objstr.c 54.2 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
#include "objstr.h"
14

15
STATIC mp_obj_t str_modulo_format(mp_obj_t pattern, uint n_args, const mp_obj_t *args);
16
17
const mp_obj_t mp_const_empty_bytes;

18
19
20
21
22
23
24
25
26
// 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; }

27
28
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);
29
STATIC mp_obj_t str_new(const mp_obj_type_t *type, const byte* data, uint len);
30
STATIC void bad_implicit_conversion(mp_obj_t self_in) __attribute__((noreturn));
xyb's avatar
xyb committed
31
32
33
34

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

35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
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");
60
61
62
63
        } else if (*s == '\r') {
            print(env, "\\r");
        } else if (*s == '\t') {
            print(env, "\\t");
64
65
66
67
68
69
70
        } else {
            print(env, "\\x%02x", *s);
        }
    }
    print(env, "%c", quote_char);
}

71
STATIC void str_print(void (*print)(void *env, const char *fmt, ...), void *env, mp_obj_t self_in, mp_print_kind_t kind) {
72
    GET_STR_DATA_LEN(self_in, str_data, str_len);
73
    bool is_bytes = MP_OBJ_IS_TYPE(self_in, &mp_type_bytes);
74
    if (kind == PRINT_STR && !is_bytes) {
75
        print(env, "%.*s", str_len, str_data);
76
    } else {
77
78
79
        if (is_bytes) {
            print(env, "b");
        }
80
        mp_str_print_quoted(print, env, str_data, str_len);
81
    }
82
83
}

84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
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
102
            if (!MP_OBJ_IS_TYPE(args[0], &mp_type_bytes)) {
103
                nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "bytes expected"));
104
105
106
            }
            GET_STR_DATA_LEN(args[0], str_data, str_len);
            GET_STR_HASH(args[0], str_hash);
107
            mp_obj_str_t *o = str_new(&mp_type_str, NULL, str_len);
108
109
110
111
112
113
            o->data = str_data;
            o->hash = str_hash;
            return o;
        }

        default:
114
            nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "str takes at most 3 arguments"));
115
116
117
    }
}

118
119
120
121
122
123
124
125
126
127
128
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);
129
        mp_obj_str_t *o = str_new(&mp_type_bytes, NULL, str_len);
130
131
132
133
134
135
136
137
138
139
140
141
142
        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;

143
        mp_obj_t o = mp_obj_str_builder_start(&mp_type_bytes, len, &data);
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
        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);
159
        o = mp_obj_str_builder_start(&mp_type_bytes, len, &data);
160
161
    }

Damien George's avatar
Damien George committed
162
    mp_obj_t iterable = mp_getiter(args[0]);
163
    mp_obj_t item;
164
    while ((item = mp_iternext(iterable)) != MP_OBJ_STOP_ITERATION) {
165
166
167
168
169
170
171
172
173
174
175
        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);
176
        o = mp_obj_str_builder_start(&mp_type_bytes, len, &data);
177
178
179
180
181
182
183
        memcpy(data, vstr_str(vstr), len);
        vstr_free(vstr);
    }

    return mp_obj_str_builder_end(o);

wrong_args:
184
        nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "wrong number of arguments"));
185
186
}

187
188
// like strstr but with specified length and allows \0 bytes
// TODO replace with something more efficient/standard
189
STATIC const byte *find_subbytes(const byte *haystack, machine_uint_t hlen, const byte *needle, machine_uint_t nlen, machine_int_t direction) {
190
    if (hlen >= nlen) {
191
192
193
194
195
196
197
198
199
200
201
202
        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;
203
            }
204
205
206
            if (str_index == str_index_end) {
                //not found
                break;
207
            }
208
            str_index += direction;
209
210
211
212
213
        }
    }
    return NULL;
}

214
STATIC mp_obj_t str_binary_op(int op, mp_obj_t lhs_in, mp_obj_t rhs_in) {
215
    GET_STR_DATA_LEN(lhs_in, lhs_data, lhs_len);
216
    switch (op) {
Damien George's avatar
Damien George committed
217
218
        case MP_BINARY_OP_ADD:
        case MP_BINARY_OP_INPLACE_ADD:
219
            if (MP_OBJ_IS_STR(rhs_in)) {
220
                // add 2 strings
221
222

                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
223
                int alloc_len = lhs_len + rhs_len;
224
225

