Issue #4258: Make it possible to use 30-bit digits for PyLongs:

- new configure option --enable-big-digits
 - new structseq sys.int_info giving information about the internal format
By default, 30-bit digits are enabled on 64-bit machines but
disabled on 32-bit machines.
This commit is contained in:
Mark Dickinson 2009-03-18 20:06:12 +00:00
parent e7f45b8e59
commit bd7926478d
15 changed files with 865 additions and 68 deletions

View File

@ -413,6 +413,23 @@ always available.
same information.
.. data:: int_info
A struct sequence that holds information about Python's
internal representation of integers. The attributes are read only.
+-------------------------+----------------------------------------------+
| attribute | explanation |
+=========================+==============================================+
| :const:`bits_per_digit` | number of bits held in each digit. Python |
| | integers are stored internally in base |
| | ``2**int_info.bits_per_digit`` |
+-------------------------+----------------------------------------------+
| :const:`sizeof_digit` | size in bytes of the C type used to |
| | represent a digit |
+-------------------------+----------------------------------------------+
.. function:: intern(string)
Enter *string* in the table of "interned" strings and return the interned string

View File

@ -87,5 +87,28 @@ Some smaller changes made to the core Python language are:
(Contributed by Fredrik Johansson and Victor Stinner; :issue:`3439`.)
* Integers are now stored internally either in base 2**15 or in base
2**30, the base being determined at build time. Previously, they
were always stored in base 2**15. Using base 2**30 gives
significant performance improvements on 64-bit machines, but
benchmark results on 32-bit machines have been mixed. Therefore,
the default is to use base 2**30 on 64-bit machines and base 2**15
on 32-bit machines; on Unix, there's a new configure option
--enable-big-digits that can be used to override this default.
Apart from the performance improvements this change should be
invisible to end users, with one exception: for testing and
debugging purposes there's a new structseq ``sys.int_info`` that
provides information about the internal format, giving the number of
bits per digit and the size in bytes of the C type used to store
each digit::
>>> import sys
>>> sys.int_info
sys.int_info(bits_per_digit=30, sizeof_digit=4)
(Contributed by Mark Dickinson; :issue:`4258`.)
.. ======================================================================

View File

@ -7,24 +7,62 @@ extern "C" {
/* This is published for the benefit of "friend" marshal.c only. */
/* Parameters of the long integer representation.
These shouldn't have to be changed as C should guarantee that a short
contains at least 16 bits, but it's made changeable anyway.
Note: 'digit' should be able to hold 2*MASK+1, and 'twodigits'
should be able to hold the intermediate results in 'mul'
(at most (BASE-1)*(2*BASE+1) == MASK*(2*MASK+3)).
Also, x_sub assumes that 'digit' is an unsigned type, and overflow
is handled by taking the result mod 2**N for some N > SHIFT.
And, at some places it is assumed that MASK fits in an int, as well.
long_pow() requires that SHIFT be divisible by 5. */
/* Parameters of the long integer representation. There are two different
sets of parameters: one set for 30-bit digits, stored in an unsigned 32-bit
integer type, and one set for 15-bit digits with each digit stored in an
unsigned short. The value of PYLONG_BITS_IN_DIGIT, defined either at
configure time or in pyport.h, is used to decide which digit size to use.
Type 'digit' should be able to hold 2*PyLong_BASE-1, and type 'twodigits'
should be an unsigned integer type able to hold all integers up to
PyLong_BASE*PyLong_BASE-1. x_sub assumes that 'digit' is an unsigned type,
and that overflow is handled by taking the result modulo 2**N for some N >
PyLong_SHIFT. The majority of the code doesn't care about the precise
value of PyLong_SHIFT, but there are some notable exceptions:
- long_pow() requires that PyLong_SHIFT be divisible by 5
- PyLong_{As,From}ByteArray require that PyLong_SHIFT be at least 8
- long_hash() requires that PyLong_SHIFT is *strictly* less than the number
of bits in an unsigned long, as do the PyLong <-> long (or unsigned long)
conversion functions
- the long <-> size_t/Py_ssize_t conversion functions expect that
PyLong_SHIFT is strictly less than the number of bits in a size_t
- the marshal code currently expects that PyLong_SHIFT is a multiple of 15
- NSMALLNEGINTS and NSMALLPOSINTS should be small enough to fit in a single
digit; with the current values this forces PyLong_SHIFT >= 9
The values 15 and 30 should fit all of the above requirements, on any
platform.
*/
#if HAVE_STDINT_H
#include <stdint.h>
#endif
#if PYLONG_BITS_IN_DIGIT == 30
#if !(defined HAVE_UINT64_T && defined HAVE_UINT32_T && \
defined HAVE_INT64_T && defined HAVE_INT32_T)
#error "30-bit long digits requested, but the necessary types are not available on this platform"
#endif
typedef PY_UINT32_T digit;
typedef PY_INT32_T sdigit; /* signed variant of digit */
typedef PY_UINT64_T twodigits;
typedef PY_INT64_T stwodigits; /* signed variant of twodigits */
#define PyLong_SHIFT 30
#elif PYLONG_BITS_IN_DIGIT == 15
typedef unsigned short digit;
typedef short sdigit; /* signed variant of digit */
#define BASE_TWODIGITS_TYPE long
typedef unsigned BASE_TWODIGITS_TYPE twodigits;
typedef BASE_TWODIGITS_TYPE stwodigits; /* signed variant of twodigits */
typedef unsigned long twodigits;
typedef long stwodigits; /* signed variant of twodigits */
#define PyLong_SHIFT 15
#else
#error "PYLONG_BITS_IN_DIGIT should be 15 or 30"
#endif
#define PyLong_BASE ((digit)1 << PyLong_SHIFT)
#define PyLong_MASK ((digit)(PyLong_BASE - 1))

