Two new private longobject API functions,
_PyLong_FromByteArray
_PyLong_AsByteArray
Untested and probably buggy -- they compile OK, but nothing calls them
yet. Will soon be called by the struct module, to implement x-platform
'q' and 'Q'.
If other people have uses for them, we could move them into the public API.
See longobject.h for usage details.
diff --git a/Include/longobject.h b/Include/longobject.h
index 7ebceec..3c6fde0 100644
--- a/Include/longobject.h
+++ b/Include/longobject.h
@@ -44,6 +44,46 @@
DL_IMPORT(PyObject *) PyLong_FromString(char *, char **, int);
DL_IMPORT(PyObject *) PyLong_FromUnicode(Py_UNICODE*, int, int);
+/* _PyLong_FromByteArray: View the n unsigned bytes as a binary integer in
+ base 256, and return a Python long with the same numeric value.
+ If n is 0, the integer is 0. Else:
+ If little_endian is 1/true, bytes[n-1] is the MSB and bytes[0] the LSB;
+ else (little_endian is 0/false) bytes[0] is the MSB and bytes[n-1] the
+ LSB.
+ If is_signed is 0/false, view the bytes as a non-negative integer.
+ If is_signed is 1/true, view the bytes as a 2's-complement integer,
+ non-negative if bit 0x80 of the MSB is clear, negative if set.
+ Error returns:
+ + Return NULL with the appropriate exception set if there's not
+ enough memory to create the Python long.
+*/
+extern DL_IMPORT(PyObject *) _PyLong_FromByteArray(
+ const unsigned char* bytes, size_t n,
+ int little_endian, int is_signed);
+
+/* _PyLong_AsByteArray: Convert the least-significant 8*n bits of long
+ v to a base-256 integer, stored in array bytes. Normally return 0,
+ return -1 on error.
+ If little_endian is 1/true, store the MSB at bytes[n-1] and the LSB at
+ bytes[0]; else (little_endian is 0/false) store the MSB at bytes[0] and
+ the LSB at bytes[n-1].
+ If is_signed is 0/false, it's an error if v < 0; else (v >= 0) n bytes
+ are filled and there's nothing special about bit 0x80 of the MSB.
+ If is_signed is 1/true, bytes is filled with the 2's-complement
+ representation of v's value. Bit 0x80 of the MSB is the sign bit.
+ Error returns (-1):
+ + is_signed is 0 and v < 0. TypeError is set in this case, and bytes
+ isn't altered.
+ + n isn't big enough to hold the full mathematical value of v. For
+ example, if is_signed is 0 and there are more digits in the v than
+ fit in n; or if is_signed is 1, v < 0, and n is just 1 bit shy of
+ being large enough to hold a sign bit. OverflowError is set in this
+ case, but bytes holds the least-signficant n bytes of the true value.
+*/
+extern DL_IMPORT(int) _PyLong_AsByteArray(PyLongObject* v,
+ unsigned char* bytes, size_t n,
+ int little_endian, int is_signed);
+
#ifdef __cplusplus
}
#endif
diff --git a/Objects/longobject.c b/Objects/longobject.c
index e762665..92f8b04 100644
--- a/Objects/longobject.c
+++ b/Objects/longobject.c
@@ -211,6 +211,219 @@
return x;
}
+PyObject *
+_PyLong_FromByteArray(const unsigned char* bytes, size_t n,
+ int little_endian, int is_signed)
+{
+ const unsigned char* pstartbyte;/* LSB of bytes */
+ int incr; /* direction to move pstartbyte */
+ const unsigned char* pendbyte; /* MSB of bytes */
+ size_t numsignificantbytes; /* number of bytes that matter */
+ size_t ndigits; /* number of Python long digits */
+ PyLongObject* v; /* result */
+ int idigit = 0; /* next free index in v->ob_digit */
+
+ if (n == 0)
+ return PyLong_FromLong(0L);
+
+ if (little_endian) {
+ pstartbyte = bytes;
+ pendbyte = bytes + n - 1;
+ incr = 1;
+ }
+ else {
+ pstartbyte = bytes + n - 1;
+ pendbyte = bytes;
+ incr = -1;
+ }
+
+ if (is_signed)
+ is_signed = *pendbyte >= 0x80;
+
+ /* Compute numsignificantbytes. This consists of finding the most
+ significant byte. Leading 0 bytes are insignficant if the number
+ is positive, and leading 0xff bytes if negative. */
+ {
+ size_t i;
+ const unsigned char* p = pendbyte;
+ const int pincr = -incr; /* search MSB to LSB */
+ const unsigned char insignficant = is_signed ? 0xff : 0x00;
+
+ for (i = 0; i < n; ++i, p += pincr) {
+ if (*p != insignficant)
+ break;
+ }
+ numsignificantbytes = n - i;
+ /* 2's-comp is a bit tricky here, e.g. 0xff00 == -0x0100, so
+ actually has 2 significant bytes. OTOH, 0xff0001 ==
+ -0x00ffff, so we wouldn't *need* to bump it there; but we
+ do for 0xffff = -0x0001. To be safe without bothering to
+ check every case, bump it regardless. */
+ if (is_signed && numsignificantbytes < n)
+ ++numsignificantbytes;
+ }
+
+ /* How many Python long digits do we need? We have
+ 8*numsignificantbytes bits, and each Python long digit has SHIFT
+ bits, so it's the ceiling of the quotient. */
+ ndigits = (numsignificantbytes * 8 + SHIFT - 1) / SHIFT;
+ if (ndigits > (size_t)INT_MAX)
+ return PyErr_NoMemory();
+ v = _PyLong_New((int)ndigits);
+ if (v == NULL)
+ return NULL;
+
+ /* Copy the bits over. The tricky parts are computing 2's-comp on
+ the fly for signed numbers, and dealing with the mismatch between
+ 8-bit bytes and (probably) 15-bit Python digits.*/
+ {
+ size_t i;
+ unsigned int carry = 1; /* for 2's-comp calculation */
+ twodigits accum = 0; /* sliding register */
+ unsigned int accumbits = 0; /* number of bits in accum */
+ const unsigned char* p = pstartbyte;
+
+ for (i = 0; i < numsignificantbytes; ++i, p += incr) {
+ unsigned int thisbyte = *p;
+ /* Compute correction for 2's comp, if needed. */
+ if (is_signed) {
+ thisbyte = (0xff ^ thisbyte) + carry;
+ carry = thisbyte >> 8;
+ thisbyte &= 0xff;
+ }
+ /* Because we're going LSB to MSB, thisbyte is
+ more significant than what's already in accum,
+ so needs to be prepended to accum. */
+ accum |= thisbyte << accumbits;
+ accumbits += 8;
+ if (accumbits >= SHIFT) {
+ /* There's enough to fill a Python digit. */
+ assert(idigit < (int)ndigits);
+ v->ob_digit[idigit] = (digit)(accum & MASK);
+ ++idigit;
+ accum >>= SHIFT;
+ accumbits -= SHIFT;
+ assert(accumbits < SHIFT);
+ }
+ }
+ assert(accumbits < SHIFT);
+ if (accumbits) {
+ assert(idigit < (int)ndigits);
+ v->ob_digit[idigit] = (digit)accum;
+ ++idigit;
+ }
+ }
+
+ v->ob_size = is_signed ? -idigit : idigit;
+ return (PyObject *)long_normalize(v);
+}
+
+int
+_PyLong_AsByteArray(PyLongObject* v,
+ unsigned char* bytes, size_t n,
+ int little_endian, int is_signed)
+{
+ int i; /* index into v->ob_digit */
+ int ndigits; /* |v->ob_size| */
+ twodigits accum; /* sliding register */
+ unsigned int accumbits; /* # bits in accum */
+ int do_twos_comp; /* store 2's-comp? is_signed and v < 0 */
+ twodigits carry; /* for computing 2's-comp */
+ size_t j; /* # bytes filled */
+ unsigned char* p; /* pointer to next byte in bytes */
+ int pincr; /* direction to move p */
+
+ assert(v != NULL && PyLong_Check(v));
+
+ if (v->ob_size < 0) {
+ ndigits = -(v->ob_size);
+ if (!is_signed) {
+ PyErr_SetString(PyExc_TypeError,
+ "can't convert negative long to unsigned");
+ return -1;
+ }
+ do_twos_comp = 1;
+ }
+ else {
+ ndigits = v->ob_size;
+ do_twos_comp = 0;
+ }
+
+ if (little_endian) {
+ p = bytes;
+ pincr = 1;
+ }
+ else {
+ p = bytes + n - 1;
+ pincr = -1;
+ }
+
+ /* Copy over all the Python digits. */
+ j = 0;
+ accum = 0;
+ accumbits = 0;
+ carry = do_twos_comp ? 1 : 0;
+ for (i = 0; i < ndigits; ++i) {
+ twodigits thisdigit = v->ob_digit[i];
+ if (do_twos_comp) {
+ thisdigit = (thisdigit ^ MASK) + carry;
+ carry = thisdigit >> SHIFT;
+ thisdigit &= MASK;
+ }
+ /* Because we're going LSB to MSB, thisdigit is more
+ significant than what's already in accum, so needs to be
+ prepended to accum. */
+ accum |= thisdigit << accumbits;
+ accumbits += SHIFT;
+ /* Store as many bytes as possible. */
+ assert(accumbits >= 8);
+ do {
+ if (j >= n)
+ goto Overflow;
+ ++j;
+ *p = (unsigned char)(accum & 0xff);
+ p += pincr;
+ accumbits -= 8;
+ accum >>= 8;
+ } while (accumbits >= 8);
+ }
+
+ /* Store the straggler (if any). */
+ assert(accumbits < 8);
+ assert(carry == 0); /* else do_twos_comp and *every* digit was 0 */
+ if (accum) {
+ if (j >= n)
+ goto Overflow;
+ ++j;
+ if (do_twos_comp) {
+ /* Fill leading bits of the byte with sign bits
+ (appropriately pretending that the long had an
+ infinite supply of sign bits). */
+ accum |= (~(twodigits)0) << accumbits;
+ }
+ *p = (unsigned char)(accum & 0xff);
+ p += pincr;
+ }
+
+ /* Fill remaining bytes with copies of the sign bit. */
+ for ( ; j < n; ++j, p += pincr)
+ *p = (unsigned char)(do_twos_comp ? 0xff : 0);
+
+ /* Check for delicate overflow (not enough room for the sign bit). */
+ if (j > 0 && is_signed) {
+ unsigned char msb = *(p - pincr);
+ int sign_bit_set = (msb & 0x80) != 0;
+ if (sign_bit_set != do_twos_comp)
+ goto Overflow;
+ }
+ return 0;
+
+Overflow:
+ PyErr_SetString(PyExc_OverflowError, "long too big to convert");
+ return -1;
+
+}
+
/* Get a C double from a long int object. */
double