This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
(No Property Editor) is shown for compilers and executors in property sheets when IDE is run by XTest harness. It seems it is a regression of issue 28536. It happens only in IDE build with new window system. To reproduce: cd java/test ant -Dxtest.run.singletest=dummy -Dxtest.ide.exit=false -Dnetbeans.home=<new winsys IDE home> Then open Tools|Options and navigate to Editing|Java Sources and you will see (No Property Editor) message instead of value of some properties. It is important to fix it because it blocks automated tests.
It can't be fixed. Options are now dialog, not TopComponent...thus changing the nodes there doesn't influence the global activated nodes... which listens the global properties window on. Change the test. It is needless now.
I think it is misunderstanding. Please, look at screen shot where is the problem.
Created attachment 11964 [details] Screen shot
Reassigning to Jesse because he fixed 28536 and 35740 and he could know where is the problem.
Issue #35470, not 35740. I can't reproduce any problem in my winsys build. I will check that all sources are updated any try again.
BTW I was using: ant -f java/test/build.xml -Dxtest.run.singletest=dummy -Dxtest.ide.exit=false -Dnetbeans.home=/space/src/winsys_29836/nbbuild/netbeans -Dxtest.jvmargs=-Xbootclasspath/p:/tmp/28536.jar The context loader was always getting set to the NB loader, and the property editors were there. I am using Linux, JDK 1.4.2. BTW the screenshot shows a problem in the editor for NbClassPath too.
Sorry, I still can't reproduce. Try running with the attached diagnostic patch to Main.java as well as the bootcp patch JAR I mentioned and see when the bad call to setContextClassLoader occurs (if that is indeed the problem). Might be that #28536's hack has to be moved even earlier in startup, since it seems that o.n.m.javahelp.HelpSetProcessor and o.n.m.java.environment.LibraryProcessor use EQ early.
Created attachment 11968 [details] Diagnostic patch for use in conjunction with 28536.jar
I run it with 28536.jar and diagnostic patch. Jesse, could you please look at ide.log?
Created attachment 11981 [details] ide.log with 28536.jar and diagnostic patch
Yeah, hack needs to be pushed further back in o.n.c.Main, I guess. No idea why it would only cause a problem under winsys2, other than that it may be stochastic - could depend on timing or GC changes.
Created attachment 11987 [details] Possible patch
Please try the attached patch and see if it fixes the problem for you. Since I cannot reproduce the bug I cannot tell if it is going to help. In theory it should solve the problem.
Great! It fixed the problem.
Good. Feel free to commit it yourself (after the usual clean build + validation tests) or reassign to me and I will.
Patch integrated both to trunk and winsys_29836 branch: /cvs/core/src/org/netbeans/core/Main.java,v <-- Main.java new revision: 1.173; previous revision: 1.172 /cvs/core/src/org/netbeans/core/Main.java,v <-- Main.java new revision: 1.157.2.12; previous revision: 1.157.2.11
Verified in winsys build where it failed before.