3:26 p.m.
Hi,
Seems like we got us a good build again after the hibernate dependencies
were straighten out, but our unit test status has not improved ;)
http://download.jboss.org/jbosstools/builds/nightly/200710231443-nightly/...
Hibernate test has 2 issues - i'll fix them tomorow.
Rob - AS test seem to fail because stuff is not packaged correctly ?
Denis - JST, seamui, tests seem to not have been fixed when the rhds
perspective were renamed ?
Sergey - VPE tests doesn't run at all ?
Seam2 core tests fails (that is expected because we don't have a seam2
available on the build server)
Seam1.2 core test fails (apparently the fix marshall made did not have
an effect ;(
Rob, Denis and Sergey - could you fix your tests ?
/max
3:43 p.m.
Max,
perspective ide updated, so we have +9 passed tests.
I'm going to look at mozilla tests.
Denis.
Max Rydahl Andersen wrote:
Hi,
Seems like we got us a good build again after the hibernate
dependencies were straighten out, but our unit test status has not
improved ;)
http://download.jboss.org/jbosstools/builds/nightly/200710231443-nightly/...
Hibernate test has 2 issues - i'll fix them tomorow.
Rob - AS test seem to fail because stuff is not packaged correctly ?
Denis - JST, seamui, tests seem to not have been fixed when the rhds
perspective were renamed ?
Sergey - VPE tests doesn't run at all ?
Seam2 core tests fails (that is expected because we don't have a seam2
available on the build server)
Seam1.2 core test fails (apparently the fix marshall made did not have
an effect ;(
Rob, Denis and Sergey - could you fix your tests ?
/max
_______________________________________________
jbosstools-dev mailing list
jbosstools-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/jbosstools-dev
4:12 p.m.
Denis Golovin wrote:
Max,
perspective ide updated, so we have +9 passed tests.
I'm going to look at mozilla tests.
Denis.
Max Rydahl Andersen wrote:
> Hi,
>
> Seems like we got us a good build again after the hibernate
> dependencies were straighten out, but our unit test status has not
> improved ;)
>
>
http://download.jboss.org/jbosstools/builds/nightly/200710231443-nightly/...
>
>
> Hibernate test has 2 issues - i'll fix them tomorow.
>
> Rob - AS test seem to fail because stuff is not packaged correctly ?
>
> Denis - JST, seamui, tests seem to not have been fixed when the rhds
> perspective were renamed ?
>
> Sergey - VPE tests doesn't run at all ?
>
> Seam2 core tests fails (that is expected because we don't have a
> seam2 available on the build server)
>
> Seam1.2 core test fails (apparently the fix marshall made did not
> have an effect ;(
>
> Rob, Denis and Sergey - could you fix your tests ?
>
> /max
>
> _______________________________________________
> jbosstools-dev mailing list
> jbosstools-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
_______________________________________________
jbosstools-dev mailing list
jbosstools-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/jbosstools-dev
I'm looking at VPE
tests right now. :)
Sergey
9:53 a.m.
Sergey Vasilyev wrote:
Denis Golovin wrote:
> Max,
>
> perspective ide updated, so we have +9 passed tests.
> I'm going to look at mozilla tests.
>
> Denis.
>
>
> Max Rydahl Andersen wrote:
>> Hi,
>>
>> Seems like we got us a good build again after the hibernate
>> dependencies were straighten out, but our unit test status has not
>> improved ;)
>>
>>
http://download.jboss.org/jbosstools/builds/nightly/200710231443-nightly/...
>>
>>
>> Hibernate test has 2 issues - i'll fix them tomorow.
>>
>> Rob - AS test seem to fail because stuff is not packaged correctly ?
>>
>> Denis - JST, seamui, tests seem to not have been fixed when the rhds
>> perspective were renamed ?
>>
>> Sergey - VPE tests doesn't run at all ?
>>
>> Seam2 core tests fails (that is expected because we don't have a
>> seam2 available on the build server)
>>
>> Seam1.2 core test fails (apparently the fix marshall made did not
>> have an effect ;(
>>
>> Rob, Denis and Sergey - could you fix your tests ?
>>
>> /max
>>
>> _______________________________________________
>> jbosstools-dev mailing list
>> jbosstools-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
>
> _______________________________________________
> jbosstools-dev mailing list
> jbosstools-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
I'm looking at VPE tests right now. :)
Sergey
_______________________________________________
jbosstools-dev mailing list
jbosstools-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/jbosstools-dev
VPE's tests fixed and
RichFaces test were added. Waiting next night. :)
Sergey
8:19 p.m.
