Re: [wip] BFD from an arbitrary object; Was: provide pass-throughvalue in bfd_elf_bfd_from_remote_memory

[DJ Delorie <dj at redhat dot com>]

> Note that it is:
> 
> +.  struct bfd_file stream;
> 
> and not:
> 
> +.  struct bfd_file *stream;
> 
> however, what ever.

Hmmm... Ok, that's more in line with the way other parts of bfd work,
I suppose.

ops should still have the length field for compatibility tests,
though.  If the backends are the ones creating the data structure,
they don't need to worry about compatibility.  We should try to keep
the ops structure private; I don't want generic users trying to bypass
the published ABI and go straight to the lowlevel stuff.

In that case, I'd prefer something less vague than "cookie".  How
about "data"?  I assume most backends will have this point to a
structure with various items in it, whereas "cookie" implies a single
value.  Or a dessert ;-)