RFC: merging MEM_ATTRS

[Joern Rennecke]

>     But that soon leads to a combinatorial explosion when you have thirty
>     such variables...  most of the time, you'll want the set that contains
>     none of these variables, sometimes you'll want to include one or more
>     likely two of them, but conceptually we could have many of them.
> 
> No, that's not what I meant: I meant one alias set for all.

For all what exactly?  One alias set to include all variables that didn't
have their address taken is not useful; most of these accesses are
already described exactly by their MEM_EXPRs.

Do you mean one alias set that excludes all of them, to be used for
indirect references?  We could use that, i.e. a MEM with a MEM_EXPR for
something that hasn't its address taken can't alias a MEM with that
alias set.
We couldn't express a merge of this alias set with just one single variable,
though, so we'd loose a bit more information when we do such a merge
than with a solution that is based solely on MEM_EXPRs.