objstr.c 51.6 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
225
226
            // TODO: need predicate to check for int-like type (bools are such for example)
            // ["no", "yes"][1 == 2] is common idiom
            if (MP_OBJ_IS_SMALL_INT(rhs_in)) {
227
                uint index = mp_get_index(mp_obj_get_type(lhs_in), lhs_len, rhs_in, false);
228
                if (MP_OBJ_IS_TYPE(lhs_in, &mp_type_bytes)) {
229
                    return MP_OBJ_NEW_SMALL_INT((mp_small_int_t)lhs_data[index]);
230
231
232
                } else {
                    return mp_obj_new_str(lhs_data + index, 1, true);
                }
233
#if MICROPY_ENABLE_SLICE
234
            } else if (MP_OBJ_IS_TYPE(rhs_in, &mp_type_slice)) {
235
                machine_uint_t start, stop;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
236
237
238
                if (!m_seq_get_fast_slice_indexes(lhs_len, rhs_in, &start, &stop)) {
                    assert(0);
                }
239
                return mp_obj_new_str(lhs_data + start, stop - start, false);
240
#endif
241
            } else {
242
243
                // Message doesn't match CPython, but we don't have so much bytes as they
                // to spend them on verbose wording
244
                nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "index must be int"));
245
            }
246

Damien George's avatar
Damien George committed
247
248
        case MP_BINARY_OP_ADD:
        case MP_BINARY_OP_INPLACE_ADD:
249
            if (MP_OBJ_IS_STR(rhs_in)) {
250
                // add 2 strings
251
252

                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
253
                int alloc_len = lhs_len + rhs_len;
254
255

                /* code for making qstr
256
257
258
259
                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);
260
261
262
263
264
                return MP_OBJ_NEW_QSTR(qstr_build_end(q_ptr));
                */

                // code for non-qstr
                byte *data;
265
                mp_obj_t s = mp_obj_str_builder_start(mp_obj_get_type(lhs_in), alloc_len, &data);
266
267
268
                memcpy(data, lhs_data, lhs_len);
                memcpy(data + lhs_len, rhs_data, rhs_len);
                return mp_obj_str_builder_end(s);
269
270
            }
            break;
271

Damien George's avatar
Damien George committed
272
        case MP_BINARY_OP_IN:
273
            /* NOTE `a in b` is `b.__contains__(a)` */
274
275
            if (MP_OBJ_IS_STR(rhs_in)) {
                GET_STR_DATA_LEN(rhs_in, rhs_data, rhs_len);
276
                return MP_BOOL(find_subbytes(lhs_data, lhs_len, rhs_data, rhs_len, 1) != NULL);
277
278
            }
            break;
279

Damien George's avatar
Damien George committed
280
        case MP_BINARY_OP_MULTIPLY:
281
282
283
284
285
        {
            if (!MP_OBJ_IS_SMALL_INT(rhs_in)) {
                return NULL;
            }
            int n = MP_OBJ_SMALL_INT_VALUE(rhs_in);
286
            byte *data;
287
            mp_obj_t s = mp_obj_str_builder_start(mp_obj_get_type(lhs_in), lhs_len * n, &data);
288
289
            mp_seq_multiply(lhs_data, sizeof(*lhs_data), lhs_len, n, data);
            return mp_obj_str_builder_end(s);
290
        }
291

292
293
294
295
296
297
298
299
300
301
302
303
304
        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
305
306
307
308
309
310
311
        // 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:
312
313
314
315
            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));
            }
316
317
318
319
320
    }

    return MP_OBJ_NULL; // op not supported
}

321
STATIC mp_obj_t str_join(mp_obj_t self_in, mp_obj_t arg) {
322
    assert(MP_OBJ_IS_STR(self_in));
323

324
    // get separation string
325
    GET_STR_DATA_LEN(self_in, sep_str, sep_len);
326
327

    // process args
328
329
    uint seq_len;
    mp_obj_t *seq_items;
330
    if (MP_OBJ_IS_TYPE(arg, &mp_type_tuple)) {
331
332
        mp_obj_tuple_get(arg, &seq_len, &seq_items);
    } else {
333
334
335
336
337
        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);
338
    }
339
340
341

    // count required length
    int required_len = 0;
342
    for (int i = 0; i < seq_len; i++) {
343
        if (!MP_OBJ_IS_STR(seq_items[i])) {
344
            nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "join expected a list of str's"));
345
        }
346
347
348
        if (i > 0) {
            required_len += sep_len;
        }
349
350
        GET_STR_LEN(seq_items[i], l);
        required_len += l;
351
352
353
    }

