Use uint16_t instead of short in audioop.

This commit is contained in:
Serhiy Storchaka 2016-09-11 14:48:16 +03:00
parent 5f1a5187f7
commit 8be17409b7
1 changed files with 14 additions and 14 deletions

View File

@ -240,7 +240,7 @@ static unsigned char
st_linear2alaw(int16_t pcm_val) /* 2's complement (13-bit range) */ st_linear2alaw(int16_t pcm_val) /* 2's complement (13-bit range) */
{ {
int16_t mask; int16_t mask;
short seg; int16_t seg;
unsigned char aval; unsigned char aval;
/* A-law using even bit inversion */ /* A-law using even bit inversion */
@ -294,7 +294,7 @@ static const int stepsizeTable[89] = {
#define GETINT8(cp, i) GETINTX(signed char, (cp), (i)) #define GETINT8(cp, i) GETINTX(signed char, (cp), (i))
#define GETINT16(cp, i) GETINTX(short, (cp), (i)) #define GETINT16(cp, i) GETINTX(int16_t, (cp), (i))
#define GETINT32(cp, i) GETINTX(int32_t, (cp), (i)) #define GETINT32(cp, i) GETINTX(int32_t, (cp), (i))
#if WORDS_BIGENDIAN #if WORDS_BIGENDIAN
@ -311,7 +311,7 @@ static const int stepsizeTable[89] = {
#define SETINT8(cp, i, val) SETINTX(signed char, (cp), (i), (val)) #define SETINT8(cp, i, val) SETINTX(signed char, (cp), (i), (val))
#define SETINT16(cp, i, val) SETINTX(short, (cp), (i), (val)) #define SETINT16(cp, i, val) SETINTX(int16_t, (cp), (i), (val))
#define SETINT32(cp, i, val) SETINTX(int32_t, (cp), (i), (val)) #define SETINT32(cp, i, val) SETINTX(int32_t, (cp), (i), (val))
#if WORDS_BIGENDIAN #if WORDS_BIGENDIAN
@ -542,7 +542,7 @@ audioop_rms_impl(PyObject *module, Py_buffer *fragment, int width)
return PyLong_FromUnsignedLong(res); return PyLong_FromUnsignedLong(res);
} }
static double _sum2(const short *a, const short *b, Py_ssize_t len) static double _sum2(const int16_t *a, const int16_t *b, Py_ssize_t len)
{ {
Py_ssize_t i; Py_ssize_t i;
double sum = 0.0; double sum = 0.0;
@ -600,7 +600,7 @@ audioop_findfit_impl(PyObject *module, Py_buffer *fragment,
Py_buffer *reference) Py_buffer *reference)
/*[clinic end generated code: output=5752306d83cbbada input=62c305605e183c9a]*/ /*[clinic end generated code: output=5752306d83cbbada input=62c305605e183c9a]*/
{ {
const short *cp1, *cp2; const int16_t *cp1, *cp2;
Py_ssize_t len1, len2; Py_ssize_t len1, len2;
Py_ssize_t j, best_j; Py_ssize_t j, best_j;
double aj_m1, aj_lm1; double aj_m1, aj_lm1;
@ -610,9 +610,9 @@ audioop_findfit_impl(PyObject *module, Py_buffer *fragment,
PyErr_SetString(AudioopError, "Strings should be even-sized"); PyErr_SetString(AudioopError, "Strings should be even-sized");
return NULL; return NULL;
} }
cp1 = (const short *)fragment->buf; cp1 = (const int16_t *)fragment->buf;
len1 = fragment->len >> 1; len1 = fragment->len >> 1;
cp2 = (const short *)reference->buf; cp2 = (const int16_t *)reference->buf;
len2 = reference->len >> 1; len2 = reference->len >> 1;
if (len1 < len2) { if (len1 < len2) {
@ -669,7 +669,7 @@ audioop_findfactor_impl(PyObject *module, Py_buffer *fragment,
Py_buffer *reference) Py_buffer *reference)
/*[clinic end generated code: output=14ea95652c1afcf8 input=816680301d012b21]*/ /*[clinic end generated code: output=14ea95652c1afcf8 input=816680301d012b21]*/
{ {
const short *cp1, *cp2; const int16_t *cp1, *cp2;
Py_ssize_t len; Py_ssize_t len;
double sum_ri_2, sum_aij_ri, result; double sum_ri_2, sum_aij_ri, result;
@ -681,8 +681,8 @@ audioop_findfactor_impl(PyObject *module, Py_buffer *fragment,
PyErr_SetString(AudioopError, "Samples should be same size"); PyErr_SetString(AudioopError, "Samples should be same size");
return NULL; return NULL;
} }
cp1 = (const short *)fragment->buf; cp1 = (const int16_t *)fragment->buf;
cp2 = (const short *)reference->buf; cp2 = (const int16_t *)reference->buf;
len = fragment->len >> 1; len = fragment->len >> 1;
sum_ri_2 = _sum2(cp2, cp2, len); sum_ri_2 = _sum2(cp2, cp2, len);
sum_aij_ri = _sum2(cp1, cp2, len); sum_aij_ri = _sum2(cp1, cp2, len);
@ -711,7 +711,7 @@ audioop_findmax_impl(PyObject *module, Py_buffer *fragment,
Py_ssize_t length) Py_ssize_t length)
/*[clinic end generated code: output=f008128233523040 input=2f304801ed42383c]*/ /*[clinic end generated code: output=f008128233523040 input=2f304801ed42383c]*/
{ {
const short *cp1; const int16_t *cp1;
Py_ssize_t len1; Py_ssize_t len1;
Py_ssize_t j, best_j; Py_ssize_t j, best_j;
double aj_m1, aj_lm1; double aj_m1, aj_lm1;
@ -721,7 +721,7 @@ audioop_findmax_impl(PyObject *module, Py_buffer *fragment,
PyErr_SetString(AudioopError, "Strings should be even-sized"); PyErr_SetString(AudioopError, "Strings should be even-sized");
return NULL; return NULL;
} }
cp1 = (const short *)fragment->buf; cp1 = (const int16_t *)fragment->buf;
len1 = fragment->len >> 1; len1 = fragment->len >> 1;
if (length < 0 || len1 < length) { if (length < 0 || len1 < length) {
@ -1122,7 +1122,7 @@ audioop_bias_impl(PyObject *module, Py_buffer *fragment, int width, int bias)
if (width == 1) if (width == 1)
val = GETINTX(unsigned char, fragment->buf, i); val = GETINTX(unsigned char, fragment->buf, i);
else if (width == 2) else if (width == 2)
val = GETINTX(unsigned short, fragment->buf, i); val = GETINTX(uint16_t, fragment->buf, i);
else if (width == 3) else if (width == 3)
val = ((unsigned int)GETINT24(fragment->buf, i)) & 0xffffffu; val = ((unsigned int)GETINT24(fragment->buf, i)) & 0xffffffu;
else { else {
@ -1137,7 +1137,7 @@ audioop_bias_impl(PyObject *module, Py_buffer *fragment, int width, int bias)
if (width == 1) if (width == 1)
SETINTX(unsigned char, ncp, i, val); SETINTX(unsigned char, ncp, i, val);
else if (width == 2) else if (width == 2)
SETINTX(unsigned short, ncp, i, val); SETINTX(uint16_t, ncp, i, val);
else if (width == 3) else if (width == 3)
SETINT24(ncp, i, (int)val); SETINT24(ncp, i, (int)val);
else { else {