feat(options): add option to pass on size missmatch #174
Conversation
|
|
|
I may be missing something but couldnt you make sure to crop the images prior to passing them in to jest-image-snapshot? Declining as I dont think this is necessary. |
|
@anescobar1991 that's not really possible. We crop our screenshots to the actual content before passing them in by removing any whitespace around our components we screenshot. Therefore we do not know the size of the screenshot beforehand and hardcode the size. Because shadows are rendered differently per browser, system and so on, the shadow might be 1 or 2px larger/smaller between systems and all the tests will fail, even though the diff is below the defined threshold (even when all missing/additional pixels are counted as a mismatch). |
|
@adrianjost I see now - I think this does make sense to add then. Can you resolve the conflicts in this PR? |
|
perfect. Yes I will resolve the conflicts later today :) |
|
Seconding that this would be a very helpful feature! 🙏 |
|
Sorry, I did something wrong during the rebase and now the tests are failing. I will fix this later this weekend. |
|
@anescobar1991 All conflicts are now resolved |
|
@anescobar1991 update the branch from the americanexpress repo and removed the |
|
🎉 This PR is included in version 3.1.0 🎉 The release is available on: Your semantic-release bot 📦🚀 |
) if the option `allowSizeMismatch` is set, a build will not always fail on images with different sizes. Missing or Added Pixel will be counted as a mismatch and respected by the set threshold. Related americanexpress#83, americanexpress#85
# [3.1.0](americanexpress/jest-image-snapshot@v3.0.1...v3.1.0) (2020-04-17) ### Features * **options:** add option to pass on size missmatch ([americanexpress#174](americanexpress#174)) ([cee46b1](americanexpress@cee46b1)), closes [americanexpress#83](americanexpress#83) [americanexpress#85](americanexpress#85)
if the option
allowSizeMismatchis set, a build will not allways fail onimages with different sizes. Missing or Added Pixel will be counted as a
missmatch and respected by the set threshold.
Related #83, #85
I am not sure if this addition is welcome because of the discussion in the issues mentioned but we needed to implement it anyway because we want our test to run on different operating systems which results in different image sizes if we crop images using 3rd party libraries. The reason for this is different shadow and text rendering of the operating systems and browsers.