Check that the hostname and
port are correct and that the postmaster is accepting TCP/IP connections.
[junit] junit.framework.AssertionFailedError: Exception thrown: Connection
refused. Check that the hostname and port are correct and that the
postmaster is accepting TCP/IP connections.
[junit]     at org.postgresql.test.ssl.SslTest.driver(SslTest.java:46)
[junit]     at org.postgresql.test.ssl.SslTest.runTest(SslTest.java:109)

I presume that these are failing as I have not enabled SSL in my pg build.
Unless someone corrects this presumption I will add SSL support to my
builds today.

HEAD is failing in a completely different way. A sample from the logs:

[junit] Testcase:
testCrossReference(org.postgresql.test.jdbc2.DatabaseMetaDataTest):
Caused an ERROR
[junit] ERROR: language "SQL" does not exist
[junit] org.postgresql.util.PSQLException: ERROR: language "SQL" does not
exist
[junit]     at
org.postgresql.core.v3.QueryExecutorImpl.receiveErrorResponse(QueryExecutorImpl.java:2159)
[junit]     at
org.postgresql.core.v3.QueryExecutorImpl.processResults(QueryExecutorImpl.java:1888)
[junit]     at
org.postgresql.core.v3.QueryExecutorImpl.execute(QueryExecutorImpl.java:257)
[junit]     at
org.postgresql.jdbc2.AbstractJdbc2Statement.execute(AbstractJdbc2Statement.java:555)
[junit]     at
org.postgresql.jdbc2.AbstractJdbc2Statement.executeWithFlags(AbstractJdbc2Statement.java:403)
[junit]     at
org.postgresql.jdbc2.AbstractJdbc2Statement.execute(AbstractJdbc2Statement.java:395)
[junit]     at
org.postgresql.test.jdbc2.DatabaseMetaDataTest.setUp(DatabaseMetaDataTest.java:50)

The full test logs can be viewed at
http://www.pgbuildfarm.org/cgi-bin/show_stage_log.pl?nm=piapiac&dt=2012-01-04%2005%3A00%3A01&stg=jdbc-test
and it should be noted that failures are only in the JDBC2 build, 2
optional, 3 and XA are all OK. Also this is results for 1.4, 1.5 and 1.6
did not run as 1.4 failed. I will run 1.5 and 1.6 seperately to see if
they fail.

Regards,
--
Mike Fowler
Registered Linux user: 379787

Re: Buildfarm now building but with test failures

From

Dave Cramer

Date:

04 January 2012, 09:56:36

Mike,

Thanks for the update.

Yes, you have to enable ssl for the tests to succeed.


Curious about HEAD failing though .

Dave Cramer

dave.cramer(at)credativ(dot)ca
http://www.credativ.ca



On Wed, Jan 4, 2012 at 5:51 AM, Mike Fowler <mike@mlfowler.com> wrote:
> Hi Dave,
>
> I've noticed that the buildfarm failures have changed from jdbc-build to
> jdbc-test so your work to fix the 1.4 problem appears good. The failures
> for 8.4, 9.0 and 9.1 are all the same and look like so:
>
> [junit] Testcase:
> sslhostnossl8-disableGG2(org.postgresql.test.ssl.SslTest):        FAILED
> [junit] Exception thrown: Connection refused. Check that the hostname and
> port are correct and that the postmaster is accepting TCP/IP connections.
> [junit] junit.framework.AssertionFailedError: Exception thrown: Connection
> refused. Check that the hostname and port are correct and that the
> postmaster is accepting TCP/IP connections.
> [junit]     at org.postgresql.test.ssl.SslTest.driver(SslTest.java:46)
> [junit]     at org.postgresql.test.ssl.SslTest.runTest(SslTest.java:109)
>
> I presume that these are failing as I have not enabled SSL in my pg build.
> Unless someone corrects this presumption I will add SSL support to my
> builds today.
>
> HEAD is failing in a completely different way. A sample from the logs:
>
> [junit] Testcase:
> testCrossReference(org.postgresql.test.jdbc2.DatabaseMetaDataTest):
> Caused an ERROR
> [junit] ERROR: language "SQL" does not exist
> [junit] org.postgresql.util.PSQLException: ERROR: language "SQL" does not
> exist
> [junit]     at
> org.postgresql.core.v3.QueryExecutorImpl.receiveErrorResponse(QueryExecutorImpl.java:2159)
> [junit]     at
> org.postgresql.core.v3.QueryExecutorImpl.processResults(QueryExecutorImpl.java:1888)
> [junit]     at
> org.postgresql.core.v3.QueryExecutorImpl.execute(QueryExecutorImpl.java:257)
> [junit]     at
> org.postgresql.jdbc2.AbstractJdbc2Statement.execute(AbstractJdbc2Statement.java:555)
> [junit]     at
> org.postgresql.jdbc2.AbstractJdbc2Statement.executeWithFlags(AbstractJdbc2Statement.java:403)
> [junit]     at
> org.postgresql.jdbc2.AbstractJdbc2Statement.execute(AbstractJdbc2Statement.java:395)
> [junit]     at
> org.postgresql.test.jdbc2.DatabaseMetaDataTest.setUp(DatabaseMetaDataTest.java:50)
>
> The full test logs can be viewed at
> http://www.pgbuildfarm.org/cgi-bin/show_stage_log.pl?nm=piapiac&dt=2012-01-04%2005%3A00%3A01&stg=jdbc-test
> and it should be noted that failures are only in the JDBC2 build, 2
> optional, 3 and XA are all OK. Also this is results for 1.4, 1.5 and 1.6
> did not run as 1.4 failed. I will run 1.5 and 1.6 seperately to see if
> they fail.
>
> Regards,
> --
> Mike Fowler
> Registered Linux user: 379787
>

Re: Buildfarm now building but with test failures

From

Mikko Tiihonen

Date:

04 January 2012, 11:43:30

On 01/04/2012 12:55 PM, Dave Cramer wrote:
> Mike,
>
> Thanks for the update.
>
> Yes, you have to enable ssl for the tests to succeed.

