Closed henryiii closed 7 years ago
Merging #230 into master will not change coverage. The diff coverage is
n/a
.
@@ Coverage Diff @@
## master #230 +/- ##
=======================================
Coverage 94.64% 94.64%
=======================================
Files 119 119
Lines 14271 14271
Branches 938 938
=======================================
Hits 13507 13507
Misses 703 703
Partials 61 61
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update ef71797...d628fe4. Read the comment docs.
As expected, the compute backend fails to build on macOS. Rebasing on master to see if that fixes that. There are lots of failing tests on macOS at the moment on OpenCL, as well. I can (temporary, hopefully) disable the failing of a macOS test from indicating an overall failure if you'd like.
The boost.compute (partial) fix in boostorg/compute#737 is not in 1.65.1, so the fail is to be expected. And OpenCL failing tests look like OpenCL compiler bugs.
But this PR looks great anyway! Yes, let's mark the MacOS tests as free to fail for now.
Some OpenCL fails seem to reproduce #92.
Thanks!
Starting a solution to #228