Hey,
Should be fixed by:
commit ba7ff5369cc602c7e0c8d788b6617f27123c2328
Author: Hugo Trippaers <[email protected]>
Date: Thu Feb 20 14:04:51 2014 +0100
Fix parameter index in getCount function.
@Talluri, Agreed. However with the current way of testing it is really hard to
see which commit actually broke something. We should be able to tell which
commits changed something when a particular test fails. Maybe we could start by
supplying the list of changes between one run of the test suite and the next.
That would at least narrow the number of changes developers have to look at to
determine which change broke what,much like the current maven build.
Cheers,
Hugo
On 20 feb. 2014, at 07:39, Srikanteswararao Talluri
<[email protected]> wrote:
> This is found by daily BVT run and sent out in daily BVT reporting mail.
> It would be great if folks could actually pick up the bugs resulted from
> BVT failures immediately.
>
> Thanks,
> ~Talluri
>
> On 20/02/14 10:20 am, "Sheng Yang" <[email protected]> wrote:
>
>> Hi Hugo/Ian,
>>
>> I suspect https://issues.apache.org/jira/browse/CLOUDSTACK-6130 is caused
>> by some of recent change of findbugs. It's likely got something wrong with
>> GeneralDaoBase, and the most recently modifications are:
>>
>> commit 443acac92792cbffff2654c3414df276efac404f
>> Author: Hugo Trippaers <[email protected]>
>> Date: Fri Feb 14 18:36:18 2014 +0100
>>
>> Fix functional issue introduced by some findbug changes.
>>
>> commit 5a6ad11fa2ca3de64b94095c628bd3ba12e618c3
>> Author: Ian Southam <[email protected]>
>> Date: Wed Feb 12 17:08:52 2014 +0100
>>
>> findbugs findings
>>
>> Signed-off-by: Hugo Trippaers <[email protected]>
>>
>> From the error message, it shows statement as:
>>
>> SELECT COUNT(*) FROM user_ip_address INNER JOIN account ON
>> user_ip_address.account_id=account.id INNER JOIN vlan ON
>> user_ip_address.vlan_db_id=vlan.id WHERE user_ip_address.account_id=5 AND
>> user_ip_address.allocated IS NOT NULL AND (account.type !=
>> 'VirtualNetwork'
>> ) AND (vlan.vlan_type = ** NOT SPECIFIED ** )
>>
>> And you can observed that some parameters have been shifted left by one
>> position, because vlan.vlan_type should compare to "virtualNetwork". It
>> should related to "i++"/"++i" change in the commit.
>>
>> After reverting these two commits, everything is fine.
>>
>> Could you take a look?
>>
>> --Sheng
>
