> From: Pops > > This might be slightly off topic, a javascript script > question, but its being applied to jQuery. :-) > > Ok, there is a different in other languages when you do this: > > var p = null; > xyz(null); > xyx(p); > > It depends on the function prototype and how a language binds > to the function.. > > In JavaScript, what is passed to the function xyz: > > The address of p or the value of p.
Object, Array, and Function arguments are passed by reference; other types such as Number and String are passed by value. null is an Object and is passed by reference. (There is only one null object.) In your example, the two calls to xyz() are identical. Both pass the same null object into the function. > I ask because in the jQuery XHR implementation, it has this > > xml.send(s.data); > > And s.data is set to null if s.type is "get" > > So I am wondering if it safe to assume the user's agent > external prototype for send() expects: > > - an asciiz string > > - OLE string (like BSTR) like its normally done in > Windows's OLE/ > ActiveX/COM/DCOM > RPC interfacing The standard calls for the argument to send() to be a DOMString, a Document, null, or omitted. Any other type is converted to a DOMString if possible, otherwise treated as null: http://www.w3.org/TR/XMLHttpRequest/#dfn-send > Of course, if you are not sure, its always to be safe and do: > > xml.send(( s.type.toLowerCase() == "get" )?null:s.data); Just as in your xyz() example, if s.data is null, then these two calls are identical: xml.send( null ); xml.send( s.data ); So the extra code isn't necessary. -Mike
