objstr.c 11.7 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
void mp_obj_str_print_qstr(void (*print)(void *env, const char *fmt, ...), void *env, qstr q, mp_print_kind_t kind) {
26
    if (kind == PRINT_STR) {
27
        print(env, "%s", qstr_str(q));
28
29
    } else {
        // TODO need to escape chars etc
30
        print(env, "'%s'", qstr_str(q));
31
    }
32
33
}

34
35
36
37
38
void str_print(void (*print)(void *env, const char *fmt, ...), void *env, mp_obj_t self_in, mp_print_kind_t kind) {
    mp_obj_str_t *self = self_in;
    mp_obj_str_print_qstr(print, env, self->qstr, kind);
}

39
40
41
42
43
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:
44
45
46
47
            // 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
48
                // TODO: Handle negative indexes.
49
                return mp_obj_new_int(lhs_str[mp_obj_get_int(rhs_in)]);
50
#if MICROPY_ENABLE_SLICE
51
            } else if (MP_OBJ_IS_TYPE(rhs_in, &slice_type)) {
52
                machine_int_t start, stop, step;
53
54
                mp_obj_slice_get(rhs_in, &start, &stop, &step);
                assert(step == 1);
55
56
57
                int len = strlen(lhs_str);
                if (start < 0) {
                    start = len + start;
58
59
60
61
62
                    if (start < 0) {
                        start = 0;
                    }
                } else if (start > len) {
                    start = len;
63
64
65
                }
                if (stop <= 0) {
                    stop = len + stop;
66
67
68
69
70
71
                    // CPython returns empty string in such case
                    if (stop < 0) {
                        stop = start;
                    }
                } else if (stop > len) {
                    stop = len;
72
                }
73
                return mp_obj_new_str(qstr_from_strn_copy(lhs_str + start, stop - start));
74
#endif
75
            } else {
76
77
                // Message doesn't match CPython, but we don't have so much bytes as they
                // to spend them on verbose wording
78
                nlr_jump(mp_obj_new_exception_msg(MP_QSTR_TypeError, "index must be int"));
79
            }
80
81
82
83
84
85

        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);
86
87
88
                size_t lhs_len = strlen(lhs_str);
                size_t rhs_len = strlen(rhs_str);
                int alloc_len = lhs_len + rhs_len + 1;
89
                char *val = m_new(char, alloc_len);
90
91
92
                memcpy(val, lhs_str, lhs_len);
                memcpy(val + lhs_len, rhs_str, rhs_len);
                val[lhs_len + rhs_len] = '\0';
93
                return mp_obj_new_str(qstr_from_str_take(val, alloc_len));
94
95
            }
            break;
96
97
98
99
100
101
102
103
104
        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;
105
106
107
108
109
    }

    return MP_OBJ_NULL; // op not supported
}

xyb's avatar
xyb committed
110
111
112
113
static mp_obj_t str_getiter(mp_obj_t o_in) {
    return mp_obj_new_str_iterator(o_in, 0);
}

114
115
116
117
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;

118
119
120
121
122
    // get separation string
    const char *sep_str = qstr_str(self->qstr);
    size_t sep_len = strlen(sep_str);

    // process args
123
124
125
126
127
128
129
130
131
    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;
    }
132
133
134

    // count required length
    int required_len = 0;
135
136
137
138
    for (int i = 0; i < seq_len; i++) {
        if (!MP_OBJ_IS_TYPE(seq_items[i], &str_type)) {
            goto bad_arg;
        }
139
140
141
        if (i > 0) {
            required_len += sep_len;
        }
142
143
144
145
146
        required_len += strlen(qstr_str(mp_obj_str_get(seq_items[i])));
    }

    // make joined string
    char *joined_str = m_new(char, required_len + 1);
147
    char *s_dest = joined_str;
148
149
    for (int i = 0; i < seq_len; i++) {
        if (i > 0) {
150
151
            memcpy(s_dest, sep_str, sep_len);
            s_dest += sep_len;
152
        }
153
154
155
156
        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;
157
    }
158
159
160
    *s_dest = '\0';

    // return joined string
161
    return mp_obj_new_str(qstr_from_str_take(joined_str, required_len + 1));
162
163

bad_arg:
164
    nlr_jump(mp_obj_new_exception_msg(MP_QSTR_TypeError, "?str.join expecting a list of str's"));
165
166
}

xbe's avatar
xbe committed
167
168
169
170
171
172
173
174
175
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;
}

176
static mp_obj_t str_find(uint n_args, const mp_obj_t *args) {
177
178
179
180
181
182
183
184
185
186
187
188
    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);

189
190
    size_t haystack_len = strlen(haystack);
    size_t needle_len = strlen(needle);
191
192
193
194
195
196
197
198
199
200
201
202

    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);
203
204
205
206
207
208
    if (p == NULL) {
        // not found
        return MP_OBJ_NEW_SMALL_INT(-1);
    } else {
        // found
        machine_int_t pos = p - haystack;
209
210
211
        if (pos + needle_len > end) {
            pos = -1;
        }
212
        return MP_OBJ_NEW_SMALL_INT(pos);
213
214
215
    }
}

216
mp_obj_t str_strip(uint n_args, const mp_obj_t *args) {
xbe's avatar
xbe committed
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
    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);
260
    memcpy(stripped_str, orig_str + first_good_char_pos, stripped_len);
xbe's avatar
xbe committed
261
262
263
264
    stripped_str[stripped_len] = '\0';
    return mp_obj_new_str(qstr_from_str_take(stripped_str, stripped_len + 1));
}

265
mp_obj_t str_format(uint n_args, const mp_obj_t *args) {
266
267
268
269
270
271
272
273
274
275
276
    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, '{');
277
278
            } else {
                while (*str != '}') str++;
279
                if (arg_i >= n_args) {
280
                    nlr_jump(mp_obj_new_exception_msg(MP_QSTR_IndexError, "tuple index out of range"));
281
                }
282
                // TODO: may be PRINT_REPR depending on formatting code
283
                mp_obj_print_helper((void (*)(void*, const char*, ...))vstr_printf, vstr, args[arg_i], PRINT_STR);
284
285
286
287
288
289
290
                arg_i++;
            }
        } else {
            vstr_add_char(vstr, *str);
        }
    }

291
    return mp_obj_new_str(qstr_from_str_take(vstr->buf, vstr->alloc));
292
293
}

294
static MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_find_obj, 2, 4, str_find);
295
static MP_DEFINE_CONST_FUN_OBJ_2(str_join_obj, str_join);
xbe's avatar
xbe committed
296
static MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_strip_obj, 1, 2, str_strip);
297
298
static MP_DEFINE_CONST_FUN_OBJ_VAR(str_format_obj, 1, str_format);

ian-v's avatar
ian-v committed
299
static const mp_method_t str_type_methods[] = {
300
    { "find", &str_find_obj },
ian-v's avatar
ian-v committed
301
    { "join", &str_join_obj },
xbe's avatar
xbe committed
302
    { "strip", &str_strip_obj },
ian-v's avatar
ian-v committed
303
304
305
    { "format", &str_format_obj },
    { NULL, NULL }, // end-of-list sentinel
};
306

307
308
309
const mp_obj_type_t str_type = {
    { &mp_const_type },
    "str",
310
311
312
    .print = str_print,
    .binary_op = str_binary_op,
    .getiter = str_getiter,
ian-v's avatar
ian-v committed
313
    .methods = str_type_methods,
314
315
316
317
318
319
320
321
322
323
324
325
326
327
};

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
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352

/******************************************************************************/
/* 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",
353
    .iternext = str_it_iternext,
xyb's avatar
xyb committed
354
355
356
357
358
359
360
361
362
};

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;
}