The Trunk: SUnit-ul.88.mcz

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

The Trunk: SUnit-ul.88.mcz

commits-2
Levente Uzonyi uploaded a new version of SUnit to project The Trunk:
http://source.squeak.org/trunk/SUnit-ul.88.mcz

==================== Summary ====================

Name: SUnit-ul.88
Author: ul
Time: 5 September 2012, 2:45:07.645 am
UUID: 37ee1c27-eaf6-8948-acfe-2f35416248b9
Ancestors: SUnit-fbs.87

- release the created classes in ClassFactoryForTestCase >> #cleanUp

=============== Diff against SUnit-fbs.87 ===============

Item was changed:
  ----- Method: ClassFactoryForTestCase>>cleanUp (in category 'cleaning') -----
  cleanUp
  | createdClassNames |
  createdClassNames := self createdClassNames.
  self deleteClasses.
  self deletePackage.
+ self cleanUpChangeSetForClassNames: createdClassNames.
+ self createdClasses: IdentitySet new!
- self cleanUpChangeSetForClassNames: createdClassNames!


Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: SUnit-ul.88.mcz

Frank Shearar-3
On 5 September 2012 01:56,  <[hidden email]> wrote:

> Levente Uzonyi uploaded a new version of SUnit to project The Trunk:
> http://source.squeak.org/trunk/SUnit-ul.88.mcz
>
> ==================== Summary ====================
>
> Name: SUnit-ul.88
> Author: ul
> Time: 5 September 2012, 2:45:07.645 am
> UUID: 37ee1c27-eaf6-8948-acfe-2f35416248b9
> Ancestors: SUnit-fbs.87
>
> - release the created classes in ClassFactoryForTestCase >> #cleanUp
>
> =============== Diff against SUnit-fbs.87 ===============

This is enormously more useful. Thanks!

Sample output:

Error Message

AnObsoleteMCMockClassA, AnObsoleteMCMockASubclass,
AnObsoleteMCMockClassB, AnObsoleteMCMockClassD,
AnObsoleteMCMockClassE, AnObsoleteMCMockClassF,
AnObsoleteMCMockClassG, AnObsoleteMCMockClassH and
AnObsoleteMCMockClassI are obsolete
Stacktrace

ReleaseTest(TestCase)>>assert:description:
ReleaseTest>>testNoObsoleteClasses
ReleaseTest(TestCase)>>performTest

frank

Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: SUnit-ul.88.mcz

Levente Uzonyi-2
On Wed, 5 Sep 2012, Frank Shearar wrote:

> On 5 September 2012 01:56,  <[hidden email]> wrote:
>> Levente Uzonyi uploaded a new version of SUnit to project The Trunk:
>> http://source.squeak.org/trunk/SUnit-ul.88.mcz
>>
>> ==================== Summary ====================
>>
>> Name: SUnit-ul.88
>> Author: ul
>> Time: 5 September 2012, 2:45:07.645 am
>> UUID: 37ee1c27-eaf6-8948-acfe-2f35416248b9
>> Ancestors: SUnit-fbs.87
>>
>> - release the created classes in ClassFactoryForTestCase >> #cleanUp
>>
>> =============== Diff against SUnit-fbs.87 ===============
>
> This is enormously more useful. Thanks!

That was the goal. Now we know, that #testUndeclared [1] is failing
because of the extra stuff loaded for the CI. And #testNoObsoleteClasses
[2] is failing because of MC tests. Btw, the latter doesn't fail on
windows.


Levente

[1]
http://www.squeakci.org/job/SqueakTrunk/lastCompletedBuild/testReport/Tests.Release/ReleaseTest/testUndeclared/
[2]
http://www.squeakci.org/job/SqueakTrunk/lastCompletedBuild/testReport/Tests.Release/ReleaseTest/testNoObsoleteClasses/

>
> Sample output:
>
> Error Message
>
> AnObsoleteMCMockClassA, AnObsoleteMCMockASubclass,
> AnObsoleteMCMockClassB, AnObsoleteMCMockClassD,
> AnObsoleteMCMockClassE, AnObsoleteMCMockClassF,
> AnObsoleteMCMockClassG, AnObsoleteMCMockClassH and
> AnObsoleteMCMockClassI are obsolete
> Stacktrace
>
> ReleaseTest(TestCase)>>assert:description:
> ReleaseTest>>testNoObsoleteClasses
> ReleaseTest(TestCase)>>performTest
>
> frank
>
>

Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: SUnit-ul.88.mcz

Frank Shearar-3
On 5 September 2012 13:49, Levente Uzonyi <[hidden email]> wrote:

> On Wed, 5 Sep 2012, Frank Shearar wrote:
>
>> On 5 September 2012 01:56,  <[hidden email]> wrote:
>>>
>>> Levente Uzonyi uploaded a new version of SUnit to project The Trunk:
>>> http://source.squeak.org/trunk/SUnit-ul.88.mcz
>>>
>>> ==================== Summary ====================
>>>
>>> Name: SUnit-ul.88
>>> Author: ul
>>> Time: 5 September 2012, 2:45:07.645 am
>>> UUID: 37ee1c27-eaf6-8948-acfe-2f35416248b9
>>> Ancestors: SUnit-fbs.87
>>>
>>> - release the created classes in ClassFactoryForTestCase >> #cleanUp
>>>
>>> =============== Diff against SUnit-fbs.87 ===============
>>
>>
>> This is enormously more useful. Thanks!
>
>
> That was the goal. Now we know, that #testUndeclared [1] is failing because
> of the extra stuff loaded for the CI. And #testNoObsoleteClasses [2] is
> failing because of MC tests. Btw, the latter doesn't fail on windows.

Yes. We (I) can solve that in one of two ways: hack out the
dependencies in HudsonBuildTools.st, or port it and install it
properly. But at least that means that the test itself is OK, which is
great.

That's a bit strange, that #testNoObsoleteClasses doesn't fail on
Windows. It must be a VM difference then, surely?

frank

> Levente
>
> [1]
> http://www.squeakci.org/job/SqueakTrunk/lastCompletedBuild/testReport/Tests.Release/ReleaseTest/testUndeclared/
> [2]
> http://www.squeakci.org/job/SqueakTrunk/lastCompletedBuild/testReport/Tests.Release/ReleaseTest/testNoObsoleteClasses/
>
>
>>
>> Sample output:
>>
>> Error Message
>>
>> AnObsoleteMCMockClassA, AnObsoleteMCMockASubclass,
>> AnObsoleteMCMockClassB, AnObsoleteMCMockClassD,
>> AnObsoleteMCMockClassE, AnObsoleteMCMockClassF,
>> AnObsoleteMCMockClassG, AnObsoleteMCMockClassH and
>> AnObsoleteMCMockClassI are obsolete
>> Stacktrace
>>
>> ReleaseTest(TestCase)>>assert:description:
>> ReleaseTest>>testNoObsoleteClasses
>> ReleaseTest(TestCase)>>performTest
>>
>> frank
>>
>>
>

Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: SUnit-ul.88.mcz

Levente Uzonyi-2
On Wed, 5 Sep 2012, Frank Shearar wrote:

> On 5 September 2012 13:49, Levente Uzonyi <[hidden email]> wrote:
>> On Wed, 5 Sep 2012, Frank Shearar wrote:
>>
>>> On 5 September 2012 01:56,  <[hidden email]> wrote:
>>>>
>>>> Levente Uzonyi uploaded a new version of SUnit to project The Trunk:
>>>> http://source.squeak.org/trunk/SUnit-ul.88.mcz
>>>>
>>>> ==================== Summary ====================
>>>>
>>>> Name: SUnit-ul.88
>>>> Author: ul
>>>> Time: 5 September 2012, 2:45:07.645 am
>>>> UUID: 37ee1c27-eaf6-8948-acfe-2f35416248b9
>>>> Ancestors: SUnit-fbs.87
>>>>
>>>> - release the created classes in ClassFactoryForTestCase >> #cleanUp
>>>>
>>>> =============== Diff against SUnit-fbs.87 ===============
>>>
>>>
>>> This is enormously more useful. Thanks!
>>
>>
>> That was the goal. Now we know, that #testUndeclared [1] is failing because
>> of the extra stuff loaded for the CI. And #testNoObsoleteClasses [2] is
>> failing because of MC tests. Btw, the latter doesn't fail on windows.
>
> Yes. We (I) can solve that in one of two ways: hack out the
> dependencies in HudsonBuildTools.st, or port it and install it
> properly. But at least that means that the test itself is OK, which is
> great.
>
> That's a bit strange, that #testNoObsoleteClasses doesn't fail on
> Windows. It must be a VM difference then, surely?

