2013-12-21 13:17:45 -05:00
|
|
|
#include <stdlib.h>
|
|
|
|
#include <stdint.h>
|
|
|
|
#include <stdarg.h>
|
|
|
|
#include <string.h>
|
|
|
|
#include <assert.h>
|
|
|
|
|
|
|
|
#include "nlr.h"
|
|
|
|
#include "misc.h"
|
|
|
|
#include "mpconfig.h"
|
2014-01-04 10:57:35 -05:00
|
|
|
#include "mpqstr.h"
|
2013-12-21 13:17:45 -05:00
|
|
|
#include "obj.h"
|
|
|
|
#include "runtime0.h"
|
|
|
|
#include "runtime.h"
|
|
|
|
|
|
|
|
typedef struct _mp_obj_str_t {
|
|
|
|
mp_obj_base_t base;
|
|
|
|
qstr qstr;
|
|
|
|
} mp_obj_str_t;
|
|
|
|
|
2014-01-05 05:47:51 -05:00
|
|
|
static mp_obj_t mp_obj_new_str_iterator(mp_obj_str_t *str, int cur);
|
|
|
|
|
|
|
|
/******************************************************************************/
|
|
|
|
/* str */
|
|
|
|
|
2013-12-21 13:17:45 -05:00
|
|
|
void str_print(void (*print)(void *env, const char *fmt, ...), void *env, mp_obj_t self_in) {
|
|
|
|
mp_obj_str_t *self = self_in;
|
|
|
|
// TODO need to escape chars etc
|
|
|
|
print(env, "'%s'", qstr_str(self->qstr));
|
|
|
|
}
|
|
|
|
|
|
|
|
mp_obj_t str_binary_op(int op, mp_obj_t lhs_in, mp_obj_t rhs_in) {
|
|
|
|
mp_obj_str_t *lhs = lhs_in;
|
|
|
|
const char *lhs_str = qstr_str(lhs->qstr);
|
|
|
|
switch (op) {
|
|
|
|
case RT_BINARY_OP_SUBSCR:
|
2014-01-02 19:51:16 -05:00
|
|
|
// TODO: need predicate to check for int-like type (bools are such for example)
|
|
|
|
// ["no", "yes"][1 == 2] is common idiom
|
|
|
|
if (MP_OBJ_IS_SMALL_INT(rhs_in)) {
|
|
|
|
// TODO: This implements byte string access for single index so far
|
2014-01-03 18:38:26 -05:00
|
|
|
// TODO: Handle negative indexes.
|
2014-01-02 19:51:16 -05:00
|
|
|
return mp_obj_new_int(lhs_str[mp_obj_get_int(rhs_in)]);
|
2014-01-03 18:34:23 -05:00
|
|
|
#if MICROPY_ENABLE_SLICE
|
2014-01-02 19:51:16 -05:00
|
|
|
} else if (MP_OBJ_IS_TYPE(rhs_in, &slice_type)) {
|
2014-01-03 20:06:10 -05:00
|
|
|
machine_int_t start, stop, step;
|
2014-01-02 19:51:16 -05:00
|
|
|
mp_obj_slice_get(rhs_in, &start, &stop, &step);
|
|
|
|
assert(step == 1);
|
2014-01-03 16:36:56 -05:00
|
|
|
int len = strlen(lhs_str);
|
|
|
|
if (start < 0) {
|
|
|
|
start = len + start;
|
2014-01-03 20:47:34 -05:00
|
|
|
if (start < 0) {
|
|
|
|
start = 0;
|
|
|
|
}
|
|
|
|
} else if (start > len) {
|
|
|
|
start = len;
|
2014-01-03 16:36:56 -05:00
|
|
|
}
|
|
|
|
if (stop <= 0) {
|
|
|
|
stop = len + stop;
|
2014-01-03 20:47:34 -05:00
|
|
|
// CPython returns empty string in such case
|
|
|
|
if (stop < 0) {
|
|
|
|
stop = start;
|
|
|
|
}
|
|
|
|
} else if (stop > len) {
|
|
|
|
stop = len;
|
2014-01-03 16:36:56 -05:00
|
|
|
}
|
2014-01-02 19:51:16 -05:00
|
|
|
return mp_obj_new_str(qstr_from_strn_copy(lhs_str + start, stop - start));
|
2014-01-03 18:34:23 -05:00
|
|
|
#endif
|
2014-01-02 19:51:16 -05:00
|
|
|
} else {
|
2014-01-03 18:38:26 -05:00
|
|
|
// Message doesn't match CPython, but we don't have so much bytes as they
|
|
|
|
