Re: [PATCH 1/3] gdb/block: make find_iterator_compunit_symtab a method of block_iterator

[Simon Marchi <simon.marchi@efficios.com>]

On 2/5/26 1:46 PM, Tom Tromey wrote:
>>>>>> "Simon" == simon marchi <simon.marchi@polymtl.ca> writes:
> 
> Simon> +compunit_symtab *
> Simon> +block_iterator::compunit_symtab () const
> Simon>  {
> Simon> -  if (iterator->idx == -1)
> Simon> -    return iterator->d.compunit_symtab;
> Simon> -  return iterator->d.compunit_symtab->includes[iterator->idx];
> Simon> +  if (this->idx == -1)
> Simon> +    return this->d.compunit_symtab;
> Simon> +  return this->d.compunit_symtab->includes[this->idx];
> 
> I just recalled that block iterators work in a funny way -- they can
> either iterate over a single block or a symtab.
> 
> And, if iterating over a single block, then it's invalid to refer to
> d.compunit_symtab.
> 
> So I think this new method should probably assert like
> 
>     gdb_assert (which != FIRST_LOCAL_BLOCK);
> 
> See the comments in block_iterator in block.h.

Yeah I saw that, but I didn't dig too much.  Because it's a functional
change, I will send it as a separate patch.

Simon