View File

@ -26,6 +26,7 @@ PyAPI_FUNC(Py_ssize_t) PyLong_AsSsize_t(PyObject *);
PyAPI_FUNC(size_t) PyLong_AsSize_t(PyObject *);
PyAPI_FUNC(unsigned long) PyLong_AsUnsignedLong(PyObject *);
PyAPI_FUNC(unsigned long) PyLong_AsUnsignedLongMask(PyObject *);
PyAPI_FUNC(PyObject *) PyLong_GetInfo(void);
/* It may be useful in the future. I've added it in the PyInt -> PyLong
cleanup to keep the extra information. [CH] */

View File

@ -69,6 +69,57 @@ Used in: PY_LONG_LONG
#endif
#endif /* HAVE_LONG_LONG */
/* a build with 30-bit digits for Python long integers needs an exact-width
* 32-bit unsigned integer type to store those digits. (We could just use
* type 'unsigned long', but that would be wasteful on a system where longs
* are 64-bits.) On Unix systems, the autoconf macro AC_TYPE_UINT32_T defines
* uint32_t to be such a type unless stdint.h or inttypes.h defines uint32_t.
* However, it doesn't set HAVE_UINT32_T, so we do that here.
*/
#if (defined UINT32_MAX || defined uint32_t)
#ifndef PY_UINT32_T
#define HAVE_UINT32_T 1
#define PY_UINT32_T uint32_t
#endif
#endif
/* Macros for a 64-bit unsigned integer type; used for type 'twodigits' in the
* long integer implementation, when 30-bit digits are enabled.
*/
#if (defined UINT64_MAX || defined uint64_t)
#ifndef PY_UINT64_T
#define HAVE_UINT64_T 1
#define PY_UINT64_T uint64_t
#endif
#endif
/* Signed variants of the above */
#if (defined INT32_MAX || defined int32_t)
#ifndef PY_INT32_T
#define HAVE_INT32_T 1
#define PY_INT32_T int32_t
#endif
#endif
#if (defined INT64_MAX || defined int64_t)
#ifndef PY_INT64_T
#define HAVE_INT64_T 1
#define PY_INT64_T int64_t
#endif
#endif
/* If PYLONG_BITS_IN_DIGIT is not defined then we'll use 30-bit digits if all
the necessary integer types are available, and we're on a 64-bit platform
(as determined by SIZEOF_VOID_P); otherwise we use 15-bit digits. */
#ifndef PYLONG_BITS_IN_DIGIT
#if (defined HAVE_UINT64_T && defined HAVE_INT64_T && \
defined HAVE_UINT32_T && defined HAVE_INT32_T && SIZEOF_VOID_P >= 8)
#define PYLONG_BITS_IN_DIGIT 30
#else
#define PYLONG_BITS_IN_DIGIT 15
#endif
#endif
/* uintptr_t is the C9X name for an unsigned integral type such that a
* legitimate void* can be cast to uintptr_t and then back to void* again
* without loss of information. Similarly for intptr_t, wrt a signed

View File

@ -15,7 +15,7 @@ class Frm(object):
return self.format % self.args
# SHIFT should match the value in longintrepr.h for best testing.
SHIFT = 15
SHIFT = sys.int_info.bits_per_digit
BASE = 2 ** SHIFT
MASK = BASE - 1
KARATSUBA_CUTOFF = 70 # from longobject.c
@ -120,6 +120,35 @@ class LongTest(unittest.TestCase):
y = self.getran(leny) or 1
self.check_division(x, y)
# specific numbers chosen to exercise corner cases of the
# current long division implementation
# 30-bit cases involving a quotient digit estimate of BASE+1
self.check_division(1231948412290879395966702881,
1147341367131428698)
self.check_division(815427756481275430342312021515587883,
707270836069027745)
self.check_division(627976073697012820849443363563599041,
643588798496057020)
self.check_division(1115141373653752303710932756325578065,
1038556335171453937726882627)
# 30-bit cases that require the post-subtraction correction step
self.check_division(922498905405436751940989320930368494,
949985870686786135626943396)
self.check_division(768235853328091167204009652174031844,
1091555541180371554426545266)
# 15-bit cases involving a quotient digit estimate of BASE+1
self.check_division(20172188947443, 615611397)
self.check_division(1020908530270155025, 950795710)
self.check_division(128589565723112408, 736393718)
self.check_division(609919780285761575, 18613274546784)
# 15-bit cases that require the post-subtraction correction step
self.check_division(710031681576388032, 26769404391308)
self.check_division(1933622614268221, 30212853348836)
def test_karatsuba(self):
digits = list(range(1, 5)) + list(range(KARATSUBA_CUTOFF,
KARATSUBA_CUTOFF + 10))

