OSX: sequential/test-http-writable-true-after-close

[mhdawson]

• Version: master
• Platform: OSX
• Subsystem: http

https://ci.nodejs.org/job/node-test-commit-osx/13264/nodes=osx1010/console

not ok 1960 sequential/test-http-writable-true-after-close
  ---
  duration_ms: 0.138
  severity: fail
  stack: |-
    Mismatched <anonymous> function calls. Expected exactly 1, actual 0.
        at Object.exports.mustCall (/Users/iojs/build/workspace/node-test-commit-osx/nodes/osx1010/test/common/index.js:488:10)
        at ClientRequest.common.mustCall (/Users/iojs/build/workspace/node-test-commit-osx/nodes/osx1010/test/sequential/test-http-writable-true-after-close.js:16:29)
        at ClientRequest.<anonymous> (/Users/iojs/build/workspace/node-test-commit-osx/nodes/osx1010/test/common/index.js:520:15)
        at Object.onceWrapper (events.js:333:30)
        at emitOne (events.js:126:13)
        at ClientRequest.emit (events.js:224:7)
        at HTTPParser.parserOnIncomingClient [as onIncoming] (_http_client.js:550:21)
        at HTTPParser.parserOnHeadersComplete (_http_common.js:117:23)
        at Socket.socketOnData (_http_client.js:439:20)

Failed in CI run for test change that was not related.

[Ethan-Arrowood]

What is the scope of this issue? With a little more explanation and guidance I'd love to try to resolve it.

[apapirovski]

@Ethan-Arrowood I'm not certain anyone truly knows since that test seems to exhibit pretty highly divergent behaviour between different systems. I'll ping @mcollina though who worked on it last and maybe he can offer some guidance.

[mcollina]

That test has been highly flaky but it tests some needed things. I've worked extensively on it without finding a good solution to make it less flaky. IMHO it might hide some sporadic bug beneath.