Max,
here are problems I found in build:
1. Requirements should be cleaned up (Marshall told me that before his
vocation). Now build copies all current and previous requirements to the
eclipse used for JUnit testing. Here is the console output
unzipEclipse:
[echo]
eclipseInstallDir=/home/hudson/jbosstools-build/200710290346-nightly/tests/eclipse-SDK-linux-gtk.tar.gz
[untar] Expanding: /home/hudson/jbosstools-build/eclipse-SDK-linux-gtk.tar.gz into
/home/hudson/jbosstools-build/200710290346-nightly/tests
[echo] Extracting drivers into
/home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding: /home/hudson/jbosstools-build/requirements/dtp/dtp-sdk_1.5.1.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding: /home/hudson/jbosstools-build/requirements/dtp/dtp-sdk_1.5.zip into
/home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding: /home/hudson/jbosstools-build/requirements/dtp/dtp-sdk_1.5RC2.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding:
/home/hudson/jbosstools-build/requirements/emf-sdo-xsd/emf-sdo-xsd-SDK-2.3.0.zip into
/home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding:
/home/hudson/jbosstools-build/requirements/emf-sdo-xsd/emf-sdo-xsd-SDK-2.3.0RC2.zip into
/home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding:
/home/hudson/jbosstools-build/requirements/emf-sdo-xsd/emf-sdo-xsd-SDK-2.3.1.zip into
/home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding: /home/hudson/jbosstools-build/requirements/gef/GEF-SDK-3.3.1.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding: /home/hudson/jbosstools-build/requirements/gef/GEF-SDK-3.3.zip into
/home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding: /home/hudson/jbosstools-build/requirements/gef/GEF-SDK-3.3RC2.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding:
/home/hudson/jbosstools-build/requirements/webtools/wtp-sdk-R-2.0-200706260303.zip into
/home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding:
/home/hudson/jbosstools-build/requirements/webtools/wtp-sdk-R-2.0.1-20070926042742.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding:
/home/hudson/jbosstools-build/requirements/webtools/wtp-sdk-S-2.0RC2-200706050057.zip into
/home/hudson/jbosstools-build/200710290346-nightly/tests
[echo] Extracting test framework
[unzip] Expanding:
/home/hudson/jbosstools-build/requirements/testframework/eclipse-test-framework-3.3.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding:
/home/hudson/jbosstools-build/requirements/testframework/eclipse-test-framework-3.3RC4.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests
[echo] Extracting build 200710290346-nightly
[unzip] Expanding:
/home/hudson/jbosstools-build/200710290346-nightly/all/JBossTools-200710290346-nightly-ALL-linux-gtk.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding:
/home/hudson/jbosstools-build/200710290346-nightly/all/JBossTools-200710290346-nightly-ALL-macosx-carbon.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding:
/home/hudson/jbosstools-build/200710290346-nightly/all/JBossTools-200710290346-nightly-ALL-win32.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests
[unzip] Expanding:
/home/hudson/jbosstools-build/200710290346-nightly/tests/HibernateTools-Tests-200710290346-nightly.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests/eclipse
[unzip] Expanding:
/home/hudson/jbosstools-build/200710290346-nightly/tests/JBossASTools-Tests-200710290346-nightly.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests/eclipse
[unzip] Expanding:
/home/hudson/jbosstools-build/200710290346-nightly/tests/JBossTools-RichFaces-Tests-200710290346-nightly.zip
into /home/hudson/jbosstools-build/200710290346-nightly/tests/eclipse
2. I look through error log and it looks like test are ran on dirty
workspace that contains projects from previous start. Here is an error
from log
[java] !ENTRY org.jboss.tools.common.model 4 0 2007-10-29 04:30:07.559
[java] !MESSAGE Resource '/JSFKickStart1/JavaSource' already exists.
[java] !STACK 1
[java] org.eclipse.core.internal.resources.ResourceException: Resource
'/JSFKickStart1/JavaSource' already exists.