View File

@ -333,6 +333,9 @@ class SysModuleTest(unittest.TestCase):
self.assert_(isinstance(sys.executable, str))
self.assertEqual(len(sys.float_info), 11)
self.assertEqual(sys.float_info.radix, 2)
self.assertEqual(len(sys.int_info), 2)
self.assert_(sys.int_info.bits_per_digit % 5 == 0)
self.assert_(sys.int_info.sizeof_digit >= 1)
self.assert_(isinstance(sys.hexversion, int))
self.assert_(isinstance(sys.maxsize, int))
self.assert_(isinstance(sys.maxunicode, int))
@ -437,6 +440,7 @@ class SizeofTest(unittest.TestCase):
if hasattr(sys, "gettotalrefcount"):
self.header += '2P'
self.vheader += '2P'
self.longdigit = sys.int_info.sizeof_digit
import _testcapi
self.gc_headsize = _testcapi.SIZEOF_PYGC_HEAD
self.file = open(test.support.TESTFN, 'wb')
@ -471,7 +475,7 @@ class SizeofTest(unittest.TestCase):
size = self.calcsize
gc_header_size = self.gc_headsize
# bool objects are not gc tracked
self.assertEqual(sys.getsizeof(True), size(vh) + self.H)
self.assertEqual(sys.getsizeof(True), size(vh) + self.longdigit)
# but lists are
self.assertEqual(sys.getsizeof([]), size(vh + 'PP') + gc_header_size)
@ -479,8 +483,8 @@ class SizeofTest(unittest.TestCase):
h = self.header
vh = self.vheader
size = self.calcsize
self.assertEqual(sys.getsizeof(True), size(vh) + self.H)
self.assertEqual(sys.getsizeof(True, -1), size(vh) + self.H)
self.assertEqual(sys.getsizeof(True), size(vh) + self.longdigit)
self.assertEqual(sys.getsizeof(True, -1), size(vh) + self.longdigit)
def test_objecttypes(self):
# check all types defined in Objects/
@ -489,7 +493,7 @@ class SizeofTest(unittest.TestCase):
size = self.calcsize
check = self.check_sizeof
# bool
check(True, size(vh) + self.H)
check(True, size(vh) + self.longdigit)
# buffer
# XXX
# builtin_function_or_method
@ -607,11 +611,12 @@ class SizeofTest(unittest.TestCase):
check(reversed([]), size(h + 'lP'))
# long
check(0, size(vh))
check(1, size(vh) + self.H)
check(-1, size(vh) + self.H)
check(32768, size(vh) + 2*self.H)
check(32768*32768-1, size(vh) + 2*self.H)
check(32768*32768, size(vh) + 3*self.H)
check(1, size(vh) + self.longdigit)
check(-1, size(vh) + self.longdigit)
PyLong_BASE = 2**sys.int_info.bits_per_digit
check(PyLong_BASE, size(vh) + 2*self.longdigit)
check(PyLong_BASE**2-1, size(vh) + 2*self.longdigit)
check(PyLong_BASE**2, size(vh) + 3*self.longdigit)
# memory
check(memoryview(b''), size(h + 'P PP2P2i7P'))
# module

View File

@ -12,6 +12,13 @@ What's New in Python 3.1 alpha 2?
Core and Builtins
-----------------
- Issue #4258: Make it possible to use base 2**30 instead of base
2**15 for the internal representation of integers, for performance
reasons. Base 2**30 is enabled by default on 64-bit machines. Add
--enable-big-digits option to configure, which overrides the
default. Add sys.int_info structseq to provide information about
the internal format.
- Issue #4474: PyUnicode_FromWideChar now converts characters outside
the BMP to surrogate pairs, on systems with sizeof(wchar_t) == 4
and sizeof(Py_UNICODE) == 2.

View File

@ -4,6 +4,7 @@
#include "Python.h"
#include "longintrepr.h"
#include "structseq.h"
#include <ctype.h>
#include <stddef.h>
@ -204,6 +205,7 @@ PyLong_FromLong(long ival)
return (PyObject*)v;
}
#if PyLONG_SHIFT==15
/* 2 digits */
if (!(abs_ival >> 2*PyLong_SHIFT)) {
v = _PyLong_New(2);
@ -216,6 +218,7 @@ PyLong_FromLong(long ival)
}
return (PyObject*)v;
}
#endif
/* Larger numbers: loop to determine number of digits */
t = abs_ival;
@ -2864,10 +2867,20 @@ long_mul(PyLongObject *a, PyLongObject *b)
CHECK_BINOP(a, b);
/* fast path for single-digit multiplication */
if (ABS(Py_SIZE(a)) <= 1 && ABS(Py_SIZE(b)) <= 1) {
PyObject *r;
r = PyLong_FromLong(MEDIUM_VALUE(a)*MEDIUM_VALUE(b));
return r;
stwodigits v = (stwodigits)(MEDIUM_VALUE(a)) * MEDIUM_VALUE(b);
#ifdef HAVE_LONG_LONG
return PyLong_FromLongLong((PY_LONG_LONG)v);
#else
/* if we don't have long long then we're almost certainly
using 15-bit digits, so v will fit in a long. In the
unlikely event that we're using 30-bit digits on a platform
without long long, a large v will just cause us to fall
through to the general multiplication code below. */
if (v >= LONG_MIN && v <= LONG_MAX)
return PyLong_FromLong((long)v);
#endif
}
z = k_mul(a, b);
@ -3991,6 +4004,45 @@ PyTypeObject PyLong_Type = {
PyObject_Del, /* tp_free */
};
static PyTypeObject Int_InfoType;
PyDoc_STRVAR(int_info__doc__,
"sys.int_info\n\
\n\
A struct sequence that holds information about Python's\n\
internal representation of integers. The attributes are read only.");
static PyStructSequence_Field int_info_fields[] = {
{"bits_per_digit", "size of a digit in bits"},
{"sizeof_digit", "size in bytes of the C type used to "
"represent a digit"},
{NULL, NULL}
};
static PyStructSequence_Desc int_info_desc = {
"sys.int_info", /* name */
int_info__doc__, /* doc */
int_info_fields, /* fields */
2 /* number of fields */
};
PyObject *
PyLong_GetInfo(void)
{
PyObject* int_info;
int field = 0;
int_info = PyStructSequence_New(&Int_InfoType);
if (int_info == NULL)
return NULL;
PyStructSequence_SET_ITEM(int_info, field++, PyLong_FromLong(PyLong_SHIFT));
PyStructSequence_SET_ITEM(int_info, field++, PyLong_FromLong(sizeof(digit)));
if (PyErr_Occurred()) {
Py_CLEAR(int_info);
return NULL;
}
return int_info;
}
int
_PyLong_Init(void)
{
@ -4023,6 +4075,10 @@ _PyLong_Init(void)
v->ob_digit[0] = abs(ival);
}
#endif
/* initialize int_info */
if (Int_InfoType.tp_name == 0)
PyStructSequence_InitType(&Int_InfoType, &int_info_desc);
return 1;
}

