Alien and special objects Array

>Could you tell us if you get that error? And more info on when it  
>crashes?

It's not a crash in the special objects array (which typically results
in closing the image).


It crashes with "TempVariableNode(Object)>>doesNotUnderstand: #code:"
while trying to compile


'Ccalloc: byteSize "<Integer> ^<Integer>"
        <primitive: ''primCalloc'' error: errorCode module: ''IA32ABI''>
        ^self primitiveFailed'


The prereq's package (which I guess allows new synatx) is loaded.

Just try it in latest #10462.

with

ScriptLoader loadLatestPackage: 'AlienLoader' from: 'http://www.squeaksource.com/Alien'.
(Smalltalk at: #AlienLoader) load


Bye
T.
--
Jetzt kostenlos herunterladen: Internet Explorer 8 und Mozilla Firefox 3.5 -
sicherer, schneller und einfacher! http://portal.gmx.net/de/go/chbrowser

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Indeed.

The error is in the 'fixTemp' of the errorCode variable of the  
primitive method.
It appears that the latest Pharo has its own #fixTemp: method on the  
encoder, which is overwritten by the one in the Alien-Prereqs package.

I removed the #fixTemp: method that comes with Alien and all seems to  
work just fine if we stick with the one that comes from Pharo (all  
tests are green). A note to make is that the #fixTemp: of Alien was  
never executed before this version of Pharo, simply because the #code:  
method on a MethodNode has never existed (which is sent by the  
#fixTemp: that comes with Alien).

However, I'm not comfortable saying that this is the right patch to  
make. Perhaps someone familiar with the compiler or John can tell us  
if this is fine?
Another issue might be the compatibility with Squeak here. I should  
take a look....

I published the patch anyway in the meantime.

On 29 Sep 2009, at 10:50, Torsten Bergmann wrote:

> Just try it in latest #10462.
>
> with
>
> ScriptLoader loadLatestPackage: 'AlienLoader' from: 'http://www.squeaksource.com/Alien' 
> .
> (Smalltalk at: #AlienLoader) load

----------------------------
Johan Brichau
[hidden email]





_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

On Sep 29, 2009, at 6:49 AM, Johan Brichau wrote:

I thought #fixTemp was for code compatibility but isn't really needed  
because of real closures?




_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

On 29 Sep 2009, at 19:38, Brian Brown wrote:

> I thought #fixTemp was for code compatibility but isn't really needed
> because of real closures?

so this change should be safe?

note that the #fixTemp: method that is used right now comes with the  
standard pharo image (which does include closures).

----------------------------
Johan Brichau
[hidden email]





_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

>>
>
> I thought #fixTemp was for code compatibility but isn't really needed
> because of real closures?

this is just for backward compatibility normally the method is empty  
on pharo

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

On 30 Sep 2009, at 12:04, Stéphane Ducasse wrote:

>>>
>>
>> I thought #fixTemp was for code compatibility but isn't really needed
>> because of real closures?
>
> this is just for backward compatibility normally the method is empty
> on pharo

In 10462  Encoder>>fixTemp: is not empty


----------------------------
Johan Brichau
[hidden email]





_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

In my pharo1.0-10451-BETAdev09.09.3   I don't have fixTemp:  neither  fixTemp

The only thing I have is BlockClosure>>fixTemps



_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

I was tlakling about BlockClosure (but I'm certainly wrong)
On Sep 30, 2009, at 4:43 PM, Johan Brichau wrote:


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project