[7/10] keep track of dependencies

[Tom Tromey]

>>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:

I made most of the changes you pointed out in this review.
Thanks.

[...]

Jan> Why is RESULT_CHILDREN used here at all?  Would not it be
Jan> sufficient to just call htab_traverse for ALL_CHILDREN?

Maybe it isn't a great reason, but it lets us easily avoid adding CUs
without symbol tables.  See recursively_compute_inclusions:

  if (get_symtab (per_cu) != NULL)
    VEC_safe_push (dwarf2_per_cu_ptr, *result, per_cu);

So, RESULT_CHILDREN and ALL_CHILDREN do not have identical contents.

Tom