Please provide a full stacktrace.

addEvent should not be called in a non-browser environment. By not having an else case you're masking an error, that's unacceptable.

@Krinkle @jzaefferer , sorry for my poor information. Here is the full stacktrace on Node.js.

/Users/takuto/work/git-sandbox/github/qunit-tap/test/compatibility/1.11.0/qunit.js:1506
        elem.attachEvent( "on" + type, fn );
             ^
TypeError: Object #<Object> has no method 'attachEvent'
    at addEvent (/Users/takuto/work/git-sandbox/github/qunit-tap/test/compatibility/1.11.0/qunit.js:1506:8)
    at /Users/takuto/work/git-sandbox/github/qunit-tap/test/compatibility/1.11.0/qunit.js:1184:1
    at Object.<anonymous> (/Users/takuto/work/git-sandbox/github/qunit-tap/test/compatibility/1.11.0/qunit.js:2152:2)
    at Module._compile (module.js:449:26)
    at Object.Module._extensions..js (module.js:467:10)
    at Module.load (module.js:356:32)
    at Function.Module._load (module.js:312:12)
    at Module.require (module.js:362:17)
    at require (module.js:378:17)
    at Object.<anonymous> (/Users/takuto/work/git-sandbox/github/qunit-tap/test/node/test_helper.js:18:13)

The root cause is calling addEvent( window, "load", QUnit.load ); on line 1184.

By not having an else case you're masking an error, that's unacceptable.

Agreed. In f249711 you said

executing it immediately is not a solution. That'll call it with the wrong arguments, with the wrong context at the wrong time.

And I respect your decision in my patch. However, not calling addEvent on non-browser enviroment is the best solution in this case. I agreed.

By not having an else case you're masking an error, that's unacceptable.

Agreed.

Well, your patch removes the else case, masking the error.

If it is okay not to call QUnit.load in a non-browser environment (which is effectively what this patch re-enforces), then the solution is to do that, not to mask it.

if ( ..browser.. ) { addEvent( .., .. QUnit.load ); seems like a more solid solution. That makes the fact that QUnit.load is not called explicit as opposed to implied behind a mask.

If it is okay not to call QUnit.load in a non-browser environment (which is effectively what this patch re-enforces), then the solution is to do that, not to mask it.

Now I totally agree with you. So I withdraw this PR now, then open another PR if I can.
Thanks.

Pull-requested again as #401