[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [cdt-debug-dev] address spaces
|
At 04:17 PM 6/22/2005, Alain Magloire wrote:
Yes, it does
look like a viable solution, at least something
that we should explore for the next version.
Can you make a proposition on how or rather what kind of
interface
that an IAddress object should have to meet you
requirements.
There will probably be some changes in the Core/UI to make
sure
we do treat addresses not like a simple string/long but
rather as
an opaque object with certain attributes.
Off the top of my head...perhaps the simplest way would be to reduce the
representation of a memory address to a string (not a string-fied integer
like BigInteger, but just a string). This of course puts the burden on
the back end to encode/decode these things every time it passes/receives
one over the fence, but it is extremely flexible in that any sort of
physical memory model could be represented--either using a generally
accepted ("standard") encoding/format or a proprietary one. In
the end it would mean that nothing above the backend layer could ever do
anything with these things except pass them around, but maybe that's
already the case (or should be).
Just an idea at this point...
John