Re: [PATCH 2/6] Make "ptype INTERNAL_FUNCTION" in Ada print like other languages

[Pedro Alves <pedro@palves.net>]

On 2023-02-14 3:30 p.m., Tom Tromey wrote:

>>> Fix this by simply adding an early check for
>>> TYPE_CODE_INTERNAL_FUNCTION.
> 
> Andrew> I confess, this is not the solution I though you'd go with.  I was
> Andrew> expecting you to handle TYPE_CODE_INTERNAL_FUNCTION in the switch, just
> Andrew> to leave things consistent.
> 
> I think this would be better.

My point with adding this check early is that these functions' type never
has anything to do with Ada, so all that code at the beginning of ada_print_type,
like decoded_type_name, ada_is_aligner_type, ada_is_constrained_packed_array_type,
etc. is always a nop for internal functions, so might as well skip it all.

I actually started out by considering moving TYPE_CODE_INTERNAL_FUNCTION printing
to common code (say, rename the virtual language_defn::print_type to do_print_type,
and add a new non-virtual wrapper language_defn::print_type method and print
TYPE_CODE_INTERNAL_FUNCTION there), and I didn't pursue that as I couldn't really convince
myself that a different language might want to print it differently (say, some other
characters instead of "<>"), and I guess that's why I ended up with putting it at the start
of the function, as that is the closest to putting it at the caller instead.