Fwd: Re: WriteBarriers and Weaklings...maybe Glorp?

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

Fwd: Re: WriteBarriers and Weaklings...maybe Glorp?

cdavidshaffer
Oops, meant to send this to the list...

-------- Original Message --------
Subject: Re: [vwnc] WriteBarriers and Weaklings...maybe Glorp?
Date: Mon, 08 Nov 2010 15:06:12 -0500
From: C. David Shaffer [hidden email]
To: Michael Lucas-Smith [hidden email]


On 11/08/10 14:43, Michael Lucas-Smith wrote:
> GlorpActiveRecord uses the WriteBarriers. We also use them in
> Seaside-Glorp. I've never seen the problem you describe though. But
> I'll keep an eye out for it.
>

I'll have a look, then.  At first I thought that this was probably a
problem localized to my particular usage pattern but the more I think
about it the more it seems like a broader problem with any application
using WriteBarriers.  Even if your application is 100% synchronous,
there are other VisualWorks processes that might signal
noModicationError asynchronously.  For applications with some level of
parallelism the cross-section for hitting that inject:into: loop is low
but you should certainly be able to produce problems with tons of
asynchronous sessions = heavily loaded Seaside app :-)

David



_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
Reply | Threaded
Open this post in threaded view
|

Re: Fwd: Re: WriteBarriers and Weaklings...maybe Glorp?

Michael Lucas-Smith-2
Thanks I'll be making an AR for this.

Michael

On Nov 8, 2010, at 12:06 PM, C. David Shaffer wrote:

Oops, meant to send this to the list...

-------- Original Message --------
Subject: Re: [vwnc] WriteBarriers and Weaklings...maybe Glorp?
Date: Mon, 08 Nov 2010 15:06:12 -0500
From: C. David Shaffer [hidden email]
To: Michael Lucas-Smith [hidden email]


On 11/08/10 14:43, Michael Lucas-Smith wrote:
> GlorpActiveRecord uses the WriteBarriers. We also use them in
> Seaside-Glorp. I've never seen the problem you describe though. But
> I'll keep an eye out for it.
>

I'll have a look, then.  At first I thought that this was probably a
problem localized to my particular usage pattern but the more I think
about it the more it seems like a broader problem with any application
using WriteBarriers.  Even if your application is 100% synchronous,
there are other VisualWorks processes that might signal
noModicationError asynchronously.  For applications with some level of
parallelism the cross-section for hitting that inject:into: loop is low
but you should certainly be able to produce problems with tons of
asynchronous sessions = heavily loaded Seaside app :-)

David


_______________________________________________
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