Closed mdesmedt closed 3 years ago
Merging #64 (c2815c6) into master (05fb0a0) will increase coverage by
0.53%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## master #64 +/- ##
==========================================
+ Coverage 79.92% 80.45% +0.53%
==========================================
Files 9 9
Lines 1529 1566 +37
==========================================
+ Hits 1222 1260 +38
+ Misses 307 306 -1
Impacted Files | Coverage Δ | |
---|---|---|
src/aabb.rs | 93.29% <100.00%> (+0.21%) |
:arrow_up: |
src/axis.rs | 43.58% <100.00%> (+1.48%) |
:arrow_up: |
src/ray.rs | 94.73% <100.00%> (+1.71%) |
:arrow_up: |
src/testbase.rs | 89.50% <100.00%> (+0.43%) |
:arrow_up: |
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 05fb0a0...c2815c6. Read the comment docs.
Switch the property testing tests to proptest. Surprisingly, a lot of fp32-related errors weren't caught by quickcheck. 8 tests initially failed:
Fixed the above tests by selectively:
float_eq
package and relative epsilon.tuplevec_large_strategy()
tuplevec_small_strategy()
. This was needed for a lot of Ray.rs tests.test_points_relative_to_center_and_size
needed to be tweaked to put the inside point inside by 0.9, and the outside point outside by 1.1. Makes the test a lot less useful, but fixing the related AABB functions needs some rethinking if they need to be robust against fp32 errors.