lexer.c 25.5 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
/*
 * This file is part of the Micro Python project, http://micropython.org/
 *
 * The MIT License (MIT)
 *
 * Copyright (c) 2013, 2014 Damien P. George
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 * THE SOFTWARE.
 */

Damien's avatar
Damien committed
27
28
29
#include <stdio.h>
#include <assert.h>

30
#include "py/lexer.h"
Damien's avatar
Damien committed
31
32
33

#define TAB_SIZE (8)

34
35
36
// TODO seems that CPython allows NULL byte in the input stream
// don't know if that's intentional or not, but we don't allow it

37
mp_uint_t mp_optimise_value;
38

39
// TODO replace with a call to a standard function
40
STATIC bool str_strn_equal(const char *str, const char *strn, mp_uint_t len) {
41
    mp_uint_t i = 0;
Damien's avatar
Damien committed
42

43
    while (i < len && *str == *strn) {
Damien's avatar
Damien committed
44
45
        ++i;
        ++str;
46
        ++strn;
Damien's avatar
Damien committed
47
48
    }

49
    return i == len && *str == 0;
Damien's avatar
Damien committed
50
51
}

52
53
#define CUR_CHAR(lex) ((lex)->chr0)

54
STATIC bool is_end(mp_lexer_t *lex) {
55
    return lex->chr0 == MP_LEXER_EOF;
Damien's avatar
Damien committed
56
57
}

58
STATIC bool is_physical_newline(mp_lexer_t *lex) {
Damien's avatar
Damien committed
59
60
61
    return lex->chr0 == '\n' || lex->chr0 == '\r';
}

62
STATIC bool is_char(mp_lexer_t *lex, char c) {
Damien's avatar
Damien committed
63
64
65
    return lex->chr0 == c;
}

66
STATIC bool is_char_or(mp_lexer_t *lex, char c1, char c2) {
Damien's avatar
Damien committed
67
68
69
    return lex->chr0 == c1 || lex->chr0 == c2;
}

70
STATIC bool is_char_or3(mp_lexer_t *lex, char c1, char c2, char c3) {
Damien's avatar
Damien committed
71
72
73
74
    return lex->chr0 == c1 || lex->chr0 == c2 || lex->chr0 == c3;
}

/*
75
STATIC bool is_char_following(mp_lexer_t *lex, char c) {
Damien's avatar
Damien committed
76
77
78
79
    return lex->chr1 == c;
}
*/

80
STATIC bool is_char_following_or(mp_lexer_t *lex, char c1, char c2) {
Damien's avatar
Damien committed
81
82
83
    return lex->chr1 == c1 || lex->chr1 == c2;
}

84
STATIC bool is_char_following_following_or(mp_lexer_t *lex, char c1, char c2) {
Damien's avatar
Damien committed
85
86
87
    return lex->chr2 == c1 || lex->chr2 == c2;
}

88
STATIC bool is_char_and(mp_lexer_t *lex, char c1, char c2) {
Damien's avatar
Damien committed
89
90
91
    return lex->chr0 == c1 && lex->chr1 == c2;
}

92
STATIC bool is_whitespace(mp_lexer_t *lex) {
93
    return unichar_isspace(lex->chr0);
Damien's avatar
Damien committed
94
95
}

96
STATIC bool is_letter(mp_lexer_t *lex) {
97
    return unichar_isalpha(lex->chr0);
Damien's avatar
Damien committed
98
99
}

100
STATIC bool is_digit(mp_lexer_t *lex) {
101
    return unichar_isdigit(lex->chr0);
Damien's avatar
Damien committed
102
103
}

104
STATIC bool is_following_digit(mp_lexer_t *lex) {
105
    return unichar_isdigit(lex->chr1);
Damien's avatar
Damien committed
106
107
}

