Bugzilla – Full Text Bug Listing
| Summary: | Run SDK tests against IBM Java 7 JRE | ||
|---|---|---|---|
| Product: | [Eclipse Project] Platform | Reporter: | Dani Megert <daniel_megert> |
| Component: | Releng | Assignee: | Platform-Releng-Inbox <platform-releng-inbox> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | enhancement | ||
| Priority: | P4 | CC: | david_williams, jarthana, kim.moir, stephan.herrmann |
| Version: | 3.7 | ||
| Target Milestone: | 3.8 | ||
| Hardware: | All | ||
| OS: | All | ||
| Whiteboard: | |||
| Bug Depends on: | 347321 | ||
| Bug Blocks: | |||
|
Description
Dani Megert
Yes, I spent some time yesterday trying to track where the latest binaries are available :-) The tests are running now on I20110425-1800. Test results are here for against IBM 7 JRE http://download.eclipse.org/eclipse/downloads/drops/I20110425-1800/testResults.php Kim, did you run them on a clean disk? It "looks" like there was already some data, e.g. testProjectCreation Error Resource '/APITests/bin' already exists. looks suspicious as this is the first test for ApiToolsPluginTestSuite. I copied the test artifacts from the build to a new machine, removed the results directory, and updated the scripts to run them with Java 7. The install is cleaned and new test bundles are installed each time. If there are other (disk) areas that aren't cleaned up, this could be the problem. (In reply to comment #5) > If there are other (disk) areas that aren't cleaned up, this could be the > problem. I'm only worried about the workspace. Dani, should I close this bug or do you want to keep it open. If so, I'll change the target to 3.7RC1 (In reply to comment #7) > Dani, should I close this bug or do you want to keep it open. If so, I'll > change the target to 3.7RC1 Yes, I'd like to keep it open until we run good on it. I've removed the target milestone as this is a low-prio effort. rerunning the tests now Did we get any results? Some, but some are missing. I'm trying to rerun them now. (Not related to Java 7, but rather properties set in the build scripts). Many of the tests rely on having J2SE-5.0, J2SE-6.0 properties set in vm.properties. I'm going to remove that restriction for 3.8 since all of our tests run on Java 6.0 anyways. I opened bug 347321 for this issue. Do you use the same JRE as last time or did you update? The latest version has some issues when launching Java stuff from inside the VM itself. I'm using the same one as last time. Latest results are here http://eclipsebuildserv.ottawa.ibm.com/downloads/master/downloads/drops/I20110526-0800/testResults.php which are replicating to http://download.eclipse.org/eclipse/downloads/drops/I20110526-0800/testResults.php Dani, do you want this bug to remain open or should I close it? It looks like the equinox.p2.tests.* are missing. Unfortunately all the individual console logs for the Java 7 JRE are not there. Can you rerun them again? (low prio). The p2 tests depend on having the J2SE-5.0 variable set. This isn't there when I run tests on Java 7. I've opened a bug 347321 to address this issue. Can this be closed now that we run tests on Oracle Java 7 each 3.8 build on Windows? (In reply to comment #18) > Can this be closed now that we run tests on Oracle Java 7 each 3.8 build on > Windows? We should run them once in a while, e.g. for each milestone. (In reply to comment #18) > Can this be closed now that we run tests on Oracle Java 7 each 3.8 build on > Windows? From JDT/Core perspective I was mainly interested in having regular tests on Java 7, no matter which vendor. -> I'm happy with the current situation. Or windows tests are always done with Java 7, so closing as "fixed" (though, we all know, tests on build.eclipse.org still need work). |