You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@geronimo.apache.org by Rick McGuire <ri...@gmail.com> on 2010/10/13 20:17:59 UTC
Bean Validation problems in trunk.
The recent changes to the bean validation support appear to be causing
failures in both the TCK and the bean validation testsuite. I've tried
to do a little problem determination, but I think Vamsi is probably
going to need to take a look at this. In the testsuite failures, it
looks to me that the validation factory configuration is not getting set
up correctly so that not all of the constraints are getting applied in
the tests. Many of the failures are just using the default constraint
classes.
In the TCK, I'm seeing a lot of the same failures where the test was
expecting constraint failures and not getting any, so this is probably
the same thing.
There's at least one other problem with the TCK. A number of tests are
getting deploy failures because of errors with the validation.xml.
Since these tests actually ARE tests for the invalid XML. this suggests
that any parsing errors discovered at deploy time should not be terminal.
Rick
Re: Bean Validation problems in trunk.
Posted by Vamsavardhana Reddy <c1...@gmail.com>.
On Wed, Oct 20, 2010 at 10:41 PM, Vamsavardhana Reddy
<c1...@gmail.com>wrote:
>
>
> On Thu, Oct 14, 2010 at 9:49 PM, Vamsavardhana Reddy <c1...@gmail.com>wrote:
>
>>
>>
>> On Thu, Oct 14, 2010 at 8:03 AM, Vamsavardhana Reddy <c1vamsi1c@gmail.com
>> > wrote:
>>
>>>
>>>
>>> On Wed, Oct 13, 2010 at 11:47 PM, Rick McGuire <ri...@gmail.com>wrote:
>>>
>>>> The recent changes to the bean validation support appear to be causing
>>>> failures in both the TCK and the bean validation testsuite. I've tried to
>>>> do a little problem determination, but I think Vamsi is probably going to
>>>> need to take a look at this. In the testsuite failures, it looks to me that
>>>> the validation factory configuration is not getting set up correctly so that
>>>> not all of the constraints are getting applied in the tests. Many of the
>>>> failures are just using the default constraint classes.
>>>>
>>> I see that only those constraint classes that are within the same test
>>> project are being used in the validation, but the default ones are not. I
>>> will take a look at what is happening.
>>>
>> As of rev 1005583 (by which the earlier test is reorganized and EJB test
>> is added) the tests are passing. Rev 1005852 is the only other rev that
>> touched plugins\bval, but the change in that rev seems to have not affected
>> the test status. I will investigate further.
>>
> When I start the server, I see one warning message indicating
> DefaultConstraints.properties could not be loaded...
>
Forgot to mention that this warning message is logged while loading
"org.apache.geronimo.configs/remote-deploy-tomcat/3.0-SNAPSHOT/car"
>
> Starting Geronimo Application Server v3.0-SNAPSHOT
> [**************** ] 42% 10s Starting
> org.apache.ger...2010-10-20 22:19:07,359 WARN [ConstraintDefaults] cannot
> find org/apache/bval/jsr303/DefaultConstraints.properties
>
> Surely this is causing the tests with default constraints to fail. I see
> that the classLoader from which the resource is to get loaded is
> org.eclipse.osgi.internal.baseadaptor.DefaultClassLoader@15a4a77 which has
> a delegate
> BundleLoader(org.apache.geronimo.framework.geronimo-system_3.0.0.SNAPSHOT)
>
> In the case where the tests are passing (before the regression),
> DefaultConstraints.properties is getting loaded from"[BundleClassLoader]
> org.apache.geronimo.framework.j2ee-security_3.0.0.SNAPSHOT [74]"
>
>
>>
>>>
>>>
>>>> In the TCK, I'm seeing a lot of the same failures where the test was
>>>> expecting constraint failures and not getting any, so this is probably the
>>>> same thing.
>>>>
>>>> There's at least one other problem with the TCK. A number of tests are
>>>> getting deploy failures because of errors with the validation.xml. Since
>>>> these tests actually ARE tests for the invalid XML. this suggests that any
>>>> parsing errors discovered at deploy time should not be terminal.
>>>>
>>> hmm... I will change the deployment exceptions to log warning messages
>>> instead.
>>>
>> I have changed the code to log debug messages instead of throwing
>> deployment exceptions. Commited in revision 1022595. This should take care
>> of the TCK tests related to invalid XML.
>>
>>
>>>
>>>
>>>>
>>>> Rick
>>>>
>>>
>>>
>>>
>>> --
>>> Vamsi
>>>
>>
>>
>>
>> --
>> Vamsi
>>
>
>
>
> --
> Vamsi
>
--
Vamsi
Re: Bean Validation problems in trunk.
Posted by Vamsavardhana Reddy <c1...@gmail.com>.
On Thu, Oct 14, 2010 at 9:49 PM, Vamsavardhana Reddy <c1...@gmail.com>wrote:
>
>
> On Thu, Oct 14, 2010 at 8:03 AM, Vamsavardhana Reddy <c1...@gmail.com>wrote:
>
>>
>>
>> On Wed, Oct 13, 2010 at 11:47 PM, Rick McGuire <ri...@gmail.com> wrote:
>>
>>> The recent changes to the bean validation support appear to be causing
>>> failures in both the TCK and the bean validation testsuite. I've tried to
>>> do a little problem determination, but I think Vamsi is probably going to
>>> need to take a look at this. In the testsuite failures, it looks to me that
>>> the validation factory configuration is not getting set up correctly so that
>>> not all of the constraints are getting applied in the tests. Many of the
>>> failures are just using the default constraint classes.