108
STATIC bool is_following_odigit(mp_lexer_t *lex) {
109
110
111
    return lex->chr1 >= '0' && lex->chr1 <= '7';
}

Damien's avatar
Damien committed
112
// TODO UNICODE include unicode characters in definition of identifiers
113
STATIC bool is_head_of_identifier(mp_lexer_t *lex) {
Damien's avatar
Damien committed
114
115
116
117
    return is_letter(lex) || lex->chr0 == '_';
}

// TODO UNICODE include unicode characters in definition of identifiers
118
STATIC bool is_tail_of_identifier(mp_lexer_t *lex) {
Damien's avatar
Damien committed
119
120
121
    return is_head_of_identifier(lex) || is_digit(lex);
}

122
STATIC void next_char(mp_lexer_t *lex) {
123
    if (lex->chr0 == MP_LEXER_EOF) {
Damien's avatar
Damien committed
124
125
126
        return;
    }

127
    mp_uint_t advance = 1;
Damien's avatar
Damien committed
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151

    if (lex->chr0 == '\n') {
        // LF is a new line
        ++lex->line;
        lex->column = 1;
    } else if (lex->chr0 == '\r') {
        // CR is a new line
        ++lex->line;
        lex->column = 1;
        if (lex->chr1 == '\n') {
            // CR LF is a single new line
            advance = 2;
        }
    } else if (lex->chr0 == '\t') {
        // a tab
        lex->column = (((lex->column - 1 + TAB_SIZE) / TAB_SIZE) * TAB_SIZE) + 1;
    } else {
        // a character worth one column
        ++lex->column;
    }

    for (; advance > 0; advance--) {
        lex->chr0 = lex->chr1;
        lex->chr1 = lex->chr2;
152
153
        lex->chr2 = lex->stream_next_byte(lex->stream_data);
        if (lex->chr2 == MP_LEXER_EOF) {
Damien's avatar
Damien committed
154
            // EOF
155
            if (lex->chr1 != MP_LEXER_EOF && lex->chr1 != '\n' && lex->chr1 != '\r') {
Damien's avatar
Damien committed
156
157
158
159
160
161
                lex->chr2 = '\n'; // insert newline at end of file
            }
        }
    }
}

162
STATIC void indent_push(mp_lexer_t *lex, mp_uint_t indent) {
Damien's avatar
Damien committed
163
    if (lex->num_indent_level >= lex->alloc_indent_level) {
164
        // TODO use m_renew_maybe and somehow indicate an error if it fails... probably by using MP_TOKEN_MEMORY_ERROR
165
166
        lex->indent_level = m_renew(uint16_t, lex->indent_level, lex->alloc_indent_level, lex->alloc_indent_level + MICROPY_ALLOC_LEXEL_INDENT_INC);
        lex->alloc_indent_level += MICROPY_ALLOC_LEXEL_INDENT_INC;
Damien's avatar
Damien committed
167
168
169
170
    }
    lex->indent_level[lex->num_indent_level++] = indent;
}

171
STATIC mp_uint_t indent_top(mp_lexer_t *lex) {
Damien's avatar
Damien committed
172
173
174
    return lex->indent_level[lex->num_indent_level - 1];
}

175
STATIC void indent_pop(mp_lexer_t *lex) {
Damien's avatar
Damien committed
176
177
178
179
180
181
182
183
184
185
    lex->num_indent_level -= 1;
}

// some tricky operator encoding:
//     <op>  = begin with <op>, if this opchar matches then begin here
//     e<op> = end with <op>, if this opchar matches then end
//     E<op> = mandatory end with <op>, this opchar must match, then end
//     c<op> = continue with <op>, if this opchar matches then continue matching
// this means if the start of two ops are the same then they are equal til the last char

186
STATIC const char *tok_enc =
Damien's avatar
Damien committed
187
188
189
190
191
192
193
194
195
196
197
198
    "()[]{},:;@~" // singles
    "<e=c<e="     // < <= << <<=
    ">e=c>e="     // > >= >> >>=
    "*e=c*e="     // * *= ** **=
    "+e="         // + +=
    "-e=e>"       // - -= ->
    "&e="         // & &=
    "|e="         // | |=
    "/e=c/e="     // / /= // //=
    "%e="         // % %=
    "^e="         // ^ ^=
    "=e="         // = ==
199
    "!E=";        // !=
Damien's avatar
Damien committed
200
201

