Validating tests for Image Processors

[antonfirsov]

Prerequisites

• I have written a descriptive pull-request title
• I have verified that there are no overlapping pull-requests open
• I have verified that I am following matches the existing coding patterns and practice as demonstrated in the repository. These follow strict Stylecop rules 👮.
• I have provided test coverage for my change (where applicable)

Description

Extending covarege to support the implementation of #897. Additionally, FileTestBase has been marked as [Obsolote] to support goals defined in #868.

[codecov]

Codecov Report

Merging #902 into master will decrease coverage by 0.1%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #902      +/-   ##
==========================================
- Coverage   89.33%   89.23%   -0.11%     
==========================================
  Files        1029     1030       +1     
  Lines       45708    45681      -27     
  Branches     3260     3260              
==========================================
- Hits        40833    40763      -70     
- Misses       4174     4178       +4     
- Partials      701      740      +39

Impacted Files	Coverage Δ
tests/ImageSharp.Tests/FileTestBase.cs	100% <ø> (ø)	⬆️
src/ImageSharp/Common/Helpers/Vector4Utils.cs	100% <ø> (ø)	⬆️
...cessing/Processors/Convolution/GaussianBlurTest.cs	100% <100%> (ø)	⬆️
.../Processors/Convolution/EdgeDetector2DProcessor.cs	100% <100%> (ø)	⬆️
...sing/Processors/Convolution/GaussianSharpenTest.cs	100% <100%> (ø)	⬆️
...ors/Convolution/Basic1ParameterConvolutionTests.cs	100% <100%> (ø)
...ocessing/Processors/Convolution/DetectEdgesTest.cs	100% <100%> (ø)	⬆️
...s/Processing/Processors/Convolution/BoxBlurTest.cs	100% <100%> (ø)	⬆️
...xelFormats/PixelOperations{TPixel}.PixelBenders.cs	89.9% <0%> (-10.1%)	⬇️
...Processing/Processors/Transforms/TransformUtils.cs	87.64% <0%> (-5.62%)	⬇️
... and 22 more

---

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 b0f6b73...c1d595e. Read the comment docs.

[codecov]

Codecov Report

Merging #902 into master will decrease coverage by 0.05%.
The diff coverage is 92.85%.

@@            Coverage Diff             @@
##           master     #902      +/-   ##
==========================================
- Coverage   89.33%   89.28%   -0.06%     
==========================================
  Files        1029     1031       +2     
  Lines       45708    45631      -77     
  Branches     3260     3261       +1     
==========================================
- Hits        40833    40741      -92     
- Misses       4174     4180       +6     
- Partials      701      710       +9

Impacted Files	Coverage Δ
tests/ImageSharp.Tests/FileTestBase.cs	100% <ø> (ø)	⬆️
...Tests/Processing/Processors/Transforms/CropTest.cs	100% <ø> (ø)	⬆️
src/ImageSharp/Common/Helpers/Vector4Utils.cs	100% <ø> (ø)	⬆️
...cessing/Processors/Convolution/GaussianBlurTest.cs	100% <100%> (ø)	⬆️
...ests/Processing/Processors/Effects/PixelateTest.cs	100% <100%> (ø)	⬆️
...sts/Processing/Processors/Overlays/VignetteTest.cs	100% <100%> (ø)	⬆️
.../Processors/Convolution/EdgeDetector2DProcessor.cs	100% <100%> (ø)	⬆️
...ocessing/Processors/Convolution/DetectEdgesTest.cs	100% <100%> (ø)	⬆️
...ests/Processing/Processors/Effects/OilPaintTest.cs	100% <100%> (ø)	⬆️
...rocessing/Processors/Transforms/AutoOrientTests.cs	100% <100%> (ø)	⬆️
... and 15 more

---

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 b0f6b73...4f41288. Read the comment docs.

[JimBobSquarePants]

What’s happening with the dither tests on 32bit?

Edit: I've seen the tests. A 5% error is massive. I think we'd be better tryting to figure out why it's failing rather than skip the tests.

The equality operator here is probably the issue.

ImageSharp/src/ImageSharp/Processing/Processors/Dithering/ErrorDiffuserBase.cs

Line 110 in b0f6b73

if (coefficient == 0)

[antonfirsov]

Expected:

32 bit (+old buggy runtime):

The dots at distributed at different positions, but otherwise the output looks good. I'd rather blame the vector expression for this behavior:

Vector4 result = ((error * coefficient) / this.divisorVector) + offsetColor;

[JimBobSquarePants]

Ah yeah, that’s a likely culprit also. Let’s get this merged in then. It’s great to have the additional tests!