                /* code for making qstr
226
227
228
229
                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);
230
231
232
233
234
                return MP_OBJ_NEW_QSTR(qstr_build_end(q_ptr));
                */

                // code for non-qstr
                byte *data;
235
                mp_obj_t s = mp_obj_str_builder_start(mp_obj_get_type(lhs_in), alloc_len, &data);
236
237
238
                memcpy(data, lhs_data, lhs_len);
                memcpy(data + lhs_len, rhs_data, rhs_len);
                return mp_obj_str_builder_end(s);
239
240
            }
            break;
241

Damien George's avatar
Damien George committed
242
        case MP_BINARY_OP_IN:
243
            /* NOTE `a in b` is `b.__contains__(a)` */
244
245
            if (MP_OBJ_IS_STR(rhs_in)) {
                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
246
                return MP_BOOL(find_subbytes(lhs_data, lhs_len, rhs_data, rhs_len, 1) != NULL);
247
248
            }
            break;
249

Damien George's avatar
Damien George committed
250
        case MP_BINARY_OP_MULTIPLY:
251
252
253
254
255
        {
            if (!MP_OBJ_IS_SMALL_INT(rhs_in)) {
                return NULL;
            }
            int n = MP_OBJ_SMALL_INT_VALUE(rhs_in);
256
            byte *data;
257
            mp_obj_t s = mp_obj_str_builder_start(mp_obj_get_type(lhs_in), lhs_len * n, &data);
258
259
            mp_seq_multiply(lhs_data, sizeof(*lhs_data), lhs_len, n, data);
            return mp_obj_str_builder_end(s);
260
        }
261

262
263
264
265
266
267
268
269
270
271
272
273
274
        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
275
276
277
278
279
280
281
        // 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:
282
283
284
285
            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));
            }
286
287
    }

288
    return MP_OBJ_NOT_SUPPORTED;
289
290
}

291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
STATIC mp_obj_t str_subscr(mp_obj_t self_in, mp_obj_t index, mp_obj_t value) {
    GET_STR_DATA_LEN(self_in, self_data, self_len);
    if (value == MP_OBJ_SENTINEL) {
        // load
#if MICROPY_ENABLE_SLICE
        if (MP_OBJ_IS_TYPE(index, &mp_type_slice)) {
            machine_uint_t start, stop;
            if (!m_seq_get_fast_slice_indexes(self_len, index, &start, &stop)) {
                assert(0);
            }
            return mp_obj_new_str(self_data + start, stop - start, false);
        }
#endif
        mp_obj_type_t *type = mp_obj_get_type(self_in);
        uint index_val = mp_get_index(type, self_len, index, false);
        if (type == &mp_type_bytes) {
            return MP_OBJ_NEW_SMALL_INT((mp_small_int_t)self_data[index_val]);
        } else {
            return mp_obj_new_str(self_data + index_val, 1, true);
        }
    } else {
        return MP_OBJ_NOT_SUPPORTED;
    }
}

316
STATIC mp_obj_t str_join(mp_obj_t self_in, mp_obj_t arg) {
317
    assert(MP_OBJ_IS_STR(self_in));
318

319
    // get separation string
320
    GET_STR_DATA_LEN(self_in, sep_str, sep_len);
321
322

    // process args
323
324
    uint seq_len;
    mp_obj_t *seq_items;
325
    if (MP_OBJ_IS_TYPE(arg, &mp_type_tuple)) {
326
327
        mp_obj_tuple_get(arg, &seq_len, &seq_items);
    } else {
328
329
        if (!MP_OBJ_IS_TYPE(arg, &mp_type_list)) {
            // arg is not a list, try to convert it to one
330
            // TODO: Try to optimize?
331
332
333
            arg = mp_type_list.make_new((mp_obj_t)&mp_type_list, 1, 0, &arg);
        }
        mp_obj_list_get(arg, &seq_len, &seq_items);
334
    }
335
336
337

    // count required length
    int required_len = 0;
338
    for (int i = 0; i < seq_len; i++) {
339
        if (!MP_OBJ_IS_STR(seq_items[i])) {
340
            nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "join expected a list of str's"));
341
        }
342
343
344
        if (i > 0) {
            required_len += sep_len;
        }
345
346
        GET_STR_LEN(seq_items[i], l);
        required_len += l;
347
348
349
    }

    // make joined string
350
    byte *data;
351
    mp_obj_t joined_str = mp_obj_str_builder_start(mp_obj_get_type(self_in), required_len, &data);
352
353
    for (int i = 0; i < seq_len; i++) {
        if (i > 0) {
354
355
            memcpy(data, sep_str, sep_len);
            data += sep_len;
356
        }
357
358
359
        GET_STR_DATA_LEN(seq_items[i], s, l);
        memcpy(data, s, l);
        data += l;
360
    }
361
362

    // return joined string
363
    return mp_obj_str_builder_end(joined_str);
364
365
}

Paul Sokolovsky's avatar
Paul Sokolovsky committed
366
367
#define is_ws(c) ((c) == ' ' || (c) == '\t')

368
STATIC mp_obj_t str_split(uint n_args, const mp_obj_t *args) {
369
    machine_int_t splits = -1;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
370
371
372
373
    mp_obj_t sep = mp_const_none;
    if (n_args > 1) {
        sep = args[1];
        if (n_args > 2) {
374
            splits = mp_obj_get_int(args[2]);
Paul Sokolovsky's avatar
Paul Sokolovsky committed
375
376
        }
    }
377

Paul Sokolovsky's avatar
Paul Sokolovsky committed
378
    mp_obj_t res = mp_obj_new_list(0, NULL);
379
380
    GET_STR_DATA_LEN(args[0], s, len);
    const byte *top = s + len;
381
382
383
384
385

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

        // Initial whitespace is not counted as split, so we pre-do it
386
        while (s < top && is_ws(*s)) s++;
387
388
389
390
391
392
393
394
395
396
397
        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
398
399
        }

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
        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
434
435
436
437
438
    }

    return res;
}

439
STATIC mp_obj_t str_finder(uint n_args, const mp_obj_t *args, machine_int_t direction, bool is_index) {
440
    assert(2 <= n_args && n_args <= 4);
441
442
    assert(MP_OBJ_IS_STR(args[0]));
    assert(MP_OBJ_IS_STR(args[1]));
443

444
445
    GET_STR_DATA_LEN(args[0], haystack, haystack_len);
    GET_STR_DATA_LEN(args[1], needle, needle_len);
446

447
448
    machine_uint_t start = 0;
    machine_uint_t end = haystack_len;
449
    if (n_args >= 3 && args[2] != mp_const_none) {
450
        start = mp_get_index(&mp_type_str, haystack_len, args[2], true);
451
452
    }
    if (n_args >= 4 && args[3] != mp_const_none) {
453
        end = mp_get_index(&mp_type_str, haystack_len, args[3], true);
454
455
    }

456
    const byte *p = find_subbytes(haystack + start, end - start, needle, needle_len, direction);
457
458
    if (p == NULL) {
        // not found
459
460
461
462
463
        if (is_index) {
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "substring not found"));
        } else {
            return MP_OBJ_NEW_SMALL_INT(-1);
        }
464
465
    } else {
        // found
466
        return MP_OBJ_NEW_SMALL_INT(p - haystack);
467
468
469
    }
}

470
STATIC mp_obj_t str_find(uint n_args, const mp_obj_t *args) {
471
    return str_finder(n_args, args, 1, false);
472
473
474
}

STATIC mp_obj_t str_rfind(uint n_args, const mp_obj_t *args) {
475
476
477
478
479
480
481
482
483
    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);
484
485
}

486
// TODO: (Much) more variety in args
487
STATIC mp_obj_t str_startswith(mp_obj_t self_in, mp_obj_t arg) {
488
489
490
491
492
493
494
495
    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);
}

496
497
498
enum { LSTRIP, RSTRIP, STRIP };

STATIC mp_obj_t str_uni_strip(int type, 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
    machine_uint_t last_good_char_pos = 0;
521
    // TODO: For RSPLIT, scan from end
522
    for (machine_uint_t i = 0; i < orig_str_len; i++) {
523
        if (find_subbytes(chars_to_del, chars_to_del_len, &orig_str[i], 1, 1) == NULL) {
xbe's avatar
xbe committed
524
525
            if (!first_good_char_pos_set) {
                first_good_char_pos = i;
526
527
528
529
                if (type == LSTRIP) {
                    last_good_char_pos = orig_str_len - 1;
                    break;
                }
xbe's avatar
xbe committed
530
531
                first_good_char_pos_set = true;
            }
532
            last_good_char_pos = i;
xbe's avatar
xbe committed
533
534
535
536
        }
    }

