KAT demo errors using Windows XP on a Dell Latitude Laptop

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

KAT demo errors using Windows XP on a Dell Latitude Laptop

RichWhite
I am running into an error message upon starting the KAT demo code using
Windows XP (KAT demo code LOVES the Mac  BTW).. The machine is a Dell Latitude
Laptop if that helps?  I grabbed a screen shot of the error... Any help would
he greatly appreciated !

Error message screen shot:
http://img81.imageshack.us/my.php?image=croqueterrormessageqh5.png

Best Regards !

Rich


============
Reply | Threaded
Open this post in threaded view
|

Re: KAT demo errors using Windows XP on a Dell Latitude Laptop

Darius Clarke
Hi Rich,
 
I'm not at home now, but I'll give it a look this evening, if you don't get an answer before then.
 
Have you used Monticello to update to the latest version? Perhaps you might have updated some other images but not this one?
 
Cheers,
Darius

 
On 10/26/07, [hidden email] <[hidden email]> wrote:
I am running into an error message upon starting the KAT demo code using
Windows XP (KAT demo code LOVES the Mac  BTW).. The machine is a Dell Latitude
Laptop if that helps?  I grabbed a screen shot of the error... Any help would
he greatly appreciated !

Error message screen shot:
http://img81.imageshack.us/my.php?image=croqueterrormessageqh5.png

Best Regards !

Rich


============

Reply | Threaded
Open this post in threaded view
|

Re: KAT demo errors using Windows XP on a Dell Latitude Laptop

Howard Stearns-3
In reply to this post by RichWhite
OpenAL is not working quite right on your Dell. Was it installed?

Here's a breakdown of what we see in the image of the debugger that  
you supplied:

* The title says MessageNotUnderstood: UndefinedObject>>isCurrent.
We can guess from the KrazyCaps that MessageNotUnderstood is the name  
of a class, and if we look it up, we see that it is the subclass of  
Error. If we look at the places where it's used (cmd-N), particularly  
Object>>doesNotUnderstand, we see that it is an error message that  
reports the class of the object (UndefinedObject) that received a  
message that it doesn't have a method for (#isCurrent).

* The first line of the debugger confirms that. (UndefinedObject
(Object)>>doesNotUnderstand).  Specifically, it is saying that an  
object of class UndefinedObject executed #doesNotUnderstand, which it  
inherited from the class Object.   If we look up UndefinedObject, we  
see that it is the class of the single object "nil".  So, someone  
sent nil isCurrent.

* The second line of the debugger tells us how that happend.  
KCroquetParticipant(CroquetParticipantWithMenu)>>mouseEnter: says  
that the above happened in the code for #mouseEnter:, as defined in  
CroquetParticipantWithMenu, where 'self' was an instance of  
KCroquetParticipant.  If we look up that code, it says:
...
        harness ifNotNil: [
                harness openAL isCurrent ifFalse: [
                        harness makeOpenALCurrent.
                ].
        ].

OK, there's a nice guard against harness being nil, but nothing about  
harness openAL being nil. The author figured that if there's a  
harness, it must have openAL initialized, but possibly not current.

If we look at 'inst var defs...'  of a suitable harness class, like  
CroquetHarnessWithmenu, we see that #initialize or  
#initializeOpenALHarness prints an error message on the transcript  
and sets the openAL inst var to nil if #makeCurrent gets an error.    
So if you try running again with a transcript open, you should see an  
error about Cannot start spatial sound.

Now, there's a couple of possibilities:

1) You have OpenAL installed, but it's not working right. That might  
be the case that was intended by the author, who chose to print a  
message but not give an error immediately when #makeCurrent failed.  
If we want to honor that and allow things to just keep on trucking  
but without sound, we should put a guard against harness openAL being  
nil.  Now, there's nothing wrong with making a change to any class in  
your own image. But if you want to share your code with others, you  
might want to contact the author (in this case pbm) and see if he  
wants to fix/change his code. Otherwise, you can make a subclass and  
fix it there. That's what I did with KCroquetParticipant>>mouseMove:,  
where I inserted a guard against harness being nil.

2) It may be that everything is working right, eventually, but  
there's something odd about the timing. I don't think that's the case  
here, but it could be. In this case, a guard like above is a quick  
and dirty fix, but it would really just be hiding a potential problem.

OpenAL has a bit of both. Originally, the concept was that it was  
just supposed to work, and nothing was guarded. But I lost so much  
time with bugy OpenAL code that I eventually ran though all my code  
with a knife putting guards around openAL being nil.  Looks like I  
missed one.


