These are apparently the same. I can reduce comparison.threshold in vizzly.config.js to fix tdd but app.vizzly.dev still reports no difference here:
I have to reduce the threshold to 0.001 to catch this difference:
But then this one fails with 7 random pixels different (just run to run variance, I didn't change anything here):
Pixelmatch only needs a 0.035 threshold for the second one so is less likely to report false positives, and vitest can pass allowedMismatchedPixels and allowedMismatchedPixelRatio to tweak this.
Even with a 0.001 threshold vitest only needs allowedMismatchedPixels=2 to filter out run to run variance, it seems like honeydiff is less sensitive to large changes but more sensitive to tiny ones.
These are apparently the same. I can reduce
comparison.thresholdin vizzly.config.js to fix tdd but app.vizzly.dev still reports no difference here:I have to reduce the threshold to 0.001 to catch this difference:
But then this one fails with 7 random pixels different (just run to run variance, I didn't change anything here):
Pixelmatch only needs a 0.035 threshold for the second one so is less likely to report false positives, and vitest can pass allowedMismatchedPixels and allowedMismatchedPixelRatio to tweak this.
Even with a 0.001 threshold vitest only needs allowedMismatchedPixels=2 to filter out run to run variance, it seems like honeydiff is less sensitive to large changes but more sensitive to tiny ones.