This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH]: begin removing size_int_type
Nathan Sidwell <nathan@codesourcery.com> writes:
| Richard Henderson wrote:
|
| > Indeed, I would prefer that we move the other direction. Virtually
| > all invocations of build_int_cst build single-word integers. It
| > would be nice if, say, build_int_cst took one HOST_WIDE_INT argument
| > and, say, build_int_cst_wide took the full two integer arguments.
| this patch changes the interface so that,
| * build_int_cst takes a single HOST_WIDE_INT that is then sign extended
| * build_int_cstu takes a single unsigned HOST_WIDE_INT that is zero extended
| * build_int_cst_wide takes a pair.
Hmm, the naming looks odd to me. Why not
* build_int_cst
* build_uint_cst
* build_wide_int_cst
?
| I decided that build_int_cst should always sign extend, rather than examine
| the signedness of the type, because this behaviour matches C itself.
Agreed.
-- Gaby