fix(testing/unstable): use the correct function frame in inline snapshot

[mochaaP]

Before the fix, it will always write to unstable_snapshot.ts:createAssertInlineSnapshot.

[CLAassistant]

All committers have signed the CLA.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.61%. Comparing base (95a1e2e) to head (4f952ad).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #7150   +/-   ##
=======================================
  Coverage   94.61%   94.61%           
=======================================
  Files         634      634           
  Lines       51843    51846    +3     
  Branches     9346     9347    +1     
=======================================
+ Hits        49050    49053    +3     
  Misses       2218     2218           
  Partials      575      575           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[bartlomieju]

Nice catch — small, well-targeted fix for a real bug in createAssertInlineSnapshot.

Before this PR, Error.captureStackTrace(stackCatcher, assertInlineSnapshot) only skipped the inner assertInlineSnapshot frame, so the wrapper returned by createAssertInlineSnapshot was reported as the caller and the lint plugin tried to update the snapshot literal inside unstable_snapshot.ts itself. The named function expression + options.frame indirection is the right idiom, and as a bonus the merge order makes nested wrappers compose (the outer wrapper's frame rides down through messageOrOptions).

A few inline nits below, plus one comment on an unrelated whitespace change in the test file. LGTM once those are addressed.

[bartlomieju]

Small grammar nit: will be the decided → will be decided. Consider rewording slightly for clarity, e.g.:

The source code location to modify is taken from the stack frame that comes after this function.

[bartlomieju]

Worth a one-line comment noting that callers (specifically createAssertInlineSnapshot) pass their own wrapper function as frame so V8 skips it during stack capture. Otherwise the ?? here looks like a generic option-passthrough and the relationship to the bug fix isn't obvious to a future reader.

[bartlomieju]

The ordering { frame, ...options, ...messageOrOptions } is load-bearing and worth a comment: it makes the wrapper's own frame the default while still allowing an outer wrapper to override frame via the forwarded options — that's what enables nested createAssertInlineSnapshot wrappers to compose correctly. Without a note, a well-meaning refactor could reorder these spreads and silently break nesting.

[bartlomieju]

Optional: a one-liner in this JSDoc mentioning that the wrapper preserves the caller's source location for --update would help discoverability of the fix.

[bartlomieju]

This trailing-whitespace change (and the mirrored one on line 177) is unrelated to the frame fix. Since the literal string here is the actual input to the "counts lines and columns like V8" test, can you confirm removing the trailing space doesn't alter what's being exercised (e.g. the column count after \u2029)? If it's a formatter-required change, fine — but worth calling out in the PR body, or splitting into its own commit.

[bartlomieju]

Good regression test — exactly the right shape. Without the fix this would fail because the wrapper would try to update the literal at unstable_snapshot.ts:createAssertInlineSnapshot's line rather than the user's test file.

[mochaaP]

all reviews resolved.