[java] at
org.eclipse.core.internal.resources.Resource.checkDoesNotExist(Resource.java:289)
[java] at
org.eclipse.core.internal.resources.Resource.checkDoesNotExist(Resource.java:276)
[java] at
org.eclipse.core.internal.resources.Folder.assertCreateRequirements(Folder.java:29)
[java] at org.eclipse.core.internal.resources.Folder.create(Folder.java:88)
[java] at org.eclipse.core.internal.resources.Folder.create(Folder.java:118)
[java] at
org.jboss.tools.common.model.filesystems.impl.FileSystemImpl.getResource(FileSystemImpl.java:74)
[java] at
org.jboss.tools.common.model.filesystems.impl.FileSystemsImpl.addChild_0(FileSystemsImpl.java:51)
[java] at
org.jboss.tools.common.model.impl.XModelObjectImpl.addChild(XModelObjectImpl.java:240)
[java] at
org.jboss.tools.common.model.util.XModelObjectLoaderUtil.loadChildren(XModelObjectLoaderUtil.java:296)
[java] at
org.jboss.tools.common.model.filesystems.impl.FileSystemsLoaderUtil.loadChildren(FileSystemsLoader.java:235)
[java] at
org.jboss.tools.common.model.util.XModelObjectLoaderUtil.load(XModelObjectLoaderUtil.java:91)
[java] at
org.jboss.tools.common.model.util.XModelObjectLoaderUtil.load(XModelObjectLoaderUtil.java:628)
[java] at
org.jboss.tools.common.model.filesystems.impl.FileSystemsLoader.load(FileSystemsLoader.java:75)
[java] at
org.jboss.tools.common.model.loaders.impl.RootLoaderImpl.load(RootLoaderImpl.java:35)
[java] at org.jboss.tools.common.model.impl.XModelImpl.load(XModelImpl.java:365)
[java] at
org.jboss.tools.common.model.XModelFactory.createInstance(XModelFactory.java:35)
[java] at
org.jboss.tools.common.model.XModelFactory.getModel(XModelFactory.java:28)
[java] at
org.jboss.tools.common.model.project.ModelNature.createProject(ModelNature.java:105)
[java] at
org.jboss.tools.common.model.project.ModelNature.setProject(ModelNature.java:50)
[java] at
org.eclipse.core.internal.resources.NatureManager.createNature(NatureManager.java:232)
[java] at org.eclipse.core.internal.resources.Project.getNature(Project.java:355)
.............
[java] at org.eclipse.test.UITestApplication.start(UITestApplication.java:210)
[java] at
org.eclipse.equinox.internal.app.EclipseAppHandle.run(EclipseAppHandle.java:169)
[java] at
org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.runApplication(EclipseAppLauncher.java:106)
[java] at
org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.start(EclipseAppLauncher.java:76)
[java] at
org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:363)
[java] at
org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:176)
[java] at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
[java] at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
[java] at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
[java] at java.lang.reflect.Method.invoke(Method.java:585)
[java] at org.eclipse.equinox.launcher.Main.invokeFramework(Main.java:508)
[java] at org.eclipse.equinox.launcher.Main.basicRun(Main.java:447)
[java] at org.eclipse.equinox.launcher.Main.run(Main.java:1173)
[java] at org.eclipse.equinox.launcher.Main.main(Main.java:1148)
[java] at org.eclipse.core.launcher.Main.main(Main.java:30)
[java] !SUBENTRY 1 org.eclipse.core.resources 4 374 2007-10-29 04:30:07.563
[java] !MESSAGE Resource '/JSFKickStart1/JavaSource' already exists.
3. test reports from public site and from QA is different:
QA -
http://reports.qa.atl.jboss.com/binaries/RHDS/nightly/200710270005-nightl...
public -
http://download.jboss.org/jbosstools/builds/nightly/200710251552-nightly/...
Denis
Max Rydahl Andersen wrote:
Hi,
Seems like we got us a good build again after the hibernate
dependencies were straighten out, but our unit test status has not
improved ;)
http://download.jboss.org/jbosstools/builds/nightly/200710231443-nightly/...
Hibernate test has 2 issues - i'll fix them tomorow.
Rob - AS test seem to fail because stuff is not packaged correctly ?
Denis - JST, seamui, tests seem to not have been fixed when the rhds
perspective were renamed ?
Sergey - VPE tests doesn't run at all ?
Seam2 core tests fails (that is expected because we don't have a seam2
available on the build server)
Seam1.2 core test fails (apparently the fix marshall made did not have
an effect ;(
Rob, Denis and Sergey - could you fix your tests ?