View File

@ -396,6 +396,42 @@ Py_NO_ENABLE_SHARED to find out. Also support MS_NO_COREDLL for b/w compat */
#endif
/* define signed and unsigned exact-width 32-bit and 64-bit types, used in the
implementation of Python long integers. */
#ifndef PY_UINT32_T
#if SIZEOF_INT == 4
#define HAVE_UINT32_T 1
#define PY_UINT32_T unsigned int
#elif SIZEOF_LONG == 4
#define HAVE_UINT32_T 1
#define PY_UINT32_T unsigned long
#endif
#endif
#ifndef PY_UINT64_T
#if SIZEOF_LONG_LONG == 8
#define HAVE_UINT64_T 1
#define PY_UINT64_T unsigned PY_LONG_LONG
#endif
#endif
#ifndef PY_INT32_T
#if SIZEOF_INT == 4
#define HAVE_INT32_T 1
#define PY_INT32_T int
#elif SIZEOF_LONG == 4
#define HAVE_INT32_T 1
#define PY_INT32_T long
#endif
#endif
#ifndef PY_INT64_T
#if SIZEOF_LONG_LONG == 8
#define HAVE_INT64_T 1
#define PY_INT64_T PY_LONG_LONG
#endif
#endif
/* Fairly standard from here! */
/* Define to 1 if you have the `copysign' function. */

