Revert "stream: defer readable and flow when sync"

[addaleax]

This reverts commit 563fff2 as it is causing failures in CITGM with dicer.

Refs: #18515

The test case that was also fixed by #18516 is being kept.

---

@mafintosh Could you take a look? If you can fix the failures without needing a revert, that would be amazing. But usually we are pretty fast to revert + open a new PR in these cases…

Example faillure:

~/src/dicer $ node test/test.js
assert.js:74
  throw new AssertionError(obj);
  ^

AssertionError [ERR_ASSERTION]: [test-multipart/Many parts]: Part #1 body mismatch
    at Dicer.<anonymous> (/home/xxxx/src/dicer/test/test-multipart.js:206:14)
    at Dicer.emit (events.js:136:15)
    at Dicer.emit (/home/xxxx/src/dicer/lib/Dicer.js:80:35)
    at PartStream.<anonymous> (/home/xxxx/src/dicer/lib/Dicer.js:212:18)
    at PartStream.emit (events.js:136:15)
    at endReadableNT (_stream_readable.js:1020:12)
    at process._tickCallback (internal/process/next_tick.js:91:19)

[addaleax]

@nodejs/streams

[mcollina]

I’ll have a look today as well.

[mcollina]

I cannot reproduce the failure locally on Mac OS X. Running CITGM now: https://ci.nodejs.org/view/Node.js-citgm/job/citgm-smoker/1261/.

[mcollina]

Seems a Linux issue (RHEL output):

 AssertionError [ERR_ASSERTION]: [test-multipart/Many parts]: Part #1 body mismatch
     at Dicer.<anonymous> (/data/iojs/build/workspace/citgm-smoker/nodes/rhel72-s390x/citgm_tmp/c72cac48-49cf-45a9-b12e-bd76883111d3/dicer/test/test-multipart.js:206:14)
     at Dicer.emit (events.js:136:15)
     at Dicer.emit (/data/iojs/build/workspace/citgm-smoker/nodes/rhel72-s390x/citgm_tmp/c72cac48-49cf-45a9-b12e-bd76883111d3/dicer/lib/Dicer.js:80:35)
     at PartStream.<anonymous> (/data/iojs/build/workspace/citgm-smoker/nodes/rhel72-s390x/citgm_tmp/c72cac48-49cf-45a9-b12e-bd76883111d3/dicer/lib/Dicer.js:212:18)
     at PartStream.emit (events.js:136:15)
     at endReadableNT (_stream_readable.js:1020:12)
     at process._tickCallback (internal/process/next_tick.js:91:19)
 npm ERR! Test failed.  See above for more details.

[addaleax]

That’s weird, dicer really doesn’t do anything platform-dependent…

[mafintosh]

@addaleax see #18615 where @mcollina fixed the issue

[mafintosh]

@addaleax the dicer tests are highly timing dependent it seems after looking at them

[addaleax]

@mafintosh Yeah, I’ll close this given that a fix is available 👍

the dicer tests are highly timing dependent it seems after looking at them

So is a lot of existing userland code, unfortunately…