I don't think so. It would be good to save the image after running the
tests if there are failing ones, so we could have a chance to see what's
holding references to those classes in this case.


Levente

>
> frank
>
>> Levente
>>
>> [1]
>> http://www.squeakci.org/job/SqueakTrunk/lastCompletedBuild/testReport/Tests.Release/ReleaseTest/testUndeclared/
>> [2]
>> http://www.squeakci.org/job/SqueakTrunk/lastCompletedBuild/testReport/Tests.Release/ReleaseTest/testNoObsoleteClasses/
>>
>>
>>>
>>> Sample output:
>>>
>>> Error Message
>>>
>>> AnObsoleteMCMockClassA, AnObsoleteMCMockASubclass,
>>> AnObsoleteMCMockClassB, AnObsoleteMCMockClassD,
>>> AnObsoleteMCMockClassE, AnObsoleteMCMockClassF,
>>> AnObsoleteMCMockClassG, AnObsoleteMCMockClassH and
>>> AnObsoleteMCMockClassI are obsolete
>>> Stacktrace
>>>
>>> ReleaseTest(TestCase)>>assert:description:
>>> ReleaseTest>>testNoObsoleteClasses
>>> ReleaseTest(TestCase)>>performTest
>>>
>>> frank
>>>
>>>
>>
>
>

Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: SUnit-ul.88.mcz

Frank Shearar-3
On 5 September 2012 14:49, Levente Uzonyi <[hidden email]> wrote:

> On Wed, 5 Sep 2012, Frank Shearar wrote:
>
>> On 5 September 2012 13:49, Levente Uzonyi <[hidden email]> wrote:
>>>
>>> On Wed, 5 Sep 2012, Frank Shearar wrote:
>>>
>>>> On 5 September 2012 01:56,  <[hidden email]> wrote:
>>>>>
>>>>>
>>>>> Levente Uzonyi uploaded a new version of SUnit to project The Trunk:
>>>>> http://source.squeak.org/trunk/SUnit-ul.88.mcz
>>>>>
>>>>> ==================== Summary ====================
>>>>>
>>>>> Name: SUnit-ul.88
>>>>> Author: ul
>>>>> Time: 5 September 2012, 2:45:07.645 am
>>>>> UUID: 37ee1c27-eaf6-8948-acfe-2f35416248b9
>>>>> Ancestors: SUnit-fbs.87
>>>>>
>>>>> - release the created classes in ClassFactoryForTestCase >> #cleanUp
>>>>>
>>>>> =============== Diff against SUnit-fbs.87 ===============
>>>>
>>>>
>>>>
>>>> This is enormously more useful. Thanks!
>>>
>>>
>>>
>>> That was the goal. Now we know, that #testUndeclared [1] is failing
>>> because
>>> of the extra stuff loaded for the CI. And #testNoObsoleteClasses [2] is
>>> failing because of MC tests. Btw, the latter doesn't fail on windows.
>>
>>
>> Yes. We (I) can solve that in one of two ways: hack out the
>> dependencies in HudsonBuildTools.st, or port it and install it
>> properly. But at least that means that the test itself is OK, which is
>> great.
>>
>> That's a bit strange, that #testNoObsoleteClasses doesn't fail on
>> Windows. It must be a VM difference then, surely?
>
>
> I don't think so. It would be good to save the image after running the tests
> if there are failing ones, so we could have a chance to see what's holding
> references to those classes in this case.

Hang on though: the CI tests don't run from a TestRunner but from the
hackported HudsonBuildTools.st. That's a big difference there.

frank

> Levente
>
>
>>
>> frank
>>
>>> Levente
>>>
>>> [1]
>>>
>>> http://www.squeakci.org/job/SqueakTrunk/lastCompletedBuild/testReport/Tests.Release/ReleaseTest/testUndeclared/
>>> [2]
>>>
>>> http://www.squeakci.org/job/SqueakTrunk/lastCompletedBuild/testReport/Tests.Release/ReleaseTest/testNoObsoleteClasses/
>>>
>>>
>>>>
>>>> Sample output:
>>>>
>>>> Error Message
>>>>
>>>> AnObsoleteMCMockClassA, AnObsoleteMCMockASubclass,
>>>> AnObsoleteMCMockClassB, AnObsoleteMCMockClassD,
>>>> AnObsoleteMCMockClassE, AnObsoleteMCMockClassF,
>>>> AnObsoleteMCMockClassG, AnObsoleteMCMockClassH and
>>>> AnObsoleteMCMockClassI are obsolete
>>>> Stacktrace
>>>>
>>>> ReleaseTest(TestCase)>>assert:description:
>>>> ReleaseTest>>testNoObsoleteClasses
>>>> ReleaseTest(TestCase)>>performTest
>>>>
>>>> frank
>>>>
>>>>
>>>
>>
>>
>

Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: SUnit-ul.88.mcz

Levente Uzonyi-2
On Wed, 5 Sep 2012, Frank Shearar wrote:

> On 5 September 2012 14:49, Levente Uzonyi <[hidden email]> wrote:
>> On Wed, 5 Sep 2012, Frank Shearar wrote:
>>
>>> On 5 September 2012 13:49, Levente Uzonyi <[hidden email]> wrote:
>>>>
>>>> On Wed, 5 Sep 2012, Frank Shearar wrote:
>>>>
>>>>> On 5 September 2012 01:56,  <[hidden email]> wrote:
>>>>>>
>>>>>>
>>>>>> Levente Uzonyi uploaded a new version of SUnit to project The Trunk:
>>>>>> http://source.squeak.org/trunk/SUnit-ul.88.mcz
>>>>>>
>>>>>> ==================== Summary ====================
>>>>>>
>>>>>> Name: SUnit-ul.88
>>>>>> Author: ul
>>>>>> Time: 5 September 2012, 2:45:07.645 am
>>>>>> UUID: 37ee1c27-eaf6-8948-acfe-2f35416248b9
>>>>>> Ancestors: SUnit-fbs.87
>>>>>>
>>>>>> - release the created classes in ClassFactoryForTestCase >> #cleanUp
>>>>>>
>>>>>> =============== Diff against SUnit-fbs.87 ===============
>>>>>
>>>>>
>>>>>
>>>>> This is enormously more useful. Thanks!
>>>>
>>>>
>>>>
>>>> That was the goal. Now we know, that #testUndeclared [1] is failing
>>>> because
>>>> of the extra stuff loaded for the CI. And #testNoObsoleteClasses [2] is
>>>> failing because of MC tests. Btw, the latter doesn't fail on windows.
>>>
>>>
>>> Yes. We (I) can solve that in one of two ways: hack out the
>>> dependencies in HudsonBuildTools.st, or port it and install it
>>> properly. But at least that means that the test itself is OK, which is
>>> great.
>>>
>>> That's a bit strange, that #testNoObsoleteClasses doesn't fail on
>>> Windows. It must be a VM difference then, surely?
>>
>>
>> I don't think so. It would be good to save the image after running the tests
>> if there are failing ones, so we could have a chance to see what's holding
>> references to those classes in this case.
>
> Hang on though: the CI tests don't run from a TestRunner but from the
> hackported HudsonBuildTools.st. That's a big difference there.

The TestResult object holds the reference to all the TestCase instances.
I'm sure that script uses TestResult too.


Levente

>
> frank
>
>> Levente
>>
>>
>>>
>>> frank
>>>
>>>> Levente
>>>>
>>>> [1]
>>>>
>>>> http://www.squeakci.org/job/SqueakTrunk/lastCompletedBuild/testReport/Tests.Release/ReleaseTest/testUndeclared/
>>>> [2]
>>>>
>>>> http://www.squeakci.org/job/SqueakTrunk/lastCompletedBuild/testReport/Tests.Release/ReleaseTest/testNoObsoleteClasses/
>>>>
>>>>
>>>>>
>>>>> Sample output:
>>>>>
>>>>> Error Message
>>>>>
>>>>> AnObsoleteMCMockClassA, AnObsoleteMCMockASubclass,
>>>>> AnObsoleteMCMockClassB, AnObsoleteMCMockClassD,
>>>>> AnObsoleteMCMockClassE, AnObsoleteMCMockClassF,
>>>>> AnObsoleteMCMockClassG, AnObsoleteMCMockClassH and
>>>>> AnObsoleteMCMockClassI are obsolete
>>>>> Stacktrace
>>>>>
>>>>> ReleaseTest(TestCase)>>assert:description:
>>>>> ReleaseTest>>testNoObsoleteClasses
>>>>> ReleaseTest(TestCase)>>performTest
>>>>>
>>>>> frank
>>>>>
>>>>>
>>>>
>>>
>>>
>>
>
>