/max
_______________________________________________
jbosstools-dev mailing list
jbosstools-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/jbosstools-dev
1:43 a.m.
here are problems I found in build:
1. Requirements should be cleaned up (Marshall told me that before his
vocation). Now build copies all current and previous requirements to the
eclipse used for JUnit testing. Here is the console output
but why did it work a week and suddenly decided to stop ? :)
Anyhow; Rob is the one that is blessed with the info about the
buildsetup etc. Rob can you clean it up ?
unzipEclipse:
[echo]
eclipseInstallDir=/home/hudson/jbosstools-build/200710290346-nightly/tests/eclipse-SDK-linux-gtk.tar.gz
2. I look through error log and it looks like test are ran on dirty
workspace that contains projects from previous start. Here is an error
from log
ok - another thing to clean up then. (note: don't our tests cleanup also
? Other independent tests shouldn't see leftovers either)
[java] !ENTRY org.jboss.tools.common.model 4 0 2007-10-29
04:30:07.559
3. test reports from public site and from QA is different:
QA -
http://reports.qa.atl.jboss.com/binaries/RHDS/nightly/200710270005-nightl...
public -
http://download.jboss.org/jbosstools/builds/nightly/200710251552-nightly/...
Ok - that's surprising. I would expect jbosstools to have more tests
than rhds since jbosstools includes more plugins...seems like the test
setup is not in sync. Not critical, but should be fixed as soon as
marshall gets back.
I've created http://jira.jboss.com/jira/browse/JBIDE-1215 for the
general fix and I hope rob gots enough info to get 1 and 2 cleaned up on
hudson.
/max
Denis
Max Rydahl Andersen wrote:
> Hi,
>
> Seems like we got us a good build again after the hibernate
> dependencies were straighten out, but our unit test status has not
> improved ;)
>
>
http://download.jboss.org/jbosstools/builds/nightly/200710231443-nightly/...
>
>
> Hibernate test has 2 issues - i'll fix them tomorow.
>
> Rob - AS test seem to fail because stuff is not packaged correctly ?
>
> Denis - JST, seamui, tests seem to not have been fixed when the rhds
> perspective were renamed ?
>
> Sergey - VPE tests doesn't run at all ?
>
> Seam2 core tests fails (that is expected because we don't have a seam2
> available on the build server)
>
> Seam1.2 core test fails (apparently the fix marshall made did not have
> an effect ;(
>
> Rob, Denis and Sergey - could you fix your tests ?
>
> /max
>
> _______________________________________________
> jbosstools-dev mailing list
> jbosstools-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
2:14 a.m.
I'll be in touch with Prabhat after the conference call (hopefully).
Max Rydahl Andersen wrote:
>
> here are problems I found in build:
> 1. Requirements should be cleaned up (Marshall told me that before
> his vocation). Now build copies all current and previous requirements
> to the eclipse used for JUnit testing. Here is the console output
but why did it work a week and suddenly decided to stop ? :)
Anyhow; Rob is the one that is blessed with the info about the
buildsetup etc. Rob can you clean it up ?
> unzipEclipse:
> [echo]
>
eclipseInstallDir=/home/hudson/jbosstools-build/200710290346-nightly/tests/eclipse-SDK-linux-gtk.tar.gz
> 2. I look through error log and it looks like test are ran on dirty
> workspace that contains projects from previous start. Here is an
> error from log
ok - another thing to clean up then. (note: don't our tests cleanup
also ? Other independent tests shouldn't see leftovers either)
> [java] !ENTRY org.jboss.tools.common.model 4 0 2007-10-29
> 04:30:07.559
> 3. test reports from public site and from QA is different:
> QA -
>
http://reports.qa.atl.jboss.com/binaries/RHDS/nightly/200710270005-nightl...
>
> public -
>
http://download.jboss.org/jbosstools/builds/nightly/200710251552-nightly/...
Ok - that's surprising. I would expect jbosstools to have more tests
than rhds since jbosstools includes more plugins...seems like the test
setup is not in sync. Not critical, but should be fixed as soon as
marshall gets back.
I've created http://jira.jboss.com/jira/browse/JBIDE-1215 for the
general fix and I hope rob gots enough info to get 1 and 2 cleaned up
on hudson.
