<AWT Dev> [9] Review request for 8149115: [hidpi] Linux: display-wise scaling factor should probably be taken into account

[Sergey Bylokhov]

On 22.08.16 11:33, Semyon Sadetsky wrote:
>> It is unclear why gdk_scale is applied on top of native scale? So if
>> both GDK_SCALE and scale-factor are set, then we will get the multiply
>> of them, is it expected?
> yes. GDK_SCALE is to scale gtklib based apps. Desktop scale is a
> separate scale.

So if the user will want to have double scale the he can set desktop 
scale=2.0(which is supported by one set of applications) and 
GDK_SCALE=2.0(which is supported by another set of applications), in 
this case the java scale will be 4, right?

>
> --Semyon
>>
>> On 08.07.16 13:27, Alexander Zvegintsev wrote:
>>> +1
>>>
>>>
>>> On 7/8/16 10:37 AM, Alexandr Scherbatiy wrote:
>>>> The fix looks good to me.
>>>>
>>>> Thanks,
>>>> Alexandr.
>>>>
>>>> On 7/8/2016 9:25 AM, Semyon Sadetsky wrote:
>>>>> I have changed variables declaration style in systemScale.c, because
>>>>> it produced warnings in Solaris build.
>>>>>
>>>>> http://cr.openjdk.java.net/~ssadetsky/8149115/webrev.03/
>>>>>
>>>>> --Semyon
>>>>>
>>>>> On 7/7/2016 9:57 AM, Alexandr Scherbatiy wrote:
>>>>>> The fix looks good to me.
>>>>>>
>>>>>> Thanks,
>>>>>> Alexandr.
>>>>>>
>>>>>> On 7/6/2016 11:46 PM, Alexander Zvegintsev wrote:
>>>>>>> Still looks good.
>>>>>>>
>>>>>>> --
>>>>>>> Thanks,
>>>>>>> Alexander.
>>>>>>>
>>>>>>> On 06.07.2016 21:42, Semyon Sadetsky wrote:
>>>>>>>> Thanks, Alexander. Please see the updated webrev
>>>>>>>> http://cr.openjdk.java.net/~ssadetsky/8149115/webrev.02/
>>>>>>>>
>>>>>>>> --Semyon
>>>>>>>>
>>>>>>>>
>>>>>>>> On 7/6/2016 9:03 PM, Alexander Zvegintsev wrote:
>>>>>>>>> The fix looks good to me.
>>>>>>>>>
>>>>>>>>> Just a minor comment: multiple NULL checks in get_schema_value
>>>>>>>>> could be wrapped in CHECK_NULL_RETURN macro, e.g.:
>>>>>>>>>
>>>>>>>>> CHECH_NULL_RETURN(fp_g_settings_schema_has_key
>>>>>>>>>             = dlsym(lib_handle, "g_settings_schema_has_key"),
>>>>>>>>> NULL);
>>>>>>>>>         CHECH_NULL_RETURN(fp_g_settings_new_full
>>>>>>>>>             = dlsym(lib_handle, "g_settings_new_full"), NULL);
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On 7/6/16 5:27 PM, Alexandr Scherbatiy wrote:
>>>>>>>>>>
>>>>>>>>>> The fix looks good to me.
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>> Alexandr.
>>>>>>>>>>
>>>>>>>>>> On 7/6/2016 4:46 PM, Semyon Sadetsky wrote:
>>>>>>>>>>> On 7/6/2016 12:26 PM, Alexandr Scherbatiy wrote:
>>>>>>>>>>>
>>>>>>>>>>>> On 7/5/2016 9:59 PM, Semyon Sadetsky wrote:
>>>>>>>>>>>>> Hello,
>>>>>>>>>>>>>
>>>>>>>>>>>>> Please review fix for JDK9:
>>>>>>>>>>>>>
>>>>>>>>>>>>> bug: https://bugs.openjdk.java.net/browse/JDK-8149115
>>>>>>>>>>>>>
>>>>>>>>>>>>> webrev:
>>>>>>>>>>>>> http://cr.openjdk.java.net/~ssadetsky/8149115/webrev.00/
>>>>>>>>>>>>>
>>>>>>>>>>>>> Currently the hidpi support on linux only reads GDK_SCALE
>>>>>>>>>>>>> environment variable to get the native scale. Although,
>>>>>>>>>>>>> Gnome3 and Unity DE use own settings to control the interface
>>>>>>>>>>>>> scale. The fix adds possibility to read those settings to
>>>>>>>>>>>>> make java apps hidpi scale similar to the native apps.
>>>>>>>>>>>>> Currently only integer scale values are supported. If native
>>>>>>>>>>>>> scale is not an integer value it is rounded to the nearest
>>>>>>>>>>>>> integer.
>>>>>>>>>>>>   systemScale.c
>>>>>>>>>>>>    Should the while loop have a break statement in the
>>>>>>>>>>>> getDesktopScale(...) method?
>>>>>>>>>>> That's make sense. Please look at the updated webrev:
>>>>>>>>>>> http://cr.openjdk.java.net/~ssadetsky/8149115/webrev.01/
>>>>>>>>>>>
>>>>>>>>>>> --Semyon
>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>   Thanks,
>>>>>>>>>>>>   Alexandr.
>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> --Semyon
>>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>>
>


-- 
Best regards, Sergey.