Making Sense of HRESULTS

re: Making Sense of HRESULTS

Jay Hugard — Fri, 02 Jul 2010 17:20:29 GMT

Thanks for responding, Eric!

Worked this out after tracing into the assembly for the JScript engine. See the following Question & Answer on StackOverflow.

stackoverflow.com/.../3167705

Man, was that a tough mystery to unravel ;-) Seriously underdocumented!

re: Making Sense of HRESULTS

Jay Hugard — Wed, 30 Jun 2010 01:25:13 GMT

I have a scriptable COM component (implements IDispatchEx) which has a method that takes a script-function as a parameter (an IDispatch* as a parameter). My COM component calls back into script via that supplied IDispatch parameter.

If the script-function happens to throw an exception, then my call to Invoke will return SCRIPT_E_PROPAGATE. If I return that HRESULT back to the caller of my COM component, then the script engine throws an error.

However, the *ACTUAL ERROR* is nowhere to be found. It's not returned in the Invoke HRESULT (that's SCRIPT_E_PROPAGATE). It's not returned via the EXCEPINFO supplied to Invoke (the structure remains empty). AND, it's not available via GetErrorInfo (the call returns S_FALSE)!

I'd *really* like to get my hands on that error, because it would be useful to cache it and return the same error on subsequent calls (getting to the error often involves an expensive operation that is defined by the script-function passed as a parameter, but I do know how to cache the error). Is there a way for a scripted COM component to get to an exception thrown during a call-back into a supplied script-function???

Hmm. I have not looked at that code for ten years (and I didn't understand it real well then; I didn't write the exception handling logic), so I'm a bit rusty. The EXCEPINFO ought to be filled in with the details. Do you know what exactly is being thrown by the script? Do you know if there are any possible handlers up the stack? The engine should interrogate the IDispatchEx to see if it can determine a call stack that has a catch block. Is there any such call stack?

re: Making Sense of HRESULTS

Jerome — Sun, 23 Sep 2007 15:23:58 GMT

Excellent, thx.

Thx to Jay Hugard too for his code

Human sacrifice, dogs and cats living together, mass hysteria! and thread model errors!

Fabulous Adventures In Coding — Tue, 08 May 2007 19:30:06 GMT

Reader Shaka comments on my post about error messages that "catastrophic failure" really does take the

re: Making Sense of HRESULTS

Alexey — Thu, 28 Sep 2006 16:55:55 GMT

What means error code 0x800a180e from Word::_Document interface?

Pelargir - Musings on software and life from Matthew Bass. » Decoding COM HRESULT error codes

Sat, 02 Sep 2006 16:25:43 GMT

PingBack from http://www.matthewbass.com/blog/2005/11/15/decoding-com-hresult-error-codes/

Integer Arithmetic in VBScript, Part Two

Fabulous Adventures In Coding — Mon, 06 Dec 2004 06:24:00 GMT

Excellent Debugging info in Visual Studio

Sohail Kadiwala — Thu, 18 Nov 2004 10:47:00 GMT

re: Typing Hard Can Trip You Up

Fabulous Adventures In Coding — Tue, 03 Aug 2004 17:31:00 GMT

RE: Making Sense of HRESULTS

Jay Hugard — Sat, 25 Oct 2003 01:44:00 GMT

Nuts... hit add before I could finish or proof the above... As I was saying, can be used as follows: function println(v) { WScript.echo(v) } try { } catch( e ) { if( e.number.toHRESULT() != 0x80070013 ) { println( e.toString() ); } }