<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Thu, Mar 26, 2015 at 10:36 PM, Chris Wilson <span dir="ltr"><<a href="mailto:chris@chris-wilson.co.uk" target="_blank">chris@chris-wilson.co.uk</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">On Thu, Mar 26, 2015 at 05:55:37PM +0100, Andrea Canciani wrote:<br>
> Commit 4e3ef57bc892b0b046c486390adc7164a1de64de added<br>
> coverage-intersecting-triangles with an incorrect reference and<br>
> generator. The test checks the rasterization of two overlapping<br>
> triangles in the following position:<br>
><br>
> . .<br>
> |\ /|<br>
> | X |<br>
> |/ \|<br>
> .---.<br>
><br>
> Since the triangles have both vertical and horizontal sides of size<br>
> x/WIDTH, the expected coverage is 3/4 (75%) of (x/WIDTH)^2. The<br>
> original code, instead, was checking for a coverage of 0.75*x/WIDTH,<br>
> as if one of the sides was always 1 unit long.<br>
><br>
> The image and xlib backends still suffer from some jitter, caused by<br>
> the approximation of the actual coverage by means of sampling, but<br>
> their result now looks consistent with the expected reference.<br>
> ---<br>
> test/coverage.c | 2 +-<br>
> .../coverage-intersecting-triangles.image.ref.png | Bin 0 -> 14444 bytes<br>
> test/reference/coverage-intersecting-triangles.ref.png | Bin 209 -> 262 bytes<br>
> .../reference/coverage-intersecting-triangles.xfail.png | Bin 14444 -> 0 bytes<br>
> .../coverage-intersecting-triangles.xlib.ref.png | Bin 0 -> 16728 bytes<br>
> .../coverage-intersecting-triangles.xlib.xfail.png | Bin 16728 -> 0 bytes<br>
> 6 files changed, 1 insertion(+), 1 deletion(-)<br>
> create mode 100644 test/reference/coverage-intersecting-triangles.image.ref.png<br>
> delete mode 100644 test/reference/coverage-intersecting-triangles.xfail.png<br>
> create mode 100644 test/reference/coverage-intersecting-triangles.xlib.ref.png<br>
> delete mode 100644 test/reference/coverage-intersecting-triangles.xlib.xfail.png<br>
<br>
</div></div>Apart from they are still xfail since they are necessarily imperfect<br>
approximations (xlib doubly so if backed by pixman due to the<br>
intersections).<br></blockquote><div><br></div><div>I just sent in a new patch which keeps xfail references for image and xlib and updates the generic reference image.<br><br></div><div>Andrea<br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<span class="HOEnZb"><font color="#888888">-Chris<br>
<br>
--<br>
Chris Wilson, Intel Open Source Technology Centre<br>
</font></span></blockquote></div><br></div></div>