Closed mateuszbaran closed 1 year ago
Merging #133 (a299890) into master (44c6c21) will decrease coverage by
0.08%
. The diff coverage is98.00%
.
@@ Coverage Diff @@
## master #133 +/- ##
==========================================
- Coverage 99.82% 99.73% -0.09%
==========================================
Files 18 18
Lines 2247 2297 +50
==========================================
+ Hits 2243 2291 +48
- Misses 4 6 +2
Impacted Files | Coverage Δ | |
---|---|---|
src/point_vector_fallbacks.jl | 99.24% <98.00%> (-0.76%) |
:arrow_down: |
:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more
The one new uncovered line looks like a false positive.
Interesting approach! I had thought we would just split Twhere into Point and TVector types and put them were necessary – but that would be breaking. Is this one a breaking change?
Besides that, allocate has an unused variable S in there (I removed it on my branch) but you could remove it here as well since it fits this PR very well.
Interesting approach! I had thought we would just split Twhere into Point and TVector types and put them were necessary – but that would be breaking. Is this one a breaking change?
The problem is we sometimes need the where
clause and sometimes we don't, and sadly I couldn't find anything that would interpolate to "this is empty" in type variable list after where
(which would simplify things significantly). Luckily this is not breaking.
Besides that, allocate has an unused variable S in there (I removed it on my branch) but you could remove it here as well since it fits this PR very well.
Sure, good idea. Weird that it didn't give me a warning.
Great, because mine would have been breaking.
Though I think in your case there might now be hidden unused variables, since Twhere still contains possibly a point and a vector type but there are functions that use only one of them. I am surprised those do not get recognised as warnings.
Resolves #132. Manifolds.jl has the same issue in a few places which have to be resolved in another PR.