    if (first_good_char_pos == 0 && last_good_char_pos == 0) {
537
538
        // string is all whitespace, return ''
        return MP_OBJ_NEW_QSTR(MP_QSTR_);
xbe's avatar
xbe committed
539
540
541
    }

    assert(last_good_char_pos >= first_good_char_pos);
542
543
544
    if (type == RSTRIP) {
        first_good_char_pos = 0;
    }
xbe's avatar
xbe committed
545
    //+1 to accomodate the last character
546
    machine_uint_t stripped_len = last_good_char_pos - first_good_char_pos + 1;
547
    return mp_obj_new_str(orig_str + first_good_char_pos, stripped_len, false);
xbe's avatar
xbe committed
548
549
}

550
551
552
553
554
555
556
557
558
559
560
561
STATIC mp_obj_t str_strip(uint n_args, const mp_obj_t *args) {
    return str_uni_strip(STRIP, n_args, args);
}

STATIC mp_obj_t str_lstrip(uint n_args, const mp_obj_t *args) {
    return str_uni_strip(LSTRIP, n_args, args);
}

STATIC mp_obj_t str_rstrip(uint n_args, const mp_obj_t *args) {
    return str_uni_strip(RSTRIP, n_args, args);
}

Dave Hylands's avatar
Dave Hylands committed
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
// 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
    ;
}

597
static mp_obj_t arg_as_int(mp_obj_t arg) {
598
599
#if MICROPY_ENABLE_FLOAT
    if (MP_OBJ_IS_TYPE(arg, &mp_type_float)) {
600
601
602
603
604

        // 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);
605
606
    }
#endif
607
    return arg;
608
609
}

610
mp_obj_t mp_obj_str_format(uint n_args, const mp_obj_t *args) {
611
    assert(MP_OBJ_IS_STR(args[0]));
612

613
    GET_STR_DATA_LEN(args[0], str, len);
Dave Hylands's avatar
Dave Hylands committed
614
    int arg_i = 0;
615
    vstr_t *vstr = vstr_new();
Dave Hylands's avatar
Dave Hylands committed
616
617
618
619
    pfenv_t pfenv_vstr;
    pfenv_vstr.data = vstr;
    pfenv_vstr.print_strn = pfenv_vstr_add_strn;

620
    for (const byte *top = str + len; str < top; str++) {
Dave Hylands's avatar
Dave Hylands committed
621
622
623
624
625
626
        if (*str == '}') {
            str++;
            if (str < top && *str == '}') {
                vstr_add_char(vstr, '}');
                continue;
            }
627
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "Single '}' encountered in format string"));
Dave Hylands's avatar
Dave Hylands committed
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
        }
        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 == '!') {
657
            str++;
Dave Hylands's avatar
Dave Hylands committed
658
659
            if (str < top && (*str == 'r' || *str == 's')) {
                conversion = *str++;
660
            } else {
661
                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
662
663
664
665
666
667
668
669
670
671
672
673
674
675
            }
        }

        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++);
676
                }
Dave Hylands's avatar
Dave Hylands committed
677
678
679
680
                vstr_add_char(format_spec, '\0');
            }
        }
        if (str >= top) {
681
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "unmatched '{' in format"));
Dave Hylands's avatar
Dave Hylands committed
682
683
        }
        if (*str != '}') {
684
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "expected ':' after format specifier"));
Dave Hylands's avatar
Dave Hylands committed
685
686
687
688
689
690
        }

        mp_obj_t arg = mp_const_none;