View File

@ -11,6 +11,8 @@
#include "code.h"
#include "marshal.h"
#define ABS(x) ((x) < 0 ? -(x) : (x))
/* High water mark to determine when the marshalled object is dangerously deep
* and risks coring the interpreter. When the object stack gets this deep,
* raise an exception instead of continuing.
@ -122,6 +124,56 @@ w_long64(long x, WFILE *p)
}
#endif
/* We assume that Python longs are stored internally in base some power of
2**15; for the sake of portability we'll always read and write them in base
exactly 2**15. */
#define PyLong_MARSHAL_SHIFT 15
#define PyLong_MARSHAL_BASE ((short)1 << PyLong_MARSHAL_SHIFT)
#define PyLong_MARSHAL_MASK (PyLong_MARSHAL_BASE - 1)
#if PyLong_SHIFT % PyLong_MARSHAL_SHIFT != 0
#error "PyLong_SHIFT must be a multiple of PyLong_MARSHAL_SHIFT"
#endif
#define PyLong_MARSHAL_RATIO (PyLong_SHIFT / PyLong_MARSHAL_SHIFT)
static void
w_PyLong(const PyLongObject *ob, WFILE *p)
{
Py_ssize_t i, j, n, l;
digit d;
w_byte(TYPE_LONG, p);
if (Py_SIZE(ob) == 0) {
w_long((long)0, p);
return;
}
/* set l to number of base PyLong_MARSHAL_BASE digits */
n = ABS(Py_SIZE(ob));
l = (n-1) * PyLong_MARSHAL_RATIO;
d = ob->ob_digit[n-1];
assert(d != 0); /* a PyLong is always normalized */
do {
d >>= PyLong_MARSHAL_SHIFT;
l++;
} while (d != 0);
w_long((long)(Py_SIZE(ob) > 0 ? l : -l), p);
for (i=0; i < n-1; i++) {
d = ob->ob_digit[i];
for (j=0; j < PyLong_MARSHAL_RATIO; j++) {
w_short(d & PyLong_MARSHAL_MASK, p);
d >>= PyLong_MARSHAL_SHIFT;
}
assert (d == 0);
}
d = ob->ob_digit[n-1];
do {
w_short(d & PyLong_MARSHAL_MASK, p);
d >>= PyLong_MARSHAL_SHIFT;
} while (d != 0);
}
static void
w_object(PyObject *v, WFILE *p)
{
@ -155,13 +207,7 @@ w_object(PyObject *v, WFILE *p)
if ((x == -1) && PyErr_Occurred()) {
PyLongObject *ob = (PyLongObject *)v;
PyErr_Clear();
w_byte(TYPE_LONG, p);
n = Py_SIZE(ob);
w_long((long)n, p);
if (n < 0)
n = -n;
for (i = 0; i < n; i++)
w_short(ob->ob_digit[i], p);
w_PyLong(ob, p);
}
else {
#if SIZEOF_LONG > 4
@ -480,6 +526,56 @@ r_long64(RFILE *p)
#endif
}
static PyObject *
r_PyLong(RFILE *p)
{
PyLongObject *ob;
int size, i, j, md;
long n;
digit d;
n = r_long(p);
if (n == 0)
return (PyObject *)_PyLong_New(0);
if (n < -INT_MAX || n > INT_MAX) {
PyErr_SetString(PyExc_ValueError,
"bad marshal data (long size out of range)");
return NULL;
}
size = 1 + (ABS(n)-1) / PyLong_MARSHAL_RATIO;
ob = _PyLong_New(size);
if (ob == NULL)
return NULL;
Py_SIZE(ob) = n > 0 ? size : -size;
for (i = 0; i < size-1; i++) {
d = 0;
for (j=0; j < PyLong_MARSHAL_RATIO; j++) {
md = r_short(p);
if (md < 0 || md > PyLong_MARSHAL_BASE)
goto bad_digit;
d += (digit)md << j*PyLong_MARSHAL_SHIFT;
}
ob->ob_digit[i] = d;
}
d = 0;
for (j=0; j < (ABS(n)-1)%PyLong_MARSHAL_RATIO + 1; j++) {
md = r_short(p);
if (md < 0 || md > PyLong_MARSHAL_BASE)
goto bad_digit;
d += (digit)md << j*PyLong_MARSHAL_SHIFT;
}
ob->ob_digit[size-1] = d;
return (PyObject *)ob;
bad_digit:
Py_DECREF(ob);
PyErr_SetString(PyExc_ValueError,
"bad marshal data (digit out of range in long)");
return NULL;
}
static PyObject *
r_object(RFILE *p)
{
@ -544,38 +640,8 @@ r_object(RFILE *p)
break;
case TYPE_LONG:
{
int size;
PyLongObject *ob;
n = r_long(p);
if (n < -INT_MAX || n > INT_MAX) {
PyErr_SetString(PyExc_ValueError,
"bad marshal data (long size out of range)");
retval = NULL;
retval = r_PyLong(p);
break;
}
size = n<0 ? -n : n;
ob = _PyLong_New(size);
if (ob == NULL) {
retval = NULL;
break;
}
Py_SIZE(ob) = n;
for (i = 0; i < size; i++) {
int digit = r_short(p);
if (digit < 0) {
Py_DECREF(ob);
PyErr_SetString(PyExc_ValueError,
"bad marshal data (negative digit in long)");
ob = NULL;
break;
}
if (ob != NULL)
ob->ob_digit[i] = digit;
}
retval = (PyObject *)ob;
break;
}
case TYPE_FLOAT:
{

View File

@ -1012,6 +1012,7 @@ PyDoc_STR(
Static objects:\n\
\n\
float_info -- a dict with information about the float implementation.\n\
int_info -- a struct sequence with information about the int implementation.\n\
maxsize -- the largest supported length of containers.\n\
maxunicode -- the largest supported character\n\
builtin_module_names -- tuple of module names built into this interpreter\n\
@ -1375,6 +1376,8 @@ _PySys_Init(void)
PyLong_FromSsize_t(PY_SSIZE_T_MAX));
SET_SYS_FROM_STRING("float_info",
PyFloat_GetInfo());
SET_SYS_FROM_STRING("int_info",
PyLong_GetInfo());
SET_SYS_FROM_STRING("maxunicode",
PyLong_FromLong(PyUnicode_GetMax()));
SET_SYS_FROM_STRING("builtin_module_names",

