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.
Continuous development build #20050219-0738 of NetBeans 4.1 Windows XP, JDK 1.5.0_02 build #05 Description: ============ There are 6 IllegalArgumentExceptions thrown during startup of the IDE with clean userdir and the same happens when shutting down. Please see the attachment.
Created attachment 20394 [details] One of the IAE exceptions thrown on startup.
The build number is #20050217-0738, of course. I apologize.
java.lang.IllegalArgumentException: URI has an authority component [file://UNKNOWN] at org.netbeans.modules.masterfs.MasterURLMapper.getFileObjects(MasterURLMapper.java:48) at org.openide.filesystems.URLMapper.findFileObject(URLMapper.java:181)
*** Issue 55121 has been marked as a duplicate of this issue. ***
I can't reproduce it on Linux, so maybe just Win problem .. As it's reported in duplicated issue , the same problem is reproducible on today's trunk build NB4.1 (200502170805)
Probably side effect of Jarda's commit fixing #38348.
What exactly subcomponent "FS extension" is ? Probably movement aproaching to replace DataSystems. If so then its wrong category in this case.
It can't be reproducible on Linux cause there is no BrokenShadow on Linux. In core/windows is defined shadow called [D-BACK_QUOTE.shadow] and its originalFile defined as an attribute is declared as methodValue with value "org.netbeans.core.windows.actions.RecentViewListAction.getStringRep4Unixes". This method return non null value just for Unix. The second problem occures when BrokenDataShadow in its constructor can't resolve URL. Then is used some fake value == url = new URL("file://UNKNOWN"); //NOI18N. Later is checked if this broken shadow is checked for validity and there is looked for for url ""file://UNKNOWN". Such broken shadows that has no valid URL won't be never valid anyway. Moreover such url contains host part "UNKNOWN" which is source of problems.
/cvs/openide/loaders/src/org/openide/loaders/BrokenDataShadow.java new revision: 1.12; previous revision: 1.11 Fixed. Loaders problem. Jarda, please verify, close and consider 4.1_beta_fix.
BTW, what are the use cases for the DataShadow and BrokenDataShadow classes in the IDE? It seems that when we don't have the "link" concept in the IDE UI, there should not be a reason to have the concept of "pointers". It is quite scary to see in the stacktrace how many layers of code are involved when calling a method such as org.netbeans.core.windows.persistence.PersistenceManager.getTopComponentForID().
Petr probably reassigned this issue to filesystem by mistake because this issue is really datsystems issue. Reassigned back.
I think this isn't issue fro beta branch. /cvs/openide/loaders/src/org/openide/loaders/BrokenDataShadow.java, new revision: 1.13; previous revision: 1.12 /cvs/openide/test/unit/src/org/openide/loaders/DataShadowTest.java, new revision: 1.10; previous revision: 1.9 I don't find this stacktrace so strange. That's current state - instances, setings are stored via SystemFileSystem and DataSystem are involved (InstanceDataObject, XMLDataObject, DataShadows). There are already some documents evaluating critically current solution.
Verified in development build #200503232205 of NetBeans 4.1 based on the fact that I have not seen anything like this again.