objstr.c 45.2 KB
Newer Older
xbe's avatar
xbe committed
1
#include <stdbool.h>
2
3
4
5
6
7
#include <string.h>
#include <assert.h>

#include "nlr.h"
#include "misc.h"
#include "mpconfig.h"
8
#include "qstr.h"
9
10
11
#include "obj.h"
#include "runtime0.h"
#include "runtime.h"
Dave Hylands's avatar
Dave Hylands committed
12
#include "pfenv.h"
13
14
15

typedef struct _mp_obj_str_t {
    mp_obj_base_t base;
16
17
    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
18
    const byte *data;
19
20
} mp_obj_str_t;

21
STATIC mp_obj_t str_modulo_format(mp_obj_t pattern, uint n_args, const mp_obj_t *args);
22
23
const mp_obj_t mp_const_empty_bytes;

24
25
26
27
28
29
30
31
32
// 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; }

33
34
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);
35
STATIC mp_obj_t str_new(const mp_obj_type_t *type, const byte* data, uint len);
xyb's avatar
xyb committed
36
37
38
39

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

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
70
71
72
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);
}

73
STATIC void str_print(void (*print)(void *env, const char *fmt, ...), void *env, mp_obj_t self_in, mp_print_kind_t kind) {
74
    GET_STR_DATA_LEN(self_in, str_data, str_len);
75
    bool is_bytes = MP_OBJ_IS_TYPE(self_in, &mp_type_bytes);
76
    if (kind == PRINT_STR && !is_bytes) {
77
        print(env, "%.*s", str_len, str_data);
78
    } else {
79
80
81
        if (is_bytes) {
            print(env, "b");
        }
82
        mp_str_print_quoted(print, env, str_data, str_len);
83
    }
84
85
}

86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
STATIC mp_obj_t str_make_new(mp_obj_t type_in, uint n_args, uint n_kw, const mp_obj_t *args) {
    switch (n_args) {
        case 0:
            return MP_OBJ_NEW_QSTR(MP_QSTR_);

        case 1:
        {
            vstr_t *vstr = vstr_new();
            mp_obj_print_helper((void (*)(void*, const char*, ...))vstr_printf, vstr, args[0], PRINT_STR);
            mp_obj_t s = mp_obj_new_str((byte*)vstr->buf, vstr->len, false);
            vstr_free(vstr);
            return s;
        }

        case 2:
        case 3:
        {
            // TODO: validate 2nd/3rd args
104
            if (!MP_OBJ_IS_TYPE(args[0], &mp_type_bytes)) {
105
106
107
108
                nlr_jump(mp_obj_new_exception_msg(&mp_type_TypeError, "bytes expected"));
            }
            GET_STR_DATA_LEN(args[0], str_data, str_len);
            GET_STR_HASH(args[0], str_hash);
109
            mp_obj_str_t *o = str_new(&mp_type_str, NULL, str_len);
110
111
112
113
114
115
116
117
118
119
            o->data = str_data;
            o->hash = str_hash;
            return o;
        }

        default:
            nlr_jump(mp_obj_new_exception_msg(&mp_type_TypeError, "str takes at most 3 arguments"));
    }
}

120
121
122
123
124
125
126
127
128
129
130
STATIC mp_obj_t bytes_make_new(mp_obj_t type_in, uint n_args, uint n_kw, const mp_obj_t *args) {
    if (n_args == 0) {
        return mp_const_empty_bytes;
    }

    if (MP_OBJ_IS_STR(args[0])) {
        if (n_args < 2 || n_args > 3) {
            goto wrong_args;
        }
        GET_STR_DATA_LEN(args[0], str_data, str_len);
        GET_STR_HASH(args[0], str_hash);
131
        mp_obj_str_t *o = str_new(&mp_type_bytes, NULL, str_len);
132
133
134
135
136
137
138
139
140
141
142
143
144
        o->data = str_data;
        o->hash = str_hash;
        return o;
    }

    if (n_args > 1) {
        goto wrong_args;
    }

    if (MP_OBJ_IS_SMALL_INT(args[0])) {
        uint len = MP_OBJ_SMALL_INT_VALUE(args[0]);
        byte *data;

145
        mp_obj_t o = mp_obj_str_builder_start(&mp_type_bytes, len, &data);
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
        memset(data, 0, len);
        return mp_obj_str_builder_end(o);
    }

    int len;
    byte *data;
    vstr_t *vstr = NULL;
    mp_obj_t o = NULL;
    // Try to create array of exact len if initializer len is known
    mp_obj_t len_in = mp_obj_len_maybe(args[0]);
    if (len_in == MP_OBJ_NULL) {
        len = -1;
        vstr = vstr_new();
    } else {
        len = MP_OBJ_SMALL_INT_VALUE(len_in);
161
        o = mp_obj_str_builder_start(&mp_type_bytes, len, &data);
162
163
    }

Damien George's avatar
Damien George committed
164
    mp_obj_t iterable = mp_getiter(args[0]);
165
    mp_obj_t item;
Damien George's avatar
Damien George committed
166
    while ((item = mp_iternext(iterable)) != MP_OBJ_NULL) {
167
168
169
170
171
172
173
174
175
176
177
        if (len == -1) {
            vstr_add_char(vstr, MP_OBJ_SMALL_INT_VALUE(item));
        } else {
            *data++ = MP_OBJ_SMALL_INT_VALUE(item);
        }
    }

    if (len == -1) {
        vstr_shrink(vstr);
        // TODO: Optimize, borrow buffer from vstr
        len = vstr_len(vstr);
178
        o = mp_obj_str_builder_start(&mp_type_bytes, len, &data);
179
180
181
182
183
184
185
186
187
188
        memcpy(data, vstr_str(vstr), len);
        vstr_free(vstr);
    }

    return mp_obj_str_builder_end(o);

wrong_args:
        nlr_jump(mp_obj_new_exception_msg(&mp_type_TypeError, "wrong number of arguments"));
}

189
190
// like strstr but with specified length and allows \0 bytes
// TODO replace with something more efficient/standard
191
STATIC const byte *find_subbytes(const byte *haystack, machine_uint_t hlen, const byte *needle, machine_uint_t nlen, machine_int_t direction) {
192
    if (hlen >= nlen) {
193
194
195
196
197
198
199
200
201
202
203
204
        machine_uint_t str_index, str_index_end;
        if (direction > 0) {
            str_index = 0;
            str_index_end = hlen - nlen;
        } else {
            str_index = hlen - nlen;
            str_index_end = 0;
        }
        for (;;) {
            if (memcmp(&haystack[str_index], needle, nlen) == 0) {
                //found
                return haystack + str_index;
205
            }
206
207
208
            if (str_index == str_index_end) {
                //not found
                break;
209
            }
210
            str_index += direction;
211
212
213
214
215
        }
    }
    return NULL;
}

216
STATIC mp_obj_t str_binary_op(int op, mp_obj_t lhs_in, mp_obj_t rhs_in) {
217
    GET_STR_DATA_LEN(lhs_in, lhs_data, lhs_len);
218
    switch (op) {
Damien George's avatar
Damien George committed
219
        case MP_BINARY_OP_SUBSCR:
220
221
222
            // 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)) {
223
                uint index = mp_get_index(mp_obj_get_type(lhs_in), lhs_len, rhs_in, false);
224
                if (MP_OBJ_IS_TYPE(lhs_in, &mp_type_bytes)) {
225
                    return MP_OBJ_NEW_SMALL_INT((mp_small_int_t)lhs_data[index]);
226
227
228
                } else {
                    return mp_obj_new_str(lhs_data + index, 1, true);
                }
229
#if MICROPY_ENABLE_SLICE
230
            } else if (MP_OBJ_IS_TYPE(rhs_in, &mp_type_slice)) {
231
                machine_uint_t start, stop;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
232
233
234
                if (!m_seq_get_fast_slice_indexes(lhs_len, rhs_in, &start, &stop)) {
                    assert(0);
                }
235
                return mp_obj_new_str(lhs_data + start, stop - start, false);
236
#endif
237
            } else {
238
239
                // Message doesn't match CPython, but we don't have so much bytes as they
                // to spend them on verbose wording
240
                nlr_jump(mp_obj_new_exception_msg(&mp_type_TypeError, "index must be int"));
241
            }
242

Damien George's avatar
Damien George committed
243
244
        case MP_BINARY_OP_ADD:
        case MP_BINARY_OP_INPLACE_ADD:
245
            if (MP_OBJ_IS_STR(rhs_in)) {
246
                // add 2 strings
247
248

                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
249
                int alloc_len = lhs_len + rhs_len;
250
251

                /* code for making qstr
252
253
254
255
                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);
256
257
258
259
260
                return MP_OBJ_NEW_QSTR(qstr_build_end(q_ptr));
                */

                // code for non-qstr
                byte *data;
261
                mp_obj_t s = mp_obj_str_builder_start(mp_obj_get_type(lhs_in), alloc_len, &data);
262
263
264
                memcpy(data, lhs_data, lhs_len);
                memcpy(data + lhs_len, rhs_data, rhs_len);
                return mp_obj_str_builder_end(s);
265
266
            }
            break;
267

Damien George's avatar
Damien George committed
268
        case MP_BINARY_OP_IN:
269
            /* NOTE `a in b` is `b.__contains__(a)` */
270
271
            if (MP_OBJ_IS_STR(rhs_in)) {
                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
272
                return MP_BOOL(find_subbytes(lhs_data, lhs_len, rhs_data, rhs_len, 1) != NULL);
273
274
            }
            break;
275

Damien George's avatar
Damien George committed
276
        case MP_BINARY_OP_MULTIPLY:
277
278
279
280
281
        {
            if (!MP_OBJ_IS_SMALL_INT(rhs_in)) {
                return NULL;
            }
            int n = MP_OBJ_SMALL_INT_VALUE(rhs_in);
282
            byte *data;
283
            mp_obj_t s = mp_obj_str_builder_start(mp_obj_get_type(lhs_in), lhs_len * n, &data);
284
285
            mp_seq_multiply(lhs_data, sizeof(*lhs_data), lhs_len, n, data);
            return mp_obj_str_builder_end(s);
286
        }
287

288
289
290
291
292
293
294
295
296
297
298
299
300
        case MP_BINARY_OP_MODULO: {
            mp_obj_t *args;
            uint n_args;
            if (MP_OBJ_IS_TYPE(rhs_in, &mp_type_tuple)) {
                // TODO: Support tuple subclasses?
                mp_obj_tuple_get(rhs_in, &n_args, &args);
            } else {
                args = &rhs_in;
                n_args = 1;
            }
            return str_modulo_format(lhs_in, n_args, args);
        }

Damien George's avatar
Damien George committed
301
302
303
304
305
306
307
        // These 2 are never passed here, dealt with as a special case in mp_binary_op().
        //case MP_BINARY_OP_EQUAL:
        //case MP_BINARY_OP_NOT_EQUAL:
        case MP_BINARY_OP_LESS:
        case MP_BINARY_OP_LESS_EQUAL:
        case MP_BINARY_OP_MORE:
        case MP_BINARY_OP_MORE_EQUAL:
308
309
310
311
            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));
            }
312
313
314
315
316
    }

    return MP_OBJ_NULL; // op not supported
}

317
STATIC mp_obj_t str_join(mp_obj_t self_in, mp_obj_t arg) {
318
    assert(MP_OBJ_IS_STR(self_in));
319

320
    // get separation string
321
    GET_STR_DATA_LEN(self_in, sep_str, sep_len);
322
323

    // process args
324
325
    uint seq_len;
    mp_obj_t *seq_items;
326
    if (MP_OBJ_IS_TYPE(arg, &mp_type_tuple)) {
327
        mp_obj_tuple_get(arg, &seq_len, &seq_items);
328
    } else if (MP_OBJ_IS_TYPE(arg, &mp_type_list)) {
329
330
331
332
        mp_obj_list_get(arg, &seq_len, &seq_items);
    } else {
        goto bad_arg;
    }
333
334
335

    // count required length
    int required_len = 0;
336
    for (int i = 0; i < seq_len; i++) {
337
        if (!MP_OBJ_IS_STR(seq_items[i])) {
338
339
            goto bad_arg;
        }
340
341
342
        if (i > 0) {
            required_len += sep_len;
        }
343
344
        GET_STR_LEN(seq_items[i], l);
        required_len += l;
345
346
347
    }

    // make joined string
348
    byte *data;
349
    mp_obj_t joined_str = mp_obj_str_builder_start(mp_obj_get_type(self_in), required_len, &data);
350
351
    for (int i = 0; i < seq_len; i++) {
        if (i > 0) {
352
353
            memcpy(data, sep_str, sep_len);
            data += sep_len;
354
        }
355
356
357
        GET_STR_DATA_LEN(seq_items[i], s, l);
        memcpy(data, s, l);
        data += l;
358
    }
359
360

    // return joined string
361
    return mp_obj_str_builder_end(joined_str);
362
363

bad_arg:
364
    nlr_jump(mp_obj_new_exception_msg(&mp_type_TypeError, "?str.join expecting a list of str's"));
365
366
}

Paul Sokolovsky's avatar
Paul Sokolovsky committed
367
368
#define is_ws(c) ((c) == ' ' || (c) == '\t')

369
STATIC mp_obj_t str_split(uint n_args, const mp_obj_t *args) {
Paul Sokolovsky's avatar
Paul Sokolovsky committed
370
371
372
373
374
375
376
377
378
    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);
379
    (void)sep; // unused; to hush compiler warning
Paul Sokolovsky's avatar
Paul Sokolovsky committed
380
    mp_obj_t res = mp_obj_new_list(0, NULL);
381
382
383
    GET_STR_DATA_LEN(args[0], s, len);
    const byte *top = s + len;
    const byte *start;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
384
385

    // Initial whitespace is not counted as split, so we pre-do it
386
387
    while (s < top && is_ws(*s)) s++;
    while (s < top && splits != 0) {
Paul Sokolovsky's avatar
Paul Sokolovsky committed
388
        start = s;
389
        while (s < top && !is_ws(*s)) s++;
390
        mp_obj_list_append(res, mp_obj_new_str(start, s - start, false));
391
        if (s >= top) {
Paul Sokolovsky's avatar
Paul Sokolovsky committed
392
393
            break;
        }
394
        while (s < top && is_ws(*s)) s++;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
395
396
397
398
399
        if (splits > 0) {
            splits--;
        }
    }

400
    if (s < top) {
401
        mp_obj_list_append(res, mp_obj_new_str(s, top - s, false));
Paul Sokolovsky's avatar
Paul Sokolovsky committed
402
403
404
405
406
    }

    return res;
}

407
STATIC mp_obj_t str_finder(uint n_args, const mp_obj_t *args, machine_int_t direction) {
408
    assert(2 <= n_args && n_args <= 4);
409
410
    assert(MP_OBJ_IS_STR(args[0]));
    assert(MP_OBJ_IS_STR(args[1]));
411

412
413
    GET_STR_DATA_LEN(args[0], haystack, haystack_len);
    GET_STR_DATA_LEN(args[1], needle, needle_len);
414

415
416
    machine_uint_t start = 0;
    machine_uint_t end = haystack_len;
417
    if (n_args >= 3 && args[2] != mp_const_none) {
418
        start = mp_get_index(&mp_type_str, haystack_len, args[2], true);
419
420
    }
    if (n_args >= 4 && args[3] != mp_const_none) {
421
        end = mp_get_index(&mp_type_str, haystack_len, args[3], true);
422
423
    }

424
    const byte *p = find_subbytes(haystack + start, end - start, needle, needle_len, direction);
425
426
427
428
429
    if (p == NULL) {
        // not found
        return MP_OBJ_NEW_SMALL_INT(-1);
    } else {
        // found
430
        return MP_OBJ_NEW_SMALL_INT(p - haystack);
431
432
433
    }
}

434
435
436
437
438
439
440
441
STATIC mp_obj_t str_find(uint n_args, const mp_obj_t *args) {
    return str_finder(n_args, args, 1);
}

STATIC mp_obj_t str_rfind(uint n_args, const mp_obj_t *args) {
    return str_finder(n_args, args, -1);
}

442
// TODO: (Much) more variety in args
443
STATIC mp_obj_t str_startswith(mp_obj_t self_in, mp_obj_t arg) {
444
445
446
447
448
449
450
451
    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);
}

452
STATIC mp_obj_t str_strip(uint n_args, const mp_obj_t *args) {
xbe's avatar
xbe committed
453
    assert(1 <= n_args && n_args <= 2);
454
455
456
457
458
    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
459
460
461

    if (n_args == 1) {
        chars_to_del = whitespace;
462
        chars_to_del_len = sizeof(whitespace);
xbe's avatar
xbe committed
463
    } else {
464
465
466
467
        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
468
469
    }

470
    GET_STR_DATA_LEN(args[0], orig_str, orig_str_len);
xbe's avatar
xbe committed
471

472
    machine_uint_t first_good_char_pos = 0;
xbe's avatar
xbe committed
473
    bool first_good_char_pos_set = false;
474
475
    machine_uint_t last_good_char_pos = 0;
    for (machine_uint_t i = 0; i < orig_str_len; i++) {
476
        if (find_subbytes(chars_to_del, chars_to_del_len, &orig_str[i], 1, 1) == NULL) {
xbe's avatar
xbe committed
477
478
479
480
481
482
483
484
485
            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) {
486
487
        // string is all whitespace, return ''
        return MP_OBJ_NEW_QSTR(MP_QSTR_);
xbe's avatar
xbe committed
488
489
490
491
    }

    assert(last_good_char_pos >= first_good_char_pos);
    //+1 to accomodate the last character
492
    machine_uint_t stripped_len = last_good_char_pos - first_good_char_pos + 1;
493
    return mp_obj_new_str(orig_str + first_good_char_pos, stripped_len, false);
xbe's avatar
xbe committed
494
495
}

Dave Hylands's avatar
Dave Hylands committed
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
// Takes an int arg, but only parses unsigned numbers, and only changes
// *num if at least one digit was parsed.
static int str_to_int(const char *str, int *num) {
    const char *s = str;
    if (unichar_isdigit(*s)) {
        *num = 0;
        do {
            *num = *num * 10 + (*s - '0');
            s++;
        }
        while (unichar_isdigit(*s));
    }
    return s - str;
}

static bool isalignment(char ch) {
    return ch && strchr("<>=^", ch) != NULL;
}

static bool istype(char ch) {
    return ch && strchr("bcdeEfFgGnosxX%", ch) != NULL;
}

static bool arg_looks_integer(mp_obj_t arg) {
    return MP_OBJ_IS_TYPE(arg, &mp_type_bool) || MP_OBJ_IS_INT(arg);
}

static bool arg_looks_numeric(mp_obj_t arg) {
    return arg_looks_integer(arg)
#if MICROPY_ENABLE_FLOAT
        || MP_OBJ_IS_TYPE(arg, &mp_type_float)
#endif
    ;
}

531
mp_obj_t str_format(uint n_args, const mp_obj_t *args) {
532
    assert(MP_OBJ_IS_STR(args[0]));
533

534
    GET_STR_DATA_LEN(args[0], str, len);
Dave Hylands's avatar
Dave Hylands committed
535
    int arg_i = 0;
536
    vstr_t *vstr = vstr_new();
Dave Hylands's avatar
Dave Hylands committed
537
538
539
540
    pfenv_t pfenv_vstr;
    pfenv_vstr.data = vstr;
    pfenv_vstr.print_strn = pfenv_vstr_add_strn;

541
    for (const byte *top = str + len; str < top; str++) {
Dave Hylands's avatar
Dave Hylands committed
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
        if (*str == '}') {
            str++;
            if (str < top && *str == '}') {
                vstr_add_char(vstr, '}');
                continue;
            }
            nlr_jump(mp_obj_new_exception_msg(&mp_type_ValueError, "Single '}' encountered in format string"));
        }
        if (*str != '{') {
            vstr_add_char(vstr, *str);
            continue;
        }

        str++;
        if (str < top && *str == '{') {
            vstr_add_char(vstr, '{');
            continue;
        }

        // replacement_field ::=  "{" [field_name] ["!" conversion] [":" format_spec] "}"

        vstr_t *field_name = NULL;
        char conversion = '\0';
        vstr_t *format_spec = NULL;

        if (str < top && *str != '}' && *str != '!' && *str != ':') {
            field_name = vstr_new();
            while (str < top && *str != '}' && *str != '!' && *str != ':') {
                vstr_add_char(field_name, *str++);
            }
            vstr_add_char(field_name, '\0');
        }

        // conversion ::=  "r" | "s"

        if (str < top && *str == '!') {
578
            str++;
Dave Hylands's avatar
Dave Hylands committed
579
580
            if (str < top && (*str == 'r' || *str == 's')) {
                conversion = *str++;
581
            } else {
Dave Hylands's avatar
Dave Hylands committed
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
                nlr_jump(mp_obj_new_exception_msg(&mp_type_ValueError, "end of format while looking for conversion specifier"));
            }
        }

        if (str < top && *str == ':') {
            str++;
            // {:} is the same as {}, which is the same as {!s}
            // This makes a difference when passing in a True or False
            // '{}'.format(True) returns 'True'
            // '{:d}'.format(True) returns '1'
            // So we treat {:} as {} and this later gets treated to be {!s}
            if (*str != '}') {
                format_spec = vstr_new(); 
                while (str < top && *str != '}') {
                    vstr_add_char(format_spec, *str++);
597
                }
Dave Hylands's avatar
Dave Hylands committed
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
                vstr_add_char(format_spec, '\0');
            }
        }
        if (str >= top) {
            nlr_jump(mp_obj_new_exception_msg(&mp_type_ValueError, "unmatched '{' in format"));
        }
        if (*str != '}') {
            nlr_jump(mp_obj_new_exception_msg(&mp_type_ValueError, "expected ':' after format specifier"));
        }

        mp_obj_t arg = mp_const_none;

        if (field_name) {
            if (arg_i > 0) {
                nlr_jump(mp_obj_new_exception_msg(&mp_type_ValueError, "cannot switch from automatic field numbering to manual field specification"));
            }
            int index;
            if (str_to_int(vstr_str(field_name), &index) != vstr_len(field_name) - 1) {
                nlr_jump(mp_obj_new_exception_msg(&mp_type_KeyError, "attributes not supported yet"));
617
            }
Dave Hylands's avatar
Dave Hylands committed
618
619
620
621
622
623
624
            if (index >= n_args - 1) {
                nlr_jump(mp_obj_new_exception_msg(&mp_type_IndexError, "tuple index out of range"));
            }
            arg = args[index + 1];
            arg_i = -1;
            vstr_free(field_name);
            field_name = NULL;
625
        } else {
Dave Hylands's avatar
Dave Hylands committed
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
            if (arg_i < 0) {
                nlr_jump(mp_obj_new_exception_msg(&mp_type_ValueError, "cannot switch from manual field specification to automatic field numbering"));
            }
            if (arg_i >= n_args - 1) {
                nlr_jump(mp_obj_new_exception_msg(&mp_type_IndexError, "tuple index out of range"));
            }
            arg = args[arg_i + 1];
            arg_i++;
        }
        if (!format_spec && !conversion) {
            conversion = 's';
        }
        if (conversion) {
            mp_print_kind_t print_kind;
            if (conversion == 's') {
                print_kind = PRINT_STR;
            } else if (conversion == 'r') {
                print_kind = PRINT_REPR;
            } else {
                nlr_jump(mp_obj_new_exception_msg_varg(&mp_type_ValueError, "Unknown conversion specifier %c", conversion));
            }
            vstr_t *arg_vstr = vstr_new();
            mp_obj_print_helper((void (*)(void*, const char*, ...))vstr_printf, arg_vstr, arg, print_kind);
            arg = mp_obj_new_str((const byte *)vstr_str(arg_vstr), vstr_len(arg_vstr), false);
            vstr_free(arg_vstr);
        }

        char sign = '\0';
        char fill = '\0';
        char align = '\0';
        int width = -1;
        int precision = -1;
        char type = '\0';
        int flags = 0;

        if (format_spec) {
            // The format specifier (from http://docs.python.org/2/library/string.html#formatspec)
            //
            // [[fill]align][sign][#][0][width][,][.precision][type]
            // fill        ::=  <any character>
            // align       ::=  "<" | ">" | "=" | "^"
            // sign        ::=  "+" | "-" | " "
            // width       ::=  integer
            // precision   ::=  integer
            // type        ::=  "b" | "c" | "d" | "e" | "E" | "f" | "F" | "g" | "G" | "n" | "o" | "s" | "x" | "X" | "%"

            const char *s = vstr_str(format_spec);
            if (isalignment(*s)) {
                align = *s++;
            } else if (*s && isalignment(s[1])) {
                fill = *s++;
                align = *s++;
            }
            if (*s == '+' || *s == '-' || *s == ' ') {
                if (*s == '+') {
                    flags |= PF_FLAG_SHOW_SIGN;
                } else if (*s == ' ') {
                    flags |= PF_FLAG_SPACE_SIGN;
                }
                sign = *s++;
            }
            if (*s == '#') {
                flags |= PF_FLAG_SHOW_PREFIX;
                s++;
            }
            if (*s == '0') {
                if (!align) {
                    align = '=';
                }
                if (!fill) {
                    fill = '0';
                }
            }
            s += str_to_int(s, &width);
            if (*s == ',') {
                flags |= PF_FLAG_SHOW_COMMA;
                s++;
            }
            if (*s == '.') {
                s++;
                s += str_to_int(s, &precision);
            }
            if (istype(*s)) {
                type = *s++;
            }
            if (*s) {
                nlr_jump(mp_obj_new_exception_msg(&mp_type_KeyError, "Invalid conversion specification"));
            }
            vstr_free(format_spec);
            format_spec = NULL;
        }
        if (!align) {
            if (arg_looks_numeric(arg)) {
                align = '>';
            } else {
                align = '<';
            }
        }
        if (!fill) {
            fill = ' ';
        }

        if (sign) {
            if (type == 's') {
                nlr_jump(mp_obj_new_exception_msg(&mp_type_ValueError, "Sign not allowed in string format specifier"));
            }
            if (type == 'c') {
                nlr_jump(mp_obj_new_exception_msg(&mp_type_ValueError, "Sign not allowed with integer format specifier 'c'"));
            }
        } else {
            sign = '-';
        }

        switch (align) {
            case '<': flags |= PF_FLAG_LEFT_ADJUST;     break;
            case '=': flags |= PF_FLAG_PAD_AFTER_SIGN;  break;
            case '^': flags |= PF_FLAG_CENTER_ADJUST;   break;
        }

        if (arg_looks_integer(arg)) {
            switch (type) {
                case 'b':
                    pfenv_print_int(&pfenv_vstr, mp_obj_get_int(arg), 1, 2, 'a', flags, fill, width);
                    continue;

                case 'c':
                {
                    char ch = mp_obj_get_int(arg);
                    pfenv_print_strn(&pfenv_vstr, &ch, 1, flags, fill, width);
                    continue;
                }

                case '\0':  // No explicit format type implies 'd'
                case 'n':   // I don't think we support locales in uPy so use 'd'
                case 'd':
                    pfenv_print_int(&pfenv_vstr, mp_obj_get_int(arg), 1, 10, 'a', flags, fill, width);
                    continue;

                case 'o':
                    pfenv_print_int(&pfenv_vstr, mp_obj_get_int(arg), 1, 8, 'a', flags, fill, width);
                    continue;

                case 'x':
                    pfenv_print_int(&pfenv_vstr, mp_obj_get_int(arg), 1, 16, 'a', flags, fill, width);
                    continue;

                case 'X':
                    pfenv_print_int(&pfenv_vstr, mp_obj_get_int(arg), 1, 16, 'A', flags, fill, width);
                    continue;

                case 'e':
                case 'E':
                case 'f':
                case 'F':
                case 'g':
                case 'G':
                case '%':
                    // The floating point formatters all work with anything that
                    // looks like an integer
                    break;

                default:
                    nlr_jump(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
                        "Unknown format code '%c' for object of type '%s'", type, mp_obj_get_type_str(arg)));
            }
791

Dave Hylands's avatar
Dave Hylands committed
792
#if MICROPY_ENABLE_FLOAT
793
        } else if (arg_looks_numeric(arg)) {
Dave Hylands's avatar
Dave Hylands committed
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
            if (!type) {

                // Even though the docs say that an unspecified type is the same
                // as 'g', there is one subtle difference, when the exponent
                // is one less than the precision.
                //  
                // '{:10.1}'.format(0.0) ==> '0e+00'
                // '{:10.1g}'.format(0.0) ==> '0'
                //
                // TODO: Figure out how to deal with this.
                //
                // A proper solution would involve adding a special flag
                // or something to format_float, and create a format_double
                // to deal with doubles. In order to fix this when using
                // sprintf, we'd need to use the e format and tweak the
                // returned result to strip trailing zeros like the g format
                // does.
                //
                // {:10.3} and {:10.2e} with 1.23e2 both produce 1.23e+02
                // but with 1.e2 you get 1e+02 and 1.00e+02
                //
                // Stripping the trailing 0's (like g) does would make the
                // e format give us the right format.
                //
                // CPython sources say:
                //   Omitted type specifier.  Behaves in the same way as repr(x)
                //   and str(x) if no precision is given, else like 'g', but with
                //   at least one digit after the decimal point. */

                type = 'g';
            }
            if (type == 'n') {
                type = 'g';
            }

            flags |= PF_FLAG_PAD_NAN_INF; // '{:06e}'.format(float('-inf')) should give '-00inf'
            switch (type) {
                case 'e':
                case 'E':
                case 'f':
                case 'F':
                case 'g':
                case 'G':
                    pfenv_print_float(&pfenv_vstr, mp_obj_get_float(arg), type, flags, fill, width, precision); 
                    break;

                case '%':
                    flags |= PF_FLAG_ADD_PERCENT;
                    pfenv_print_float(&pfenv_vstr, mp_obj_get_float(arg) * 100.0F, 'f', flags, fill, width, precision);
                    break;

                default:
                    nlr_jump(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
                        "Unknown format code '%c' for object of type 'float'",
                        type, mp_obj_get_type_str(arg)));
            }
#endif
851

Dave Hylands's avatar
Dave Hylands committed
852
        } else {
853
854
            // arg doesn't look like a number

Dave Hylands's avatar
Dave Hylands committed
855
856
857
            if (align == '=') {
                nlr_jump(mp_obj_new_exception_msg(&mp_type_ValueError, "'=' alignment not allowed in string format specifier"));
            }
858

Dave Hylands's avatar
Dave Hylands committed
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
            switch (type) {
                case '\0':
                    mp_obj_print_helper((void (*)(void*, const char*, ...))vstr_printf, vstr, arg, PRINT_STR);
                    break;

                case 's':
                {
                    uint len;
                    const char *s = mp_obj_str_get_data(arg, &len);
                    if (precision < 0) {
                        precision = len;
                    }
                    if (len > precision) {
                        len = precision;
                    }
                    pfenv_print_strn(&pfenv_vstr, s, len, flags, fill, width);
                    break;
                }

                default:
                    nlr_jump(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
                        "Unknown format code '%c' for object of type 'str'",
                        type, mp_obj_get_type_str(arg)));
            }
883
884
885
        }
    }

886
887
888
    mp_obj_t s = mp_obj_new_str((byte*)vstr->buf, vstr->len, false);
    vstr_free(vstr);
    return s;
889
890
}

891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
STATIC mp_obj_t str_modulo_format(mp_obj_t pattern, uint n_args, const mp_obj_t *args) {
    assert(MP_OBJ_IS_STR(pattern));

    GET_STR_DATA_LEN(pattern, str, len);
    int arg_i = 0;
    vstr_t *vstr = vstr_new();
    for (const byte *top = str + len; str < top; str++) {
        if (*str == '%') {
            if (++str >= top) {
                break;
            }
            if (*str == '%') {
                vstr_add_char(vstr, '%');
            } else {
                if (arg_i >= n_args) {
                    nlr_jump(mp_obj_new_exception_msg(&mp_type_TypeError, "not enough arguments for format string"));
                }
                switch (*str) {
                    case 's':
                        mp_obj_print_helper((void (*)(void*, const char*, ...))vstr_printf, vstr, args[arg_i], PRINT_STR);
                        break;
                    case 'r':
                        mp_obj_print_helper((void (*)(void*, const char*, ...))vstr_printf, vstr, args[arg_i], PRINT_REPR);
                        break;
                }
                arg_i++;
            }
        } else {
            vstr_add_char(vstr, *str);
        }
    }

    if (arg_i != n_args) {
        nlr_jump(mp_obj_new_exception_msg(&mp_type_TypeError, "not all arguments converted during string formatting"));
    }

    mp_obj_t s = mp_obj_new_str((byte*)vstr->buf, vstr->len, false);
    vstr_free(vstr);
    return s;
}

932
STATIC mp_obj_t str_replace(uint n_args, const mp_obj_t *args) {
933
934
935
936
    assert(MP_OBJ_IS_STR(args[0]));
    assert(MP_OBJ_IS_STR(args[1]));
    assert(MP_OBJ_IS_STR(args[2]));

937
    machine_int_t max_rep = 0;
938
    if (n_args == 4) {
939
940
941
942
943
944
945
        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;
        }
946
    }
947
948

    // if max_rep is still 0 by this point we will need to do all possible replacements
949
950
951
952

    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);
953
954

    // old won't exist in str if it's longer, so nothing to replace
955
    if (old_len > str_len) {
956
        return args[0];
957
958
    }

959
960
961
962
963
964
965
966
967
968
969
970
971
    // 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;
972
        while ((old_occurrence = find_subbytes(offset_ptr, str_len - offset_num, old, old_len, 1)) != NULL) {
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
            // 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;
        }
1011
    }
1012

1013
1014
1015
    return mp_obj_str_builder_end(replaced_str);
}

1016
1017
1018
1019
1020
1021
1022
1023
STATIC mp_obj_t str_count(uint n_args, const mp_obj_t *args) {
    assert(2 <= n_args && n_args <= 4);
    assert(MP_OBJ_IS_STR(args[0]));
    assert(MP_OBJ_IS_STR(args[1]));

    GET_STR_DATA_LEN(args[0], haystack, haystack_len);
    GET_STR_DATA_LEN(args[1], needle, needle_len);

1024
1025
    machine_uint_t start = 0;
    machine_uint_t end = haystack_len;
1026
    if (n_args >= 3 && args[2] != mp_const_none) {
1027
        start = mp_get_index(&mp_type_str, haystack_len, args[2], true);
1028
1029
    }
    if (n_args >= 4 && args[3] != mp_const_none) {
1030
        end = mp_get_index(&mp_type_str, haystack_len, args[3], true);
1031
1032
    }

1033
1034
1035
    // if needle_len is zero then we count each gap between characters as an occurrence
    if (needle_len == 0) {
        return MP_OBJ_NEW_SMALL_INT(end - start + 1);
1036
1037
    }

1038
1039
    // count the occurrences
    machine_int_t num_occurrences = 0;
xbe's avatar
xbe committed
1040
1041
1042
1043
1044
    for (machine_uint_t haystack_index = start; haystack_index + needle_len <= end; haystack_index++) {
        if (memcmp(&haystack[haystack_index], needle, needle_len) == 0) {
            num_occurrences++;
            haystack_index += needle_len - 1;
        }
1045
1046
1047
1048
1049
    }

    return MP_OBJ_NEW_SMALL_INT(num_occurrences);
}

1050
STATIC mp_obj_t str_partitioner(mp_obj_t self_in, mp_obj_t arg, machine_int_t direction) {
1051
1052
1053
1054
1055
    assert(MP_OBJ_IS_STR(self_in));
    if (!MP_OBJ_IS_STR(arg)) {
        nlr_jump(mp_obj_new_exception_msg_varg(&mp_type_TypeError,
                                               "Can't convert '%s' object to str implicitly", mp_obj_get_type_str(arg)));
    }
1056

1057
1058
1059
1060
1061
1062
    GET_STR_DATA_LEN(self_in, str, str_len);
    GET_STR_DATA_LEN(arg, sep, sep_len);

    if (sep_len == 0) {
        nlr_jump(mp_obj_new_exception_msg(&mp_type_ValueError, "empty separator"));
    }
1063
1064
1065
1066
1067

    mp_obj_t result[] = {MP_OBJ_NEW_QSTR(MP_QSTR_), MP_OBJ_NEW_QSTR(MP_QSTR_), MP_OBJ_NEW_QSTR(MP_QSTR_)};

    if (direction > 0) {
        result[0] = self_in;
1068
    } else {
1069
        result[2] = self_in;
1070
    }
1071

1072
1073
1074
1075
1076
1077
    const byte *position_ptr = find_subbytes(str, str_len, sep, sep_len, direction);
    if (position_ptr != NULL) {
        machine_uint_t position = position_ptr - str;
        result[0] = mp_obj_new_str(str, position, false);
        result[1] = arg;
        result[2] = mp_obj_new_str(str + position + sep_len, str_len - position - sep_len, false);
1078
    }
1079

1080
    return mp_obj_new_tuple(3, result);
1081
1082
}

1083
1084
STATIC mp_obj_t str_partition(mp_obj_t self_in, mp_obj_t arg) {
    return str_partitioner(self_in, arg, 1);
1085
}
1086

1087
1088
STATIC mp_obj_t str_rpartition(mp_obj_t self_in, mp_obj_t arg) {
    return str_partitioner(self_in, arg, -1);
1089
1090
}

1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
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;
    }
}

1105
STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_find_obj, 2, 4, str_find);
1106
STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_rfind_obj, 2, 4, str_rfind);
1107
1108
1109
1110
1111
1112
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);
1113
STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(str_count_obj, 2, 4, str_count);
1114
STATIC MP_DEFINE_CONST_FUN_OBJ_2(str_partition_obj, str_partition);
1115
STATIC MP_DEFINE_CONST_FUN_OBJ_2(str_rpartition_obj, str_rpartition);
1116

1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
STATIC const mp_map_elem_t str_locals_dict_table[] = {
    { MP_OBJ_NEW_QSTR(MP_QSTR_find), (mp_obj_t)&str_find_obj },
    { MP_OBJ_NEW_QSTR(MP_QSTR_rfind), (mp_obj_t)&str_rfind_obj },
    { MP_OBJ_NEW_QSTR(MP_QSTR_join), (mp_obj_t)&str_join_obj },
    { MP_OBJ_NEW_QSTR(MP_QSTR_split), (mp_obj_t)&str_split_obj },
    { MP_OBJ_NEW_QSTR(MP_QSTR_startswith), (mp_obj_t)&str_startswith_obj },
    { MP_OBJ_NEW_QSTR(MP_QSTR_strip), (mp_obj_t)&str_strip_obj },
    { MP_OBJ_NEW_QSTR(MP_QSTR_format), (mp_obj_t)&str_format_obj },
    { MP_OBJ_NEW_QSTR(MP_QSTR_replace), (mp_obj_t)&str_replace_obj },
    { MP_OBJ_NEW_QSTR(MP_QSTR_count), (mp_obj_t)&str_count_obj },
    { MP_OBJ_NEW_QSTR(MP_QSTR_partition), (mp_obj_t)&str_partition_obj },
    { MP_OBJ_NEW_QSTR(MP_QSTR_rpartition), (mp_obj_t)&str_rpartition_obj },
ian-v's avatar
ian-v committed
1129
};
1130

1131
1132
STATIC MP_DEFINE_CONST_DICT(str_locals_dict, str_locals_dict_table);

1133
const mp_obj_type_t mp_type_str = {
1134
    { &mp_type_type },
1135
    .name = MP_QSTR_str,
1136
    .print = str_print,
1137
    .make_new = str_make_new,
1138
    .binary_op = str_binary_op,
1139
    .getiter = mp_obj_new_str_iterator,
1140
    .buffer_p = { .get_buffer = str_get_buffer },
1141
    .locals_dict = (mp_obj_t)&str_locals_dict,
1142
1143
1144
};

// Reuses most of methods from str
1145
const mp_obj_type_t mp_type_bytes = {
1146
    { &mp_type_type },
1147
    .name = MP_QSTR_bytes,
1148
    .print = str_print,
1149
    .make_new = bytes_make_new,
1150
1151
    .binary_op = str_binary_op,
    .getiter = mp_obj_new_bytes_iterator,
1152
    .locals_dict = (mp_obj_t)&str_locals_dict,
1153
1154
};

1155
// the zero-length bytes
1156
STATIC const mp_obj_str_t empty_bytes_obj = {{&mp_type_bytes}, 0, 0, NULL};
1157
1158
const mp_obj_t mp_const_empty_bytes = (mp_obj_t)&empty_bytes_obj;

1159
mp_obj_t mp_obj_str_builder_start(const mp_obj_type_t *type, uint len, byte **data) {
1160
    mp_obj_str_t *o = m_new_obj(mp_obj_str_t);
1161
    o->base.type = type;
1162
    o->len = len;
1163
1164
1165
    byte *p = m_new(byte, len + 1);
    o->data = p;
    *data = p;
1166
1167
1168
1169
1170
1171
    return o;
}

mp_obj_t mp_obj_str_builder_end(mp_obj_t o_in) {
    mp_obj_str_t *o = o_in;
    o->hash = qstr_compute_hash(o->data, o->len);
1172
1173
    byte *p = (byte*)o->data;
    p[o->len] = '\0'; // for now we add null for compatibility with C ASCIIZ strings
1174
1175
1176
    return o;
}

1177
STATIC mp_obj_t str_new(const mp_obj_type_t *type, const byte* data, uint len) {
1178
    mp_obj_str_t *o = m_new_obj(mp_obj_str_t);
1179
1180
    o->base.type = type;
    o->len = len;
1181
1182
1183
1184
1185
1186
1187
    if (data) {
        o->hash = qstr_compute_hash(data, len);
        byte *p = m_new(byte, len + 1);
        o->data = p;
        memcpy(p, data, len * sizeof(byte));
        p[len] = '\0'; // for now we add null for compatibility with C ASCIIZ strings
    }
1188
1189
1190
    return o;
}

1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
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
1201
        return str_new(&mp_type_str, data, len);
1202
1203
1204
    }
}

1205
mp_obj_t mp_obj_new_bytes(const byte* data, uint len) {
1206
    return str_new(&mp_type_bytes, data, len);
1207
1208
}

1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
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;
        }
1223
        return memcmp(d1, d2, l1) == 0;
1224
1225
1226
    }
}

1227
1228
void bad_implicit_conversion(mp_obj_t self_in) __attribute__((noreturn));
void bad_implicit_conversion(mp_obj_t self_in) {
1229
    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)));
1230
1231
}

1232
1233
1234
1235
1236
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 {
1237
        bad_implicit_conversion(self_in);
1238
    }
1239
1240
1241
1242
1243
1244
1245
}

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 {
1246
1247
1248
1249
1250
1251
1252
1253
1254
        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);
1255
    } else if (MP_OBJ_IS_TYPE(self_in, &mp_type_str)) {
1256
1257
1258
1259
        mp_obj_str_t *self = self_in;
        return qstr_from_strn((char*)self->data, self->len);
    } else {
        bad_implicit_conversion(self_in);
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
    }
}

// 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 {
1271
        bad_implicit_conversion(self_in);
1272
1273
1274
    }
}

1275
const char *mp_obj_str_get_data(mp_obj_t self_in, uint *len) {
1276
1277
1278
    if (MP_OBJ_IS_STR(self_in)) {
        GET_STR_DATA_LEN(self_in, s, l);
        *len = l;
1279
        return (const char*)s;
1280
    } else {
1281
        bad_implicit_conversion(self_in);
1282
    }