It does sound like a bug. Moved the web incidents into a JDK bug (8014558)
Naoto
On 5/14/13 5:30 AM, Anthony Petrov wrote:
Hi Deven,
I'm copying i18n-dev@ because they manage the IM code.
As to a test, jtreg supports manual tests. See [1] for a lot of tips
about using manual tests. An example of a manual test is at:
jdk/test/java/awt/event/MouseEvent/SpuriousExitEnter/SpuriousExitEnter.java
[1] http://openjdk.java.net/jtreg/faq.html
--
best regards,
Anthony
On 05/09/2013 11:33 AM, Deven You wrote:
Hi All,
I found there is a problem with our IMF(Input Methods Framework) when
dealing with DBCS input like Chinese and Japanese on Windows 7 32bit env.
The simple scenario is if your application has 2 windows and then:
1: focus on first window, change to microsoft pinyin input method
Chinese mode.
2: Input some words, leave some words in preedit string not committed.
3: Move focus to the other window, change to English mode and input
something
4: Move back to the first window. It is in English mode, but the preedit
string is still there(this is the bug).
5: Delete the preedit string with backspace and press enter, it appears
again.
I have raised a sunbug for this issue, the internal ID is: 9002399
I have written a test case[1] to reproduce this problem on windows 7 32
bit machine with latest OpenJDK 8.
Since I am not clear how to run this test case as a jtreg, I just put it
to cr.openjdk.java.net. There are several files within this test case:
actual-step*.png Steps reproducing the bug. (can not )
JTextAreaTest3.java The test case.
Readme.txt Simple description
IMF_04_69120.txt More detail.
I also made possible patch[2] for this problem.
Please anyone take a look at this issue and give your suggestion.
[1] http://cr.openjdk.java.net/~youdwei/ojdk-687/IMF4/
<http://cr.openjdk.java.net/%7Eyoudwei/ojdk-687/IMF4/>
[2] http://cr.openjdk.java.net/~youdwei/ojdk-687/webrev/
<http://cr.openjdk.java.net/%7Eyoudwei/ojdk-687/webrev/>
Thanks a lot!
