objstr.c 52.1 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);
36
STATIC void bad_implicit_conversion(mp_obj_t self_in) __attribute__((noreturn));
xyb's avatar
xyb committed
37
38
39
40

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

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
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");
66
67
68
69
        } else if (*s == '\r') {
            print(env, "\\r");
        } else if (*s == '\t') {
            print(env, "\\t");
70
71
72
73
74
75
76
        } else {
            print(env, "\\x%02x", *s);
        }
    }
    print(env, "%c", quote_char);
}

77
STATIC void str_print(void (*print)(void *env, const char *fmt, ...), void *env, mp_obj_t self_in, mp_print_kind_t kind) {
78
    GET_STR_DATA_LEN(self_in, str_data, str_len);
79
    bool is_bytes = MP_OBJ_IS_TYPE(self_in, &mp_type_bytes);
80
    if (kind == PRINT_STR && !is_bytes) {
81
        print(env, "%.*s", str_len, str_data);
82
    } else {
83
84
85
        if (is_bytes) {
            print(env, "b");
        }
86
        mp_str_print_quoted(print, env, str_data, str_len);
87
    }
88
89
}

90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
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
108
            if (!MP_OBJ_IS_TYPE(args[0], &mp_type_bytes)) {
109
                nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "bytes expected"));
110
111
112
            }
            GET_STR_DATA_LEN(args[0], str_data, str_len);
            GET_STR_HASH(args[0], str_hash);
113
            mp_obj_str_t *o = str_new(&mp_type_str, NULL, str_len);
114
115
116
117
118
119
            o->data = str_data;
            o->hash = str_hash;
            return o;
        }

        default:
120
            nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "str takes at most 3 arguments"));
121
122
123
    }
}

124
125
126
127
128
129
130
131
132
133
134
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);
135
        mp_obj_str_t *o = str_new(&mp_type_bytes, NULL, str_len);
136
137
138
139
140
141
142
143
144
145
146
147
148
        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;

149
        mp_obj_t o = mp_obj_str_builder_start(&mp_type_bytes, len, &data);
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
        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);
165
        o = mp_obj_str_builder_start(&mp_type_bytes, len, &data);
166
167
    }

Damien George's avatar
Damien George committed
168
    mp_obj_t iterable = mp_getiter(args[0]);
169
    mp_obj_t item;
Damien George's avatar
Damien George committed
170
    while ((item = mp_iternext(iterable)) != MP_OBJ_NULL) {
171
172
173
174
175
176
177
178
179
180
181
        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);
182
        o = mp_obj_str_builder_start(&mp_type_bytes, len, &data);
183
184
185
186
187
188
189
        memcpy(data, vstr_str(vstr), len);
        vstr_free(vstr);
    }

    return mp_obj_str_builder_end(o);

wrong_args:
190
        nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "wrong number of arguments"));
191
192
}

193
194
// like strstr but with specified length and allows \0 bytes
// TODO replace with something more efficient/standard
195
STATIC const byte *find_subbytes(const byte *haystack, machine_uint_t hlen, const byte *needle, machine_uint_t nlen, machine_int_t direction) {
196
    if (hlen >= nlen) {
197
198
199
200
201
202
203
204
205
206
207
208
        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;
209
            }
210
211
212
            if (str_index == str_index_end) {
                //not found
                break;
213
            }
214
            str_index += direction;
215
216
217
218
219
        }
    }
    return NULL;
}

220
STATIC mp_obj_t str_binary_op(int op, mp_obj_t lhs_in, mp_obj_t rhs_in) {
221
    GET_STR_DATA_LEN(lhs_in, lhs_data, lhs_len);
222
    switch (op) {
Damien George's avatar
Damien George committed
223
        case MP_BINARY_OP_SUBSCR:
224
            if (mp_obj_is_integer(rhs_in)) {
225
                uint index = mp_get_index(mp_obj_get_type(lhs_in), lhs_len, rhs_in, false);
226
                if (MP_OBJ_IS_TYPE(lhs_in, &mp_type_bytes)) {
227
                    return MP_OBJ_NEW_SMALL_INT((mp_small_int_t)lhs_data[index]);
228
229
230
                } else {
                    return mp_obj_new_str(lhs_data + index, 1, true);
                }
231
#if MICROPY_ENABLE_SLICE
232
            } else if (MP_OBJ_IS_TYPE(rhs_in, &mp_type_slice)) {
233
                machine_uint_t start, stop;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
234
235
236
                if (!m_seq_get_fast_slice_indexes(lhs_len, rhs_in, &start, &stop)) {
                    assert(0);
                }
237
                return mp_obj_new_str(lhs_data + start, stop - start, false);
238
#endif
239
            } else {
240
241
                // Message doesn't match CPython, but we don't have so much bytes as they
                // to spend them on verbose wording
242
                nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "index must be int"));
243
            }
244

