objstr.c 9.62 KB
Newer Older
1
2
3
4
5
6
7
8
9
#include <stdlib.h>
#include <stdint.h>
#include <stdarg.h>
#include <string.h>
#include <assert.h>

#include "nlr.h"
#include "misc.h"
#include "mpconfig.h"
10
#include "mpqstr.h"
11
12
13
14
15
16
17
18
19
#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;

xyb's avatar
xyb committed
20
21
22
23
24
static mp_obj_t mp_obj_new_str_iterator(mp_obj_str_t *str, int cur);

/******************************************************************************/
/* str                                                                        */

25
26
27
28
29
30
31
32
33
34
35
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:
36
37
38
39
            // 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
40
                // TODO: Handle negative indexes.
41
                return mp_obj_new_int(lhs_str[mp_obj_get_int(rhs_in)]);
42
#if MICROPY_ENABLE_SLICE
43
            } else if (MP_OBJ_IS_TYPE(rhs_in, &slice_type)) {
44
                machine_int_t start, stop, step;
45
46
                mp_obj_slice_get(rhs_in, &start, &stop, &step);
                assert(step == 1);
47
48
49
                int len = strlen(lhs_str);
                if (start < 0) {
                    start = len + start;
50
51
52
53
54
                    if (start < 0) {
                        start = 0;
                    }
                } else if (start > len) {
                    start = len;
55
56
57
                }
                if (stop <= 0) {
                    stop = len + stop;
58
59
60
61
62
63
                    // CPython returns empty string in such case
                    if (stop < 0) {
                        stop = start;
                    }
                } else if (stop > len) {
                    stop = len;
64
                }
65
                return mp_obj_new_str(qstr_from_strn_copy(lhs_str + start, stop - start));
66
#endif
67
            } else {
68
69
                // Message doesn't match CPython, but we don't have so much bytes as they
                // to spend them on verbose wording
70
                nlr_jump(mp_obj_new_exception_msg(MP_QSTR_TypeError, "index must be int"));
71
            }
72
73
74
75
76
77

        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);
78
79
80
                size_t lhs_len = strlen(lhs_str);
                size_t rhs_len = strlen(rhs_str);
                int alloc_len = lhs_len + rhs_len + 1;
81
                char *val = m_new(char, alloc_len);
82
83
84
                memcpy(val, lhs_str, lhs_len);
                memcpy(val + lhs_len, rhs_str, rhs_len);
                val[lhs_len + rhs_len] = '\0';
85
                return mp_obj_new_str(qstr_from_str_take(val, alloc_len));
86
87
88
89
90
91
92
            }
            break;
    }

    return MP_OBJ_NULL; // op not supported
}

xyb's avatar
xyb committed
93
94
95
96
static mp_obj_t str_getiter(mp_obj_t o_in) {
    return mp_obj_new_str_iterator(o_in, 0);
}

97
98
99
100
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;

101
102
103
104
105
    // get separation string
    const char *sep_str = qstr_str(self->qstr);
    size_t sep_len = strlen(sep_str);

    // process args
106
107
108
109
110
111
112
113
114
    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;
    }
115
116
117

    // count required length
    int required_len = 0;
118
119
120
121
    for (int i = 0; i < seq_len; i++) {
        if (!MP_OBJ_IS_TYPE(seq_items[i], &str_type)) {
            goto bad_arg;
        }
122
123
124
        if (i > 0) {
            required_len += sep_len;
        }
125
126
127
128
129
        required_len += strlen(qstr_str(mp_obj_str_get(seq_items[i])));
    }

    // make joined string
    char *joined_str = m_new(char, required_len + 1);
130
    char *s_dest = joined_str;
131
132
    for (int i = 0; i < seq_len; i++) {
        if (i > 0) {
133
134
            memcpy(s_dest, sep_str, sep_len);
            s_dest += sep_len;
135
        }
136
137
138
139
        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;
140
    }
141
142
143
    *s_dest = '\0';

    // return joined string
144
    return mp_obj_new_str(qstr_from_str_take(joined_str, required_len + 1));
145
146

bad_arg:
147
    nlr_jump(mp_obj_new_exception_msg(MP_QSTR_TypeError, "?str.join expecting a list of str's"));
148
149
}

xbe's avatar
xbe committed
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
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;
}

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);
203
    memcpy(stripped_str, orig_str + first_good_char_pos, stripped_len);
xbe's avatar
xbe committed
204
205
206
207
    stripped_str[stripped_len] = '\0';
    return mp_obj_new_str(qstr_from_str_take(stripped_str, stripped_len + 1));
}

208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
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) {
229
                    nlr_jump(mp_obj_new_exception_msg(MP_QSTR_IndexError, "tuple index out of range"));
230
231
232
233
234
235
236
237
238
                }
                mp_obj_print_helper(vstr_printf_wrapper, vstr, args[arg_i]);
                arg_i++;
            }
        } else {
            vstr_add_char(vstr, *str);
        }
    }

239
    return mp_obj_new_str(qstr_from_str_take(vstr->buf, vstr->alloc));
240
241
242
}

static MP_DEFINE_CONST_FUN_OBJ_2(str_join_obj, str_join);
xbe's avatar
xbe committed
243
static MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_strip_obj, 1, 2, str_strip);
244
245
static MP_DEFINE_CONST_FUN_OBJ_VAR(str_format_obj, 1, str_format);

ian-v's avatar
ian-v committed
246
static const mp_method_t str_type_methods[] = {
ian-v's avatar
ian-v committed
247
    { "join", &str_join_obj },
xbe's avatar
xbe committed
248
    { "strip", &str_strip_obj },
ian-v's avatar
ian-v committed
249
250
251
    { "format", &str_format_obj },
    { NULL, NULL }, // end-of-list sentinel
};
252

253
254
255
const mp_obj_type_t str_type = {
    { &mp_const_type },
    "str",
256
257
258
    .print = str_print,
    .binary_op = str_binary_op,
    .getiter = str_getiter,
ian-v's avatar
ian-v committed
259
    .methods = str_type_methods,
260
261
262
263
264
265
266
267
268
269
270
271
272
273
};

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;
}
xyb's avatar
xyb committed
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298

/******************************************************************************/
/* 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",
299
    .iternext = str_it_iternext,
xyb's avatar
xyb committed
300
301
302
303
304
305
306
307
308
};

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;
}