298336 - xpconnect js out argument conversion error is NS_ERROR_FAILURE

Status: RESOLVED INACTIVE

(Core :: XPConnect, defect)

Description

[timeless]

in this simple example, you can easily see what's wrong and why, unfortunately
with real code, it was a different story.

js>
SIP=Components.Constructor("@mozilla.org/supports-interface-pointer;1",Components.interfaces.nsISupportsInterfacePointer)
[object nsXPCConstructor]
js> sip=new SIP;
[interface pointer]
js> a={iids:[Components.interfaces.nsISupports,Components.interfaces.nsIFile],
QueryInterface:function (iid) { for (var i in this.iids) if
(this.iids[i].equals(iid)) return this; throw
Components.results.NS_ERROR_NO_INTERFACE}, get parent(){return
10}};sip.data=a;sip.data instanceof a.iids[1];sip.data
[xpconnect wrapped nsIFile]
js> sip.data.parent
uncaught exception: [Exception... "Component returned failure code: 0x80004005
(NS_ERROR_FAILURE) [nsIFile.parent]"  nsresult: "0x80004005 (NS_ERROR_FAILURE)"
 location: "JS frame :: typein :: <TOP_LEVEL> :: line 5"  data: no]

Comment 1

[timeless]

Attachment: be more descriptive about the problem

js> sip.data.parent
typein:32: Component returned failure code: 0x8057003b
(NS_ERROR_XPC_BAD_CONVERT_RETURN) [nsIFile.parent]

Comment 2

[David Bradley]

Comment on attachment 187111 [details] [diff] [review]
be more descriptive about the problem

I think we should try and presever the error value from
XPCConvert::JSData2Native interface. It might provide more specific
information, though, from the looks of it, it doesn't. But that's another
issue.

Ideally it would be nice to report the parameter's index. I've cursed Microsoft
many times for similar lack in IDispatch. That's kind of where you were heading
with the retval vs out param variation of error codes anyway.

So I think this should just use NS_ERROR_XPC_BAD_CONVERT_JS for the
non-JSData2Native cases and then pass in the error return variable for the
JSData2Native case and return that. Knowing that its a return vs out parameter
would be nice, but I'd like to see a solution that gives people real
information about the location of problem

Comment 3

[BMO Automation]

Per policy at https://wiki.mozilla.org/Bug_Triage/Projects/Bug_Handling/Bug_Husbandry#Inactive_Bugs. If this bug is not an enhancement request or a bug not present in a supported release of Firefox, then it may be reopened.