Hello, Vivi. >> Is it possible to write a test (manual or automated) for the fix? > Two test files (One for JTree, one for JTable) attached to the bug for manual > test, comments added in each file about the how to do the test.
Is is possible to make an automatic test? You could show the JTree, manually get it’s AccessibleContext and call methods that reproduce the problem. Adding a manual test to the JBS wouldn’t help, because nobody would ever run this test. So the best is to make an automatic jtreg test, or at least convert your manual tests to jtreg and push them into the repo together with the fix. With best regards. Petr. 10 апр. 2014 г., в 6:04 после полудня, Vivi An <vivi...@oracle.com> написал(а): > Updated webrev: > > http://cr.openjdk.java.net/~dmarkov/8036983/jdk9/webrev.01/ > > Comments as below > > Thanks > > Vivi > > > > On 4/9/2014 8:08 AM, Alexander Scherbatiy wrote: >> On 4/8/2014 9:19 PM, Vivi An wrote: >>> Hello, >>> >>> Could you please review the fix for JDK 9? >>> >>> This bug is JAB related. ActivateDescenderPropertyChanged event for JTree >>> and JTable were not sent properly in case SHIFT+CursorDown or >>> Ctrl+CursorUp/Down are used. Fix made mainly uses lead path (acctive child >>> path) instead of selection path to check if an event needs to be fired. >>> >>> Bug: https://bugs.openjdk.java.net/browse/JDK-8036983 >>> Webrev: http://cr.openjdk.java.net/~dmarkov/8036983/jdk9/webrev.00/ >>> >> + public void fireActiveDescendantPropertyChange(TreePath oldPath, >> TreePath newPath) { >> Is it possible to make the method package access instead of public? > Yes, good idea, fixed >> >> + int focusedRow = >> JTable.this.getSelectionModel().getLeadSelectionIndex();; >> There is one more semicolon at the end. >> > Fixed >> Is it possible to write a test (manual or automated) for the fix? > Two test files (One for JTree, one for JTable) attached to the bug for manual > test, comments added in each file about the how to do the test. >> >> Thanks, >> Alexandr. >>> Thanks >>> >>> ~ Vivi >> >
