Closed vlads11 closed 2 years ago
Hi, the @2
is the DPI from the browser that took the screenshots, if no @2
is present that means the DPI on that browser (may be headless?) is just 1, so we're not adding the suffix since it's the default.
Why and for what purpose? That value helps us to determine the DPI from images to be compared, because different DPI yields different image resolutions, so we need to know the DPI from both images to scale them properly and then run the diff.
TY @pateketrueke for the explanation on the @2. Learned something new =)
I can see everything working locally now which is awesome but running the our test suite in a docker container via Jenkins still only captures the base.png.
`node ./zwift-blinkdiff.js -e production --browser firefox:headless -s ./Tests-Visual/screenshots/vlad --testMeta.robinTest --take-snapshot actual Executing TestCafe runner. Credentials - Environment data found. Running tests in:
Firefox 106.0 / Alpine Linux 3.17_alpha20220715
Zwift Hub Snapshots ✓ Home page (screenshots: /home/jenkins/agent/workspace/TestCafe/Blinkdiff-Visual-Regression-Actual/Zwift-TestCafe/Tests-Visual/screenshots/vlad/Zwift_Hub_Snapshots/Full_Home_Page/base.png)`
Any thoughts or ideas on why running in the cloud/docker would be a problem?
I am curious about your ./zwift-blinkdiff.js
script, what it does exactly?
Please, format your code properly or upload it to a gist instead, it's bit hard to read and follow here, thank you!
Thanks for the help here: https://gist.github.com/vlads11/4eae63c855a12edefc61bd9493be9520
The custom flags (e.g. --take-screenshot
) from this package CLI are not supported by Testcafé itself, in fact, they're hacked and extracted when the package is used on the test files: https://github.com/tacoss/testcafe-blink-diff/blob/master/lib/index.js#L10-L12
If you don't propagate those arguments through the process.argv
object, you may try using the process.env
variables instead, that way you can tell this module how to behave properly.
Thanks for this, will give it a try and see how far I get.
Thanks @pateketrueke - The issue was ultimately with process.env TAKE_SNAPSHOT variable that we were over writing in our Jenkins Job Defintiion. 100% appreciate the assist though, the lines of code you shared led me to the right solution and I learned much in the process.
Hi BlinkDiff team,
Reaching out for some help and guidance on 2 issues we're experiencing when trying to integrate blinkdiff into our automation repo.
1) Zwift Hub Snapshots ✓ Home page (screenshots: /Tests-Visual/screenshots/vlad/Zwift_Hub_Snapshots/Full_Home_Page/actual@2.png) -- When actual and base images are saved they @2 is appended to the image name and I am not sure why. 2) Issues running blinkdiff test via Jenkins in a docker container, same call in jenkins results in base.png being saved (no @2 appended).
Pretty sure the 2 issues are related but not sure what could be causing the inconsistency with the file-saving system. Any help is much appreciated.