Damien George's avatar
Damien George committed
245
246
        case MP_BINARY_OP_ADD:
        case MP_BINARY_OP_INPLACE_ADD:
247
            if (MP_OBJ_IS_STR(rhs_in)) {
248
                // add 2 strings
249
250

                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
251
                int alloc_len = lhs_len + rhs_len;
252
253

                /* code for making qstr
254
255
256
257
                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);
258
259
260
261
262
                return MP_OBJ_NEW_QSTR(qstr_build_end(q_ptr));
                */

                // code for non-qstr
                byte *data;
263
                mp_obj_t s = mp_obj_str_builder_start(mp_obj_get_type(lhs_in), alloc_len, &data);
264
265
266
                memcpy(data, lhs_data, lhs_len);
                memcpy(data + lhs_len, rhs_data, rhs_len);
                return mp_obj_str_builder_end(s);
267
268
            }
            break;
269

Damien George's avatar
Damien George committed
270
        case MP_BINARY_OP_IN:
271
            /* NOTE `a in b` is `b.__contains__(a)` */
272
273
            if (MP_OBJ_IS_STR(rhs_in)) {
                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
274
                return MP_BOOL(find_subbytes(lhs_data, lhs_len, rhs_data, rhs_len, 1) != NULL);
275
276
            }
            break;
277

Damien George's avatar
Damien George committed
278
        case MP_BINARY_OP_MULTIPLY:
279
280
281
282
283
        {
            if (!MP_OBJ_IS_SMALL_INT(rhs_in)) {
                return NULL;
            }
            int n = MP_OBJ_SMALL_INT_VALUE(rhs_in);
284
            byte *data;
285
            mp_obj_t s = mp_obj_str_builder_start(mp_obj_get_type(lhs_in), lhs_len * n, &data);
286
287
            mp_seq_multiply(lhs_data, sizeof(*lhs_data), lhs_len, n, data);
            return mp_obj_str_builder_end(s);
288
        }
289

290
291
292
293
294
295
296
297
298
299
300
301
302
        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
303
304
305
306
307
308
309
        // 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:
310
311
312
313
            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));
            }
314
315
316
317
318
    }

    return MP_OBJ_NULL; // op not supported
}

319
STATIC mp_obj_t str_join(mp_obj_t self_in, mp_obj_t arg) {
320
    assert(MP_OBJ_IS_STR(self_in));
321

322
    // get separation string
323
    GET_STR_DATA_LEN(self_in, sep_str, sep_len);
324
325

    // process args
326
327
    uint seq_len;
    mp_obj_t *seq_items;
328
    if (MP_OBJ_IS_TYPE(arg, &mp_type_tuple)) {
329
330
        mp_obj_tuple_get(arg, &seq_len, &seq_items);
    } else {
331
332
333
334
335
        if (!MP_OBJ_IS_TYPE(arg, &mp_type_list)) {
            // arg is not a list, try to convert it to one
            arg = mp_type_list.make_new((mp_obj_t)&mp_type_list, 1, 0, &arg);
        }
        mp_obj_list_get(arg, &seq_len, &seq_items);
336
    }
337
338
339

    // count required length
    int required_len = 0;
340
    for (int i = 0; i < seq_len; i++) {
341
        if (!MP_OBJ_IS_STR(seq_items[i])) {
342
            nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "join expected a list of str's"));
343
        }
344
345
346
        if (i > 0) {
            required_len += sep_len;
        }
347
348
        GET_STR_LEN(seq_items[i], l);
        required_len += l;
349
350
351
    }

    // make joined string
352
    byte *data;
353
    mp_obj_t joined_str = mp_obj_str_builder_start(mp_obj_get_type(self_in), required_len, &data);
354
355
    for (int i = 0; i < seq_len; i++) {
        if (i > 0) {
356
357
            memcpy(data, sep_str, sep_len);
            data += sep_len;
358
        }
359
360
361
        GET_STR_DATA_LEN(seq_items[i], s, l);
        memcpy(data, s, l);
        data += l;
362
    }
363
364

    // return joined string
365
    return mp_obj_str_builder_end(joined_str);
366
367
}

Paul Sokolovsky's avatar
Paul Sokolovsky committed
368
369
#define is_ws(c) ((c) == ' ' || (c) == '\t')

370
STATIC mp_obj_t str_split(uint n_args, const mp_obj_t *args) {
371
    machine_int_t splits = -1;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
372
373
374
375
    mp_obj_t sep = mp_const_none;
    if (n_args > 1) {
        sep = args[1];
        if (n_args > 2) {
376
            splits = mp_obj_get_int(args[2]);
Paul Sokolovsky's avatar
Paul Sokolovsky committed
377
378
        }
    }
379

Paul Sokolovsky's avatar
Paul Sokolovsky committed
380
    mp_obj_t res = mp_obj_new_list(0, NULL);
381
382
    GET_STR_DATA_LEN(args[0], s, len);
    const byte *top = s + len;
383
384
385
386
387

    if (sep == mp_const_none) {
        // sep not given, so separate on whitespace

        // Initial whitespace is not counted as split, so we pre-do it
388
        while (s < top && is_ws(*s)) s++;
389
390
391
392
393
394
395
396
397
398
399
        while (s < top && splits != 0) {
            const byte *start = s;
            while (s < top && !is_ws(*s)) s++;
            mp_obj_list_append(res, mp_obj_new_str(start, s - start, false));
            if (s >= top) {
                break;
            }
            while (s < top && is_ws(*s)) s++;
            if (splits > 0) {
                splits--;
            }
Paul Sokolovsky's avatar
Paul Sokolovsky committed
400
401
        }

402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
        if (s < top) {
            mp_obj_list_append(res, mp_obj_new_str(s, top - s, false));
        }

    } else {
        // sep given

        uint sep_len;
        const char *sep_str = mp_obj_str_get_data(sep, &sep_len);

        if (sep_len == 0) {
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "empty separator"));
        }

        for (;;) {
            const byte *start = s;
            for (;;) {
                if (splits == 0 || s + sep_len > top) {
                    s = top;
                    break;
                } else if (memcmp(s, sep_str, sep_len) == 0) {
                    break;
                }
                s++;
            }
            mp_obj_list_append(res, mp_obj_new_str(start, s - start, false));
            if (s >= top) {
                break;
            }
            s += sep_len;
            if (splits > 0) {
                splits--;
            }
        }
Paul Sokolovsky's avatar
Paul Sokolovsky committed
436
437
438
439
440
    }

    return res;
}

441
STATIC mp_obj_t str_finder(uint n_args, const mp_obj_t *args, machine_int_t direction, bool is_index) {
442
    assert(2 <= n_args && n_args <= 4);
443
444
    assert(MP_OBJ_IS_STR(args[0]));
    assert(MP_OBJ_IS_STR(args[1]));
445

446
447
    GET_STR_DATA_LEN(args[0], haystack, haystack_len);
    GET_STR_DATA_LEN(args[1], needle, needle_len);
448

449
450
    machine_uint_t start = 0;
    machine_uint_t end = haystack_len;
451
    if (n_args >= 3 && args[2] != mp_const_none) {
452
        start = mp_get_index(&mp_type_str, haystack_len, args[2], true);
453
454
    }
    if (n_args >= 4 && args[3] != mp_const_none) {
455
        end = mp_get_index(&mp_type_str, haystack_len, args[3], true);
456
457
    }

458
    const byte *p = find_subbytes(haystack + start, end - start, needle, needle_len, direction);
459
460
    if (p == NULL) {
        // not found
461
462
463
464
465
        if (is_index) {
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "substring not found"));
        } else {
            return MP_OBJ_NEW_SMALL_INT(-1);
        }
466
467
    } else {
        // found
468
        return MP_OBJ_NEW_SMALL_INT(p - haystack);
469
470
471
    }
}

472
STATIC mp_obj_t str_find(uint n_args, const mp_obj_t *args) {
473
    return str_finder(n_args, args, 1, false);
474
475
476
}

STATIC mp_obj_t str_rfind(uint n_args, const mp_obj_t *args) {
477
478
479
480
481
482
483
484
485
    return str_finder(n_args, args, -1, false);
}

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

STATIC mp_obj_t str_rindex(uint n_args, const mp_obj_t *args) {
    return str_finder(n_args, args, -1, true);
486
487
}

488
// TODO: (Much) more variety in args
489
STATIC mp_obj_t str_startswith(mp_obj_t self_in, mp_obj_t arg) {
490
491
492
493
494
495
496
497
    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);
}

498
STATIC mp_obj_t str_strip(uint n_args, const mp_obj_t *args) {
xbe's avatar
xbe committed
499
    assert(1 <= n_args && n_args <= 2);
500
501
502
503
504
    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
505
506
507

    if (n_args == 1) {
        chars_to_del = whitespace;
508
        chars_to_del_len = sizeof(whitespace);
xbe's avatar
xbe committed
509
    } else {
510
511
512
513
        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
514
515
    }

516
    GET_STR_DATA_LEN(args[0], orig_str, orig_str_len);
xbe's avatar
xbe committed
517

518
    machine_uint_t first_good_char_pos = 0;
xbe's avatar
xbe committed
519
    bool first_good_char_pos_set = false;
520
521
    machine_uint_t last_good_char_pos = 0;
    for (machine_uint_t i = 0; i < orig_str_len; i++) {
522
        if (find_subbytes(chars_to_del, chars_to_del_len, &orig_str[i], 1, 1) == NULL) {
xbe's avatar
xbe committed
523
524
525
526
527
528
529
530
531
            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) {
532
533
        // string is all whitespace, return ''
        return MP_OBJ_NEW_QSTR(MP_QSTR_);
xbe's avatar
xbe committed
534
535
536
537
    }

    assert(last_good_char_pos >= first_good_char_pos);
    //+1 to accomodate the last character
538
    machine_uint_t stripped_len = last_good_char_pos - first_good_char_pos + 1;
539
    return mp_obj_new_str(orig_str + first_good_char_pos, stripped_len, false);
xbe's avatar
xbe committed
540
541
}

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
// 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
    ;
}

577
static mp_obj_t arg_as_int(mp_obj_t arg) {
578
579
#if MICROPY_ENABLE_FLOAT
    if (MP_OBJ_IS_TYPE(arg, &mp_type_float)) {
580
581
582
583
584

        // TODO: Needs a way to construct an mpz integer from a float

        mp_small_int_t num = mp_obj_get_float(arg);
        return MP_OBJ_NEW_SMALL_INT(num);
585
586
    }
#endif
587
    return arg;
588
589
}

590
mp_obj_t str_format(uint n_args, const mp_obj_t *args) {
591
    assert(MP_OBJ_IS_STR(args[0]));
592

593
    GET_STR_DATA_LEN(args[0], str, len);
Dave Hylands's avatar
Dave Hylands committed
594
    int arg_i = 0;
595
    vstr_t *vstr = vstr_new();
Dave Hylands's avatar
Dave Hylands committed
596
597
598
599
    pfenv_t pfenv_vstr;
    pfenv_vstr.data = vstr;
    pfenv_vstr.print_strn = pfenv_vstr_add_strn;

600
    for (const byte *top = str + len; str < top; str++) {
Dave Hylands's avatar
Dave Hylands committed
601
602
603
604
605
606
        if (*str == '}') {
            str++;
            if (str < top && *str == '}') {
                vstr_add_char(vstr, '}');
                continue;
            }
607
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "Single '}' encountered in format string"));
Dave Hylands's avatar
Dave Hylands committed
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
        }
        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 == '!') {
637
            str++;
Dave Hylands's avatar
Dave Hylands committed
638
639
            if (str < top && (*str == 'r' || *str == 's')) {
                conversion = *str++;
640
            } else {
641
                nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "end of format while looking for conversion specifier"));
Dave Hylands's avatar
Dave Hylands committed
642
643
644
645
646
647
648
649
650
651
652
653
654
655
            }
        }