// to spend them on verbose wording
|
2014-01-04 10:57:35 -05:00
|
|
|
nlr_jump(mp_obj_new_exception_msg(MP_QSTR_TypeError, "index must be int"));
|
2014-01-02 19:51:16 -05:00
|
|
|
}
|
2013-12-21 13:17:45 -05:00
|
|
|
|
|
|
|
case RT_BINARY_OP_ADD:
|
|
|
|
case RT_BINARY_OP_INPLACE_ADD:
|
|
|
|
if (MP_OBJ_IS_TYPE(rhs_in, &str_type)) {
|
|
|
|
// add 2 strings
|
|
|
|
const char *rhs_str = qstr_str(((mp_obj_str_t*)rhs_in)->qstr);
|
2014-01-02 11:36:09 -05:00
|
|
|
size_t lhs_len = strlen(lhs_str);
|
|
|
|
size_t rhs_len = strlen(rhs_str);
|
|
|
|
int alloc_len = lhs_len + rhs_len + 1;
|
2013-12-29 14:33:23 -05:00
|
|
|
char *val = m_new(char, alloc_len);
|
2014-01-02 11:36:09 -05:00
|
|
|
memcpy(val, lhs_str, lhs_len);
|
|
|
|
memcpy(val + lhs_len, rhs_str, rhs_len);
|
|
|
|
val[lhs_len + rhs_len] = '\0';
|
2013-12-29 14:33:23 -05:00
|
|
|
return mp_obj_new_str(qstr_from_str_take(val, alloc_len));
|
2013-12-21 13:17:45 -05:00
|
|
|
}
|
|
|
|
break;
|
2014-01-11 07:39:33 -05:00
|
|
|
case RT_COMPARE_OP_IN:
|
|
|
|
case RT_COMPARE_OP_NOT_IN:
|
|
|
|
/* NOTE `a in b` is `b.__contains__(a)` */
|
|
|
|
if (MP_OBJ_IS_TYPE(rhs_in, &str_type)) {
|
|
|
|
const char *rhs_str = qstr_str(((mp_obj_str_t*)rhs_in)->qstr);
|
|
|
|
/* FIXME \0 in strs */
|
|
|
|
return MP_BOOL((op == RT_COMPARE_OP_IN) ^ (strstr(lhs_str, rhs_str) == NULL));
|
|
|
|
}
|
|
|
|
break;
|
2013-12-21 13:17:45 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
return MP_OBJ_NULL; // op not supported
|
|
|
|
}
|
|
|
|
|
2014-01-05 05:47:51 -05:00
|
|
|
static mp_obj_t str_getiter(mp_obj_t o_in) {
|
|
|
|
return mp_obj_new_str_iterator(o_in, 0);
|
|
|
|
}
|
|
|
|
|
2013-12-21 13:17:45 -05:00
|
|
|
mp_obj_t str_join(mp_obj_t self_in, mp_obj_t arg) {
|
|
|
|
assert(MP_OBJ_IS_TYPE(self_in, &str_type));
|
|
|
|
mp_obj_str_t *self = self_in;
|
|
|
|
|
2014-01-02 11:36:09 -05:00
|
|
|
// get separation string
|
|
|
|
const char *sep_str = qstr_str(self->qstr);
|
|
|
|
size_t sep_len = strlen(sep_str);
|
|
|
|
|
|
|
|
// process args
|
2013-12-21 13:17:45 -05:00
|
|
|
uint seq_len;
|
|
|
|
mp_obj_t *seq_items;
|
|
|
|
if (MP_OBJ_IS_TYPE(arg, &tuple_type)) {
|
|
|
|
mp_obj_tuple_get(arg, &seq_len, &seq_items);
|
|
|
|
} else if (MP_OBJ_IS_TYPE(arg, &list_type)) {
|
|
|
|
mp_obj_list_get(arg, &seq_len, &seq_items);
|
|
|
|
} else {
|
|
|
|
goto bad_arg;
|
|
|
|
}
|
2014-01-02 11:36:09 -05:00
|
|
|
|
|
|
|
// count required length
|
|
|
|
int required_len = 0;
|
2013-12-21 13:17:45 -05:00
|
|
|
for (int i = 0; i < seq_len; i++) {
|
|
|
|
if (!MP_OBJ_IS_TYPE(seq_items[i], &str_type)) {
|
|
|
|
goto bad_arg;
|
|
|
|
}
|
2014-01-02 11:36:09 -05:00
|
|
|
if (i > 0) {
|
|
|
|
required_len += sep_len;
|
|
|
|
}
|
2013-12-21 13:17:45 -05:00
|
|
|
required_len += strlen(qstr_str(mp_obj_str_get(seq_items[i])));
|
|
|
|
}
|
|
|
|
|
|
|
|
// make joined string
|
|
|
|
char *joined_str = m_new(char, required_len + 1);
|