413
configure vendored
View File

@ -1320,6 +1320,8 @@ Optional Features:
--enable-profiling enable C-level code profiling
--enable-ipv6 Enable ipv6 (with ipv4) support
--disable-ipv6 Disable ipv6 support
--enable-big-digits[=BITS]
use big digits for Python longs [BITS=30]
Optional Packages:
--with-PACKAGE[=ARG] use PACKAGE [ARG=yes]
@ -6965,6 +6967,386 @@ _ACEOF
fi
{ echo "$as_me:$LINENO: checking for uint32_t" >&5
echo $ECHO_N "checking for uint32_t... $ECHO_C" >&6; }
if test "${ac_cv_c_uint32_t+set}" = set; then
echo $ECHO_N "(cached) $ECHO_C" >&6
else
ac_cv_c_uint32_t=no
for ac_type in 'uint32_t' 'unsigned int' 'unsigned long int' \
'unsigned long long int' 'unsigned short int' 'unsigned char'; do
cat >conftest.$ac_ext <<_ACEOF
/* confdefs.h. */
_ACEOF
cat confdefs.h >>conftest.$ac_ext
cat >>conftest.$ac_ext <<_ACEOF
/* end confdefs.h. */
$ac_includes_default
int
main ()
{
static int test_array [1 - 2 * !(($ac_type) -1 >> (32 - 1) == 1)];
test_array [0] = 0
;
return 0;
}
_ACEOF
rm -f conftest.$ac_objext
if { (ac_try="$ac_compile"
case "(($ac_try" in
*\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
*) ac_try_echo=$ac_try;;
esac
eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
(eval "$ac_compile") 2>conftest.er1
ac_status=$?
grep -v '^ *+' conftest.er1 >conftest.err
rm -f conftest.er1
cat conftest.err >&5
echo "$as_me:$LINENO: \$? = $ac_status" >&5
(exit $ac_status); } && {
test -z "$ac_c_werror_flag" ||
test ! -s conftest.err
} && test -s conftest.$ac_objext; then
case $ac_type in
uint32_t) ac_cv_c_uint32_t=yes ;;
*) ac_cv_c_uint32_t=$ac_type ;;
esac
else
echo "$as_me: failed program was:" >&5
sed 's/^/| /' conftest.$ac_ext >&5
fi
rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
test "$ac_cv_c_uint32_t" != no && break
done
fi
{ echo "$as_me:$LINENO: result: $ac_cv_c_uint32_t" >&5
echo "${ECHO_T}$ac_cv_c_uint32_t" >&6; }
case $ac_cv_c_uint32_t in #(
no|yes) ;; #(
*)
cat >>confdefs.h <<\_ACEOF
#define _UINT32_T 1
_ACEOF
cat >>confdefs.h <<_ACEOF
#define uint32_t $ac_cv_c_uint32_t
_ACEOF
;;
esac
{ echo "$as_me:$LINENO: checking for uint64_t" >&5
echo $ECHO_N "checking for uint64_t... $ECHO_C" >&6; }
if test "${ac_cv_c_uint64_t+set}" = set; then
echo $ECHO_N "(cached) $ECHO_C" >&6
else
ac_cv_c_uint64_t=no
for ac_type in 'uint64_t' 'unsigned int' 'unsigned long int' \
'unsigned long long int' 'unsigned short int' 'unsigned char'; do
cat >conftest.$ac_ext <<_ACEOF
/* confdefs.h. */
_ACEOF
cat confdefs.h >>conftest.$ac_ext
cat >>conftest.$ac_ext <<_ACEOF
/* end confdefs.h. */
$ac_includes_default
int
main ()
{
static int test_array [1 - 2 * !(($ac_type) -1 >> (64 - 1) == 1)];
test_array [0] = 0
;
return 0;
}
_ACEOF
rm -f conftest.$ac_objext
if { (ac_try="$ac_compile"
case "(($ac_try" in
*\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
*) ac_try_echo=$ac_try;;
esac
eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
(eval "$ac_compile") 2>conftest.er1
ac_status=$?
grep -v '^ *+' conftest.er1 >conftest.err
rm -f conftest.er1
cat conftest.err >&5
echo "$as_me:$LINENO: \$? = $ac_status" >&5
(exit $ac_status); } && {
test -z "$ac_c_werror_flag" ||
test ! -s conftest.err
} && test -s conftest.$ac_objext; then
case $ac_type in
uint64_t) ac_cv_c_uint64_t=yes ;;
*) ac_cv_c_uint64_t=$ac_type ;;
esac
else
echo "$as_me: failed program was:" >&5
sed 's/^/| /' conftest.$ac_ext >&5
fi
rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
test "$ac_cv_c_uint64_t" != no && break
done
fi
{ echo "$as_me:$LINENO: result: $ac_cv_c_uint64_t" >&5
echo "${ECHO_T}$ac_cv_c_uint64_t" >&6; }
case $ac_cv_c_uint64_t in #(
no|yes) ;; #(
*)
cat >>confdefs.h <<\_ACEOF
#define _UINT64_T 1
_ACEOF
cat >>confdefs.h <<_ACEOF
#define uint64_t $ac_cv_c_uint64_t
_ACEOF
;;
esac
{ echo "$as_me:$LINENO: checking for int32_t" >&5
echo $ECHO_N "checking for int32_t... $ECHO_C" >&6; }
if test "${ac_cv_c_int32_t+set}" = set; then
echo $ECHO_N "(cached) $ECHO_C" >&6
else
ac_cv_c_int32_t=no
for ac_type in 'int32_t' 'int' 'long int' \
'long long int' 'short int' 'signed char'; do
cat >conftest.$ac_ext <<_ACEOF
/* confdefs.h. */
_ACEOF
cat confdefs.h >>conftest.$ac_ext
cat >>conftest.$ac_ext <<_ACEOF
/* end confdefs.h. */
$ac_includes_default
int
main ()
{
static int test_array [1 - 2 * !(0 < ($ac_type) (((($ac_type) 1 << (32 - 2)) - 1) * 2 + 1))];
test_array [0] = 0
;
return 0;
}
_ACEOF
rm -f conftest.$ac_objext
if { (ac_try="$ac_compile"
case "(($ac_try" in
*\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
*) ac_try_echo=$ac_try;;
esac
eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
(eval "$ac_compile") 2>conftest.er1
ac_status=$?
grep -v '^ *+' conftest.er1 >conftest.err
rm -f conftest.er1
cat conftest.err >&5
echo "$as_me:$LINENO: \$? = $ac_status" >&5
(exit $ac_status); } && {
test -z "$ac_c_werror_flag" ||
test ! -s conftest.err
} && test -s conftest.$ac_objext; then
cat >conftest.$ac_ext <<_ACEOF
/* confdefs.h. */
_ACEOF
cat confdefs.h >>conftest.$ac_ext
cat >>conftest.$ac_ext <<_ACEOF
/* end confdefs.h. */
$ac_includes_default
int
main ()
{
static int test_array [1 - 2 * !(($ac_type) (((($ac_type) 1 << (32 - 2)) - 1) * 2 + 1)
