Skip to main content

Notice: this Wiki will be going read only early in 2024 and edits will no longer be possible. Please see: https://gitlab.eclipse.org/eclipsefdn/helpdesk/-/wikis/Wiki-shutdown-plan for the plan.

Jump to: navigation, search

Difference between revisions of "WTP/Build/WTP Build Break Board"

< WTP‎ | Build
(M6 02/28)
(I20080229065303)
Line 98: Line 98:
 
! WS
 
! WS
 
| Respin requested based on bad bug (regression) found in smoke test.  
 
| Respin requested based on bad bug (regression) found in smoke test.  
|  
+
|The Web Services smoke test failed due to a change in the Tomcat server tools [https://bugs.eclipse.org/bugs/show_bug.cgi?id=220879 220879].  The problem was fixed by Server tooling team and was dropped in I20080229195548.  Based on the comments in the bug, this was due to a change in publishing in Tomcat server tools recently and was not found because testing was not done in a new workspace.  They had commented that new JUnit would be included in the next I-build to catch this in the future.
 
|}
 
|}
  

Revision as of 10:00, 1 March 2008

WTP weekly I build breakages

The purpose of this board is to track the list of breakages that we have on our weekly I builds. Compile, JUnit, respin requests that cause a delay in smoketest and thus declaration of our builds. We hope to find a pattern with these breakages so that we can learn from them and the same mistakes can be avoided in future.

M6 02/28

I20080228113529

Project Error Type Compile/JUnit/Respin Analysis/Initials
Server 3 compile errors in org.eclipse.jst.server.ui.tests Didn't run local junit test after removing a class. / AV
JPT DNF in Junit for org.eclipse.jpt.core.tests.internal.JptCoreTests Deadlock in the tests, deleting resources. Seems to be a test issue, still investigating root cause / njh
Server 2 Junit failures in org.eclipse.jst.server.ui.tests.AllTests Related to the compilation errors / AV
XML 1 Junit failure in org.eclipse.wst.xml.core.tests.SSEModelXMLTestSuite Changes to our line delimiter handling invalidated one of our tests. It used to test if the default delimiter was the platform delimiter, but now the default is the one already in the document. The document being tested was using the Windows line delimiter, so the test passed fine on Windows but would break on any other platform. The test has been updated. /ND
Releng 1 Junit failure in org.eclipse.wtp.releng.tests.TestBuild Versioning Errors. Plugin versions went down compared to 202 release. Appeared this week since reference data was just updated, but the error has probably existed for a while. One contributing factor was perhaps that the "build failed" notifications were broken for several days (not being sent) so may have not been noticed by development teams.

I20080228162856

Project Error Type Compile/JUnit/Respin Analysis/Initials
JEE 4 JUnit failures. testAddListener_Web24_Defaults - random failure - WebProject does not exist

testWebPublishJEE5 - random failure - Component TestUtil should have 2 component classpath dependencies, only has: 0 /ccc

XML 1 Junit failure in org.eclipse.wst.xml.core.tests.SSEModelXMLTestSuite Above update missed the build./ND
Releng 1 Junit failure in org.eclipse.wtp.releng.tests.TestBuild Same Versioning Errors (in JSF, Common Facet). JSF can not easily update, since other changes have been made to HEAD.

I20080228214133

Project Error Type Compile/JUnit/Respin Analysis/Initials
JEE 2 JUnit failures. testClassFolderDependency - random failure - Component TestWeb should have 1 component classpath dependencies, only has: 0/ccc
JPT Massive compile failure. Seems a number of source files are missing? Releasing process failure. Sync error displayed at end of release process, manual inspection looked OK but there were apparently problems (files left untagged that had moved). Not sure what went wrong with the process (standard releng tool was used, perhaps a CVS glitch?). Manually tagging and re-releasing changed plugins to fix the problem /njh

I also want to note this release was done only to update JavaDoc. The JPT team felt this was critical to do this week, but normally, JavaDoc changes should not be released after candidate produced, since normally that would not be a blocking or critical problem ... so, remember ... every release is an opportunity to break something. dw

Releng 1 Junit failure in org.eclipse.wtp.releng.tests.TestBuild Same Versioning Errors (in JSF, Common Facet). JSF can not easily update, since other changes have been made to HEAD.

I20080229031047

Project Error Type Compile/JUnit/Respin Analysis/Initials
JPT Same massive compile failure as previous build. Assuming fix isn't in yet Releasing process failure. Sync error displayed at end of release process, manual inspection looked OK but there were apparently problems (files left untagged that had moved). Not sure what went wrong with the process (standard releng tool was used, perhaps a CVS glitch?). Manually tagging and re-releasing changed plugins to fix the problem /njh

I20080229065303

Project Error Type Compile/JUnit/Respin Analysis/Initials
WS Respin requested based on bad bug (regression) found in smoke test. The Web Services smoke test failed due to a change in the Tomcat server tools 220879. The problem was fixed by Server tooling team and was dropped in I20080229195548. Based on the comments in the bug, this was due to a change in publishing in Tomcat server tools recently and was not found because testing was not done in a new workspace. They had commented that new JUnit would be included in the next I-build to catch this in the future.

I20080229121121

Project Error Type Compile/JUnit/Respin Analysis/Initials
releng WS fix not in yet. Unclear what kicked off this build. No change in map files from previous. Either someone did "manually", or, bug in CruiseControl. action: will keep an eye on this, and if continues to happen will experiment with CC settings, or upgrate to 2.7.2 EA, or disable manual force build. (If others committers do 'manually', for any reason, that's fine ... just let Helen or David know). dw

I20080229195548

Project Error Type Compile/JUnit/Respin Analysis/Initials
JSF JUnit breakage due to core exception while deleting project tearDown. The failure in I20080229195548 is new and indeed spurious. It stems from the standard class of problem we have identified where project contents are being file locked in ways that previously they either weren't or in ways that are more susceptible to race conditions. Once we get past code review I will endeavour to convert to the busy loop idiom for deleting projects across all of our JUnit tests. CB.

I20080301004930

Project Error Type Compile/JUnit/Respin Analysis/Initials
SSE (JSP) Another, apparently 'random' error in JSP JUnits, no change in map files from previous build.

Back to the top