<Swing Dev> [9] Review request for 8142966 Wrong cursor position in text components on HiDPI display
Alexandr Scherbatiy
alexandr.scherbatiy at oracle.com
Tue May 31 20:36:54 UTC 2016
On 5/31/2016 10:54 PM, Phil Race wrote:
>
>
> This looks reasonable except I am not sure when gc would be null here :-
>
> 1062 private static FontRenderContext getFRCProperty(JComponent c) {
> 1063 if (c != null) {
> 1064
> 1065 GraphicsConfiguration gc = c.getGraphicsConfiguration();
> 1066 AffineTransform tx = (gc == null) ? null :
> gc.getDefaultTransform();
>
>
> if it is only in some case such as where Swing is rendering in
> headless mode that should be OK
It happens when a size of a component is requested but the
component has not been added to a frame yet. This is a case where a text
component size can differ before it is added to a frame and after that.
Thanks,
Alexandr.
>
> -phil.
>
> On 05/04/2016 10:42 PM, Alexandr Scherbatiy wrote:
>>
>> Hello,
>>
>> Could you review the fix:
>> bug: https://bugs.openjdk.java.net/browse/JDK-8142966
>> webrev: http://cr.openjdk.java.net/~alexsch/8142966/webrev.02
>>
>> Swing uses a font rendering context which does not take the
>> graphics configuration transform into account.
>> This lead to the situation when text size is measured by one font
>> rendering context and drawn by another. As the result the calculated
>> text size can differ from the drawn text size.
>>
>> For this issue it means that a cursor position can be farther away
>> from the char position.
>>
>> This is the initial fix where graphics configuration transform is
>> included into the Swing font rendering context.
>>
>> There is one more issue related to the fact that glyph advances are
>> calculated in dev space and then transformed to the user space. Char
>> width can have fractional value in user space if dev transform is not
>> identity.
>> It means that Swing should be able to draw a cursor at or select a
>> text from a fractional position.
>> These issues will be considered in separated fixes.
>>
>> Thanks,
>> Alexandr.
>>
>
More information about the swing-dev
mailing list