    // make joined string
354
    byte *data;
355
    mp_obj_t joined_str = mp_obj_str_builder_start(mp_obj_get_type(self_in), required_len, &data);
356
357
    for (int i = 0; i < seq_len; i++) {
        if (i > 0) {
358
359
            memcpy(data, sep_str, sep_len);
            data += sep_len;
360
        }
361
362
363
        GET_STR_DATA_LEN(seq_items[i], s, l);
        memcpy(data, s, l);
        data += l;
364
    }
365
366

    // return joined string
367
    return mp_obj_str_builder_end(joined_str);
368
369
}

Paul Sokolovsky's avatar
Paul Sokolovsky committed
370
371
#define is_ws(c) ((c) == ' ' || (c) == '\t')

372
STATIC mp_obj_t str_split(uint n_args, const mp_obj_t *args) {
373
    machine_int_t splits = -1;
Paul Sokolovsky's avatar
Paul Sokolovsky committed
374
375
376
377
    mp_obj_t sep = mp_const_none;
    if (n_args > 1) {
        sep = args[1];
        if (n_args > 2) {
378
            splits = mp_obj_get_int(args[2]);
Paul Sokolovsky's avatar
Paul Sokolovsky committed
379
380
        }
    }
381

Paul Sokolovsky's avatar
Paul Sokolovsky committed
382
    mp_obj_t res = mp_obj_new_list(0, NULL);
383
384
    GET_STR_DATA_LEN(args[0], s, len);
    const byte *top = s + len;
385
386
387
388
389

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

        // Initial whitespace is not counted as split, so we pre-do it
390
        while (s < top && is_ws(*s)) s++;
391
392
393
394
395
396
397
398
399
400
401
        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
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
436
437
        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
438
439
440
441
442
    }

    return res;
}

443
STATIC mp_obj_t str_finder(uint n_args, const mp_obj_t *args, machine_int_t direction) {
444
    assert(2 <= n_args && n_args <= 4);
445
446
    assert(MP_OBJ_IS_STR(args[0]));
    assert(MP_OBJ_IS_STR(args[1]));
447

448
449
    GET_STR_DATA_LEN(args[0], haystack, haystack_len);
    GET_STR_DATA_LEN(args[1], needle, needle_len);
450

451
452
    machine_uint_t start = 0;
    machine_uint_t end = haystack_len;
453
    if (n_args >= 3 && args[2] != mp_const_none) {
454
        start = mp_get_index(&mp_type_str, haystack_len, args[2], true);
455
456
    }
    if (n_args >= 4 && args[3] != mp_const_none) {
457
        end = mp_get_index(&mp_type_str, haystack_len, args[3], true);
458
459
    }

460
    const byte *p = find_subbytes(haystack + start, end - start, needle, needle_len, direction);
461
462
463
464
465
    if (p == NULL) {
        // not found
        return MP_OBJ_NEW_SMALL_INT(-1);
    } else {
        // found
466
        return MP_OBJ_NEW_SMALL_INT(p - haystack);
467
468
469
    }
}

470
471
472
473
474
475
476
477
STATIC mp_obj_t str_find(uint n_args, const mp_obj_t *args) {
    return str_finder(n_args, args, 1);
}

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

478
// TODO: (Much) more variety in args
479
STATIC mp_obj_t str_startswith(mp_obj_t self_in, mp_obj_t arg) {
480
481
482
483
484
485
486
487
    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);
}

488
STATIC mp_obj_t str_strip(uint n_args, const mp_obj_t *args) {
xbe's avatar
xbe committed
489
    assert(1 <= n_args && n_args <= 2);
490
491
492
493
494
    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
495
496
497

    if (n_args == 1) {
        chars_to_del = whitespace;
498
        chars_to_del_len = sizeof(whitespace);
xbe's avatar
xbe committed
499
    } else {
500
501
502
503
        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
504
505
    }

506
    GET_STR_DATA_LEN(args[0], orig_str, orig_str_len);
xbe's avatar
xbe committed
507

508
    machine_uint_t first_good_char_pos = 0;
xbe's avatar
xbe committed
509
    bool first_good_char_pos_set = false;
510
511
    machine_uint_t last_good_char_pos = 0;
    for (machine_uint_t i = 0; i < orig_str_len; i++) {
512
        if (find_subbytes(chars_to_del, chars_to_del_len, &orig_str[i], 1, 1) == NULL) {
xbe's avatar
xbe committed
513
514
515
516
517
518
519
520
521
            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) {
522
523
        // string is all whitespace, return ''
        return MP_OBJ_NEW_QSTR(MP_QSTR_);
xbe's avatar
xbe committed
524
525
526
527
    }

    assert(last_good_char_pos >= first_good_char_pos);
    //+1 to accomodate the last character
528
    machine_uint_t stripped_len = last_good_char_pos - first_good_char_pos + 1;
529
    return mp_obj_new_str(orig_str + first_good_char_pos, stripped_len, false);
xbe's avatar
xbe committed
530
531
}

Dave Hylands's avatar
Dave Hylands committed
532
533
534
535
536
537
538
539
540
541
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
// 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
    ;
}

567
static mp_obj_t arg_as_int(mp_obj_t arg) {
568
569
#if MICROPY_ENABLE_FLOAT
    if (MP_OBJ_IS_TYPE(arg, &mp_type_float)) {
570
571
572
573
574

        // 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);
575
576
    }
#endif
577
    return arg;
578
579
}

580
mp_obj_t str_format(uint n_args, const mp_obj_t *args) {
581
    assert(MP_OBJ_IS_STR(args[0]));
582

583
    GET_STR_DATA_LEN(args[0], str, len);
Dave Hylands's avatar
Dave Hylands committed
584
    int arg_i = 0;
585
    vstr_t *vstr = vstr_new();
Dave Hylands's avatar
Dave Hylands committed
586
587
588
589
    pfenv_t pfenv_vstr;
    pfenv_vstr.data = vstr;
    pfenv_vstr.print_strn = pfenv_vstr_add_strn;

590
    for (const byte *top = str + len; str < top; str++) {
Dave Hylands's avatar
Dave Hylands committed
591
592
593
594
595
596
        if (*str == '}') {
            str++;
            if (str < top && *str == '}') {
                vstr_add_char(vstr, '}');
                continue;
            }
597
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "Single '}' encountered in format string"));
Dave Hylands's avatar
Dave Hylands committed
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
        }
        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 == '!') {
627
            str++;
Dave Hylands's avatar
Dave Hylands committed
628
629
            if (str < top && (*str == 'r' || *str == 's')) {
                conversion = *str++;
630
            } else {
631
                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
632
633
634
635
636
637
638
639
640
641
642
643
644
645
            }
        }

        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++);
646
                }
Dave Hylands's avatar
Dave Hylands committed
647
648
649
650
                vstr_add_char(format_spec, '\0');
            }
        }
        if (str >= top) {
651
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "unmatched '{' in format"));
Dave Hylands's avatar
Dave Hylands committed
652
653
        }
        if (*str != '}') {
654
            nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "expected ':' after format specifier"));
Dave Hylands's avatar
Dave Hylands committed
655
656
657
658
659
660
        }

        mp_obj_t arg = mp_const_none;

        if (field_name) {
            if (arg_i > 0) {
661
                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
662
663
664
            }
            int index;
            if (str_to_int(vstr_str(field_name), &index) != vstr_len(field_name) - 1) {
665
                nlr_raise(mp_obj_new_exception_msg(&mp_type_KeyError, "attributes not supported yet"));
666
            }
Dave Hylands's avatar
Dave Hylands committed
667
            if (index >= n_args - 1) {
668
                nlr_raise(mp_obj_new_exception_msg(&mp_type_IndexError, "tuple index out of range"));
Dave Hylands's avatar
Dave Hylands committed
669
670
671
672
673
            }
            arg = args[index + 1];
            arg_i = -1;
            vstr_free(field_name);
            field_name = NULL;
674
        } else {
Dave Hylands's avatar
Dave Hylands committed
675
            if (arg_i < 0) {
676
                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
677
678
            }
            if (arg_i >= n_args - 1) {
679
                nlr_raise(mp_obj_new_exception_msg(&mp_type_IndexError, "tuple index out of range"));
Dave Hylands's avatar
Dave Hylands committed
680
681
682
683
684
685
686
687
688
689
690
691
692
693
            }
            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 {
694
                nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError, "Unknown conversion specifier %c", conversion));
Dave Hylands's avatar
Dave Hylands committed
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
            }
            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) {
761
                nlr_raise(mp_obj_new_exception_msg(&mp_type_KeyError, "Invalid conversion specification"));
Dave Hylands's avatar
Dave Hylands committed
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
            }
            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') {
779
                nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "Sign not allowed in string format specifier"));
Dave Hylands's avatar
Dave Hylands committed
780
781
            }
            if (type == 'c') {
782
                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
783
784
785
786
787
788
789
790
791
792
793
794
795
796
            }
        } 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':
797
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 2, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
798
799
800
801
802
803
804
805
806
807
808
809
                    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':
810
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 10, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
811
812
813
                    continue;

                case 'o':
814
815
816
817
                    if (flags & PF_FLAG_SHOW_PREFIX) {
                        flags |= PF_FLAG_SHOW_OCTAL_LETTER;
                    }

818
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 8, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
819
820
821
                    continue;

                case 'x':
822
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 16, 'a', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
823
824
825
                    continue;

                case 'X':
826
                    pfenv_print_mp_int(&pfenv_vstr, arg, 1, 16, 'A', flags, fill, width);
Dave Hylands's avatar
Dave Hylands committed
827
828
829
830
831
832
833
834
835
836
837
838
839
840
                    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:
841
                    nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
842
843
                        "Unknown format code '%c' for object of type '%s'", type, mp_obj_get_type_str(arg)));
            }
844
        }
845

846
847
        // NOTE: no else here. We need the e, f, g etc formats for integer
        //       arguments (from above if) to take this if.
848
        if (arg_looks_numeric(arg)) {
Dave Hylands's avatar
Dave Hylands committed
849
850
851
852
853
854
855
856
857
858
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
            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) {
886
#if MICROPY_ENABLE_FLOAT
Dave Hylands's avatar
Dave Hylands committed
887
888
889
890
891
892
893
894
895
896
897
898
899
                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;
900
#endif
Dave Hylands's avatar
Dave Hylands committed
901
902

                default:
903
                    nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
904
905
906
907
                        "Unknown format code '%c' for object of type 'float'",
                        type, mp_obj_get_type_str(arg)));
            }
        } else {
908
909
            // arg doesn't look like a number

Dave Hylands's avatar
Dave Hylands committed
910
            if (align == '=') {
911
                nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "'=' alignment not allowed in string format specifier"));
Dave Hylands's avatar
Dave Hylands committed
912
            }
913

Dave Hylands's avatar
Dave Hylands committed
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
            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:
934
                    nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError,
Dave Hylands's avatar
Dave Hylands committed
935
936
937
                        "Unknown format code '%c' for object of type 'str'",
                        type, mp_obj_get_type_str(arg)));
            }
938
939
940
        }
    }

941
942
943
    mp_obj_t s = mp_obj_new_str((byte*)vstr->buf, vstr->len, false);
    vstr_free(vstr);
    return s;
944
945
}

946
947
948
949
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
950
    const byte *start_str = str;
951
952
    int arg_i = 0;
    vstr_t *vstr = vstr_new();
Dave Hylands's avatar
Dave Hylands committed
953
954
955
956
    pfenv_t pfenv_vstr;
    pfenv_vstr.data = vstr;
    pfenv_vstr.print_strn = pfenv_vstr_add_strn;

957
    for (const byte *top = str + len; str < top; str++) {
Dave Hylands's avatar
Dave Hylands committed
958
959
960
961
962
963
964
        if (*str != '%') {
            vstr_add_char(vstr, *str);
            continue;
        }
        if (++str >= top) {
            break;
        }
965
        if (*str == '%') {
Dave Hylands's avatar
Dave Hylands committed
966
967
968
969
            vstr_add_char(vstr, '%');
            continue;
        }
        if (arg_i >= n_args) {
970
            nlr_raise(mp_obj_new_exception_msg(&mp_type_TypeError, "not enough arguments for format string"));
Dave Hylands's avatar
Dave Hylands committed
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
        }
        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++;
992
            } else {
Dave Hylands's avatar
Dave Hylands committed
993
994
                for (; str < top && '0' <= *str && *str <= '9'; str++) {
                    width = width * 10 + *str - '0';
995
                }
Dave Hylands's avatar
Dave Hylands committed
996
997
998
999
1000
            }
        }
        int prec = -1;
        if (str < top && *str == '.') {
            if (++str < top) {
For faster browsing, not all history is shown. View entire blame