        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++);
656
                }
Dave Hylands's avatar
Dave Hylands committed
657
658
659
660
                vstr_add_char(format_spec, '\0');
            }
        }
        if (str >= top) {
661
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "unmatched '{' in format"));
Dave Hylands's avatar
Dave Hylands committed
662
663
        }
        if (*str != '}') {
664
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "expected ':' after format specifier"));
Dave Hylands's avatar
Dave Hylands committed
665
666
667
668
669
670
        }

        mp_obj_t arg = mp_const_none;

        if (field_name) {
            if (arg_i > 0) {
671
                nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "cannot switch from automatic field numbering to manual field specification"));
Dave Hylands's avatar
Dave Hylands committed
672
673
674
            }
            int index;
            if (str_to_int(vstr_str(field_name), &index) != vstr_len(field_name) - 1) {
675
                nlr_raise(mp_obj_new_exception_msg(&mp_type_KeyError, "attributes not supported yet"));
676
            }
Dave Hylands's avatar
Dave Hylands committed
677
            if (index >= n_args - 1) {
678
                nlr_raise(mp_obj_new_exception_msg(&mp_type_IndexError, "tuple index out of range"));
Dave Hylands's avatar
Dave Hylands committed
679
680
681
682
683
            }
            arg = args[index + 1];
            arg_i = -1;
            vstr_free(field_name);
            field_name = NULL;
684
        } else {
Dave Hylands's avatar
Dave Hylands committed
685
            if (arg_i < 0) {
686
                nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "cannot switch from manual field specification to automatic field numbering"));
Dave Hylands's avatar
Dave Hylands committed
687
688
            }
            if (arg_i >= n_args - 1) {
689
                nlr_raise(mp_obj_new_exception_msg(&mp_type_IndexError, "tuple index out of range"));
Dave Hylands's avatar
Dave Hylands committed
690
691
692
693
694
695
696
697
698
699
700
701
702
703
            }
            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 {
704
                nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError, "Unknown conversion specifier %c", conversion));
Dave Hylands's avatar
Dave Hylands committed
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
            }
            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) {
771
                nlr_raise(mp_obj_new_exception_msg(&mp_type_KeyError, "Invalid conversion specification"));
Dave Hylands's avatar
Dave Hylands committed
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
            }
            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') {
789
                nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "Sign not allowed in string format specifier"));
Dave Hylands's avatar
Dave Hylands committed
790
791
            }
            if (type == 'c') {
792
                nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "Sign not allowed with integer format specifier 'c'"));
Dave Hylands's avatar
Dave Hylands committed
793
794
795
796
797
798
799
800
801
802
803
804
805
806
            }
        } 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':
807
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 2, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
808
809
810
811
812
813
814
815
816
817
818
819
                    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':
820
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 10, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
821
822
823
                    continue;

                case 'o':
824
825
826
827
                    if (flags & PF_FLAG_SHOW_PREFIX) {
                        flags |= PF_FLAG_SHOW_OCTAL_LETTER;
                    }

828
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 8, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
829
830
831
                    continue;

                case 'x':
832
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 16, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
833
834
835
                    continue;

                case 'X':
836
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 16, 'A', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
837
838
839
840
841
842
843
844
845
846
847
848
849
850
                    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:
851
                    nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
852
853
                        "Unknown format code '%c' for object of type '%s'", type, mp_obj_get_type_str(arg)));
            }
854
        }
855

856
857
        // NOTE: no else here. We need the e, f, g etc formats for integer
        //       arguments (from above if) to take this if.
858
        if (arg_looks_numeric(arg)) {
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
883
884
885
886
887
888
889
890
891
892
893
894
895
            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) {
896
#if MICROPY_ENABLE_FLOAT
Dave Hylands's avatar
Dave Hylands committed
897
898
899
900
901
902
903
904
905
906
907
908
909
                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;
910
#endif
Dave Hylands's avatar
Dave Hylands committed
911
912

                default:
913
                    nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
914
915
916
917
                        "Unknown format code '%c' for object of type 'float'",
                        type, mp_obj_get_type_str(arg)));
            }
        } else {
918
919
            // arg doesn't look like a number

Dave Hylands's avatar
Dave Hylands committed
920
            if (align == '=') {
921
                nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "'=' alignment not allowed in string format specifier"));
Dave Hylands's avatar
Dave Hylands committed
922
            }
923

Dave Hylands's avatar
Dave Hylands committed
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
            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:
944
                    nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
945
946
947
                        "Unknown format code '%c' for object of type 'str'",
                        type, mp_obj_get_type_str(arg)));
            }
948
949
950
        }
    }

951
952
953
    mp_obj_t s = mp_obj_new_str((byte*)vstr->buf, vstr->len, false);
    vstr_free(vstr);
    return s;
954
955
}

956
957
958
959
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);
Dave Hylands's avatar
Dave Hylands committed
960
    const byte *start_str = str;
961
962
    int arg_i = 0;
    vstr_t *vstr = vstr_new();
Dave Hylands's avatar
Dave Hylands committed
963
964
965
966
    pfenv_t pfenv_vstr;
    pfenv_vstr.data = vstr;
    pfenv_vstr.print_strn = pfenv_vstr_add_strn;

967
    for (const byte *top = str + len; str < top; str++) {
Dave Hylands's avatar
Dave Hylands committed
968
969
970
971
972
973
974
        if (*str != '%') {
            vstr_add_char(vstr, *str);
            continue;
        }
        if (++str >= top) {
            break;
        }
975
        if (*str == '%') {
Dave Hylands's avatar
Dave Hylands committed
976
977
978
979
            vstr_add_char(vstr, '%');
            continue;
        }
        if (arg_i >= n_args) {
980
            nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "not enough arguments for format string"));
Dave Hylands's avatar
Dave Hylands committed
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
        }
        int flags = 0;
        char fill = ' ';
        bool alt = false;
        while (str < top) {
            if (*str == '-')      flags |= PF_FLAG_LEFT_ADJUST;
            else if (*str == '+') flags |= PF_FLAG_SHOW_SIGN;
            else if (*str == ' ') flags |= PF_FLAG_SPACE_SIGN;
            else if (*str == '#') alt = true;
            else if (*str == '0') {
                flags |= PF_FLAG_PAD_AFTER_SIGN;
                fill = '0';
            } else break;
            str++;
        }
        // parse width, if it exists
        int width = 0; 
        if (str < top) {
            if (*str == '*') {
                width = mp_obj_get_int(args[arg_i++]);
                str++;
1002
            } else {
Dave Hylands's avatar
Dave Hylands committed
1003
1004
                for (; str < top && '0' <= *str && *str <= '9'; str++) {
                    width = width * 10 + *str - '0';
1005
                }
Dave Hylands's avatar
Dave Hylands committed
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
            }
        }
        int prec = -1;
        if (str < top && *str == '.') {
            if (++str < top) {
                if (*str == '*') {
                    prec = mp_obj_get_int(args[arg_i++]);
                    str++;
                } else {
                    prec = 0;
                    for (; str < top && '0' <= *str && *str <= '9'; str++) {
                        prec = prec * 10 + *str - '0';
                    }
                }
            }
        }

        if (str >= top) {
1024
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "incomplete format"));
Dave Hylands's avatar
Dave Hylands committed
1025
1026
1027
1028
1029
1030
1031
1032
        }
        mp_obj_t arg = args[arg_i];
        switch (*str) {
            case 'c':
                if (MP_OBJ_IS_STR(arg)) {
                    uint len;
                    const char *s = mp_obj_str_get_data(arg, &len);
                    if (len != 1) {
1033
                        nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "%c requires int or char")); 
1034
                        break;
Dave Hylands's avatar
Dave Hylands committed
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
                    }
                    pfenv_print_strn(&pfenv_vstr, s, 1, flags, ' ', width);
                    break;
                }
                if (arg_looks_integer(arg)) {
                    char ch = mp_obj_get_int(arg);
                    pfenv_print_strn(&pfenv_vstr, &ch, 1, flags, ' ', width);
                    break;
                }
#if MICROPY_ENABLE_FLOAT
                // This is what CPython reports, so we report the same.
                if (MP_OBJ_IS_TYPE(arg, &mp_type_float)) {
1047
                    nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "integer argument expected, got float")); 
Dave Hylands's avatar
Dave Hylands committed
1048
1049
1050

                }
#endif
1051
                nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "an integer is required")); 
Dave Hylands's avatar
Dave Hylands committed
1052
1053
1054
1055
1056
                break; 

            case 'd':
            case 'i':
            case 'u':
1057
                pfenv_print_mp_int(&pfenv_vstr, arg_as_int(arg), 1, 10, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
                break;

