| Summary: | PDE chooses wrong JRE / EE to launch JUnit plugin tests | ||
|---|---|---|---|
| Product: | [Eclipse Project] PDE | Reporter: | Michael Rennie <Michael_Rennie> |
| Component: | UI | Assignee: | Curtis Windatt <curtis.windatt.public> |
| Status: | VERIFIED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | CC: | curtis.windatt.public, daniel_megert, david_williams, jarthana, pwebster, stephan.herrmann |
| Version: | 3.7 | ||
| Target Milestone: | 4.3 M3 | ||
| Hardware: | PC | ||
| OS: | Windows 7 | ||
| Whiteboard: | |||
|
Description
Michael Rennie
This occurs alot, see bug 282562 and bug 341612 I'm not sure what the answer is. One option (we're using) is to define EEs for our launch configs. PW I wrote some code for PDE to calculate the default EE and use that in the launch over a default JRE. Still a couple of problems to fix: 1) Whether fragments in the launch should be checked 2) Checking all the BREEs ends up too restrictive 3) Need to check performance as we end up checking the BREEs of every bundle in the workspace. http://git.eclipse.org/c/pde/eclipse.pde.ui.git/?h=cwindatt%2Fbug344633_PDE_chooses_wrong_JRE Fixed in master: http://git.eclipse.org/c/pde/eclipse.pde.ui.git/commit/?id=66dce08a807338fc4928522be86f7c4e38017ece 1) Fragments are skipped 2) BREE checking is fixed, tests against bundles with 1.5, 1.6, 1.7, CDC, etc. 3) Checking EE against the 520 bundles in my launch takes 15ms. If the config is opened in the LCD we store the setting so no more lookups happen. Build I20121029-2000 exhibits failures in JDT/Core tests, which cannot readily be connected to the only fix in JDT/Core that went into that specific build: http://download.eclipse.org/eclipse/downloads/drops4/I20121029-2000/testresults/html/org.eclipse.jdt.core.tests.compiler_macosx.cocoa.x86_5.0.html The pattern of failures is: tests are supposed to execute a small test program and check the printout of the program, which is unexpectedly empty. For lack of a better theory let me ask if this bug (also released for the same I-Build) could possibly have an influence on executing programs from within a JUnit-Plug-in-Test on Mac? (In reply to comment #4) > Build I20121029-2000 exhibits failures in JDT/Core tests, which cannot > readily be connected to the only fix in JDT/Core that went into that > specific build: > http://download.eclipse.org/eclipse/downloads/drops4/I20121029-2000/ > testresults/html/org.eclipse.jdt.core.tests.compiler_macosx.cocoa.x86_5.0. > html > > The pattern of failures is: tests are supposed to execute a small > test program and check the printout of the program, which is > unexpectedly empty. > > For lack of a better theory let me ask if this bug (also released for the > same I-Build) could possibly have an influence on executing programs from > within a JUnit-Plug-in-Test on Mac? It could be related to the launching of the junit plug-in test. The default JRE used to launch the plug-in test may differ. Cool! Now I can run new JUnit tests again without manually adjusting the launch config. *** Bug 297191 has been marked as a duplicate of this bug. *** |