summaryrefslogtreecommitdiffstatshomepage
path: root/py/objstr.c
diff options
context:
space:
mode:
Diffstat (limited to 'py/objstr.c')
-rw-r--r--py/objstr.c50
1 files changed, 25 insertions, 25 deletions
diff --git a/py/objstr.c b/py/objstr.c
index b13517b63d..63d394e0a9 100644
--- a/py/objstr.c
+++ b/py/objstr.c
@@ -226,9 +226,9 @@ wrong_args:
// like strstr but with specified length and allows \0 bytes
// TODO replace with something more efficient/standard
-STATIC const byte *find_subbytes(const byte *haystack, machine_uint_t hlen, const byte *needle, machine_uint_t nlen, machine_int_t direction) {
+STATIC const byte *find_subbytes(const byte *haystack, mp_uint_t hlen, const byte *needle, mp_uint_t nlen, mp_int_t direction) {
if (hlen >= nlen) {
- machine_uint_t str_index, str_index_end;
+ mp_uint_t str_index, str_index_end;
if (direction > 0) {
str_index = 0;
str_index_end = hlen - nlen;
@@ -348,7 +348,7 @@ uncomparable:
// objstrunicode defines own version
const byte *str_index_to_ptr(const mp_obj_type_t *type, const byte *self_data, uint self_len,
mp_obj_t index, bool is_slice) {
- machine_uint_t index_val = mp_get_index(type, self_len, index, is_slice);
+ mp_uint_t index_val = mp_get_index(type, self_len, index, is_slice);
return self_data + index_val;
}
#endif
@@ -370,7 +370,7 @@ STATIC mp_obj_t str_subscr(mp_obj_t self_in, mp_obj_t index, mp_obj_t value) {
#endif
const byte *p = str_index_to_ptr(type, self_data, self_len, index, false);
if (type == &mp_type_bytes) {
- return MP_OBJ_NEW_SMALL_INT((mp_small_int_t)*p);
+ return MP_OBJ_NEW_SMALL_INT((mp_int_t)*p);
} else {
return mp_obj_new_str((char*)p, 1, true);
}
@@ -435,7 +435,7 @@ STATIC mp_obj_t str_join(mp_obj_t self_in, mp_obj_t arg) {
STATIC mp_obj_t str_split(uint n_args, const mp_obj_t *args) {
const mp_obj_type_t *self_type = mp_obj_get_type(args[0]);
- machine_int_t splits = -1;
+ mp_int_t splits = -1;
mp_obj_t sep = mp_const_none;
if (n_args > 1) {
sep = args[1];
@@ -515,8 +515,8 @@ STATIC mp_obj_t str_rsplit(uint n_args, const mp_obj_t *args) {
mp_obj_t sep = args[1];
GET_STR_DATA_LEN(args[0], s, len);
- machine_int_t splits = mp_obj_get_int(args[2]);
- machine_int_t org_splits = splits;
+ mp_int_t splits = mp_obj_get_int(args[2]);
+ mp_int_t org_splits = splits;
// Preallocate list to the max expected # of elements, as we
// will fill it from the end.
mp_obj_list_t *res = mp_obj_new_list(splits + 1, NULL);
@@ -566,7 +566,7 @@ STATIC mp_obj_t str_rsplit(uint n_args, const mp_obj_t *args) {
return res;
}
-STATIC mp_obj_t str_finder(uint n_args, const mp_obj_t *args, machine_int_t direction, bool is_index) {
+STATIC mp_obj_t str_finder(uint n_args, const mp_obj_t *args, mp_int_t direction, bool is_index) {
const mp_obj_type_t *self_type = mp_obj_get_type(args[0]);
assert(2 <= n_args && n_args <= 4);
assert(MP_OBJ_IS_STR(args[0]));
@@ -670,16 +670,16 @@ STATIC mp_obj_t str_uni_strip(int type, uint n_args, const mp_obj_t *args) {
GET_STR_DATA_LEN(args[0], orig_str, orig_str_len);
- machine_uint_t first_good_char_pos = 0;
+ mp_uint_t first_good_char_pos = 0;
bool first_good_char_pos_set = false;
- machine_uint_t last_good_char_pos = 0;
- machine_uint_t i = 0;
- machine_int_t delta = 1;
+ mp_uint_t last_good_char_pos = 0;
+ mp_uint_t i = 0;
+ mp_int_t delta = 1;
if (type == RSTRIP) {
i = orig_str_len - 1;
delta = -1;
}
- for (machine_uint_t len = orig_str_len; len > 0; len--) {
+ for (mp_uint_t len = orig_str_len; len > 0; len--) {
if (find_subbytes(chars_to_del, chars_to_del_len, &orig_str[i], 1, 1) == NULL) {
if (!first_good_char_pos_set) {
first_good_char_pos_set = true;
@@ -705,7 +705,7 @@ STATIC mp_obj_t str_uni_strip(int type, uint n_args, const mp_obj_t *args) {