/max
>
>
> Denis
>
> Max Rydahl Andersen wrote:
>> Hi,
>>
>> Seems like we got us a good build again after the hibernate
>> dependencies were straighten out, but our unit test status has not
>> improved ;)
>>
>>
http://download.jboss.org/jbosstools/builds/nightly/200710231443-nightly/...
>>
>>
>> Hibernate test has 2 issues - i'll fix them tomorow.
>>
>> Rob - AS test seem to fail because stuff is not packaged correctly ?
>>
>> Denis - JST, seamui, tests seem to not have been fixed when the rhds
>> perspective were renamed ?
>>
>> Sergey - VPE tests doesn't run at all ?
>>
>> Seam2 core tests fails (that is expected because we don't have a
>> seam2 available on the build server)
>>
>> Seam1.2 core test fails (apparently the fix marshall made did not
>> have an effect ;(
>>
>> Rob, Denis and Sergey - could you fix your tests ?
>>
>> /max
>>
>> _______________________________________________
>> jbosstools-dev mailing list
>> jbosstools-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
>
>
_______________________________________________
jbosstools-dev mailing list
jbosstools-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/jbosstools-dev
11:50 a.m.
Ok. The requirements folders have been nuked. This solves issue 1.
As for issue two, looking through the log, it seems the same workspace
is used for all tests in each nightly build. For example, the workspace
/home/hudson/jbosstools-build/200710300938-nightly/tests/workspace
is used for all tests for that specific build.
So my gut instinct tells me there's a test out there that's not cleaning
up after itself, and the next test is running into problems because of it.
The options to fix, again IMO, are:
1) Find the test case that's leaving a mess and make it behave more
appropriately, or
2) Get Marshall to ensure a nuked / clean workspace between running
tests for each plugin ?
It does *not* seem to be the case that the workspace used for a test is
colliding with the workspace used in a previous build's test.
- Rob Stryker
My gut tells me there
Rob Stryker wrote:
I'll be in touch with Prabhat after the conference call
(hopefully).
Max Rydahl Andersen wrote:
>>
>> here are problems I found in build:
>> 1. Requirements should be cleaned up (Marshall told me that before
>> his vocation). Now build copies all current and previous
>> requirements to the eclipse used for JUnit testing. Here is the
>> console output
>
> but why did it work a week and suddenly decided to stop ? :)
>
> Anyhow; Rob is the one that is blessed with the info about the
> buildsetup etc. Rob can you clean it up ?
>
>> unzipEclipse:
>> [echo]
>>
eclipseInstallDir=/home/hudson/jbosstools-build/200710290346-nightly/tests/eclipse-SDK-linux-gtk.tar.gz
>
>
>
>> 2. I look through error log and it looks like test are ran on dirty
>> workspace that contains projects from previous start. Here is an
>> error from log
>
> ok - another thing to clean up then. (note: don't our tests cleanup
> also ? Other independent tests shouldn't see leftovers either)
>
>> [java] !ENTRY org.jboss.tools.common.model 4 0 2007-10-29
>> 04:30:07.559
>
>> 3. test reports from public site and from QA is different:
>> QA -
>>
http://reports.qa.atl.jboss.com/binaries/RHDS/nightly/200710270005-nightl...
>>
>> public -
>>
http://download.jboss.org/jbosstools/builds/nightly/200710251552-nightly/...
>
>
>
> Ok - that's surprising. I would expect jbosstools to have more tests
> than rhds since jbosstools includes more plugins...seems like the
> test setup is not in sync. Not critical, but should be fixed as soon
> as marshall gets back.
>
> I've created http://jira.jboss.com/jira/browse/JBIDE-1215 for the
> general fix and I hope rob gots enough info to get 1 and 2 cleaned up
> on hudson.
>
> /max
>
>>
>>
>> Denis
>>
>> Max Rydahl Andersen wrote:
>>> Hi,
>>>
>>> Seems like we got us a good build again after the hibernate
>>> dependencies were straighten out, but our unit test status has not
>>> improved ;)
>>>
>>>
http://download.jboss.org/jbosstools/builds/nightly/200710231443-nightly/...
>>>
>>>
>>> Hibernate test has 2 issues - i'll fix them tomorow.
>>>
>>> Rob - AS test seem to fail because stuff is not packaged correctly ?
>>>
>>> Denis - JST, seamui, tests seem to not have been fixed when the
>>> rhds perspective were renamed ?
>>>
>>> Sergey - VPE tests doesn't run at all ?
