Pre-inline optimization, version 3

[Richard Kenner]

> The problem is more stubble - the reference to dead SSA name is caused
> by Ada language specific part of declaration holding pointers to random
> parts of GIMPLE statements.
> It would be great if those pointers was cleared after gimplification by
> Ada, but I am not familiar enough with that frontend to try that.

Can you give some examples of where these are?

There certainly are GIMPLE variables in type sizes and offsets and it
may indeed be the case that they are never referenced elsewhere and
hence the only usage is in the type tree, but those should be
considered "live" by virtue of the BIND_EXPR.

It would indeed be good to get rid of these, but (1) you have to be
concerned about debugging data and (2) the fact that they're dead
isn't known until way later than the front end.