2014-01-02 11:36:09 -05:00
|
|
|
char *s_dest = joined_str;
|
2013-12-21 13:17:45 -05:00
|
|
|
for (int i = 0; i < seq_len; i++) {
|
|
|
|
if (i > 0) {
|
2014-01-02 11:36:09 -05:00
|
|
|
memcpy(s_dest, sep_str, sep_len);
|
|
|
|
s_dest += sep_len;
|
2013-12-21 13:17:45 -05:00
|
|
|
}
|
2014-01-02 11:36:09 -05:00
|
|
|
const char *s2 = qstr_str(mp_obj_str_get(seq_items[i]));
|
|
|
|
size_t s2_len = strlen(s2);
|
|
|
|
memcpy(s_dest, s2, s2_len);
|
|
|
|
s_dest += s2_len;
|
2013-12-21 13:17:45 -05:00
|
|
|
}
|
2014-01-02 11:36:09 -05:00
|
|
|
*s_dest = '\0';
|
|
|
|
|
|
|
|
// return joined string
|
2013-12-29 14:33:23 -05:00
|
|
|
return mp_obj_new_str(qstr_from_str_take(joined_str, required_len + 1));
|
2013-12-21 13:17:45 -05:00
|
|
|
|
|
|
|
bad_arg:
|
2014-01-04 10:57:35 -05:00
|
|
|
nlr_jump(mp_obj_new_exception_msg(MP_QSTR_TypeError, "?str.join expecting a list of str's"));
|
2013-12-21 13:17:45 -05:00
|
|
|
}
|
|
|
|
|
2014-01-08 17:23:45 -05:00
|
|
|
static bool chr_in_str(const char* const str, const size_t str_len, const char c) {
|
|
|
|
for (size_t i = 0; i < str_len; i++) {
|
|
|
|
if (str[i] == c) {
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2014-01-12 16:53:52 -05:00
|
|
|
static mp_obj_t str_find(int n_args, const mp_obj_t *args) {
|
|
|
|
assert(2 <= n_args && n_args <= 4);
|
|
|
|
assert(MP_OBJ_IS_TYPE(args[0], &str_type));
|
|
|
|
if (!MP_OBJ_IS_TYPE(args[1], &str_type)) {
|
|
|
|
nlr_jump(mp_obj_new_exception_msg_1_arg(
|
|
|
|
MP_QSTR_TypeError,
|
|
|
|
"Can't convert '%s' object to str implicitly",
|
|
|
|
mp_obj_get_type_str(args[1])));
|
|
|
|
}
|
|
|
|
|
|
|
|
const char* haystack = qstr_str(((mp_obj_str_t*)args[0])->qstr);
|
|
|
|
const char* needle = qstr_str(((mp_obj_str_t*)args[1])->qstr);
|
|
|
|
|
2014-01-13 14:39:01 -05:00
|
|
|
size_t haystack_len = strlen(haystack);
|
|
|
|
size_t needle_len = strlen(needle);
|
2014-01-12 16:53:52 -05:00
|
|
|
|
|
|
|
size_t start = 0;
|
|
|
|
size_t end = haystack_len;
|
|
|
|
/* TODO use a non-exception-throwing mp_get_index */
|
|
|
|
if (n_args >= 3 && args[2] != mp_const_none) {
|
|
|
|
start = mp_get_index(&str_type, haystack_len, args[2]);
|
|
|
|
}
|
|
|
|
if (n_args >= 4 && args[3] != mp_const_none) {
|
|
|
|
end = mp_get_index(&str_type, haystack_len, args[3]);
|
|
|
|
}
|
|
|
|
|
|
|
|
char *p = strstr(haystack + start, needle);
|
2014-01-13 14:39:01 -05:00
|
|
|
if (p == NULL) {
|
|
|
|
// not found
|
|
|
|
return MP_OBJ_NEW_SMALL_INT(-1);
|
|
|
|
} else {
|
|
|
|
// found
|
|
|
|
machine_int_t pos = p - haystack;
|
2014-01-12 16:53:52 -05:00
|
|
|
if (pos + needle_len > end) {
|
|
|
|
pos = -1;
|
|
|
|
}
|
2014-01-13 14:39:01 -05:00
|
|
|
return MP_OBJ_NEW_SMALL_INT(pos);
|
2014-01-12 16:53:52 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-01-08 17:23:45 -05:00
|
|
|
mp_obj_t str_strip(int n_args, const mp_obj_t *args) {
|
|
|
|
assert(1 <= n_args && n_args <= 2);
|
|
|
|
