Yet another segmentation fault in node-v4.0.0-rc.1

[pmq20]

var cp = require('child_process');
cp.spawn('/bin/ls', [], {
  stdio: [
    process.stdin,
    process.stdout,
    process.stderr,
  ]
});

[silverwind]

Confirmed, I'm almost certain this is the one seen in #2552

(lldb) target create "node"
Current executable set to 'node' (x86_64).
(lldb) settings set -- target.run-args  "seg"
(lldb) r
Process 37974 launched: '/usr/local/bin/node' (x86_64)
Process 37974 stopped
* thread #1: tid = 0x1a1f80, 0x000000010015e1ff node`v8::Value::IntegerValue(v8::Local<v8::Context>) const + 415, queue = 'com.apple.main-thread', stop reason = EXC_BAD_ACCESS (code=EXC_I386_GPFLT)
    frame #0: 0x000000010015e1ff node`v8::Value::IntegerValue(v8::Local<v8::Context>) const + 415
node`v8::Value::IntegerValue:
->  0x10015e1ff <+415>: cvttsd2si 0x7(%rdx), %rdx
    0x10015e205 <+421>: jmp    0x10015e20b               ; <+427>
    0x10015e207 <+423>: sarq   $0x20, %rdx
    0x10015e20b <+427>: movb   $0x1, %al
(lldb) bt
* thread #1: tid = 0x1a1f80, 0x000000010015e1ff node`v8::Value::IntegerValue(v8::Local<v8::Context>) const + 415, queue = 'com.apple.main-thread', stop reason = EXC_BAD_ACCESS (code=EXC_I386_GPFLT)
  * frame #0: 0x000000010015e1ff node`v8::Value::IntegerValue(v8::Local<v8::Context>) const + 415
    frame #1: 0x000000010015e2c9 node`v8::Value::IntegerValue() const + 169
    frame #2: 0x00000001006b2af7 node`node::ProcessWrap::ParseStdioOptions(node::Environment*, v8::Local<v8::Object>, uv_process_options_s*) + 483
    frame #3: 0x00000001006b22e9 node`node::ProcessWrap::Spawn(v8::FunctionCallbackInfo<v8::Value> const&) + 1109
    frame #4: 0x000000010017fcfe node`v8::internal::FunctionCallbackArguments::Call(void (*)(v8::FunctionCallbackInfo<v8::Value> const&)) + 158
    frame #5: 0x00000001001a9735 node`v8::internal::MaybeHandle<v8::internal::Object> v8::internal::HandleApiCallHelper<false>(v8::internal::Isolate*, v8::internal::(anonymous namespace)::BuiltinArguments<(v8::internal::BuiltinExtraArguments)1>&) + 1125
    frame #6: 0x00000001001ac5bd node`v8::internal::Builtin_HandleApiCall(int, v8::internal::Object**, v8::internal::Isolate*) + 61
    frame #7: 0x00001b54705060bb
    frame #8: 0x00001b54705db6f1
    frame #9: 0x00001b54705d998d
    frame #10: 0x00001b5470519494
    frame #11: 0x00001b54705bae73
    frame #12: 0x00001b547051f5c7
    frame #13: 0x00001b54705ba497
    frame #14: 0x00001b54705b4325
    frame #15: 0x00001b54705b0dd2
    frame #16: 0x00001b54705a6bd0
    frame #17: 0x00001b54705a6303
    frame #18: 0x00001b5470577965
    frame #19: 0x00001b5470575e12
    frame #20: 0x00001b5470519f7d
    frame #21: 0x00001b54705189e2
    frame #22: 0x00000001002ecaec node`v8::internal::Invoke(bool, v8::internal::Handle<v8::internal::JSFunction>, v8::internal::Handle<v8::internal::Object>, int, v8::internal::Handle<v8::internal::Object>*) + 732
    frame #23: 0x0000000100166724 node`v8::Function::Call(v8::Local<v8::Context>, v8::Local<v8::Value>, int, v8::Local<v8::Value>*) + 276
    frame #24: 0x0000000100685954 node`node::LoadEnvironment(node::Environment*) + 478
    frame #25: 0x0000000100686cfc node`node::Start(int, char**) + 474
    frame #26: 0x0000000100001034 node`start + 52

[Fishrock123]

@silverwind could you get the JavaScript stack out of that too? You can use https://github.com/thlorenz/lldb-jbt for that.

[silverwind]

@Fishrock123 hmm, can't get that to work:

$ source jbt
Now debug your node process with lldb and ensure to use the '--perf-basic-prof' flag.
In lldb run 'command script import jbt' to initialize the command.
$ lldb -- node --perf-basic-prof seg
(lldb) target create "node"
Current executable set to 'node' (x86_64).
(lldb) settings set -- target.run-args  "--perf-basic-prof" "seg"
(lldb) command script import jbt
error: module importing failed: No module named jbt
  File "temp.py", line 1, in <module>

Anyways, for all I know, #2722 fixes the issue.

[Fishrock123]

@silverwind you installed it globally, right? Maybe try watching the screencast?
cc @thlorenz

[silverwind]

Yeah, of course. I'm doing it like in the screencast. Maybe try for yourself, you should see the segfault with current master or the 4.0 RCs.

[evanlucas]

Ah this is actually a bug in child_process. PR incoming

[brodycj]

Please make sure you include a test. I issued #2725 with the test that was included in PR #2722, with the original author preserved in the commit.

[silverwind]

Fixed in evanlucas@3bc7e58