Re: (patch) hpjyg23: gdbtypes.[ch] & values.c

[Andrew Cagney <ac131313 at cygnus dot com>]

Jimmy Guo wrote:
> 
> >> + #ifdef BFD64
> >> +   builtin_type_CORE_ADDR = builtin_type_unsigned_long_long;
> >> + #else
> >> +   builtin_type_CORE_ADDR = builtin_type_unsigned_long;
> >> + #endif
> >
> >Can you expand a little - looks like there isn't a builtin_type_* for
> >pointer :-(  That makes makes the introduction of a builtin_type for
> >target pointers a pretty good idea.
> 
> builtin_type_CORE_ADDR is used in cp-valprint.c and valops.c in offset
> calculation to skip over entries of pointer type.

Hmm, this makes the new builtin_type sound more like the target language
pointer (C's void*) rather than the target architectures address.  That
would suggest the need for a different name and a comment explaining why
it isn't CORE_ADDR.

Is the list of cases:

	o	CORE_ADDR
		builtin_type_CORE_ADDR?

	o	object file addr
		builtin_type_bfd_addr

	o	language addr
		builtin_type_C_pointer?
		hmm, GDB's current schemea would
		make it just builtin_type_pointer.

complete?  If we added all of them you can then just pick and choose :-)

Stan,
now you're the stylistic expert, thoughts?

	enjoy,
		Andrew

PS: I like your move to add builtin_type_*.  I think this is going to
finally cleanup/fix another of GDB's confusions.