Polycephaly

Hi folks,

today I've tried to use the Polycephaly package to get some heavy  
compuations done by drone images. However, my application generates some  
code (namespaces, classes and shared vars) during runtime _before_ I start  
any drone image, but I ran into some OSTransferError. It seems that  
everything has to be defined in the applications startup image _on_disk_.  
I've been able to circumvent this by taking a snapshot right before  
starting up the drones.

However, ss there another way to propagate the image's current state to  
these drones instead of snapshotting to disk?

Btw: great package - just KISS!

Cheeers, Steffen
_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Steffen,

on some operating systems, mainly Unix, there is a system call named fork.
This call duplicates a running program, see
http://www.opengroup.org/onlinepubs/007908799/xsh/fork.html. We have done
experiences with it and it applies also to Smalltalk. We used to do a
snapshot without losing the reactiveness of the system, and it worked (on
Linux). I think a similar approach should also apply to Polycephaly and the
way you want to use it.

Georg


Georg Heeg eK, Dortmund und Köthen, HR Dortmund A 12812
Tel. +49-3496-214328, Fax +49-3496-214712
compuations
> done by drone images. However, my application generates some code
> (namespaces, classes and shared vars) during runtime _before_ I start any
drone
> image, but I ran into some OSTransferError. It seems that everything has
to be
> defined in the applications startup image _on_disk_.
> I've been able to circumvent this by taking a snapshot right before
starting up
> the drones.
>
> However, ss there another way to propagate the image's current state to
these
> drones instead of snapshotting to disk?
>
> Btw: great package - just KISS!
>
> Cheeers, Steffen
> _______________________________________________
> vwnc mailing list
> [hidden email]
> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc


_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Ok, I'll look into this.
Since all classes and namespaces are generated in one package, it should  
work to file out that as parcel. But I wonder whether this works for  
instances stored in shared variables package too, since they are  
referenced elsewhere. Are the object identities preserved? This is  
important for my application because it needs to pass some of that  
instances as arguments to the block / code to be run in the drones.
Imagine a structure like:

GeneratedCodePackage
        NamespaceA
        NamespaceA.ClassA
        NamespaceA at: SharedVariable put: InstanceOfClassA

vm := VirtualMachine new.
vm do: [:ins | Computation on: ins] with: someDataStructure.

... where someDataStructure may contain other objects referencing  
InstanceOfClassA.

Regards,
Steffen




Am 19.02.2010, 18:00 Uhr, schrieb Dave Stevenson  
<[hidden email]>:

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Hi,

I've been testing a bit more and ran into another small problem. If I  
build a runtime image, polycephaly won't work (error.log attached).  
Actually this tracks down to "DeploymentOptionsSystem current  
startInRuntime: true", but I do not understand how this affects this  
concurrency mechanism. Can anybody help?

Thanks in advance,
Steffen Märcker



Am 19.02.2010, 17:56 Uhr, schrieb Steffen Märcker <[hidden email]>:

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Hi,

I've been testing a bit more and ran into another small problem. If I
build a runtime image, polycephaly won't work (error.log attached).
Actually this tracks down to "DeploymentOptionsSystem current
startInRuntime: true", but I do not understand how this affects this
concurrency mechanism. Can anybody help?

Thanks in advance,
Steffen Märcker



Am 19.02.2010, 17:56 Uhr, schrieb Steffen Märcker <[hidden email]>:
_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Hi,

I've been testing a bit more and ran into another small problem. If I
build a runtime image, polycephaly won't work (error.log attached).
Actually this tracks down to "DeploymentOptionsSystem current
startInRuntime: true", but I do not understand how this affects this
concurrency mechanism. Can anybody help?

Thanks in advance,
Steffen Märcker



Am 19.02.2010, 17:56 Uhr, schrieb Steffen Märcker <[hidden email]>:
_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc