objstr.c 23.5 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 "qstr.h"
11
12
13
14
15
16
#include "obj.h"
#include "runtime0.h"
#include "runtime.h"

typedef struct _mp_obj_str_t {
    mp_obj_base_t base;
17
18
19
    machine_uint_t hash : 16; // XXX here we assume the hash size is 16 bits (it is at the moment; see qstr.c)
    machine_uint_t len : 16; // len == number of bytes used in data, alloc = len + 1 because (at the moment) we also append a null byte
    byte data[];
20
21
} mp_obj_str_t;

22
23
24
25
26
27
28
29
30
// use this macro to extract the string hash
#define GET_STR_HASH(str_obj_in, str_hash) uint str_hash; if (MP_OBJ_IS_QSTR(str_obj_in)) { str_hash = qstr_hash(MP_OBJ_QSTR_VALUE(str_obj_in)); } else { str_hash = ((mp_obj_str_t*)str_obj_in)->hash; }

// use this macro to extract the string length
#define GET_STR_LEN(str_obj_in, str_len) uint str_len; if (MP_OBJ_IS_QSTR(str_obj_in)) { str_len = qstr_len(MP_OBJ_QSTR_VALUE(str_obj_in)); } else { str_len = ((mp_obj_str_t*)str_obj_in)->len; }

// use this macro to extract the string data and length
#define GET_STR_DATA_LEN(str_obj_in, str_data, str_len) const byte *str_data; uint str_len; if (MP_OBJ_IS_QSTR(str_obj_in)) { str_data = qstr_data(MP_OBJ_QSTR_VALUE(str_obj_in), &str_len); } else { str_len = ((mp_obj_str_t*)str_obj_in)->len; str_data = ((mp_obj_str_t*)str_obj_in)->data; }

31
32
STATIC mp_obj_t mp_obj_new_str_iterator(mp_obj_t str);
STATIC mp_obj_t mp_obj_new_bytes_iterator(mp_obj_t str);
xyb's avatar
xyb committed
33
34
35
36

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

37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
void mp_str_print_quoted(void (*print)(void *env, const char *fmt, ...), void *env, const byte *str_data, uint str_len) {
    // this escapes characters, but it will be very slow to print (calling print many times)
    bool has_single_quote = false;
    bool has_double_quote = false;
    for (const byte *s = str_data, *top = str_data + str_len; (!has_single_quote || !has_double_quote) && s < top; s++) {
        if (*s == '\'') {
            has_single_quote = true;
        } else if (*s == '"') {
            has_double_quote = true;
        }
    }
    int quote_char = '\'';
    if (has_single_quote && !has_double_quote) {
        quote_char = '"';
    }
    print(env, "%c", quote_char);
    for (const byte *s = str_data, *top = str_data + str_len; s < top; s++) {
        if (*s == quote_char) {
            print(env, "\\%c", quote_char);
        } else if (*s == '\\') {
            print(env, "\\\\");
        } else if (32 <= *s && *s <= 126) {
            print(env, "%c", *s);
        } else if (*s == '\n') {
            print(env, "\\n");
        // TODO add more escape codes here if we want to match CPython
        } else {
            print(env, "\\x%02x", *s);
        }
    }
    print(env, "%c", quote_char);
}

70
STATIC void str_print(void (*print)(void *env, const char *fmt, ...), void *env, mp_obj_t self_in, mp_print_kind_t kind) {
71
    GET_STR_DATA_LEN(self_in, str_data, str_len);
72
73
    bool is_bytes = MP_OBJ_IS_TYPE(self_in, &bytes_type);
    if (kind == PRINT_STR && !is_bytes) {
74
        print(env, "%.*s", str_len, str_data);
75
    } else {
76
77
78
        if (is_bytes) {
            print(env, "b");
        }
79
        mp_str_print_quoted(print, env, str_data, str_len);
80
    }
81
82
}

83
84
// like strstr but with specified length and allows \0 bytes
// TODO replace with something more efficient/standard
85
STATIC const byte *find_subbytes(const byte *haystack, uint hlen, const byte *needle, uint nlen) {
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
    if (hlen >= nlen) {
        for (uint i = 0; i <= hlen - nlen; i++) {
            bool found = true;
            for (uint j = 0; j < nlen; j++) {
                if (haystack[i + j] != needle[j]) {
                    found = false;
                    break;
                }
            }
            if (found) {
                return haystack + i;
            }
        }
    }
    return NULL;
}

103
STATIC mp_obj_t str_binary_op(int op, mp_obj_t lhs_in, mp_obj_t rhs_in) {
104
    GET_STR_DATA_LEN(lhs_in, lhs_data, lhs_len);
105
106
    switch (op) {
        case RT_BINARY_OP_SUBSCR:
107
108
109
            // 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)) {
110
                uint index = mp_get_index(mp_obj_get_type(lhs_in), lhs_len, rhs_in);
111
                if (MP_OBJ_IS_TYPE(lhs_in, &bytes_type)) {
112
                    return MP_OBJ_NEW_SMALL_INT((mp_small_int_t)lhs_data[index]);
113
114
115
                } else {
                    return mp_obj_new_str(lhs_data + index, 1, true);
                }
116
#if MICROPY_ENABLE_SLICE
117
            } else if (MP_OBJ_IS_TYPE(rhs_in, &slice_type)) {
118
                machine_uint_t start, stop;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
119
120
121
                if (!m_seq_get_fast_slice_indexes(lhs_len, rhs_in, &start, &stop)) {
                    assert(0);
                }
122
                return mp_obj_new_str(lhs_data + start, stop - start, false);
123
#endif
124
            } else {
125
126
                // Message doesn't match CPython, but we don't have so much bytes as they
                // to spend them on verbose wording
127
                nlr_jump(mp_obj_new_exception_msg(&mp_type_TypeError, "index must be int"));
128
            }
129
130
131

        case RT_BINARY_OP_ADD:
        case RT_BINARY_OP_INPLACE_ADD:
132
            if (MP_OBJ_IS_STR(rhs_in)) {
133
                // add 2 strings
134
135

                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
136
                int alloc_len = lhs_len + rhs_len;
137
138

                /* code for making qstr
139
140
141
142
                byte *q_ptr;
                byte *val = qstr_build_start(alloc_len, &q_ptr);
                memcpy(val, lhs_data, lhs_len);
                memcpy(val + lhs_len, rhs_data, rhs_len);
143
144
145
146
147
                return MP_OBJ_NEW_QSTR(qstr_build_end(q_ptr));
                */

                // code for non-qstr
                byte *data;
148
                mp_obj_t s = mp_obj_str_builder_start(mp_obj_get_type(lhs_in), alloc_len, &data);
149
150
151
                memcpy(data, lhs_data, lhs_len);
                memcpy(data + lhs_len, rhs_data, rhs_len);
                return mp_obj_str_builder_end(s);
152
153
            }
            break;
154

155
        case RT_BINARY_OP_IN:
156
            /* NOTE `a in b` is `b.__contains__(a)` */
157
158
            if (MP_OBJ_IS_STR(rhs_in)) {
                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
159
                return MP_BOOL(find_subbytes(lhs_data, lhs_len, rhs_data, rhs_len) != NULL);
160
161
            }
            break;
162

163
164
165
166
167
168
        case RT_BINARY_OP_MULTIPLY:
        {
            if (!MP_OBJ_IS_SMALL_INT(rhs_in)) {
                return NULL;
            }
            int n = MP_OBJ_SMALL_INT_VALUE(rhs_in);
169
            byte *data;
170
            mp_obj_t s = mp_obj_str_builder_start(mp_obj_get_type(lhs_in), lhs_len * n, &data);
171
172
            mp_seq_multiply(lhs_data, sizeof(*lhs_data), lhs_len, n, data);
            return mp_obj_str_builder_end(s);
173
        }
174
175
176
177
178
179
180
181
182
183
184
185

        // These 2 are never passed here, dealt with as a special case in rt_binary_op().
        //case RT_BINARY_OP_EQUAL:
        //case RT_BINARY_OP_NOT_EQUAL:
        case RT_BINARY_OP_LESS:
        case RT_BINARY_OP_LESS_EQUAL:
        case RT_BINARY_OP_MORE:
        case RT_BINARY_OP_MORE_EQUAL:
            if (MP_OBJ_IS_STR(rhs_in)) {
                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
                return MP_BOOL(mp_seq_cmp_bytes(op, lhs_data, lhs_len, rhs_data, rhs_len));
            }
186
187
188
189
190
    }

    return MP_OBJ_NULL; // op not supported
}

191
STATIC mp_obj_t str_join(mp_obj_t self_in, mp_obj_t arg) {
192
    assert(MP_OBJ_IS_STR(self_in));
193

194
    // get separation string
195
    GET_STR_DATA_LEN(self_in, sep_str, sep_len);
196
197

    // process args
198
199
200
201
202
203
204
205
206
    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;
    }
207
208
209

    // count required length
    int required_len = 0;
210
    for (int i = 0; i < seq_len; i++) {
211
        if (!MP_OBJ_IS_STR(seq_items[i])) {
212
213
            goto bad_arg;
        }
214
215
216
        if (i > 0) {
            required_len += sep_len;
        }
217
218
        GET_STR_LEN(seq_items[i], l);
        required_len += l;
219
220
221
    }

    // make joined string
222
    byte *data;
223
    mp_obj_t joined_str = mp_obj_str_builder_start(mp_obj_get_type(self_in), required_len, &data);
224
225
    for (int i = 0; i < seq_len; i++) {
        if (i > 0) {
226
227
            memcpy(data, sep_str, sep_len);
            data += sep_len;
228
        }
229
230
231
        GET_STR_DATA_LEN(seq_items[i], s, l);
        memcpy(data, s, l);
        data += l;
232
    }
233
234

    // return joined string
235
    return mp_obj_str_builder_end(joined_str);
236
237

bad_arg:
238
    nlr_jump(mp_obj_new_exception_msg(&mp_type_TypeError, "?str.join expecting a list of str's"));
239
240
}

Paul Sokolovsky's avatar
Paul Sokolovsky committed
241
242
#define is_ws(c) ((c) == ' ' || (c) == '\t')

243
STATIC mp_obj_t str_split(uint n_args, const mp_obj_t *args) {
Paul Sokolovsky's avatar
Paul Sokolovsky committed
244
245
246
247
248
249
250
251
252
    int splits = -1;
    mp_obj_t sep = mp_const_none;
    if (n_args > 1) {
        sep = args[1];
        if (n_args > 2) {
            splits = MP_OBJ_SMALL_INT_VALUE(args[2]);
        }
    }
    assert(sep == mp_const_none);
253
    (void)sep; // unused; to hush compiler warning
Paul Sokolovsky's avatar
Paul Sokolovsky committed
254
    mp_obj_t res = mp_obj_new_list(0, NULL);
255
256
257
    GET_STR_DATA_LEN(args[0], s, len);
    const byte *top = s + len;
    const byte *start;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
258
259

    // Initial whitespace is not counted as split, so we pre-do it
260
261
    while (s < top && is_ws(*s)) s++;
    while (s < top && splits != 0) {
Paul Sokolovsky's avatar
Paul Sokolovsky committed
262
        start = s;
263
264
265
        while (s < top && !is_ws(*s)) s++;
        rt_list_append(res, mp_obj_new_str(start, s - start, false));
        if (s >= top) {
Paul Sokolovsky's avatar
Paul Sokolovsky committed
266
267
            break;
        }
268
        while (s < top && is_ws(*s)) s++;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
269
270
271
272
273
        if (splits > 0) {
            splits--;
        }
    }

274
275
    if (s < top) {
        rt_list_append(res, mp_obj_new_str(s, top - s, false));
Paul Sokolovsky's avatar
Paul Sokolovsky committed
276
277
278
279
280
    }

    return res;
}

281
STATIC mp_obj_t str_find(uint n_args, const mp_obj_t *args) {
282
    assert(2 <= n_args && n_args <= 4);
283
284
    assert(MP_OBJ_IS_STR(args[0]));
    assert(MP_OBJ_IS_STR(args[1]));
285

286
287
    GET_STR_DATA_LEN(args[0], haystack, haystack_len);
    GET_STR_DATA_LEN(args[1], needle, needle_len);
288
289
290
291
292
293
294
295
296
297
298

    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]);
    }

299
    const byte *p = find_subbytes(haystack + start, haystack_len - start, needle, needle_len);
300
301
302
303
304
305
    if (p == NULL) {
        // not found
        return MP_OBJ_NEW_SMALL_INT(-1);
    } else {
        // found
        machine_int_t pos = p - haystack;
306
307
308
        if (pos + needle_len > end) {
            pos = -1;
        }
309
        return MP_OBJ_NEW_SMALL_INT(pos);
310
311
312
    }
}

313
// TODO: (Much) more variety in args
314
STATIC mp_obj_t str_startswith(mp_obj_t self_in, mp_obj_t arg) {
315
316
317
318
319
320
321
322
    GET_STR_DATA_LEN(self_in, str, str_len);
    GET_STR_DATA_LEN(arg, prefix, prefix_len);
    if (prefix_len > str_len) {
        return mp_const_false;
    }
    return MP_BOOL(memcmp(str, prefix, prefix_len) == 0);
}

323
STATIC bool chr_in_str(const byte* const str, const size_t str_len, int c) {
324
325
326
327
328
329
330
331
    for (size_t i = 0; i < str_len; i++) {
        if (str[i] == c) {
            return true;
        }
    }
    return false;
}

332
STATIC mp_obj_t str_strip(uint n_args, const mp_obj_t *args) {
xbe's avatar
xbe committed
333
    assert(1 <= n_args && n_args <= 2);
334
335
336
337
338
    assert(MP_OBJ_IS_STR(args[0]));

    const byte *chars_to_del;
    uint chars_to_del_len;
    static const byte whitespace[] = " \t\n\r\v\f";
xbe's avatar
xbe committed
339
340
341

    if (n_args == 1) {
        chars_to_del = whitespace;
342
        chars_to_del_len = sizeof(whitespace);
xbe's avatar
xbe committed
343
    } else {
344
345
346
347
        assert(MP_OBJ_IS_STR(args[1]));
        GET_STR_DATA_LEN(args[1], s, l);
        chars_to_del = s;
        chars_to_del_len = l;
xbe's avatar
xbe committed
348
349
    }

350
    GET_STR_DATA_LEN(args[0], orig_str, orig_str_len);
xbe's avatar
xbe committed
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365

    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) {
366
367
        // string is all whitespace, return ''
        return MP_OBJ_NEW_QSTR(MP_QSTR_);
xbe's avatar
xbe committed
368
369
370
371
372
    }

    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;
373
    return mp_obj_new_str(orig_str + first_good_char_pos, stripped_len, false);
xbe's avatar
xbe committed
374
375
}

376
mp_obj_t str_format(uint n_args, const mp_obj_t *args) {
377
    assert(MP_OBJ_IS_STR(args[0]));
378

379
    GET_STR_DATA_LEN(args[0], str, len);
380
381
    int arg_i = 1;
    vstr_t *vstr = vstr_new();
382
    for (const byte *top = str + len; str < top; str++) {
383
384
        if (*str == '{') {
            str++;
385
            if (str < top && *str == '{') {
386
                vstr_add_char(vstr, '{');
387
            } else {
388
                while (str < top && *str != '}') str++;
389
                if (arg_i >= n_args) {
390
                    nlr_jump(mp_obj_new_exception_msg(&mp_type_IndexError, "tuple index out of range"));
391
                }
392
                // TODO: may be PRINT_REPR depending on formatting code
393
                mp_obj_print_helper((void (*)(void*, const char*, ...))vstr_printf, vstr, args[arg_i], PRINT_STR);
394
395
396
397
398
399
400
                arg_i++;
            }
        } else {
            vstr_add_char(vstr, *str);
        }
    }

401
402
403
    mp_obj_t s = mp_obj_new_str((byte*)vstr->buf, vstr->len, false);
    vstr_free(vstr);
    return s;
404
405
}

406
STATIC mp_obj_t str_replace(uint n_args, const mp_obj_t *args) {
407
408
409
410
    assert(MP_OBJ_IS_STR(args[0]));
    assert(MP_OBJ_IS_STR(args[1]));
    assert(MP_OBJ_IS_STR(args[2]));

411
    machine_int_t max_rep = 0;
412
    if (n_args == 4) {
413
414
415
416
417
418
419
        assert(MP_OBJ_IS_SMALL_INT(args[3]));
        max_rep = MP_OBJ_SMALL_INT_VALUE(args[3]);
        if (max_rep == 0) {
            return args[0];
        } else if (max_rep < 0) {
            max_rep = 0;
        }
420
    }
421
422

    // if max_rep is still 0 by this point we will need to do all possible replacements
423
424
425
426

    GET_STR_DATA_LEN(args[0], str, str_len);
    GET_STR_DATA_LEN(args[1], old, old_len);
    GET_STR_DATA_LEN(args[2], new, new_len);
427
428

    // old won't exist in str if it's longer, so nothing to replace
429
    if (old_len > str_len) {
430
        return args[0];
431
432
    }

433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
    // data for the replaced string
    byte *data = NULL;
    mp_obj_t replaced_str = MP_OBJ_NULL;

    // do 2 passes over the string:
    //   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;
        const byte *old_occurrence;
        const byte *offset_ptr = str;
        machine_uint_t offset_num = 0;
        while ((old_occurrence = find_subbytes(offset_ptr, str_len - offset_num, old, old_len)) != NULL) {
            // copy from just after end of last occurrence of to-be-replaced string to right before start of next occurrence
            if (data != NULL) {
                memcpy(data + replaced_str_index, offset_ptr, old_occurrence - offset_ptr);
            }
            replaced_str_index += old_occurrence - offset_ptr;
            // copy the replacement string
            if (data != NULL) {
                memcpy(data + replaced_str_index, new, new_len);
            }
            replaced_str_index += new_len;
            offset_ptr = old_occurrence + old_len;
            offset_num = offset_ptr - str;

            num_replacements_done++;
            if (max_rep != 0 && num_replacements_done == max_rep){
                break;
            }
        }

        // copy from just after end of last occurrence of to-be-replaced string to end of old string
        if (data != NULL) {
            memcpy(data + replaced_str_index, offset_ptr, str_len - offset_num);
        }
        replaced_str_index += str_len - offset_num;

        if (data == NULL) {
            // first pass
            if (num_replacements_done == 0) {
                // no substr found, return original string
                return args[0];
            } else {
                // substr found, allocate new string
                replaced_str = mp_obj_str_builder_start(mp_obj_get_type(args[0]), replaced_str_index, &data);
            }
        } else {
            // second pass, we are done
            break;
        }
485
    }
486

487
488
489
    return mp_obj_str_builder_end(replaced_str);
}

490
491
492
493
494
495
496
497
498
499
500
501
502
503
STATIC machine_int_t str_get_buffer(mp_obj_t self_in, buffer_info_t *bufinfo, int flags) {
    if (flags == BUFFER_READ) {
        GET_STR_DATA_LEN(self_in, str_data, str_len);
        bufinfo->buf = (void*)str_data;
        bufinfo->len = str_len;
        return 0;
    } else {
        // can't write to a string
        bufinfo->buf = NULL;
        bufinfo->len = 0;
        return 1;
    }
}

504
505
506
507
508
509
510
STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_find_obj, 2, 4, str_find);
STATIC MP_DEFINE_CONST_FUN_OBJ_2(str_join_obj, str_join);
STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_split_obj, 1, 3, str_split);
STATIC MP_DEFINE_CONST_FUN_OBJ_2(str_startswith_obj, str_startswith);
STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_strip_obj, 1, 2, str_strip);
STATIC MP_DEFINE_CONST_FUN_OBJ_VAR(str_format_obj, 1, str_format);
STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_replace_obj, 3, 4, str_replace);
511

512
STATIC const mp_method_t str_type_methods[] = {
513
    { "find", &str_find_obj },
ian-v's avatar
ian-v committed
514
    { "join", &str_join_obj },
Paul Sokolovsky's avatar
Paul Sokolovsky committed
515
    { "split", &str_split_obj },
516
    { "startswith", &str_startswith_obj },
xbe's avatar
xbe committed
517
    { "strip", &str_strip_obj },
ian-v's avatar
ian-v committed
518
    { "format", &str_format_obj },
519
    { "replace", &str_replace_obj },
ian-v's avatar
ian-v committed
520
521
    { NULL, NULL }, // end-of-list sentinel
};
522

523
const mp_obj_type_t str_type = {
524
    { &mp_type_type },
525
    .name = MP_QSTR_str,
526
527
    .print = str_print,
    .binary_op = str_binary_op,
528
529
    .getiter = mp_obj_new_str_iterator,
    .methods = str_type_methods,
530
    .buffer_p = { .get_buffer = str_get_buffer },
531
532
533
534
};

// Reuses most of methods from str
const mp_obj_type_t bytes_type = {
535
    { &mp_type_type },
536
    .name = MP_QSTR_bytes,
537
538
539
    .print = str_print,
    .binary_op = str_binary_op,
    .getiter = mp_obj_new_bytes_iterator,
ian-v's avatar
ian-v committed
540
    .methods = str_type_methods,
541
542
};

543
mp_obj_t mp_obj_str_builder_start(const mp_obj_type_t *type, uint len, byte **data) {
544
    mp_obj_str_t *o = m_new_obj_var(mp_obj_str_t, byte, len + 1);
545
    o->base.type = type;
546
547
548
549
550
551
552
553
554
555
    o->len = len;
    *data = o->data;
    return o;
}

mp_obj_t mp_obj_str_builder_end(mp_obj_t o_in) {
    assert(MP_OBJ_IS_STR(o_in));
    mp_obj_str_t *o = o_in;
    o->hash = qstr_compute_hash(o->data, o->len);
    o->data[o->len] = '\0'; // for now we add null for compatibility with C ASCIIZ strings
556
557
558
    return o;
}

559
STATIC mp_obj_t str_new(const mp_obj_type_t *type, const byte* data, uint len) {
560
561
562
563
564
565
566
567
568
    mp_obj_str_t *o = m_new_obj_var(mp_obj_str_t, byte, len + 1);
    o->base.type = type;
    o->hash = qstr_compute_hash(data, len);
    o->len = len;
    memcpy(o->data, data, len * sizeof(byte));
    o->data[len] = '\0'; // for now we add null for compatibility with C ASCIIZ strings
    return o;
}

569
570
571
572
573
574
575
576
577
578
mp_obj_t mp_obj_new_str(const byte* data, uint len, bool make_qstr_if_not_already) {
    qstr q = qstr_find_strn(data, len);
    if (q != MP_QSTR_NULL) {
        // qstr with this data already exists
        return MP_OBJ_NEW_QSTR(q);
    } else if (make_qstr_if_not_already) {
        // no existing qstr, make a new one
        return MP_OBJ_NEW_QSTR(qstr_from_strn((const char*)data, len));
    } else {
        // no existing qstr, don't make one
579
        return str_new(&str_type, data, len);
580
581
582
    }
}

583
584
585
586
mp_obj_t mp_obj_new_bytes(const byte* data, uint len) {
    return str_new(&bytes_type, data, len);
}

587
588
589
590
591
592
593
594
595
596
597
598
599
600
bool mp_obj_str_equal(mp_obj_t s1, mp_obj_t s2) {
    if (MP_OBJ_IS_QSTR(s1) && MP_OBJ_IS_QSTR(s2)) {
        return s1 == s2;
    } else {
        GET_STR_HASH(s1, h1);
        GET_STR_HASH(s2, h2);
        if (h1 != h2) {
            return false;
        }
        GET_STR_DATA_LEN(s1, d1, l1);
        GET_STR_DATA_LEN(s2, d2, l2);
        if (l1 != l2) {
            return false;
        }
601
        return memcmp(d1, d2, l1) == 0;
602
603
604
    }
}

605
606
void bad_implicit_conversion(mp_obj_t self_in) __attribute__((noreturn));
void bad_implicit_conversion(mp_obj_t self_in) {
607
    nlr_jump(mp_obj_new_exception_msg_varg(&mp_type_TypeError, "Can't convert '%s' object to str implicitly", mp_obj_get_type_str(self_in)));
608
609
}

610
611
612
613
614
uint mp_obj_str_get_hash(mp_obj_t self_in) {
    if (MP_OBJ_IS_STR(self_in)) {
        GET_STR_HASH(self_in, h);
        return h;
    } else {
615
        bad_implicit_conversion(self_in);
616
    }
617
618
619
620
621
622
623
}

uint mp_obj_str_get_len(mp_obj_t self_in) {
    if (MP_OBJ_IS_STR(self_in)) {
        GET_STR_LEN(self_in, l);
        return l;
    } else {
624
625
626
627
628
629
630
631
632
633
634
635
636
637
        bad_implicit_conversion(self_in);
    }
}

// use this if you will anyway convert the string to a qstr
// will be more efficient for the case where it's already a qstr
qstr mp_obj_str_get_qstr(mp_obj_t self_in) {
    if (MP_OBJ_IS_QSTR(self_in)) {
        return MP_OBJ_QSTR_VALUE(self_in);
    } else if (MP_OBJ_IS_TYPE(self_in, &str_type)) {
        mp_obj_str_t *self = self_in;
        return qstr_from_strn((char*)self->data, self->len);
    } else {
        bad_implicit_conversion(self_in);
638
639
640
641
642
643
644
645
646
647
648
    }
}

// only use this function if you need the str data to be zero terminated
// at the moment all strings are zero terminated to help with C ASCIIZ compatibility
const char *mp_obj_str_get_str(mp_obj_t self_in) {
    if (MP_OBJ_IS_STR(self_in)) {
        GET_STR_DATA_LEN(self_in, s, l);
        (void)l; // len unused
        return (const char*)s;
    } else {
649
        bad_implicit_conversion(self_in);
650
651
652
    }
}

653
const char *mp_obj_str_get_data(mp_obj_t self_in, uint *len) {
654
655
656
    if (MP_OBJ_IS_STR(self_in)) {
        GET_STR_DATA_LEN(self_in, s, l);
        *len = l;
657
        return (const char*)s;
658
    } else {
659
        bad_implicit_conversion(self_in);
660
    }
661
}
xyb's avatar
xyb committed
662
663
664
665
666
667

/******************************************************************************/
/* str iterator                                                               */

typedef struct _mp_obj_str_it_t {
    mp_obj_base_t base;
668
    mp_obj_t str;
xyb's avatar
xyb committed
669
670
671
    machine_uint_t cur;
} mp_obj_str_it_t;

672
STATIC mp_obj_t str_it_iternext(mp_obj_t self_in) {
xyb's avatar
xyb committed
673
    mp_obj_str_it_t *self = self_in;
674
675
676
    GET_STR_DATA_LEN(self->str, str, len);
    if (self->cur < len) {
        mp_obj_t o_out = mp_obj_new_str(str + self->cur, 1, true);
xyb's avatar
xyb committed
677
678
679
680
681
682
683
        self->cur += 1;
        return o_out;
    } else {
        return mp_const_stop_iteration;
    }
}

684
STATIC const mp_obj_type_t str_it_type = {
685
    { &mp_type_type },
686
    .name = MP_QSTR_iterator,
687
    .iternext = str_it_iternext,
xyb's avatar
xyb committed
688
689
};

690
STATIC mp_obj_t bytes_it_iternext(mp_obj_t self_in) {
691
692
693
    mp_obj_str_it_t *self = self_in;
    GET_STR_DATA_LEN(self->str, str, len);
    if (self->cur < len) {
694
        mp_obj_t o_out = MP_OBJ_NEW_SMALL_INT((mp_small_int_t)str[self->cur]);
695
696
697
698
699
700
701
        self->cur += 1;
        return o_out;
    } else {
        return mp_const_stop_iteration;
    }
}

702
STATIC const mp_obj_type_t bytes_it_type = {
703
    { &mp_type_type },
704
    .name = MP_QSTR_iterator,
705
706
707
708
    .iternext = bytes_it_iternext,
};

mp_obj_t mp_obj_new_str_iterator(mp_obj_t str) {
xyb's avatar
xyb committed
709
710
711
    mp_obj_str_it_t *o = m_new_obj(mp_obj_str_it_t);
    o->base.type = &str_it_type;
    o->str = str;
712
713
714
715
716
717
718
719
720
    o->cur = 0;
    return o;
}

mp_obj_t mp_obj_new_bytes_iterator(mp_obj_t str) {
    mp_obj_str_it_t *o = m_new_obj(mp_obj_str_it_t);
    o->base.type = &bytes_it_type;
    o->str = str;
    o->cur = 0;
xyb's avatar
xyb committed
721
722
    return o;
}