Sorry Gabe,

I ran all the other variants, I just don't have files, etc for x86.  
Perhaps you should place some on zizzer. It is possible to turn your  
changesets back into patches. The first thing I would do is clone the  
repository though. That way you will have something to go back to if  
something goes wrong. All parameters should have a default or a user  
set value. If the parameter is left blank, you're still getting a  
value in the params struct, it just happens to be NULL. If NULL is ok,  
then it should be the default, if it's not OK then it was going to  
generate an error anyway (later in the execution). To answer your  
question more explicitly, you can set something to NULL.

Ali


On Feb 1, 2009, at 4:32 AM, Gabe Black wrote:

>    The recent change which added a fatal for not having a default or
> user set value for a parameter broke X86_FS. There wasn't any way to
> really know that was going to happen since it wouldn't show up until  
> run
> time and X86_FS isn't generally in a known condition, so I'm not all
> that upset about the actual breakage. The fix is relatively simple,  
> but
> unfortunately I didn't realize my recent pull broke things until I'd
> moved a considerable number of patches into regular changesets. I  
> could
> push as is and fix the problem later in the history, but then all of  
> the
> steps between the param change and now will be broken. I think  
> there's a
> way to turn changesets back into patches which I'll try. Is there a
> better way to handle this? Also, is it really correct for all  
> parameters
> to have to have a default value or a user set value? What about things
> that are optional? It's not possible to set a default value of  
> "None" as
> far as I know.
>
> Gabe
> _______________________________________________
> m5-dev mailing list
> [email protected]
> http://m5sim.org/mailman/listinfo/m5-dev
>

_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev

Reply via email to