You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
With npx playwright test, if a snapshot fails and traces are being saved, the trace will have nice attachments for debugging - actual.png, expected.png, diff.png.
But with npx playwright test --update-snapshots, these attachments are not saved. It would be nice to have these when snapshots are being updated, so that I can still see the pixel-precise differences. Potentially saving these in this case could be based on a config or CLI option.
I'm using a workflow where I assume that snapshot diffs are intentional, so that when there are diffs I don't have to run the tests a second time just to fix them. I'm also running this in CI, and automatically creating a commit with the diffs.
Most of the time, I expect this to be good because the diffs are intentional, and usually using github/vscode git diff to compare before/after is good enough, but sometimes there have been small pixel diffs that are hard to tell without the playwright-generated diff file.
Since I'm using version control, I can always run the tests locally without --update-snapshots in order to see the diffs (or maybe manually use the same algorithm that playwright uses to create a diff file), but that seems like more effort than is necessary.
The text was updated successfully, but these errors were encountered:
🚀 Feature Request
With
npx playwright test
, if a snapshot fails and traces are being saved, the trace will have nice attachments for debugging - actual.png, expected.png, diff.png.But with
npx playwright test --update-snapshots
, these attachments are not saved. It would be nice to have these when snapshots are being updated, so that I can still see the pixel-precise differences. Potentially saving these in this case could be based on a config or CLI option.Initial thread on Discord
Example
No response
Motivation
I'm using a workflow where I assume that snapshot diffs are intentional, so that when there are diffs I don't have to run the tests a second time just to fix them. I'm also running this in CI, and automatically creating a commit with the diffs.
Most of the time, I expect this to be good because the diffs are intentional, and usually using github/vscode git diff to compare before/after is good enough, but sometimes there have been small pixel diffs that are hard to tell without the playwright-generated diff file.
Since I'm using version control, I can always run the tests locally without
--update-snapshots
in order to see the diffs (or maybe manually use the same algorithm that playwright uses to create a diff file), but that seems like more effort than is necessary.The text was updated successfully, but these errors were encountered: