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.
Summary: | [TTV] Bugs in selections in TreeTreeView | ExplorerManager | ||
---|---|---|---|
Product: | platform | Reporter: | Ondrej Rypacek <rypacek> |
Component: | Explorer | Assignee: | Jiri Rechtacek <jrechtacek> |
Status: | VERIFIED FIXED | ||
Severity: | blocker | Keywords: | REGRESSION |
Priority: | P2 | ||
Version: | 3.x | ||
Hardware: | PC | ||
OS: | Linux | ||
Issue Type: | DEFECT | Exception Reporter: | |
Attachments: |
a demo program
a help code for investigate |
Description
Ondrej Rypacek
2003-05-02 10:52:37 UTC
Ondro, I evaluate ASAP what's wrong. Created attachment 10439 [details]
a demo program
I don't see what's wrong. The selection changes in JTree are correctly propagated to ExplorerManager and followup fired outside, for me it looks ok. I attached a simple program, it listens on ExplorerManager and prints a content of this property. Could you reproduce this problem with the attached program? Or could attach any program which can demonstrate it? I left it open for now. Pls. attach a comment. Thanks I didn't reproduced it in your demo test, but I did in JarPackager but also in the Options dialog in projects. I didn't manage to find any regularity of when it appears, but when I click it for some time, it appears. Just give me a call when it suits you and I'll show you. I've probably got the right clicking pattern ;-) Created attachment 10551 [details]
a help code for investigate
I attached a code for deeply investigation. There are println from TreeView's TreeSelectionListener and from ExplorerManager's PropertyChangeListener. You could see the content of native JTree's selection paths and compare with value of PROP_SELECTED_NODES in ExplorerManager. Pls. send a result of your scrutiny. Thanks I tried it with your patch and found out that *selection in TreeView and ExplorerManager are the same, but it the nodes don't appear to be selected*. At some point , even that the TV correctly deselects as I click , the internal state ov TV and EM incorrectly indicates that the node is still selected. Simply put, there's a difference between what is kept internally (inside TV and EM) and what is shown on the screen. The logic of selections is correct with respect to *what is seen*, only from time to time the internal state of the selection is displayed -- that appears as *one click -> many selections at once*. But internally it is wildly out of sync. I'm sure. The prints in TreeView display a state from TreeSelectionListener attched on JTree, prints the content of TreeSelectionEvent. Do you want to say that JTree wrongly fires selection changes? I suppose: not. Could you more precise problem or call me home? There are discrepancies in visible selection in TreeTableView and content of PROP_SELECTED_NODES. It is maybe regression of issue 32462. I reverted fix of 32462; now it should bahave same as in nb3.5. verified |