Test failing - CppunitTest_cppcanvas_emfplus

[Chris Sherlock]

Hi all,

It appears that the test is failing when I build with SAL_FORCEGL=1

If I had to guess, I’d say that the anti-aliasing method is different - I’m using software GL via llvm-pipe though. 

The image is correct, however - perhaps it would be better to try to test the color of a pixel closer to the middle of the image to ensure it is green? 

Chris

> On 16 Feb 2015, at 11:53 pm, Chris Sherlock <chris.sherlock79 at gmail.com> wrote:
> 
> Hi Miklos,
> 
> I’m using Ubuntu Linux, running Gnome. 
> 
> It’s odd, I believe that this pixel is right next to a line that has been anti-aliased. 
> 
> I’m wondering if we could get a dump of the PNG that is generated from a non-affected system and I’ll compare it against the one I’m looking at. 
> 
> Chris
> 
>> On 16 Feb 2015, at 7:40 pm, Miklos Vajna <vmiklos at collabora.co.uk> wrote:
>> 
>> Hi Chris,
>> 
>> On Mon, Feb 16, 2015 at 01:47:51PM +1100, Chris Sherlock <chris.sherlock79 at gmail.com> wrote:
>>> I guess it would also be helpful to note:
>>> 
>>> When I open it in oletoy, it's very definitely meant to be #00ff00 (green). There is some anti-aliasing, but that's not at position (140, 142)... and actually (140, 140) is the same as (140, 142). 
>> 
>> Hm yes, indeed -- even if I use a color picker to see what is the color
>> on the screen (when the EMF+ file is opened in Draw) I get 00FF00. On
>> what system do you see the failure? I'm asking as I see none of the
>> tinderboxes at http://tinderbox.libreoffice.org/MASTER/status.html
>> failing due to this test (including my local Linux build) -- i.e. if you
>> would change it as-is, you would probably break all of them. :-)
>> 
>> So it would be good to find out why 00fe00 works in the majority of the
>> cases.
>> 
>> Regards,
>> 
>> Miklos
>