Path Rendering in 8u40

[Jim Graham]

Where are the attached files?  When you say you zoomed them in, is that 
with a screen pixel scaler, or by applying a scaling transform to the 
SVG path?

I'm guessing that you are correct about RT-39439, but the fix there 
should have made the paths more accurate?  Could the kinks have been in 
the path, but not rendered correctly until we upgraded our BND constants 
as part of RT-39439?

Can you make a standalone sample with just the path segments in question?

			...jim

On 1/27/15 5:28 PM, Scott Palmer wrote:
> AH! Thankfully that was it.. I had never intended for the paths to no be
> smooth, but I suspect I was experimenting trying to get better performance
> at some point and the call to setSmooth(false) was left in.
>
> However, the issue with the SVG paths from CSS remains.  You can see it
> more clearly in the attached image files, where I've zoomed in.
>
>
> Cheers,
>
> Scott
>
> On Tue, Jan 27, 2015 at 5:18 PM, Kevin Rushforth <kevin.rushforth at oracle.com
>> wrote:
>
>> Do you explicitly set smooth=false on your path? If so, that would be due
>> to:
>>
>> https://javafx-jira.kenai.com/browse/RT-39468
>>
>> We used to ignore the smooth flag (it's a rasterization hint) but now
>> honor it and disable AA if set to false.
>>
>> -- Kevin
>>
>>
>>
>> Scott Palmer wrote:
>>
>>> I'm seeing a regression where Paths are not rendered with anti-aliasing on
>>> 8u40.  On 8u20 they are fine.  There are also glitches in SVG Paths from
>>> CSS.  The SVG paths are smooth, but the end points are different and they
>>> have 'kinks'.  There is also a difference in the position of my
>>> "arrowhead"
>>> on the end of the Path which is done using SVG in the CSS for a Region.
>>>
>>> I've attached images of the exact same code running with 8u40 and 8u20.
>>>   (not sure if they will make it to the mailing list)
>>>
>>> I searched JIRA but didn't see any issues that seemed relevant, though I
>>> wonder if work done for RT-39439 may be related.
>>> Has anyone else seen something like this?
>>>
>>> Scott
>>>
>>>
>>