On Oct 26, 2007, at 6:04 PM, [hidden email] wrote:

> I am running into an error message upon starting the KAT demo code  
> using
> Windows XP (KAT demo code LOVES the Mac  BTW).. The machine is a  
> Dell Latitude
> Laptop if that helps?  I grabbed a screen shot of the error... Any  
> help would
> he greatly appreciated !
>
> Error message screen shot:
> http://img81.imageshack.us/my.php?image=croqueterrormessageqh5.png
>
> Best Regards !
>
> Rich
>
>
> ============

Reply | Threaded
Open this post in threaded view
|

Re: KAT demo errors using Windows XP on a Dell Latitude Laptop

RichWhite
Thank You Howard !   Sure enough That fixed it !

Best Regards,
Rich
============

On 10/26/07, Howard Stearns <[hidden email]> wrote:

> OpenAL is not working quite right on your Dell. Was it installed?
>
> Here's a breakdown of what we see in the image of the debugger that
> you supplied:
>
> * The title says MessageNotUnderstood: UndefinedObject>>isCurrent.
> We can guess from the KrazyCaps that MessageNotUnderstood is the name
> of a class, and if we look it up, we see that it is the subclass of
> Error. If we look at the places where it's used (cmd-N), particularly
> Object>>doesNotUnderstand, we see that it is an error message that
> reports the class of the object (UndefinedObject) that received a
> message that it doesn't have a method for (#isCurrent).
>
> * The first line of the debugger confirms that. (UndefinedObject
> (Object)>>doesNotUnderstand).  Specifically, it is saying that an
> object of class UndefinedObject executed #doesNotUnderstand, which it
> inherited from the class Object.   If we look up UndefinedObject, we
> see that it is the class of the single object "nil".  So, someone
> sent nil isCurrent.
>
> * The second line of the debugger tells us how that happend.
> KCroquetParticipant(CroquetParticipantWithMenu)>>mouseEnter: says
> that the above happened in the code for #mouseEnter:, as defined in
> CroquetParticipantWithMenu, where 'self' was an instance of
> KCroquetParticipant.  If we look up that code, it says:
> ...
>         harness ifNotNil: [
>                 harness openAL isCurrent ifFalse: [
>                         harness makeOpenALCurrent.
>                 ].
>         ].
>
> OK, there's a nice guard against harness being nil, but nothing about
> harness openAL being nil. The author figured that if there's a
> harness, it must have openAL initialized, but possibly not current.
>
> If we look at 'inst var defs...'  of a suitable harness class, like
> CroquetHarnessWithmenu, we see that #initialize or
> #initializeOpenALHarness prints an error message on the transcript
> and sets the openAL inst var to nil if #makeCurrent gets an error.
> So if you try running again with a transcript open, you should see an
> error about Cannot start spatial sound.
>
> Now, there's a couple of possibilities:
>
> 1) You have OpenAL installed, but it's not working right. That might
> be the case that was intended by the author, who chose to print a
> message but not give an error immediately when #makeCurrent failed.
> If we want to honor that and allow things to just keep on trucking
> but without sound, we should put a guard against harness openAL being
> nil.  Now, there's nothing wrong with making a change to any class in
> your own image. But if you want to share your code with others, you
> might want to contact the author (in this case pbm) and see if he
> wants to fix/change his code. Otherwise, you can make a subclass and
> fix it there. That's what I did with KCroquetParticipant>>mouseMove:,
> where I inserted a guard against harness being nil.
>
> 2) It may be that everything is working right, eventually, but
> there's something odd about the timing. I don't think that's the case
> here, but it could be. In this case, a guard like above is a quick
> and dirty fix, but it would really just be hiding a potential problem.
>
> OpenAL has a bit of both. Originally, the concept was that it was
> just supposed to work, and nothing was guarded. But I lost so much
> time with bugy OpenAL code that I eventually ran though all my code
> with a knife putting guards around openAL being nil.  Looks like I
> missed one.
>
>
> On Oct 26, 2007, at 6:04 PM, [hidden email] wrote:
>
> > I am running into an error message upon starting the KAT demo code
> > using
> > Windows XP (KAT demo code LOVES the Mac  BTW).. The machine is a
> > Dell Latitude
> > Laptop if that helps?  I grabbed a screen shot of the error... Any
> > help would
> > he greatly appreciated !
> >
> > Error message screen shot:
> > http://img81.imageshack.us/my.php?image=croqueterrormessageqh5.png
> >
> > Best Regards !
> >
> > Rich
> >
> >
> > ============
>
>