< ($ac_type) (((($ac_type) 1 << (32 - 2)) - 1) * 2 + 2))];
test_array [0] = 0
;
return 0;
}
_ACEOF
rm -f conftest.$ac_objext
if { (ac_try="$ac_compile"
case "(($ac_try" in
*\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
*) ac_try_echo=$ac_try;;
esac
eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
(eval "$ac_compile") 2>conftest.er1
ac_status=$?
grep -v '^ *+' conftest.er1 >conftest.err
rm -f conftest.er1
cat conftest.err >&5
echo "$as_me:$LINENO: \$? = $ac_status" >&5
(exit $ac_status); } && {
test -z "$ac_c_werror_flag" ||
test ! -s conftest.err
} && test -s conftest.$ac_objext; then
:
else
echo "$as_me: failed program was:" >&5
sed 's/^/| /' conftest.$ac_ext >&5
case $ac_type in
int32_t) ac_cv_c_int32_t=yes ;;
*) ac_cv_c_int32_t=$ac_type ;;
esac
fi
rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
else
echo "$as_me: failed program was:" >&5
sed 's/^/| /' conftest.$ac_ext >&5
fi
rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
test "$ac_cv_c_int32_t" != no && break
done
fi
{ echo "$as_me:$LINENO: result: $ac_cv_c_int32_t" >&5
echo "${ECHO_T}$ac_cv_c_int32_t" >&6; }
case $ac_cv_c_int32_t in #(
no|yes) ;; #(
*)
cat >>confdefs.h <<_ACEOF
#define int32_t $ac_cv_c_int32_t
_ACEOF
;;
esac
{ echo "$as_me:$LINENO: checking for int64_t" >&5
echo $ECHO_N "checking for int64_t... $ECHO_C" >&6; }
if test "${ac_cv_c_int64_t+set}" = set; then
echo $ECHO_N "(cached) $ECHO_C" >&6
else
ac_cv_c_int64_t=no
for ac_type in 'int64_t' 'int' 'long int' \
'long long int' 'short int' 'signed char'; do
cat >conftest.$ac_ext <<_ACEOF
/* confdefs.h. */
_ACEOF
cat confdefs.h >>conftest.$ac_ext
cat >>conftest.$ac_ext <<_ACEOF
/* end confdefs.h. */
$ac_includes_default
int
main ()
{
static int test_array [1 - 2 * !(0 < ($ac_type) (((($ac_type) 1 << (64 - 2)) - 1) * 2 + 1))];
test_array [0] = 0
;
return 0;
}
_ACEOF
rm -f conftest.$ac_objext
if { (ac_try="$ac_compile"
case "(($ac_try" in
*\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
*) ac_try_echo=$ac_try;;
esac
eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
(eval "$ac_compile") 2>conftest.er1
ac_status=$?
grep -v '^ *+' conftest.er1 >conftest.err
rm -f conftest.er1
cat conftest.err >&5
echo "$as_me:$LINENO: \$? = $ac_status" >&5
(exit $ac_status); } && {
test -z "$ac_c_werror_flag" ||
test ! -s conftest.err
} && test -s conftest.$ac_objext; then
cat >conftest.$ac_ext <<_ACEOF
/* confdefs.h. */
_ACEOF
cat confdefs.h >>conftest.$ac_ext
cat >>conftest.$ac_ext <<_ACEOF
/* end confdefs.h. */
$ac_includes_default
int
main ()
{
static int test_array [1 - 2 * !(($ac_type) (((($ac_type) 1 << (64 - 2)) - 1) * 2 + 1)
< ($ac_type) (((($ac_type) 1 << (64 - 2)) - 1) * 2 + 2))];
test_array [0] = 0
;
return 0;
}
_ACEOF
rm -f conftest.$ac_objext
if { (ac_try="$ac_compile"
case "(($ac_try" in
*\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
*) ac_try_echo=$ac_try;;
esac
eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
(eval "$ac_compile") 2>conftest.er1
ac_status=$?
grep -v '^ *+' conftest.er1 >conftest.err
rm -f conftest.er1
cat conftest.err >&5
echo "$as_me:$LINENO: \$? = $ac_status" >&5
(exit $ac_status); } && {
test -z "$ac_c_werror_flag" ||
test ! -s conftest.err
} && test -s conftest.$ac_objext; then
:
else
echo "$as_me: failed program was:" >&5
sed 's/^/| /' conftest.$ac_ext >&5
case $ac_type in
int64_t) ac_cv_c_int64_t=yes ;;
*) ac_cv_c_int64_t=$ac_type ;;
esac
fi
rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
else
echo "$as_me: failed program was:" >&5
sed 's/^/| /' conftest.$ac_ext >&5
fi
rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
test "$ac_cv_c_int64_t" != no && break
done
fi
{ echo "$as_me:$LINENO: result: $ac_cv_c_int64_t" >&5
echo "${ECHO_T}$ac_cv_c_int64_t" >&6; }
case $ac_cv_c_int64_t in #(
no|yes) ;; #(
*)
cat >>confdefs.h <<_ACEOF
#define int64_t $ac_cv_c_int64_t
_ACEOF
;;
esac
{ echo "$as_me:$LINENO: checking for ssize_t" >&5
echo $ECHO_N "checking for ssize_t... $ECHO_C" >&6; }
if test "${ac_cv_type_ssize_t+set}" = set; then
@ -21687,6 +22069,37 @@ fi
LIBS=$LIBS_SAVE
# determine what size digit to use for Python's longs
{ echo "$as_me:$LINENO: checking digit size for Python's longs" >&5
echo $ECHO_N "checking digit size for Python's longs... $ECHO_C" >&6; }
# Check whether --enable-big-digits was given.
if test "${enable_big_digits+set}" = set; then
enableval=$enable_big_digits; case $enable_big_digits in
yes)
enable_big_digits=30 ;;
no)
enable_big_digits=15 ;;
15|30)
;;
*)
{ { echo "$as_me:$LINENO: error: bad value $enable_big_digits for --enable-big-digits; value should be 15 or 30" >&5
echo "$as_me: error: bad value $enable_big_digits for --enable-big-digits; value should be 15 or 30" >&2;}
{ (exit 1); exit 1; }; } ;;
esac
{ echo "$as_me:$LINENO: result: $enable_big_digits" >&5
echo "${ECHO_T}$enable_big_digits" >&6; }
cat >>confdefs.h <<_ACEOF
#define PYLONG_BITS_IN_DIGIT $enable_big_digits
_ACEOF
else
{ echo "$as_me:$LINENO: result: no value specified" >&5
echo "${ECHO_T}no value specified" >&6; }
fi
# check for wchar.h
if test "${ac_cv_header_wchar_h+set}" = set; then
{ echo "$as_me:$LINENO: checking for wchar.h" >&5