>>>
>>> Seam2 core tests fails (that is expected because we don't have a
>>> seam2 available on the build server)
>>>
>>> Seam1.2 core test fails (apparently the fix marshall made did not
>>> have an effect ;(
>>>
>>> Rob, Denis and Sergey - could you fix your tests ?
>>>
>>> /max
>>>
>>> _______________________________________________
>>> jbosstools-dev mailing list
>>> jbosstools-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
>>
>>
>
> _______________________________________________
> jbosstools-dev mailing list
> jbosstools-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
_______________________________________________
jbosstools-dev mailing list
jbosstools-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/jbosstools-dev
12:27 p.m.
Rob Stryker wrote:
Ok. The requirements folders have been nuked. This solves issue 1.
As for issue two, looking through the log, it seems the same workspace
is used for all tests in each nightly build. For example, the workspace
/home/hudson/jbosstools-build/200710300938-nightly/tests/workspace
is used for all tests for that specific build.
So my gut instinct tells me there's a test out there that's not cleaning
up after itself, and the next test is running into problems because of it.
The options to fix, again IMO, are:
1) Find the test case that's leaving a mess and make it behave more
appropriately, or
+100
2) Get Marshall to ensure a nuked / clean workspace between running
tests for each plugin ?
that would make the tests take ages to complete afaik.
Our tests should clean up after them selves.
/max
It does *not* seem to be the case that the workspace used for a test
is
colliding with the workspace used in a previous build's test.
- Rob Stryker
My gut tells me there
Rob Stryker wrote:
> I'll be in touch with Prabhat after the conference call (hopefully).
>
> Max Rydahl Andersen wrote:
>>>
>>> here are problems I found in build:
>>> 1. Requirements should be cleaned up (Marshall told me that before
>>> his vocation). Now build copies all current and previous
>>> requirements to the eclipse used for JUnit testing. Here is the
>>> console output
>>
>> but why did it work a week and suddenly decided to stop ? :)
>>
>> Anyhow; Rob is the one that is blessed with the info about the
>> buildsetup etc. Rob can you clean it up ?
>>
>>> unzipEclipse:
>>> [echo]
>>>
eclipseInstallDir=/home/hudson/jbosstools-build/200710290346-nightly/tests/eclipse-SDK-linux-gtk.tar.gz
>>
>>
>>
>>
>>> 2. I look through error log and it looks like test are ran on dirty
>>> workspace that contains projects from previous start. Here is an
>>> error from log
>>
>> ok - another thing to clean up then. (note: don't our tests cleanup
>> also ? Other independent tests shouldn't see leftovers either)
>>
>>> [java] !ENTRY org.jboss.tools.common.model 4 0 2007-10-29
>>> 04:30:07.559
>>
>>> 3. test reports from public site and from QA is different:
>>> QA -
>>>
http://reports.qa.atl.jboss.com/binaries/RHDS/nightly/200710270005-nightl...
>>>
>>> public -
>>>
http://download.jboss.org/jbosstools/builds/nightly/200710251552-nightly/...
>>
>>
>>
>>
>> Ok - that's surprising. I would expect jbosstools to have more tests
>> than rhds since jbosstools includes more plugins...seems like the
>> test setup is not in sync. Not critical, but should be fixed as soon
>> as marshall gets back.
>>
>> I've created http://jira.jboss.com/jira/browse/JBIDE-1215 for the
>> general fix and I hope rob gots enough info to get 1 and 2 cleaned up
>> on hudson.
>>
>> /max
>>
>>>
>>>
>>> Denis
>>>
>>> Max Rydahl Andersen wrote:
>>>> Hi,
>>>>
>>>> Seems like we got us a good build again after the hibernate
>>>> dependencies were straighten out, but our unit test status has not
>>>> improved ;)
>>>>
>>>>
http://download.jboss.org/jbosstools/builds/nightly/200710231443-nightly/...
>>>>
>>>>
>>>> Hibernate test has 2 issues - i'll fix them tomorow.
>>>>
>>>> Rob - AS test seem to fail because stuff is not packaged correctly ?
>>>>
>>>> Denis - JST, seamui, tests seem to not have been fixed when the
>>>> rhds perspective were renamed ?
>>>>
>>>> Sergey - VPE tests doesn't run at all ?
>>>>
>>>> Seam2 core tests fails (that is expected because we don't have a
>>>> seam2 available on the build server)
>>>>
>>>> Seam1.2 core test fails (apparently the fix marshall made did not
>>>> have an effect ;(
>>>>
>>>> Rob, Denis and Sergey - could you fix your tests ?
>>>>
>>>> /max
>>>>
>>>> _______________________________________________
>>>> jbosstools-dev mailing list
>>>> jbosstools-dev(a)lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
>>>
>>>
>>
>> _______________________________________________
>> jbosstools-dev mailing list
>> jbosstools-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
>
> _______________________________________________
> jbosstools-dev mailing list
> jbosstools-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
12:34 p.m.
I don't mean Nuke the build. I mean nuke the workspace between plugins.
Each plugin (or feature, not sure) seems to get its own launch in the
test workspace. I think it's fair to expect a plugin clean up after
itself, but I also think its fair that should a plugin forget to clean
up after itself, some *other* plugin does not experience the effects of
the collision.
So if as.core(or as.feature) has 10 tests, they ideally would all clean
up after themselves. But if a test forgets to clean up after itself, it
should only affect the other as.* tests. It should *not* affect seam.*
which happens to try to copy a similarly-named project into the workspace.
I think its difficult to debug or figure out which test isn't cleaning
up after itself when it could be any of dozens of plugins that could
have left the project in the workspace. It would aid this tremendously
to limit this between instantiations of the runtime workbench.
- Rob
Max Rydahl Andersen wrote:
> 2) Get Marshall to ensure a nuked / clean workspace between
running
> tests for each plugin ?
that would make the tests take ages to complete afaik.
Our tests should clean up after them selves.
/max
12:54 p.m.
I don't mean Nuke the build. I mean nuke the workspace between
plugins.
ah ok - I thought the tests ran in one long testsuite and didn't launch
per suite. If that is the case your suggestion makes perfect sense to me
too then.
/max
Each plugin (or feature, not sure) seems to get its own launch in the
test workspace. I think it's fair to expect a plugin clean up after
itself, but I also think its fair that should a plugin forget to clean
up after itself, some *other* plugin does not experience the effects of
the collision.
So if as.core(or as.feature) has 10 tests, they ideally would all clean
up after themselves. But if a test forgets to clean up after itself, it
should only affect the other as.* tests. It should *not* affect seam.*
which happens to try to copy a similarly-named project into the workspace.
I think its difficult to debug or figure out which test isn't cleaning
up after itself when it could be any of dozens of plugins that could
have left the project in the workspace. It would aid this tremendously
to limit this between instantiations of the runtime workbench.
- Rob
Max Rydahl Andersen wrote:
>> 2) Get Marshall to ensure a nuked / clean workspace between running
>> tests for each plugin ?
>
> that would make the tests take ages to complete afaik.
>
> Our tests should clean up after them selves.
>
> /max
9:21 a.m.
Our nightly builds unittests are still not running - whatsup with it ?
/max
> I don't mean Nuke the build. I mean nuke the workspace between plugins.
ah ok - I thought the tests ran in one long testsuite and didn't launch
per suite. If that is the case your suggestion makes perfect sense to me
too then.
/max
> Each plugin (or feature, not sure) seems to get its own launch in the
> test workspace. I think it's fair to expect a plugin clean up after
> itself, but I also think its fair that should a plugin forget to clean
> up after itself, some *other* plugin does not experience the effects
> of the collision.
>
> So if as.core(or as.feature) has 10 tests, they ideally would all
> clean up after themselves. But if a test forgets to clean up after
> itself, it should only affect the other as.* tests. It should *not*
> affect seam.* which happens to try to copy a similarly-named project
> into the workspace.
>
> I think its difficult to debug or figure out which test isn't cleaning
> up after itself when it could be any of dozens of plugins that could
> have left the project in the workspace. It would aid this tremendously
> to limit this between instantiations of the runtime workbench.
>
> - Rob
>
> Max Rydahl Andersen wrote:
>>> 2) Get Marshall to ensure a nuked / clean workspace between running
>>> tests for each plugin ?
>>
>> that would make the tests take ages to complete afaik.
>>
>> Our tests should clean up after them selves.
>>
>> /max
>
_______________________________________________
jbosstools-dev mailing list
jbosstools-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/jbosstools-dev
6230
days inactive
6239
days old
jbosstools-dev@lists.jboss.org
11 comments
4 participants
participants (4)
-
Denis Golovin -
Max Rydahl Andersen -
Rob Stryker -
Sergey Vasilyev