The marker-multi-policy visual test keeps failing for me, reporing 1700 different pixels and showing a blank (fully transparent) image as the obtained image. But when I run the style with nik2img.py I do see the expected image instead.
Any idea about what could be going wrong ?
Maybe a missing input plugin during "make test" ?
This is 2.1.x branch.
Found: it's a real failure, due to the default rendering bounding box being used, and containing NOthing to render.
I got confused by #1611 by thinking it worked. TravisCI also isn't useful due to #1611 ...
What's the rationale for such a small default bbox, btw ?
Just wondering... mapnik.Box2d(-0.05, -0.01, 0.95, 0.01) ?
It's not even a square, but a 1.0 x 0.2 box ?
mapnik.Box2d(-0.05, -0.01, 0.95, 0.01)
1.0 x 0.2
Fix marker-multi-policy test. Closes #1612
Pull #1612 is for 2.1.x branch, working on a pull for master too...
And #1615 is for master, closing this as it's superceeded by the other.
fix marker-multi-policy test - closes #1613 and closes #1612