[#QB-500] Ant specific JUnit code can lead to failing pseudo tests that cause subsequent problems in test history tracking

QuickBuild

Ant specific JUnit code can lead to failing pseudo tests that cause subsequent problems in test history tracking

Created: 13/Jan/10 08:50 AM Updated: 19/Jan/10 09:03 AM

Component/s:

None

Affects Version/s:

None

Fix Version/s:

None

Original Estimate:

Unknown

Remaining Estimate:

Unknown

Time Spent:

Unknown

File Attachments:

tests.xml (7 kb)

Image Attachments:

1. BeforeFirstTest1.png (81 kb)	2. BeforeFirstTest2.png (35 kb)	3. BeforeFirstTest3.png (61 kb)
4. BeforeFirstTest4.png (40 kb)

Environment:

Ant + JUnit

Description

« Hide

Note: I have submitted an unexpected exception before (cf. attached screenshot #4) -- please ignore that and focus instead on this one.

"BeforeFirstTest" stems from Ant: see http://mail-archives.apache.org/mod_mbox/ant-dev/200607.mbox/%3C20060708211941.792401A981A@eris.apache.org%3E
...
} else if (args[i].startsWith("crashfile=")) {
+ crashFile = args[i].substring(12);
+ registerTestCase("BeforeFirstTest");
...

And presumably the JVM of the JUnit test execution crashed once before, e.g. leading to this error message:
junit.framework.AssertionFailedError: Forked Java VM exited abnormally. Please note the time in the report does not reflect the time until the VM exit.

All

Comments

Work Log

Change History

Sort Order:

Change by AlSt [13/Jan/10 08:50 AM]
Field	Original Value	New Value
Attachment		BeforeFirstTest1.png [ 10134 ]

Change by AlSt [13/Jan/10 08:51 AM]
Attachment		BeforeFirstTest2.png [ 10135 ]

Change by AlSt [13/Jan/10 08:51 AM]
Attachment		BeforeFirstTest3.png [ 10136 ]

Change by AlSt [13/Jan/10 08:51 AM]
Attachment		BeforeFirstTest4.png [ 10137 ]

[ Permlink | « Hide ]

AlSt [13/Jan/10 08:55 AM]

From qb20\storage\builds\136040\reports\junit\DEFAULT: this shows why I think it is a test history problem

Change by AlSt [13/Jan/10 08:55 AM]
Attachment		tests.xml [ 10138 ]

[ Permlink | « Hide ]

AlSt [13/Jan/10 09:23 AM]

Note: After the next run this problem has gone again -- most likely because the history/trend is only shown between two consecutive builds.

=> Thus, I would say the priority of this issue should be lower, i.e. normal.

Change by Robin Shen [13/Jan/10 09:42 AM]
Assignee	Robin Shine [ robinshine ]	Steve Luo [ steve ]

Change by Steve Luo [19/Jan/10 09:03 AM]
Status	Open [ 1 ]	Resolved [ 5 ]
Fixed In Version		2.0.15 [ 11113 ]
Fix Version/s		2.0.15 [ 11113 ]
Resolution		Fixed [ 1 ]