Skip to content

Confirm e2e failures from backend source-formatting deploy - #10657

Merged
markerikson merged 11 commits into
mainfrom
bugfix/e2e-test-source-formatting
Aug 23, 2026
Merged

Confirm e2e failures from backend source-formatting deploy#10657
markerikson merged 11 commits into
mainfrom
bugfix/e2e-test-source-formatting

Conversation

@markerikson

Copy link
Copy Markdown
Collaborator

Draft PR with an empty commit, purely to run e2e CI.

  • The Aug 20 backend controllers deploy shipped the source-worker formatting rewrite (backend #13141): isMinified long-line threshold went 500 → 200 chars, so regeneratorRuntime.js (and a few other CRA sources) are now classified minified and get pretty-printed.
  • Expected result on this PR, with zero devtools changes:
    • sourcemap_stacktrace fails: frames report pretty-printed lines regeneratorRuntime.js:136 / :72 instead of the raw :125 / :69 the test expects.
    • jump-to-code-01_basic fails at the first check: "Jump to code" where the test expects "No results".
  • Both failed identically on four backend-pin bisect runs (e2e #2547–Include webpack stats in deploy #2550), including one pinned to the runtime config that passed on Aug 19 — so the cause is the prod deploy, not runtime changes.

If CI here shows the same two failures, the fix is updating these test expectations (follow-up commits on this branch).

Confirming that the two deterministic e2e failures (sourcemap_stacktrace,
jump-to-code-01_basic) reproduce on any new PR after the Aug 20 backend
controllers deploy, independent of devtools changes.
@vercel

vercel Bot commented Aug 21, 2026

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
devtools Ready Ready Preview Aug 23, 2026 9:14pm

Request Review

@markerikson

Copy link
Copy Markdown
Collaborator Author

Confirmed. This PR is an empty commit on top of main (2c49b9a), and e2e fails, while the Nightly runs on that same SHA were green on Aug 19 and Aug 20 08:12 UTC — before the backend controllers deploy at 09:42 UTC.

Failed (all retries):

  • jump-to-code-01_basic — "Jump to code" where the test expects "No results" (same symptom as the Buildkite runtime e2e runs)
  • regression-fe-1875-cypress-jump-to-code
  • stepping-05 — stepping in pretty-printed code
  • stepping-06 — async frames / async call stacks
  • inspector-computed-02
  • playwright-05 — DOM node previews on step hover

Note sourcemap_stacktrace passes here but fails deterministically in the Buildkite runtime e2e suite. Likely difference: pretty-print artifacts are cached in S3 per recording, so tests running against a stored example recording still see pre-deploy formatting, while freshly recorded examples get the new isMinified classification (long-line threshold 500 → 200 chars, backend #13141), which pretty-prints regeneratorRuntime.js and friends and shifts frame lines.

Next step: update the affected test expectations (or revisit the threshold) — follow-ups on this branch.

@replay-io

replay-io Bot commented Aug 22, 2026

Copy link
Copy Markdown

E2E Tests

Status Starting
Commit 626d9ae

@markerikson
markerikson marked this pull request as ready for review August 23, 2026 21:21
@markerikson
markerikson merged commit cd64c09 into main Aug 23, 2026
23 checks passed
@markerikson
markerikson deleted the bugfix/e2e-test-source-formatting branch August 23, 2026 21:25
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants