1) have an element: child1[i]=document.createElement('div'); 2) create an object: child6 = new Object; 3) have object contain a text node: child6[i]=document.createTextNode('test'); 4) do an append child of the textNode to the element works fine: child1[i].appendChild(child6[i]); 5) attempting an invalid append child with the object itself instead of the textnode crashes browser. It shouldn't work, obviously, but it sure shouldn't crash. child1[id].appendChild(child6); Note that steps four and five are not necessary, but are there for explanatory purposes, and to address the "what the heck are you doing that for" question. It was an accident, but quite hard to debug what was going on when every test crashed the browser.

*** Bug 2026 has been marked as a duplicate of this bug. ***

Fixed on 1/5/1999. Generic jsval to nsISupports conversion routine now checks for validity of object.

xpviewer is no more but setting qa contact to reporter for verify against viewer

Hey tupshin, Can you provide an html document with this testcase in it? That would be awesome! Thanks.

test case is at http://www.tupshin.com/domtests/badappendtest1.html 2/8/98 windows build worked did not crash. instead of crashing the console logged: Parameter is not an object. Note that this did not log the line number of the Javascript error(line 16 in the example). Also, I don't believe that the "Parameter is not an object" message is a correct message. It seems to be saying that the parameter(child6) is not an object, but child6 was just created with a "new Object" command, and alert(child6) generates [object Object] to the console. Should these be logged as new bugs, or should the old bug be reopened with a different summary?

marking as verified. the bug for #1956 is that the application crashed. Since that is no longer the case, I am closing out this bug and writing a new one specific to the new javascript error.