Two test failures for v0.4.12 on OSX Lion

[telemachus]

The http-full-response test has been failing since v0.4.10 or so, but the second failure is new. Either way, I thought a report might be useful.

(The machine is OSX Lion 10.7.1. If any other information about the environment would be useful, let me know.)

python tools/test.py --mode=release simple message
=== release test-http-full-response ===                                    
Path: simple/test-http-full-response
node.js:134
        throw e; // process.nextTick error, or 'error' event on first tick
        ^
AssertionError: 0 == 3
    at EventEmitter.<anonymous> (/Users/telemachus/Downloads/node-v0.4.12/test/simple/test-http-full-response.js:92:10)
    at EventEmitter.emit (events.js:81:20)
    at EventEmitter.exit (node.js:199:15)
    at /Users/telemachus/Downloads/node-v0.4.12/test/simple/test-http-full-response.js:53:15
    at ChildProcess.exithandler (child_process.js:106:7)
    at ChildProcess.emit (events.js:67:17)
    at Socket.<anonymous> (child_process.js:172:12)
    at Socket.emit (events.js:64:17)
    at Array.<anonymous> (net.js:831:12)
    at EventEmitter._tickCallback (node.js:126:26)
Command: build/default/node /Users/telemachus/Downloads/node-v0.4.12/test/simple/test-http-full-response.js
=== release test-sendfd ===                                                
Path: simple/test-sendfd
DEBUG: CHILD: node.js:134
DEBUG: CHILD:         throw e; // process.nextTick error, or 'error' event on first tick
DEBUG: CHILD:         ^
DEBUG: CHILD: Error: ENOENT, No such file or directory
DEBUG: CHILD:     at doConnect (net.js:549:5)
DEBUG: CHILD:     at Socket.connect (net.js:709:5)
DEBUG: CHILD:     at Object.<anonymous> (/Users/telemachus/Downloads/node-v0.4.12/test/fixtures/recvfd.js:75:3)
DEBUG: CHILD:     at Module._compile (module.js:411:26)
DEBUG: CHILD:     at Object..js (module.js:417:10)
DEBUG: CHILD:     at Module.load (module.js:343:31)
DEBUG: CHILD:     at Function._load (module.js:302:12)
DEBUG: CHILD:     at Array.<anonymous> (module.js:430:10)
DEBUG: CHILD:     at EventEmitter._tickCallback (node.js:126:26)

node.js:134
        throw e; // process.nextTick error, or 'error' event on first tick
        ^
AssertionError: 0 == 1
    at ChildProcess.<anonymous> (/Users/telemachus/Downloads/node-v0.4.12/test/simple/test-sendfd.js:142:10)
    at ChildProcess.emit (events.js:67:17)
    at Socket.<anonymous> (child_process.js:172:12)
    at Socket.emit (events.js:64:17)
    at Array.<anonymous> (net.js:831:12)
    at EventEmitter._tickCallback (node.js:126:26)
Command: build/default/node /Users/telemachus/Downloads/node-v0.4.12/test/simple/test-sendfd.js
[01:14|% 100|+ 219|-   2]: Done                                                
make: *** [test] Error 1

[arlolra]

I can't confirm the problem w/ sendfd but the first one, test-http-full-response, is a bug in apache bench, discussed here: https://groups.google.com/forum/#!msg/nodejs/TZU5H7MdoII/yivu0d4LMaAJ

You can probably close this.

[telemachus]

@arlolra Maybe not close it until the sendfd failure is accounted for?

[arlolra]

@telemachus is that still a problem in v0.5.10?

[telemachus]

@arlolra I only follow the stable branch, but I can try v0.5.10.

Oh, wow: 0.5.10 produces 54 errors:

[02:06|% 100|+ 216|-  54]: Done

Let me know if it's useful, and I'll gist the entire test output.

[bnoordhuis]

Closing. The 0.4.x branch is dead, we're working (quite successfully) on making all tests pass on all platforms.

[telemachus]

@bnoordhuis Fair enough. Is it still useful to post test failures on current branches? (I still get two failures with 0.6.14 on OSX.)

[bnoordhuis]

@telemachus: Sure. But I'd prefer results for releases from the master branch (0.7.x) as that's where most of the work goes into.