fix: preserve utf8 decoding across response chunks

[pupuking723]

This relates to...

Fixes #5002

Rationale

BodyReadable#setEncoding() only stored the encoding on the readable state. That made streamed consumers decode each buffered chunk independently, so an incomplete UTF-8 sequence at a response chunk boundary produced replacement characters.

The internal body consumers still need raw Buffer chunks so body.text() and body.json() can aggregate and decode the full payload.

Changes

Features

N/A

Bug Fixes

• Keep raw Buffer chunks for internal body consumption.
• Use a StringDecoder when BodyReadable is consumed through the Readable API after setEncoding().
• Add a regression test for async iteration over a response body where a 3-byte UTF-8 character spans response chunks.

Breaking Changes and Deprecations

N/A

Status

• I have read and agreed to the Developer's Certificate of Origin
• Tested
• Benchmarked (optional)
• Documented
• Review ready
• In review
• Merge ready

Verification:

• node --test --test-name-pattern "request multibyte (json|text) with setEncoding|async iteration and setEncoding" test/client-request.js
• npx borp --timeout 180000 -p test/client-request.js
• npm run lint
• npm run test:typescript
• git diff --check

Note: npm ci --ignore-scripts was used locally because npm install exited with "Exit handler never called!" after dependency extraction. The test/lint commands above were run against the installed dependency tree. Local Node.js is v22.17.0 while the package currently requires >=22.19.0.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.47%. Comparing base (0f1f890) to head (83fbbf6).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #5414   +/-   ##
=======================================
  Coverage   93.46%   93.47%           
=======================================
  Files         110      110           
  Lines       37106    37106           
=======================================
+ Hits        34682    34684    +2     
+ Misses       2424     2422    -2     

☔ View full report in Codecov by Harness.
📢 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.

[mcollina]

I'm relatively certain that Readable implements this logic. I think the problem is in the override of setEncoding, we should also call super.setEncoding

[metcoder95]

also PR has conflicts

[pupuking723]

Thanks for the review. After rebasing, BodyReadable#setEncoding() now delegates to super.setEncoding() in the current implementation, so this PR only keeps the focused regression test covering async iteration after body.setEncoding("utf8") across multibyte chunk boundaries.

The current GitHub checks are passing across the CI matrix. Could you please take another look when you have a chance?