        if (field_name) {
            if (arg_i > 0) {
691
                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
692
            }
693
            int index = 0;
Dave Hylands's avatar
Dave Hylands committed
694
            if (str_to_int(vstr_str(field_name), &index) != vstr_len(field_name) - 1) {
695
                nlr_raise(mp_obj_new_exception_msg(&mp_type_KeyError, "attributes not supported yet"));
696
            }
Dave Hylands's avatar
Dave Hylands committed
697
            if (index >= n_args - 1) {
698
                nlr_raise(mp_obj_new_exception_msg(&mp_type_IndexError, "tuple index out of range"));
Dave Hylands's avatar
Dave Hylands committed
699
700
701
702
703
            }
            arg = args[index + 1];
            arg_i = -1;
            vstr_free(field_name);
            field_name = NULL;
704
        } else {
Dave Hylands's avatar
Dave Hylands committed
705
            if (arg_i < 0) {
706
                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
707
708
            }
            if (arg_i >= n_args - 1) {
709
                nlr_raise(mp_obj_new_exception_msg(&mp_type_IndexError, "tuple index out of range"));
Dave Hylands's avatar
Dave Hylands committed
710
711
712
713
714
715
716
717
718
719
720
721
722
723
            }
            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 {
724
                nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError, "Unknown conversion specifier %c", conversion));
Dave Hylands's avatar
Dave Hylands committed
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
            }
            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) {
791
                nlr_raise(mp_obj_new_exception_msg(&mp_type_KeyError, "Invalid conversion specification"));
Dave Hylands's avatar
Dave Hylands committed
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
            }
            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') {
809
                nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "Sign not allowed in string format specifier"));
Dave Hylands's avatar
Dave Hylands committed
810
811
            }
            if (type == 'c') {
812
                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
813
814
815
816
817
818
819
820
821
822
823
824
825
826
            }
        } 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':
827
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 2, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
828
829
830
831
832
833
834
835
836
837
838
839
                    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':
840
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 10, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
841
842
843
                    continue;

                case 'o':
844
845
846
847
                    if (flags & PF_FLAG_SHOW_PREFIX) {
                        flags |= PF_FLAG_SHOW_OCTAL_LETTER;
                    }

848
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 8, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
849
850
851
                    continue;

                case 'x':
852
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 16, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
853
854
855
                    continue;

                case 'X':
856
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 16, 'A', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
857
858
859
860
861
862
863
864
865
866
867
868
869
870
                    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:
871
                    nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
872
873
                        "Unknown format code '%c' for object of type '%s'", type, mp_obj_get_type_str(arg)));
            }
874
        }
875

876
877
        // NOTE: no else here. We need the e, f, g etc formats for integer
        //       arguments (from above if) to take this if.
878
        if (arg_looks_numeric(arg)) {
Dave Hylands's avatar
Dave Hylands committed
879
880
881
882
883
884
885
886
887
888
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
            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) {
916
#if MICROPY_ENABLE_FLOAT
Dave Hylands's avatar
Dave Hylands committed
917
918
919
920
921
922
923
924
925
926
927
928
929
                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;
930
#endif
Dave Hylands's avatar
Dave Hylands committed
931
932

                default:
933
                    nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
934
935
936
937
                        "Unknown format code '%c' for object of type 'float'",
                        type, mp_obj_get_type_str(arg)));
            }
        } else {
938
939
            // arg doesn't look like a number

Dave Hylands's avatar
Dave Hylands committed
940
            if (align == '=') {
941
                nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "'=' alignment not allowed in string format specifier"));
Dave Hylands's avatar
Dave Hylands committed
942
            }
943

Dave Hylands's avatar
Dave Hylands committed
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
            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:
964
                    nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
965
966
967
                        "Unknown format code '%c' for object of type 'str'",
                        type, mp_obj_get_type_str(arg)));
            }
968
969
970
        }
    }

971
972
973
    mp_obj_t s = mp_obj_new_str((byte*)vstr->buf, vstr->len, false);
    vstr_free(vstr);
    return s;
974
975
}

976
977
978
979
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
980
    const byte *start_str = str;
981
982
    int arg_i = 0;
    vstr_t *vstr = vstr_new();
Dave Hylands's avatar
Dave Hylands committed
983
984
985
986
    pfenv_t pfenv_vstr;
    pfenv_vstr.data = vstr;
    pfenv_vstr.print_strn = pfenv_vstr_add_strn;

987
    for (const byte *top = str + len; str < top; str++) {
Dave Hylands's avatar
Dave Hylands committed
988
989
990
991
992
993
994
        if (*str != '%') {
            vstr_add_char(vstr, *str);
            continue;
        }
        if (++str >= top) {
            break;
        }
995
        if (*str == '%') {
Dave Hylands's avatar
Dave Hylands committed
996
997
998
999
            vstr_add_char(vstr, '%');
            continue;
        }
        if (arg_i >= n_args) {
1000
            nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "not enough arguments for format string"));
Dave Hylands's avatar
Dave Hylands committed
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
        }
        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++;
1022
            } else {
Dave Hylands's avatar
Dave Hylands committed
1023
1024
                for (; str < top && '0' <= *str && *str <= '9'; str++) {
                    width = width * 10 + *str - '0';
1025
                }
Dave Hylands's avatar
Dave Hylands committed
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
            }
        }
        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) {
1044
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "incomplete format"));
Dave Hylands's avatar
Dave Hylands committed
1045
1046
1047
1048
1049
1050
1051
1052
        }
        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) {
1053
                        nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "%c requires int or char")); 
1054
                        break;
Dave Hylands's avatar
Dave Hylands committed
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
                    }
                    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)) {
1067
                    nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "integer argument expected, got float")); 
Dave Hylands's avatar
Dave Hylands committed
1068
1069
1070

