Just use 'get_compound_{ident,name}()' instead.

It was only used by the class casst optimization, which was removed almost four years ago.

Just check that the size of the array is 0 instead.

'align_all' always is at least 1.

It is unused.

The length is now set on creation.
Neither length nor element type can be changed later on.

This is a bigger rewrite of stack frame handling in the backend:

- Do the stack frame type layout late, after all the spill slots have
  been created. This way we can avoid (and remove) the brittle
  frame_alloc_area().
- Standardize the meaning of stack entity offsets. After stack frame type
  layout they are relative to the stack pointer position at the beginning
  of the function.  It is all in one type now, no splitting into "arg",
  "between" and "stack_frame" type anymore.
- Generalize the stack pointer simulation code to work with a single
  callback. Represent stack state as current offset and align_padding
  number. Now that callbacks can access and modify them both we do not
  need custom code in the sparc backend anymore.
- Remove alignment specification on IncSP, only keep a ignore_align flag
  used for the last IncSP that has to remove all existing sp adjustments
  and may end up on a misaligned SP on ia32/amd64.
- Align stack by default on ia32+amd64 backends.

They are not used yet as we need an adapted layout algorithm for them.

The soon to be introduced spillslot entities can be at negative offset,
so the previously used -1 as convention for invalid offsets is not a
good idea anymore.

The added unittest was failing and also x10i. Frontend moves global
entity into class for native method calls. liboo lowering moves it back
during vtable creation.

The issue is that "pmap_insert(map,key,NULL)" does not delete an entry,
so "pmap_contains(map,key)" afterwards does not return false. This
changes add_compound_member to assert a NULL value instead of a NULL
entry.

A similar error might be in ir/be/bechordal.c.

This is done by introducing a new new_global_entity() function for
constructing globals/adding a visibility argument to new_alias_entity().

The map maps global symbols to the respective entities in the segment
types. This should avoid hacks in some frontends, avoids a sepcial map
for compilerlib entities and ensures that frontend and compilerlib
entities are in sync.

This is cleaner than having them being class types with a tf_segment
flag set.

The simple tp_opcode enum is enough, no need to have dispatch tables and
information structs behind each type kind.

There is no need for dispatch through tpop, they all use the same
underlying data structure anyway, so just work on that in
get_compound_xxx()

Do not use the type address as number because that makes things
nondeterministic.

For stabs debug info it was used to suppress debug info, so set ir_visibility_private instead.

You can simply use set_type_mode() instead.