<Swing Dev> Questionable Rendering of Disabled Text
Semyon Sadetsky
semyon.sadetsky at oracle.com
Mon Nov 13 21:20:32 UTC 2017
+1
--Semyon
On 11/09/2017 10:50 AM, Phil Race wrote:
> Protocol is that there first be a bug report and then a "RFR" (request
> for review) is sent
> to the list referencing the bug + the fix
>
> I have filed https://bugs.openjdk.java.net/browse/JDK-8191041
>
> and verified that it works. So we can perhaps short-circuit that this
> time .. but not next time ..
>
> As to the blueprints path it looks safer to leave it since I don't
> know how to test it
> so I would be happy to commit the patch as is but I think this opinion
> needs to be
> reviewed by someone else on the list first ..
>
> -phil.
>
> On 11/03/2017 09:53 PM, Michael D wrote:
>> Apologies, noticed I broke the patch between testing and submission.
>>
>> Correct and re-tested version
>>
>>
>> ---
>> a/src/java.desktop/share/classes/com/sun/java/swing/plaf/gtk/GTKGraphicsUtils.java
>> +++
>> b/src/java.desktop/share/classes/com/sun/java/swing/plaf/gtk/GTKGraphicsUtils.java
>> @@ -47,34 +47,17 @@ class GTKGraphicsUtils extends SynthGraphicsUtils {
>> return;
>> }
>> int componentState = context.getComponentState();
>> - if ((componentState & SynthConstants.DISABLED) ==
>> - SynthConstants.DISABLED){
>> - if (!GTKLookAndFeel.is3()) {
>> - Color orgColor = g.getColor();
>> - g.setColor(context.getStyle().getColor(context,
>> - GTKColorType.WHITE));
>> - x += 1;
>> - y += 1;
>> - super.paintText(context, g, text, x, y, mnemonicIndex);
>>
>> - g.setColor(orgColor);
>> - x -= 1;
>> - y -= 1;
>> - }
>> - super.paintText(context, g, text, x, y, mnemonicIndex);
>> - }
>> - else {
>> - String themeName = GTKLookAndFeel.getGtkThemeName();
>> - if (themeName != null &&
>> themeName.startsWith("blueprint") &&
>> - shouldShadowText(context.getRegion(),
>> componentState)) {
>> + String themeName = GTKLookAndFeel.getGtkThemeName();
>> + if (themeName != null && themeName.startsWith("blueprint") &&
>> + shouldShadowText(context.getRegion(), componentState)) {
>>
>> - g.setColor(Color.BLACK);
>> - super.paintText(context, g, text, x+1, y+1,
>> mnemonicIndex);
>> - g.setColor(Color.WHITE);
>> - }
>> -
>> - super.paintText(context, g, text, x, y, mnemonicIndex);
>> + g.setColor(Color.BLACK);
>> + super.paintText(context, g, text, x+1, y+1, mnemonicIndex);
>> + g.setColor(Color.WHITE);
>> }
>> +
>> + super.paintText(context, g, text, x, y, mnemonicIndex);
>> }
>>
>> /**
>>
>> On 4 November 2017 at 07:13, Michael D <me at md-5.net> wrote:
>>> My bad,
>>>
>>> Patch just looks like this (bit messy because of indent change).
>>> Thanks
>>> Michael
>>>
>>>
>>>
>>>
>>> ---
>>> a/src/java.desktop/share/classes/com/sun/java/swing/plaf/gtk/GTKGraphicsUtils.java
>>> +++
>>> b/src/java.desktop/share/classes/com/sun/java/swing/plaf/gtk/GTKGraphicsUtils.java
>>> @@ -46,35 +46,17 @@ class GTKGraphicsUtils extends SynthGraphicsUtils {
>>> // ignore this.
