This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: rs6000 3.1 Ada bootstrap problem
- From: David Edelsohn <dje at watson dot ibm dot com>
- To: kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)
- Cc: gcc at gcc dot gnu dot org
- Date: Fri, 15 Mar 2002 14:30:22 -0500
- Subject: Re: rs6000 3.1 Ada bootstrap problem
>>>>> Richard Kenner writes:
> gnatpsta is referencing symbols which reasonably can come from
> that source, but does not link against it. I do not know how the Ada
> developers want to address this problem.
Richard> Probably the simplest would be for rs6000.h to define
Richard> WIDEST_FP_HARDWARE_SIZE.
Define it to what? WIDEST_FP_HARDWARE_SIZE is used in
Pos
get_target_long_double_size ()
{
return WIDEST_HARDWARE_FP_SIZE;
}
In comparison to other functions in the file, get_target_bits_per_unit()
DOES NOT return MAX_BITS_PER_UNIT. get_target_long_size() DOES NOT return
MAX_LONG_TYPE_SIZE.
rs6000.h already defines MAX_LONG_DOUBLE_TYPE_SIZE, but
WIDEST_FP_HARDWARE_SIZE is not defined in terms of that.
I do not know how Ada eventually uses these values, but the
context and name implies that Ada wants the active, in-use sizes, not the
maximum size the value can take. Redefining WIDEST_FP_HARDWARE_SIZE does
not get around the fact that one needs rs6000_long_double_type_size to
know the size.
David