fix: do not throw unhandled exception when data is undefined in interceptor.reply

[frederikprijck]

This relates to...

N/A

Rationale

See below (Bug Fixes)

Changes

See below (Bug Fixes)

Features

N/A

Bug Fixes

When interceptor.reply would return { data: undefined }, it blows up in getResponseData:

TypeError [Error]: Cannot read properties of undefined (reading 'toString')
      at getResponseData (/Users/frederikprijck/Development/frederikprijck/undici/lib/mock/mock-utils.js:128:17)
      at TestContext.<anonymous> (/Users/frederikprijck/Development/frederikprijck/undici/test/mock-utils.js:180:26)
      at Test.runInAsyncScope (node:async_hooks:206:9)
      at Test.run (node:internal/test_runner/test:639:25)
      at Suite.processPendingSubtests (node:internal/test_runner/test:382:18)
      at Test.postRun (node:internal/test_runner/test:730:19)
      at Test.run (node:internal/test_runner/test:688:12)
      at async Suite.processPendingSubtests (node:internal/test_runner/test:382:7)

This got introduced here, more specifcally, this change

More specifically, in the before and after, data is not the same when it was undefined originally:

const { statusCode, data = '', responseOptions = {} } = {data: undefined}
console.log(data); // ''

vs

const replyParameters = { data: '', responseOptions: {}, ...{data: undefined} }
console.log(replyParameters.data) // undefined

As previously, when data was undefined, '' would be passed to getResponseData, and getResponseData would return '', I ensured we now also return '' when data being passed to getResponseData is undefined.

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

[mcollina]

lgtm

[mcollina]

CI seems failing

[frederikprijck]

I am not sure the CI failure relates to my changes? Or at least, it isn't clear for me that it would. Any pointers?

[metcoder95]

Failures seems unrelated, are the same ones failing in other PRs (e.g. #4044)

[frederikprijck]

Hey @mcollina and @metcoder95, is there anything I can do to help progress this PR?

[mcollina]

It seems windows CI is utterly broken. I'm not keen on landing any changes until that's sorted, and I haven't had time yet.

[frederikprijck]

No worries at all. I entirely understand and I am not rushed in getting this out (we have worked around the behavior in our code-base, and don't need this change to unblock ourselves. I mainly just wanted to solve this for others that may encounter it). Just trying to follow up.

Take your time, I might drive by to follow up a few times, with no bad intentions to put any kind of pressure to land this, I entirly understand and appreciate you even willing to put the time to look into this eventually.