assert(last_good_char_pos >= first_good_char_pos);
//+1 to accomodate the last character
- machine_uint_t stripped_len = last_good_char_pos - first_good_char_pos + 1;
+ mp_uint_t stripped_len = last_good_char_pos - first_good_char_pos + 1;
if (stripped_len == orig_str_len) {
// If nothing was stripped, don't bother to dup original string
// TODO: watch out for this case when we'll get to bytearray.strip()
@@ -768,7 +768,7 @@ static mp_obj_t arg_as_int(mp_obj_t arg) {
// TODO: Needs a way to construct an mpz integer from a float
- mp_small_int_t num = mp_obj_get_float(arg);
+ mp_int_t num = mp_obj_get_float(arg);
return MP_OBJ_NEW_SMALL_INT(num);
}
#endif
@@ -1329,7 +1329,7 @@ not_enough_args:
STATIC mp_obj_t str_replace(uint n_args, const mp_obj_t *args) {
assert(MP_OBJ_IS_STR(args[0]));
- machine_int_t max_rep = -1;
+ mp_int_t max_rep = -1;
if (n_args == 4) {
max_rep = mp_obj_get_int(args[3]);
if (max_rep == 0) {
@@ -1370,11 +1370,11 @@ STATIC mp_obj_t str_replace(uint n_args, const mp_obj_t *args) {
// first pass computes the required length of the replaced string
// second pass does the replacements
for (;;) {
- machine_uint_t replaced_str_index = 0;
- machine_uint_t num_replacements_done = 0;
+ mp_uint_t replaced_str_index = 0;
+ mp_uint_t num_replacements_done = 0;
const byte *old_occurrence;
const byte *offset_ptr = str;
- machine_uint_t str_len_remain = str_len;
+ mp_uint_t str_len_remain = str_len;
if (old_len == 0) {
// if old_str is empty, copy new_str to start of replaced string
// copy the replacement string
@@ -1452,7 +1452,7 @@ STATIC mp_obj_t str_count(uint n_args, const mp_obj_t *args) {
}
// count the occurrences
- machine_int_t num_occurrences = 0;
+ mp_int_t num_occurrences = 0;
for (const byte *haystack_ptr = start; haystack_ptr + needle_len <= end;) {
if (memcmp(haystack_ptr, needle, needle_len) == 0) {
num_occurrences++;
@@ -1465,7 +1465,7 @@ STATIC mp_obj_t str_count(uint n_args, const mp_obj_t *args) {
return MP_OBJ_NEW_SMALL_INT(num_occurrences);
}
-STATIC mp_obj_t str_partitioner(mp_obj_t self_in, mp_obj_t arg, machine_int_t direction) {
+STATIC mp_obj_t str_partitioner(mp_obj_t self_in, mp_obj_t arg, mp_int_t direction) {
if (!is_str_or_bytes(self_in)) {
assert(0);
}
@@ -1491,7 +1491,7 @@ STATIC mp_obj_t str_partitioner(mp_obj_t self_in, mp_obj_t arg, machine_int_t di
const byte *position_ptr = find_subbytes(str, str_len, sep, sep_len, direction);
if (position_ptr != NULL) {
- machine_uint_t position = position_ptr - str;
+ mp_uint_t position = position_ptr - str;
result[0] = mp_obj_new_str_of_type(self_type, str, position);
result[1] = arg;
result[2] = mp_obj_new_str_of_type(self_type, str + position + sep_len, str_len - position - sep_len);
@@ -1609,7 +1609,7 @@ STATIC mp_obj_t str_encode(uint n_args, const mp_obj_t *args) {
}
#endif
-machine_int_t mp_obj_str_get_buffer(mp_obj_t self_in, mp_buffer_info_t *bufinfo, int flags) {
+mp_int_t mp_obj_str_get_buffer(mp_obj_t self_in, mp_buffer_info_t *bufinfo, int flags) {
if (flags == MP_BUFFER_READ) {
GET_STR_DATA_LEN(self_in, str_data, str_len);
bufinfo->buf = (void*)str_data;
@@ -1871,7 +1871,7 @@ const char *mp_obj_str_get_data(mp_obj_t self_in, uint *len) {
typedef struct _mp_obj_str_it_t {
mp_obj_base_t base;
mp_obj_t str;
- machine_uint_t cur;
+ mp_uint_t cur;
} mp_obj_str_it_t;
#if !MICROPY_PY_BUILTINS_STR_UNICODE
@@ -1907,7 +1907,7 @@ STATIC mp_obj_t bytes_it_iternext(mp_obj_t self_in) {
mp_obj_str_it_t *self = self_in;
GET_STR_DATA_LEN(self->str, str, len);
if (self->cur < len) {
- mp_obj_t o_out = MP_OBJ_NEW_SMALL_INT((mp_small_int_t)str[self->cur]);
+ mp_obj_t o_out = MP_OBJ_NEW_SMALL_INT((mp_int_t)str[self->cur]);
self->cur += 1;
return o_out;
} else {