                }
#endif
1071
                nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "an integer is required")); 
Dave Hylands's avatar
Dave Hylands committed
1072
1073
1074
1075
1076
                break; 

            case 'd':
            case 'i':
            case 'u':
1077
                pfenv_print_mp_int(&pfenv_vstr, arg_as_int(arg), 1, 10, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
                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) {
1093
                    flags |= (PF_FLAG_SHOW_PREFIX | PF_FLAG_SHOW_OCTAL_LETTER);
Dave Hylands's avatar
Dave Hylands committed
1094
                }
1095
                pfenv_print_mp_int(&pfenv_vstr, arg_as_int(arg), 1, 8, 'a', flags, fill, width); 
Dave Hylands's avatar
Dave Hylands committed
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
                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;
1110
                }
Dave Hylands's avatar
Dave Hylands committed
1111
1112
1113
                pfenv_print_strn(&pfenv_vstr, vstr_str(arg_vstr), len, flags, ' ', width);
                vstr_free(arg_vstr);
                break;
1114
            }
Dave Hylands's avatar
Dave Hylands committed
1115
1116
1117
1118
1119

            case 'x':
                if (alt) {
                    flags |= PF_FLAG_SHOW_PREFIX;
                }
1120
                pfenv_print_mp_int(&pfenv_vstr, arg_as_int(arg), 1, 16, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
1121
1122
1123
1124
1125
1126
                break;

            case 'X':
                if (alt) {
                    flags |= PF_FLAG_SHOW_PREFIX;
                }
1127
                pfenv_print_mp_int(&pfenv_vstr, arg_as_int(arg), 1, 16, 'A', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
1128
                break;
1129

Dave Hylands's avatar
Dave Hylands committed
1130
            default:
1131
                nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
1132
1133
                    "unsupported format character '%c' (0x%x) at index %d",
                    *str, *str, str - start_str));
1134
        }
Dave Hylands's avatar
Dave Hylands committed
1135
        arg_i++;
1136
1137
1138
    }

    if (arg_i != n_args) {
1139
        nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "not all arguments converted during string formatting"));
1140
1141
1142
1143
1144
1145
1146
    }

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

1147
STATIC mp_obj_t str_replace(uint n_args, const mp_obj_t *args) {
1148
1149
    assert(MP_OBJ_IS_STR(args[0]));

1150
    machine_int_t max_rep = -1;
1151
    if (n_args == 4) {
1152
        max_rep = mp_obj_get_int(args[3]);
1153
1154
1155
        if (max_rep == 0) {
            return args[0];
        } else if (max_rep < 0) {
1156
            max_rep = -1;
1157
        }
1158
    }
1159

xbe's avatar
xbe committed
1160
    // if max_rep is still -1 by this point we will need to do all possible replacements
1161

1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
    // 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

1174
1175
1176
    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);
1177
1178

    // old won't exist in str if it's longer, so nothing to replace
1179
    if (old_len > str_len) {
1180
        return args[0];
1181
1182
    }

1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
    // 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;
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
        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;
            }
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
            // 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;
1220
            str_len_remain = str + str_len - offset_ptr;
1221
1222
1223
1224
1225
            num_replacements_done++;
        }

        // copy from just after end of last occurrence of to-be-replaced string to end of old string
        if (data != NULL) {
1226
            memcpy(data + replaced_str_index, offset_ptr, str_len_remain);
1227
        }
1228
        replaced_str_index += str_len_remain;
1229
1230
1231
1232
1233
1234
1235
1236
1237

        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);
1238
                assert(data != NULL);
1239
1240
1241
1242
1243
            }
        } else {
            // second pass, we are done
            break;
        }
1244
    }
1245

1246
1247
1248
    return mp_obj_str_builder_end(replaced_str);
}

1249
1250
1251
1252
1253
1254
1255
1256
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]));