>>>
>> I see that only those constraint classes that are within the same test
>> project are being used in the validation, but the default ones are not. I
>> will take a look at what is happening.
>>
> As of rev 1005583 (by which the earlier test is reorganized and EJB test is
> added) the tests are passing. Rev 1005852 is the only other rev that
> touched plugins\bval, but the change in that rev seems to have not affected
> the test status. I will investigate further.
>
When I start the server, I see one warning message indicating
DefaultConstraints.properties could not be loaded...
Starting Geronimo Application Server v3.0-SNAPSHOT
[**************** ] 42% 10s Starting
org.apache.ger...2010-10-20 22:19:07,359 WARN [ConstraintDefaults] cannot
find org/apache/bval/jsr303/DefaultConstraints.properties
Surely this is causing the tests with default constraints to fail. I see
that the classLoader from which the resource is to get loaded is
org.eclipse.osgi.internal.baseadaptor.DefaultClassLoader@15a4a77 which has a
delegate
BundleLoader(org.apache.geronimo.framework.geronimo-system_3.0.0.SNAPSHOT)
In the case where the tests are passing (before the regression),
DefaultConstraints.properties is getting loaded from"[BundleClassLoader]
org.apache.geronimo.framework.j2ee-security_3.0.0.SNAPSHOT [74]"
>
>>
>>
>>> In the TCK, I'm seeing a lot of the same failures where the test was
>>> expecting constraint failures and not getting any, so this is probably the
>>> same thing.
>>>
>>> There's at least one other problem with the TCK. A number of tests are
>>> getting deploy failures because of errors with the validation.xml. Since
>>> these tests actually ARE tests for the invalid XML. this suggests that any
>>> parsing errors discovered at deploy time should not be terminal.
>>>
>> hmm... I will change the deployment exceptions to log warning messages
>> instead.
>>
> I have changed the code to log debug messages instead of throwing
> deployment exceptions. Commited in revision 1022595. This should take care
> of the TCK tests related to invalid XML.
>
>
>>
>>
>>>
>>> Rick
>>>
>>
>>
>>
>> --
>> Vamsi
>>
>
>
>
> --
> Vamsi
>
--
Vamsi
Re: Bean Validation problems in trunk.
Posted by Vamsavardhana Reddy <c1...@gmail.com>.
On Thu, Oct 14, 2010 at 8:03 AM, Vamsavardhana Reddy <c1...@gmail.com>wrote:
>
>
> On Wed, Oct 13, 2010 at 11:47 PM, Rick McGuire <ri...@gmail.com> wrote:
>
>> The recent changes to the bean validation support appear to be causing
>> failures in both the TCK and the bean validation testsuite. I've tried to
>> do a little problem determination, but I think Vamsi is probably going to
>> need to take a look at this. In the testsuite failures, it looks to me that
>> the validation factory configuration is not getting set up correctly so that
>> not all of the constraints are getting applied in the tests. Many of the
>> failures are just using the default constraint classes.
>>
> I see that only those constraint classes that are within the same test
> project are being used in the validation, but the default ones are not. I
> will take a look at what is happening.
>
As of rev 1005583 (by which the earlier test is reorganized and EJB test is
added) the tests are passing. Rev 1005852 is the only other rev that
touched plugins\bval, but the change in that rev seems to have not affected
the test status. I will investigate further.
>
>
>> In the TCK, I'm seeing a lot of the same failures where the test was
>> expecting constraint failures and not getting any, so this is probably the
>> same thing.
>>
>> There's at least one other problem with the TCK. A number of tests are
>> getting deploy failures because of errors with the validation.xml. Since
>> these tests actually ARE tests for the invalid XML. this suggests that any
>> parsing errors discovered at deploy time should not be terminal.
>>
> hmm... I will change the deployment exceptions to log warning messages
> instead.
>
I have changed the code to log debug messages instead of throwing deployment
exceptions. Commited in revision 1022595. This should take care of the TCK
tests related to invalid XML.
>
>
>>
>> Rick
>>
>
>
>
> --
> Vamsi
>
--
Vamsi
Re: Bean Validation problems in trunk.
Posted by Vamsavardhana Reddy <c1...@gmail.com>.
On Wed, Oct 13, 2010 at 11:47 PM, Rick McGuire <ri...@gmail.com> wrote:
> The recent changes to the bean validation support appear to be causing
> failures in both the TCK and the bean validation testsuite. I've tried to
> do a little problem determination, but I think Vamsi is probably going to
> need to take a look at this. In the testsuite failures, it looks to me that
> the validation factory configuration is not getting set up correctly so that
> not all of the constraints are getting applied in the tests. Many of the
> failures are just using the default constraint classes.
>
I see that only those constraint classes that are within the same test
project are being used in the validation, but the default ones are not. I
will take a look at what is happening.
> In the TCK, I'm seeing a lot of the same failures where the test was
> expecting constraint failures and not getting any, so this is probably the
> same thing.
>
> There's at least one other problem with the TCK. A number of tests are
> getting deploy failures because of errors with the validation.xml. Since
> these tests actually ARE tests for the invalid XML. this suggests that any
> parsing errors discovered at deploy time should not be terminal.
>
hmm... I will change the deployment exceptions to log warning messages
instead.
>
> Rick
>
--
Vamsi