Re: [PATCHv2] gdb/python: add gdb.Architecture.format_address

[Simon Marchi <simon.marchi@polymtl.ca>]

> @@ -391,6 +416,12 @@ group GROUP-NAME." },
>      METH_NOARGS,
>      "register_groups () -> Iterator.\n\
>  Return an iterator over all of the register groups in this architecture." },
> +  { "format_address", (PyCFunction) archpy_format_address,
> +    METH_VARARGS | METH_KEYWORDS,
> +    "format_address (ADDRESS) -> String.\n\
> +Format ADDRESS, an address within the currently selected inferior's\n\
> +address space, as a string.  The format of the returned string is\n\

I didn't have much time to think about this, but I was just wondering if
there would be a way to avoid relying on the currently selected
inferior.

My understanding is that this function depends on both an architecture,
for formatting the number part of the address, and an inferior for the symbolic
part of the the address.  Or is it a program space instead of inferior?

So the possibilities I see are, to let the user specify both:

  arch_obj.format_address(addr, inf_obj)
  arch_obj.format_address(addr, pspace_obj)
  inf_obj.format_address(addr, arch_obj)
  pspace_obj.format_address(addr, arch_obj)
  gdb.format_address(addr, inf_obj, arch_obj)
  gdb.format_address(addr, pspace_obj, arch_obj)

Putting the format_address on one of the objects (arch, inf, pspace)
seems strange, because it's not a arch-specific operation more than it
is an inf-specific (or pspace-specific operation).  So I'm more leaning
towards gdb.format_address.

And even if the user can pass in objects, we can offer sensible defaults
that will make format_address use "the current things".  Let's say that
format_address is declared as:

  def format_address(addr, pspace_obj=None, arch_obj=None)

If pspace_obj is None, we can fetch it from the currently selected
inferior.  And same for arch_obj.

Also, maybe that doing:

  format_address(addr, inf_obj)

could be a shortcut for:

  format_address(addr, inf_obj.progspace, inf_obj.architecture())

Not sure about that one though.

Simon