View File

@ -1288,6 +1288,10 @@ AC_TYPE_PID_T
AC_TYPE_SIGNAL
AC_TYPE_SIZE_T
AC_TYPE_UID_T
AC_TYPE_UINT32_T
AC_TYPE_UINT64_T
AC_TYPE_INT32_T
AC_TYPE_INT64_T
AC_CHECK_TYPE(ssize_t,
AC_DEFINE(HAVE_SSIZE_T, 1, Define if your compiler provides ssize_t),,)
@ -3125,6 +3129,25 @@ AC_CHECK_DECLS([isinf, isnan, isfinite], [], [], [[#include <math.h>]])
LIBS=$LIBS_SAVE
# determine what size digit to use for Python's longs
AC_MSG_CHECKING([digit size for Python's longs])
AC_ARG_ENABLE(big-digits,
AC_HELP_STRING([--enable-big-digits@<:@=BITS@:>@],[use big digits for Python longs [[BITS=30]]]),
[case $enable_big_digits in
yes)
enable_big_digits=30 ;;
no)
enable_big_digits=15 ;;
[15|30])
;;
*)
AC_MSG_ERROR([bad value $enable_big_digits for --enable-big-digits; value should be 15 or 30]) ;;
esac
AC_MSG_RESULT($enable_big_digits)
AC_DEFINE_UNQUOTED(PYLONG_BITS_IN_DIGIT, $enable_big_digits, [Define as the preferred size in bits of long digits])
],
[AC_MSG_RESULT(no value specified)])
# check for wchar.h
AC_CHECK_HEADER(wchar.h, [
AC_DEFINE(HAVE_WCHAR_H, 1,

View File

@ -855,6 +855,9 @@
/* Defined if PTHREAD_SCOPE_SYSTEM supported. */
#undef PTHREAD_SYSTEM_SCHED_SUPPORTED
/* Define as the preferred size in bits of long digits */
#undef PYLONG_BITS_IN_DIGIT
/* Define to printf format modifier for Py_ssize_t */
#undef PY_FORMAT_SIZE_T
@ -1036,6 +1039,16 @@
/* Define to force use of thread-safe errno, h_errno, and other functions */
#undef _REENTRANT
/* Define for Solaris 2.5.1 so the uint32_t typedef from <sys/synch.h>,
<pthread.h>, or <semaphore.h> is not used. If the typedef was allowed, the
#define below would cause a syntax error. */
#undef _UINT32_T
/* Define for Solaris 2.5.1 so the uint64_t typedef from <sys/synch.h>,
<pthread.h>, or <semaphore.h> is not used. If the typedef was allowed, the
#define below would cause a syntax error. */
#undef _UINT64_T
/* Define to the level of X/Open that your system supports */
#undef _XOPEN_SOURCE
@ -1062,6 +1075,14 @@
/* Define to `int' if <sys/types.h> doesn't define. */
#undef gid_t
/* Define to the type of a signed integer type of width exactly 32 bits if
such a type exists and the standard includes do not define it. */
#undef int32_t
/* Define to the type of a signed integer type of width exactly 64 bits if
such a type exists and the standard includes do not define it. */
#undef int64_t
/* Define to `int' if <sys/types.h> does not define. */
#undef mode_t
@ -1083,6 +1104,14 @@
/* Define to `int' if <sys/types.h> doesn't define. */
#undef uid_t
/* Define to the type of an unsigned integer type of width exactly 32 bits if
such a type exists and the standard includes do not define it. */
#undef uint32_t
/* Define to the type of an unsigned integer type of width exactly 64 bits if
such a type exists and the standard includes do not define it. */
#undef uint64_t
/* Define to empty if the keyword does not work. */
#undef volatile