On 12/23/2015 7:08 PM, Semyon Sadetsky wrote:
Alexander, please clarify
1. will this fix the problem?
It will still fails for example for the scale factor 3.
2. What if user uses value > Integer.MAX_VALUE / SCALE_FACTOR for
window dimensions? If it should be prohibited now with your change,
then we need to add this to all javadocs because window size does not
meet its native limit in this case.
The sizes are scaled on the native system. I think that values that
are greater than Integer.MAX_VALUE should be truncated to the
Integer.MAX_VALUE.
If it does not checked it should be considered as a bug.
The Window.setSize() method have already had the documentation to
cover such cases: "The method changes the geometry-related data.
Therefore, the native windowing system may ignore such requests, or it
may modify the requested data, so that the Window object is placed and
sized in a way that corresponds closely to the desktop settings."
Thanks,
Alexandr.
--Semyon
On 12/23/2015 6:41 PM, Alexander Scherbatiy wrote:
On 12/22/2015 6:41 PM, Semyon Sadetsky wrote:
Right. This is regression of 8073320.
Fix is annulled. Thanks.
It would be better to update the test to take the scale factor into
the account.
Something like A_BIG_SIZE = Integer.MAX_VALUE / SCALE_FACTOR.
Thanks,
Alexandr.
--Semyon
On 12/22/2015 6:03 PM, Yuri Nesterenko wrote:
Semyon,
but earlier builds of jdk9 could handle this bordercase properly,
didn't they?. E.g. b95?
-yan
On 12/22/2015 04:22 PM, Semyon Sadetsky wrote:
Hello,
Please review fix for JDK9:
bug: https://bugs.openjdk.java.net/browse/JDK-8145795
webrev: http://cr.openjdk.java.net/~ssadetsky/8145795/webrev.00/
Windows do not accept Integer.MAX_VALUE as dimension for internal
reasons. Twice less value works.
--Semyon
