Re: [dsdp-dd-dev] Re: [platform-debug-dev] Variable Access Size in a Mem

[John Cortell <john.cortell@xxxxxxxxxxxxx>]

At 04:47 PM 11/14/2006, Samantha Chan wrote:
> Hi John,
>
> > There's the ability to customize the add-monitor action that was
> > added in 3.2, but I suspect this requirement is pervasive enough that
> > we'd not want to impose that high-cost in order to get that
> > capability. Ideally, the platform would expose that parameter if the
> > debugger adapter says it supports/needs the capability.
>
> Do you mean that you still think that the platform should have the UI to
> allow the user to enter this information?
> We need interfaces to ask a debug adapter if it supports this capability
> and for passing the access size to the adapter?
>
> I was suggesting that all this work should not be done in the platform.
> Instead, a debug adapter, like DSF or CDT, should have interfaces to ask
> its engine if this capability is supported.  If so, then the adapter will
> customize the Add Memory Monitor action to allow the user to enter the
> information.
>
> I see this similar to our "Address Space" problem.  The platform should
> have enough hooks to allow clients to customize the "Add Memory Monitor"
> actions and related dialogs.  It's up to a debug adapter to implement the
> concept of address space.  The platform does not understand anything about
> address spaces.

Fair enough. I guess I was putting this in a different league than 
the memory space issue because of its relative simplicity. For access 
size, we're talking about exposing a text field--maybe a combo box, 
where as exposing memory space choices could be trickier. But I get 
your point, and the fact that this sort of special behavior really 
can and probably should reside completely on the adapter end. I've 
already added a memory-space aware add-monitor handler in CDT. I 
guess this would just be an extension of that work.

John