#if MICROPY_ENABLE_FLOAT
            case 'e':
            case 'E':
            case 'f':
            case 'F':
            case 'g':
            case 'G':
                pfenv_print_float(&pfenv_vstr, mp_obj_get_float(arg), *str, flags, fill, width, prec);
                break;
#endif

            case 'o':
                if (alt) {
1073
                    flags |= (PF_FLAG_SHOW_PREFIX | PF_FLAG_SHOW_OCTAL_LETTER);
Dave Hylands's avatar
Dave Hylands committed
1074
                }
1075
                pfenv_print_mp_int(&pfenv_vstr, arg_as_int(arg), 1, 8, 'a', flags, fill, width); 
Dave Hylands's avatar
Dave Hylands committed
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
                break;

            case 'r':
            case 's':
            {
                vstr_t *arg_vstr = vstr_new();
                mp_obj_print_helper((void (*)(void*, const char*, ...))vstr_printf,
                                    arg_vstr, arg, *str == 'r' ? PRINT_REPR : PRINT_STR);
                uint len = vstr_len(arg_vstr);
                if (prec < 0) {
                    prec = len;
                }
                if (len > prec) {
                    len = prec;
1090
                }
Dave Hylands's avatar
Dave Hylands committed
1091
1092
1093
                pfenv_print_strn(&pfenv_vstr, vstr_str(arg_vstr), len, flags, ' ', width);
                vstr_free(arg_vstr);
                break;
1094
            }
Dave Hylands's avatar
Dave Hylands committed
1095
1096
1097
1098
1099

            case 'x':
                if (alt) {
                    flags |= PF_FLAG_SHOW_PREFIX;
                }
1100
                pfenv_print_mp_int(&pfenv_vstr, arg_as_int(arg), 1, 16, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
1101
1102
1103
1104
1105
1106
                break;

            case 'X':
                if (alt) {
                    flags |= PF_FLAG_SHOW_PREFIX;
                }
1107
                pfenv_print_mp_int(&pfenv_vstr, arg_as_int(arg), 1, 16, 'A', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
1108
                break;
1109

Dave Hylands's avatar
Dave Hylands committed
1110
            default:
1111
                nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
1112
1113
                    "unsupported format character '%c' (0x%x) at index %d",
                    *str, *str, str - start_str));
1114
        }
Dave Hylands's avatar
Dave Hylands committed
1115
        arg_i++;
1116
1117
1118
    }

    if (arg_i != n_args) {
1119
        nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "not all arguments converted during string formatting"));
1120
1121
1122
1123
1124
1125
1126
    }

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

1127
STATIC mp_obj_t str_replace(uint n_args, const mp_obj_t *args) {
1128
1129
    assert(MP_OBJ_IS_STR(args[0]));

1130
    machine_int_t max_rep = -1;
1131
    if (n_args == 4) {
1132
        max_rep = mp_obj_get_int(args[3]);
1133
1134
1135
        if (max_rep == 0) {
            return args[0];
        } else if (max_rep < 0) {
1136
            max_rep = -1;
1137
        }
1138
    }
1139

xbe's avatar
xbe committed
1140
    // if max_rep is still -1 by this point we will need to do all possible replacements
1141

1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
    // check argument types

    if (!MP_OBJ_IS_STR(args[1])) {
        bad_implicit_conversion(args[1]);
    }

    if (!MP_OBJ_IS_STR(args[2])) {
        bad_implicit_conversion(args[2]);
    }

    // extract string data

1154
1155
1156
    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);
1157
1158

    // old won't exist in str if it's longer, so nothing to replace
1159
    if (old_len > str_len) {
1160
        return args[0];
1161
1162
    }

1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
    // 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;
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
        machine_uint_t str_len_remain = str_len;
        if (old_len == 0) {
            // if old_str is empty, copy new_str to start of replaced string
            // copy the replacement string
            if (data != NULL) {
                memcpy(data, new, new_len);
            }
            replaced_str_index += new_len;
            num_replacements_done++;
        }
        while (num_replacements_done != max_rep && str_len_remain > 0 && (old_occurrence = find_subbytes(offset_ptr, str_len_remain, old, old_len, 1)) != NULL) {
            if (old_len == 0) {
                old_occurrence += 1;
            }
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
            // 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;
1200
            str_len_remain = str + str_len - offset_ptr;
1201
1202
1203
1204
1205
            num_replacements_done++;
        }

        // copy from just after end of last occurrence of to-be-replaced string to end of old string
        if (data != NULL) {
1206
            memcpy(data + replaced_str_index, offset_ptr, str_len_remain);
1207
        }
1208
        replaced_str_index += str_len_remain;
1209
1210
1211
1212
1213
1214
1215
1216
1217

        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);
1218
                assert(data != NULL);
1219
1220
1221
1222
1223
            }
        } else {
            // second pass, we are done
            break;
        }
1224
    }
1225

1226
1227
1228
    return mp_obj_str_builder_end(replaced_str);
}

1229
1230
1231
1232
1233
1234
1235
1236
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