Should we change test suite so that tests that require SSL are skipped
if the feature is not enabled in server.

-Mikko

Re: Buildfarm now building but with test failures

From

Dave Cramer

Date:

04 January 2012, 12:56:15

Makes sense. Although it may mean that ssl never gets tested

Dave Cramer

dave.cramer(at)credativ(dot)ca
http://www.credativ.ca



On Wed, Jan 4, 2012 at 7:43 AM, Mikko Tiihonen
<mikko.tiihonen@nitorcreations.com> wrote:
> On 01/04/2012 12:55 PM, Dave Cramer wrote:
>>
>> Mike,
>>
>> Thanks for the update.
>>
>> Yes, you have to enable ssl for the tests to succeed.
>
>
> Should we change test suite so that tests that require SSL are skipped
> if the feature is not enabled in server.
>
> -Mikko
>
> --
> Sent via pgsql-jdbc mailing list (pgsql-jdbc@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-jdbc

Re: Buildfarm now building but with test failures

From

Tom Lane

Date:

04 January 2012, 15:26:56

"Mike Fowler" <mike@mlfowler.com> writes:
> HEAD is failing in a completely different way. A sample from the logs:
> [junit] ERROR: language "SQL" does not exist

I believe that's something Robert changed recently.  We used to forcibly
downcase language names even if quoted, but he successfully argued that
that was inconsistent and unnecessary.  So I'm guessing your test suite
contains CREATE FUNCTION ... LANGUAGE "SQL", and you should remove the
quotes.

[ digs in logs... ]  Ah, here we are:

Author: Robert Haas <rhaas@postgresql.org>
Branch: master [67dc4eed4] 2011-11-17 14:25:18 -0500

    Remove ancient downcasing code from procedural language operations.

    A very long time ago, language names were specified as literals rather
    than identifiers, so this code was added to do case-folding.  But that
    style has ben deprecated for many years so this isn't needed any more.
    Language names will still be downcased when specified as unquoted
    identifiers, but quoted identifiers or the old style using string
    literals will be left as-is.

            regards, tom lane

Re: Buildfarm now building but with test failures

From

Dave Cramer

Date:

04 January 2012, 20:38:17

Tom,

At what version will the server accept non quoted language identifier ?

We need the tests to pass on at least 8.4 I would think.

Dave Cramer

dave.cramer(at)credativ(dot)ca
http://www.credativ.ca



On Wed, Jan 4, 2012 at 11:26 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> "Mike Fowler" <mike@mlfowler.com> writes:
>> HEAD is failing in a completely different way. A sample from the logs:
>> [junit] ERROR: language "SQL" does not exist
>
> I believe that's something Robert changed recently.  We used to forcibly
> downcase language names even if quoted, but he successfully argued that
> that was inconsistent and unnecessary.  So I'm guessing your test suite
> contains CREATE FUNCTION ... LANGUAGE "SQL", and you should remove the
> quotes.
>
> [ digs in logs... ]  Ah, here we are:
>
> Author: Robert Haas <rhaas@postgresql.org>
> Branch: master [67dc4eed4] 2011-11-17 14:25:18 -0500
>
>    Remove ancient downcasing code from procedural language operations.
>
>    A very long time ago, language names were specified as literals rather
>    than identifiers, so this code was added to do case-folding.  But that
>    style has ben deprecated for many years so this isn't needed any more.
>    Language names will still be downcased when specified as unquoted
>    identifiers, but quoted identifiers or the old style using string
>    literals will be left as-is.
>
>                        regards, tom lane

Re: Buildfarm now building but with test failures

From

Tom Lane

Date:

04 January 2012, 20:46:55

Dave Cramer <pg@fastcrypt.com> writes:
> At what version will the server accept non quoted language identifier ?

A quick test shows that 7.1 was the last version that insisted on a
string literal there.  7.1 didn't have the CREATE OR REPLACE FUNCTION
syntax either ...

            regards, tom lane

Re: Buildfarm now building but with test failures

From

Dave Cramer

Date:

05 January 2012, 00:13:54

Thanks,

committed a fix for this.


Dave Cramer

dave.cramer(at)credativ(dot)ca
http://www.credativ.ca



On Wed, Jan 4, 2012 at 4:46 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Dave Cramer <pg@fastcrypt.com> writes:
>> At what version will the server accept non quoted language identifier ?
>
> A quick test shows that 7.1 was the last version that insisted on a
> string literal there.  7.1 didn't have the CREATE OR REPLACE FUNCTION
> syntax either ...
>
>                        regards, tom lane

Re: Buildfarm now building but with test failures

From

Bodor András

Date:

05 January 2012, 16:50:28

I am afraid, simply enabling ssl will not be enough
to pass the test. Different databases with different
pg_hba.conf configurations must be set up, as
documented in certdir/README.
 I do not think, it can be expected from every buildfarm
host, so ssl tests might be switched off by default, or
we must streamline its customization.
           Andras

> Mike,
>
> Thanks for the update.
>
> Yes, you have to enable ssl for the tests to succeed.
>
>
> Curious about HEAD failing though .
>
> Dave Cramer
>
> dave.cramer(at)credativ(dot)ca
> http://www.credativ.ca