This series contains some small adjustments to the builtin types
and their predefined macros, for all supported architectures.
This should fully cover now all the integer types. The next step
would be the ones for floating-point numbers but it's not exactly
on the high-priority list.
Luc Van Oostenryck (9):
sparc: add 'sparcv8' predefines for sparc32
alpha: has 64-bit long double & int128
ppc: add predefines __LONGDOUBLE128 & __LONG_DOUBLE_128__
arch: add predefines __INT_LEAST${N}_TYPE__
arch: add predefines __INT_FAST${N}_TYPE__
predefine: teach sparse about __SIG_ATOMIC_TYPE__
arch: allow target specific [u]intptr_t & ptrdiff_t
x86-x32: fix it by defining a separate target for it
predefine: let predefine_width() take the usual interface
predefine.c | 31 +++++++++++++++++++-----
symbol.c | 7 ++++++
symbol.h | 3 ---
target-alpha.c | 3 +++
target-arm.c | 7 ++++++
target-h8300.c | 7 ++++++
target-m68k.c | 9 +++++++
target-microblaze.c | 9 +++++++
target-nds32.c | 5 ++++
target-nios2.c | 9 +++++++
target-openrisc.c | 5 ++++
target-ppc.c | 13 ++++++++++
target-riscv.c | 12 ++++++++-
target-s390.c | 12 +++++++++
target-sh.c | 8 ++++++
target-sparc.c | 24 +++++++++++++++---
target-x86.c | 59 +++++++++++++++++++++++++++++++++++++++++++++
target-xtensa.c | 5 ++++
target.c | 33 ++++++++++++++++++++++---
target.h | 21 ++++++++++++++++
20 files changed, 266 insertions(+), 16 deletions(-)
base-commit: d2947a933574da18bab1ef7b7534199f44cf9e98
--
2.27.0
