Hi Pankaj,
Some comments about the test:
I guess since we are testing for Platform.isLinux() in the test, there's
no need of "requires" tag. But, I think it's better to use
|System.getProperty("os.name") so that this test can be run in
standalone mode also(without using Platform class), otherwise it needs
to be run in conjunction with jtreg(in which case @requires tag is better)|
Also, I guess it will be good if you use
UIManager.setLookAndFeel(...GtkLookAndFeel..) instead of passing through
command line as user who is running standalone may forget to test in GTKL&F
Also, you are not disposing the frame if it times out so the instruction
frame will linger even when the test fails.
Regards
Prasanta
On 02-Dec-19 4:45 PM, Pankaj Bansal wrote:
Hello Prasanta/Phil/Sergey,
<<1) I suppose this fix is for 14, not 13.
Yes, that was by mistake. I am not changes the subject line, as this
would create separate mail thread. This will be pushed to 14 only.
2) It is "OL", not "OEL"
Ok.
3) Can we create a reliable automated test for this ? If not we should
at least create a manual one.
Done. I have added a manual test case.
4) A separate issue but we should consider adding a JSpinner demo to
SwingSet2
Yes, I was also thinking that we should do it. Do we need separate
approvals for changing the demos. I mean approval from some place
other than swing-dev like demo team etc.
<< Does it mean that this is the behavior of the gtk3 native spinner
component?
Yes, this behavior happens in native spinner. This change makes our
Spinner much more similar to native gtk3 spinner
webrev: http://cr.openjdk.java.net/~pbansal/8234733/webrev01/
Regards,
Pankaj
*From:*Philip Race
*Sent:* Wednesday, November 27, 2019 9:49 PM
*To:* Prasanta Sadhukhan
*Cc:* [email protected]
*Subject:* Re: <Swing Dev> [13] RFR JDK-8234733: We can't distinguish
if the spinner button is pressed or unpressed
1) I suppose this fix is for 14, not 13.
2) It is "OL", not "OEL"
3) Can we create a reliable automated test for this ?
If not we should at least create a manual one.
4) A separate issue but we should consider adding a JSpinner demo to
SwingSet2
-phil.
On 11/27/19, 2:31 AM, Prasanta Sadhukhan wrote:
It works for oel8 too. So the fix looks fine to me.
Regards
Prasanta
On 27-Nov-19 3:54 PM, Prasanta Sadhukhan wrote:
Hi Pankaj,
I can see the issue resolved in ubuntu18.04 but not in OEL8.
Also, SwingSet2 does not have JSpinner demo so we cannot put
noreg-demo in JBS, so we probably need a regression test, even
manual, for this.
Regards
Prasanta
On 25-Nov-19 4:13 PM, Pankaj Bansal wrote:
Hi All,
Please review the following fix for jdk14.
Bug:
https://bugs.openjdk.java.net/browse/JDK-8234733
webrev:
http://cr.openjdk.java.net/~pbansal/8234733/webrev00/
<http://cr.openjdk.java.net/%7Epbansal/8234733/webrev00/>
Issue:
In GTKL&F, the spinner’s up/down buttons are drawn in same
way whether the button is pressed or not. We can't
distinguish visually whether a button is pressed or not.
The buttons should be highlighted when they are in pressed
state. The issue is due to style and theme changes done in
gtk3 in version gtk3.20.
Fix:
The fix sets the style properly when drawing the spinner
buttons with gtk3.20 or higher. All works as it is if the
gtk3 version is lower than gtk3.20. The current fix is in
continuity of the fixes done for issues like
https://bugs.openjdk.java.net/browse/JDK-8218469
Testing:
The fix can be verified by running SwingSet2 or any other
demo/sample of JSpinner (one is attached in the JBS) and
it has to be verified manually. I have verified this on
Ubuntu 18.04, 19.04, 19.10 and OEL 7.5. The issue is
solved on all platforms.
Regards,
Pankaj Bansal