>>> return;
>>> }
>>> - int componentState = context.getComponentState();
>>> - if ((componentState & SynthConstants.DISABLED) ==
>>> - SynthConstants.DISABLED){
>>> - if (!GTKLookAndFeel.is3()) {
>>> - Color orgColor = g.getColor();
>>> - g.setColor(context.getStyle().getColor(context,
>>> - GTKColorType.WHITE));
>>> - x += 1;
>>> - y += 1;
>>> - super.paintText(context, g, text, x, y,
>>> mnemonicIndex);
>>>
>>> - g.setColor(orgColor);
>>> - x -= 1;
>>> - y -= 1;
>>> - }
>>> - super.paintText(context, g, text, x, y, mnemonicIndex);
>>> - }
>>> - else {
>>> - String themeName = GTKLookAndFeel.getGtkThemeName();
>>> - if (themeName != null &&
>>> themeName.startsWith("blueprint") &&
>>> - shouldShadowText(context.getRegion(),
>>> componentState)) {
>>> + String themeName = GTKLookAndFeel.getGtkThemeName();
>>> + if (themeName != null && themeName.startsWith("blueprint") &&
>>> + shouldShadowText(context.getRegion(), componentState)) {
>>>
>>> - g.setColor(Color.BLACK);
>>> - super.paintText(context, g, text, x+1, y+1,
>>> mnemonicIndex);
>>> - g.setColor(Color.WHITE);
>>> - }
>>> -
>>> - super.paintText(context, g, text, x, y, mnemonicIndex);
>>> + g.setColor(Color.BLACK);
>>> + super.paintText(context, g, text, x+1, y+1,
>>> mnemonicIndex);
>>> + g.setColor(Color.WHITE);
>>> }
>>> +
>>> + super.paintText(context, g, text, x, y, mnemonicIndex);
>>> }
>>>
>>> /**
>>>
>>> On 4 November 2017 at 06:39, Philip Race <philip.race at oracle.com>
>>> wrote:
>>>> Yes these mailing lists strip attachements
>>>> So an in-line patch .. or send directly to one of us to host it for
>>>> you on
>>>> cr.openjdk.java.net.
>>>>
>>>> -phil.
>>>>
>>>> On 11/3/17, 12:34 PM, Sergey Bylokhov wrote:
>>>>> Hi, Michael.
>>>>> It seems that the attached patch was removed from the email, can you
>>>>> please provide a link to it or inline in email.
>>>>> Thank you.
>>>>>
>>>>> On 03/11/2017 03:09, Michael D wrote:
>>>>>> On 3 November 2017 at 21:07, Michael D <me at md-5.net> wrote:
>>>>>>> Hi All,
>>>>>>>
>>>>>>> Last couple of days I've been trying to root out this bug -
>>>>>>> https://bugs.launchpad.net/ubuntu/+source/openjdk-8/+bug/1729558
>>>>>>> which
>>>>>>> has resulted in poor looking rendering on my machine.
>>>>>>> It turns out that the cause of this behavior is that Swing tries
>>>>>>> to do
>>>>>>> its own magic disabled item rendering by turning the text white and
>>>>>>> shifting it a pixel in each direction. I have no doubt there was
>>>>>>> probably a good reason for this, however the file has remained
>>>>>>> unchanged since its initial import in 2007! (Maybe someone from
>>>>>>> Oracle
>>>>>>> can find out more)
>>>>>>> Removing this code (patch attached) makes disabled menu items
>>>>>>> render
>>>>>>> correctly in both light and dark GTK themes. Its completely
>>>>>>> trivial,
>>>>>>> but I've nonetheless provided the patch for context. There is also
>>>>>>> some adjacent code that references a metacity theme called
>>>>>>> "blueprint"
>>>>>>> that also seems pre 2007 era, but I've left that in as I have no
>>>>>>> idea
>>>>>>> of its context.
>>>>>>>
>>>>>>> Appreciate if this change / bug could be reviewed & discussed.
>>>>>>> In the
>>>>>>> meantime I will also forward it downstream to Ubuntu as this is
>>>>>>> where
>>>>>>> I encountered it / affects my usage.
>>>>>>>
>>>>>>> Thanks
>>>>>>> Michael
>>>>>
>>>>>
>
More information about the swing-dev
mailing list