assert(MP_OBJ_IS_TYPE(args[0], &str_type));
|
|
|
|
const char *chars_to_del;
|
|
|
|
static const char whitespace[] = " \t\n\r\v\f";
|
|
|
|
|
|
|
|
if (n_args == 1) {
|
|
|
|
chars_to_del = whitespace;
|
|
|
|
} else {
|
|
|
|
assert(MP_OBJ_IS_TYPE(args[1], &str_type));
|
|
|
|
mp_obj_str_t *chars_to_del_obj = args[1];
|
|
|
|
chars_to_del = qstr_str(chars_to_del_obj->qstr);
|
|
|
|
}
|
|
|
|
|
|
|
|
const size_t chars_to_del_len = strlen(chars_to_del);
|
|
|
|
mp_obj_str_t *self = args[0];
|
|
|
|
const char *orig_str = qstr_str(self->qstr);
|
|
|
|
const size_t orig_str_len = strlen(orig_str);
|
|
|
|
|
|
|
|
size_t first_good_char_pos = 0;
|
|
|
|
bool first_good_char_pos_set = false;
|
|
|
|
size_t last_good_char_pos = 0;
|
|
|
|
for (size_t i = 0; i < orig_str_len; i++) {
|
|
|
|
if (!chr_in_str(chars_to_del, chars_to_del_len, orig_str[i])) {
|
|
|
|
last_good_char_pos = i;
|
|
|
|
if (!first_good_char_pos_set) {
|
|
|
|
first_good_char_pos = i;
|
|
|
|
first_good_char_pos_set = true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (first_good_char_pos == 0 && last_good_char_pos == 0) {
|
|
|
|
//string is all whitespace, return '\0'
|
|
|
|
char *empty = m_new(char, 1);
|
|
|
|
empty[0] = '\0';
|
|
|
|
return mp_obj_new_str(qstr_from_str_take(empty, 1));
|
|
|
|
}
|
|
|
|
|
|
|
|
assert(last_good_char_pos >= first_good_char_pos);
|
|
|
|
//+1 to accomodate the last character
|
|
|
|
size_t stripped_len = last_good_char_pos - first_good_char_pos + 1;
|
|
|
|
//+1 to accomodate '\0'
|
|
|
|
char *stripped_str = m_new(char, stripped_len + 1);
|
2014-01-09 17:04:45 -05:00
|
|
|
memcpy(stripped_str, orig_str + first_good_char_pos, stripped_len);
|
2014-01-08 17:23:45 -05:00
|
|
|
stripped_str[stripped_len] = '\0';
|
|
|
|
return mp_obj_new_str(qstr_from_str_take(stripped_str, stripped_len + 1));
|
|
|
|
}
|
|
|
|
|
2013-12-21 13:17:45 -05:00
|
|
|
void vstr_printf_wrapper(void *env, const char *fmt, ...) {
|
|
|
|
va_list args;
|
|
|
|
va_start(args, fmt);
|
|
|
|
vstr_vprintf(env, fmt, args);
|
|
|
|
va_end(args);
|
|
|
|
}
|
|
|
|
|
|
|
|
mp_obj_t str_format(int n_args, const mp_obj_t *args) {
|
|
|
|
assert(MP_OBJ_IS_TYPE(args[0], &str_type));
|
|
|
|
mp_obj_str_t *self = args[0];
|
|
|
|
|
|
|
|
const char *str = qstr_str(self->qstr);
|
|
|
|
int arg_i = 1;
|
|
|
|
vstr_t *vstr = vstr_new();
|
|
|
|
for (; *str; str++) {
|
|
|
|
if (*str == '{') {
|
|
|
|
str++;
|
|
|
|
if (*str == '{') {
|
|
|
|
vstr_add_char(vstr, '{');
|
|
|
|
} else if (*str == '}') {
|
|
|
|
if (arg_i >= n_args) {
|
2014-01-04 10:57:35 -05:00
|
|
|
nlr_jump(mp_obj_new_exception_msg(MP_QSTR_IndexError, "tuple index out of range"));
|
2013-12-21 13:17:45 -05:00
|
|
|
}
|
|
|
|
mp_obj_print_helper(vstr_printf_wrapper, vstr, args[arg_i]);
|
|
|
|
arg_i++;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
vstr_add_char(vstr, *str);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-12-29 14:33:23 -05:00
|
|
|
return mp_obj_new_str(qstr_from_str_take(vstr->buf, vstr->alloc));
|
2013-12-21 13:17:45 -05:00
|
|
|
}
|
|
|
|
|
2014-01-12 16:53:52 -05:00
|
|
|
static MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_find_obj, 2, 4, str_find);
|
2013-12-21 13:17:45 -05:00
|
|
|
static MP_DEFINE_CONST_FUN_OBJ_2(str_join_obj, str_join);
|
2014-01-08 17:23:45 -05:00
|
|
|
static MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_strip_obj, 1, 2, str_strip);
|
2013-12-21 13:17:45 -05:00
|
|
|
static MP_DEFINE_CONST_FUN_OBJ_VAR(str_format_obj, 1, str_format);
|
|
|
|
|
2014-01-06 17:14:11 -05:00
|
|
|
static const mp_method_t str_type_methods[] = {
|
2014-01-12 16:53:52 -05:00
|
|
|
{ "find", &str_find_obj },
|
2014-01-06 12:52:29 -05:00
|
|
|
{ "join", &str_join_obj },
|
2014-01-08 17:23:45 -05:00
|
|
|
{ "strip", &str_strip_obj },
|
2014-01-06 12:52:29 -05:00
|
|
|
{ "format", &str_format_obj },
|
|
|
|
{ NULL, NULL }, // end-of-list sentinel
|
|
|
|
};
|
2014-01-07 10:58:30 -05:00
|
|
|
|
2013-12-21 13:17:45 -05:00
|
|
|
const mp_obj_type_t str_type = {
|
|
|
|
{ &mp_const_type },
|
|
|
|
"str",
|
2014-01-05 15:34:09 -05:00
|
|
|
.print = str_print,
|
|
|
|
.binary_op = str_binary_op,
|
|
|
|
.getiter = str_getiter,
|
2014-01-06 12:52:29 -05:00
|
|
|
.methods = str_type_methods,
|
2013-12-21 13:17:45 -05:00
|
|
|
};
|
|
|
|
|
|
|
|
mp_obj_t mp_obj_new_str(qstr qstr) {
|
|
|
|
mp_obj_str_t *o = m_new_obj(mp_obj_str_t);
|
|
|
|
o->base.type = &str_type;
|
|
|
|
o->qstr = qstr;
|
|
|
|
return o;
|
|
|
|
}
|
|
|
|
|
|
|
|
qstr mp_obj_str_get(mp_obj_t self_in) {
|
|
|
|
assert(MP_OBJ_IS_TYPE(self_in, &str_type));
|
|
|
|
mp_obj_str_t *self = self_in;
|
|
|
|
return self->qstr;
|
|
|
|
}
|
2014-01-05 05:47:51 -05:00
|
|
|
|
|
|
|
/******************************************************************************/
|
|
|
|
/* str iterator */
|
|
|
|
|
|
|
|
typedef struct _mp_obj_str_it_t {
|
|
|
|
mp_obj_base_t base;
|
|
|
|
mp_obj_str_t *str;
|
|
|
|
machine_uint_t cur;
|
|
|
|
} mp_obj_str_it_t;
|
|
|
|
|
|
|
|
mp_obj_t str_it_iternext(mp_obj_t self_in) {
|
|
|
|
mp_obj_str_it_t *self = self_in;
|
|
|
|
const char *str = qstr_str(self->str->qstr);
|
|
|
|
if (self->cur < strlen(str)) {
|
|
|
|
mp_obj_t o_out = mp_obj_new_str(qstr_from_strn_copy(str + self->cur, 1));
|
|
|
|
self->cur += 1;
|
|
|
|
return o_out;
|
|
|
|
} else {
|
|
|
|
return mp_const_stop_iteration;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static const mp_obj_type_t str_it_type = {
|
|
|
|
{ &mp_const_type },
|
|
|
|
"str_iterator",
|
2014-01-05 15:34:09 -05:00
|
|
|
.iternext = str_it_iternext,
|
2014-01-05 05:47:51 -05:00
|
|
|
};
|
|
|
|
|
|
|
|
mp_obj_t mp_obj_new_str_iterator(mp_obj_str_t *str, int cur) {
|
|
|
|
mp_obj_str_it_t *o = m_new_obj(mp_obj_str_it_t);
|
|
|
|
o->base.type = &str_it_type;
|
|
|
|
o->str = str;
|
|
|
|
o->cur = cur;
|
|
|
|
return o;
|
|
|
|
}
|