Perhaps the test can be simplified further(some flags can be removed as well as the number of try/catch may be minimized), but this version loos fine. Thanks.
----- sreeprakas...@oracle.com wrote: > Thanks for the inputs Sergey. > > I have changed the test case to use CountDownLatch and Swing > components. I have also moved the test location from awt/TextField to > swing/JTextField > > Updated Webrev : > http://cr.openjdk.java.net/~rpatil/8180370/webrev.03/ > > I have also tested out the cases that you had mentioned (JDK-8148555, > JDK-8132503,JDK-8073008 & JDK-8068283) and they all worked fine. > > Regards, > Sreeprakash > > -----Original Message----- > From: Sergey Bylokhov > Sent: Tuesday, June 6, 2017 9:57 AM > To: Sreeprakash Sreedharan <sreeprakas...@oracle.com> > Cc: Philip Race <philip.r...@oracle.com>; awt-dev@openjdk.java.net > Subject: Re: <AWT Dev> [10] Review request for JDK-8180370: Characters > are skipped on input of Korean text on OS X > > Hi, > I suggest to check the Swing component first because they have more > priority, so the swing test will be useful. > A few notes about the fix: > - Can you please double check that the cases described in > JDK-8148555, JDK-8132503,JDK-8073008,JDK-8068283 still works as > expected. We had some regressions here so it will be useful to double > check. > - Its possible to tweak the test a little bit, and use CountDownLatch > instead of Thread.sleep + interrupt()+TimerTask +Timer. Just await() > the main method and countDown() then the user press the button. In > this case you can move all logic from time to the main method. I guess > this will simplify the test. > - Note that in case of swing you will need to call validateInput() on > EDT. > > > > > As the term "Special character" was ambiguous, I have renamed the > test case and the appropriate comments and also removed some swing > dependencies. > > Updated Webrev : > http://cr.openjdk.java.net/~rpatil/8180370/webrev.02/ > > > > Also, I have written the test case for the AWT component (TextField) > only. > > Is it required to write a duplicate test case for the equivalent > swing component (JTextField) ? > > > > Regards, > > Sreeprakash > > > > -----Original Message----- > > From: Phil Race > > Sent: Saturday, June 3, 2017 2:29 AM > > To: Sreeprakash Sreedharan <sreeprakas...@oracle.com>; > > awt-dev@openjdk.java.net > > Subject: Re: <AWT Dev> [10] Review request for JDK-8180370: > Characters > > are skipped on input of Korean text on OS X > > > > OK .. although I think this may need bake time before being > backported. > > but since I suppose hardly anyone is using or testing 10 at the > moment that is going to be tricky. > > > > -phil. > > > > On 06/02/2017 01:47 PM, Sreeprakash Sreedharan wrote: > >> Thanks for the inputs Phil. > >> > >> I have updated the code and the comments. > >> > >> Updated Webrev : > >> http://cr.openjdk.java.net/~rpatil/8180370/webrev.01/ > >> > >> The issue was that any non-alpha character(numbers and symbols) > typed immediately after a Korean character was getting skipped. > >> The non-alpha character will come if you type again. > >> For example, an exclamation mark (!) entered after the Korean > character for q will not show up. > >> However, if you type it again it will appear. But instead of having > 2 exclamation marks we have just one. > >> > >> I had tested it out with 2-Set Korean, Wubi Xing (Chinese) and ABC > - Extended keyboard layouts on mac with different input methods . > >> > >> I had run all the manual and automatic regression tests for all > text based controls in awt (like TextField, TextArea etc) and swing > (like JTextField, JTextArea, JTextPane, JEditorPane etc) and did not > find any issues. > >> > >> Thanks, > >> Sreeprakash > >> > >> > >> From: Phil Race > >> Sent: Friday, June 2, 2017 9:44 PM > >> To: Sreeprakash Sreedharan <sreeprakas...@oracle.com>; > >> awt-dev@openjdk.java.net > >> Subject: Re: <AWT Dev> [10] Review request for JDK-8180370: > >> Characters are skipped on input of Korean text on OS X > >> > >> I am not familiar with this code but I have a few comments anyway > >> > >> 1. I dislike cluttering the source with bug ids. If we did that for > every fix > >> quite soon the source code would be a mess of semi-random > numbers. > >> Anyone who really wants to know when this change was made has > the > >> history > >> > >> 2. if is not a function. So "if(" -> "if (" > >> > >> 3. When a marked text -> When marked text > >> > >> 4. "!" is not a "special character" .. its quite ordinary .. so > what do you mean ? > >> > >> 5. What testing have you done to make sure no other cases are > broken by this change ? > >> The new test is manual and I'd bet that most tests that might > cover this are manual > >> I'd expect to hear that you have tested different input > scenarios such as a couple > >> of input methods/locales, and AWT and Swing input with a > representative set of > >> input as well as running all the relevant regression tests. > >> > >> -phil. > >> On 06/02/2017 07:48 AM, Sreeprakash Sreedharan wrote: > >> Hi All, > >> > >> Kindly review the fix for JDK10. > >> Bug: https://bugs.openjdk.java.net/browse/JDK-8180370 > >> Webrev: http://cr.openjdk.java.net/~rpatil/8180370/webrev.00/ > >> Issue: Special characters (like !,/\<> ) were getting skipped when > immediately entered after a marked text on MacOSX. > >> Fix: Made sure that fKeyEventsNeeded flag is reset, when a > non-marked text is encountered, so that it doesn't get ignored by key > down. > >> > >> Note: Since the keyboard layout has to be changed to Korean, I have > added a manual test case wherein the user is prompted to change to > Korean keyboard layout and then execute the test. > >> > >> > >> Regards, > >> Sreeprakash > >