Our proposed approach is to -- by default -- assume that "g" may access all of "b". However, in the event that the corresponding parameter to "g" has an attribute (name TBD, possibly the same as the one that appears in Danny's recent patch), then we may assume that "g" (and its callees) do not use the pointer to obtain access to any fields of "b".
[ I did not discuss this with Kenny, but another option is to have a -fassume-X switch, off by default, which treats your code as if you had the magic attribute everywhere. ]
I have no problem with this resolution, but as i previously explained, it does require the patch i submitted (which actually removes the code entirely for the moment) be applied, unless you are going to go the flag route.
If you are doing it just by attribute, the code to handle this the above way needs to be reinserted somewhere else when we can use it (IE without redoing call clobbering a little bit, it's not useful).