// TODO static assert that number of tokens is less than 256 so we can safely make this table with byte sized entries
202
STATIC const uint8_t tok_enc_kind[] = {
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
    MP_TOKEN_DEL_PAREN_OPEN, MP_TOKEN_DEL_PAREN_CLOSE,
    MP_TOKEN_DEL_BRACKET_OPEN, MP_TOKEN_DEL_BRACKET_CLOSE,
    MP_TOKEN_DEL_BRACE_OPEN, MP_TOKEN_DEL_BRACE_CLOSE,
    MP_TOKEN_DEL_COMMA, MP_TOKEN_DEL_COLON, MP_TOKEN_DEL_SEMICOLON, MP_TOKEN_DEL_AT, MP_TOKEN_OP_TILDE,

    MP_TOKEN_OP_LESS, MP_TOKEN_OP_LESS_EQUAL, MP_TOKEN_OP_DBL_LESS, MP_TOKEN_DEL_DBL_LESS_EQUAL,
    MP_TOKEN_OP_MORE, MP_TOKEN_OP_MORE_EQUAL, MP_TOKEN_OP_DBL_MORE, MP_TOKEN_DEL_DBL_MORE_EQUAL,
    MP_TOKEN_OP_STAR, MP_TOKEN_DEL_STAR_EQUAL, MP_TOKEN_OP_DBL_STAR, MP_TOKEN_DEL_DBL_STAR_EQUAL,
    MP_TOKEN_OP_PLUS, MP_TOKEN_DEL_PLUS_EQUAL,
    MP_TOKEN_OP_MINUS, MP_TOKEN_DEL_MINUS_EQUAL, MP_TOKEN_DEL_MINUS_MORE,
    MP_TOKEN_OP_AMPERSAND, MP_TOKEN_DEL_AMPERSAND_EQUAL,
    MP_TOKEN_OP_PIPE, MP_TOKEN_DEL_PIPE_EQUAL,
    MP_TOKEN_OP_SLASH, MP_TOKEN_DEL_SLASH_EQUAL, MP_TOKEN_OP_DBL_SLASH, MP_TOKEN_DEL_DBL_SLASH_EQUAL,
    MP_TOKEN_OP_PERCENT, MP_TOKEN_DEL_PERCENT_EQUAL,
    MP_TOKEN_OP_CARET, MP_TOKEN_DEL_CARET_EQUAL,
    MP_TOKEN_DEL_EQUAL, MP_TOKEN_OP_DBL_EQUAL,
    MP_TOKEN_OP_NOT_EQUAL,
Damien's avatar
Damien committed
220
221
222
};

// must have the same order as enum in lexer.h
223
STATIC const char *tok_kw[] = {
Damien's avatar
Damien committed
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
    "False",
    "None",
    "True",
    "and",
    "as",
    "assert",
    "break",
    "class",
    "continue",
    "def",
    "del",
    "elif",
    "else",
    "except",
    "finally",
    "for",
    "from",
    "global",
    "if",
    "import",
    "in",
    "is",
    "lambda",
    "nonlocal",
    "not",
    "or",
    "pass",
    "raise",
    "return",
    "try",
    "while",
    "with",
    "yield",
257
    "__debug__",
Damien's avatar
Damien committed
258
259
};

260
STATIC mp_uint_t hex_digit(unichar c) {
261
    // c is assumed to be hex digit
262
    mp_uint_t n = c - '0';
263
264
265
266
267
268
269
270
271
    if (n > 9) {
        n &= ~('a' - 'A');
        n -= ('A' - ('9' + 1));
    }
    return n;
}

// This is called with CUR_CHAR() before first hex digit, and should return with
// it pointing to last hex digit
272
273
274
// num_digits must be greater than zero
STATIC bool get_hex(mp_lexer_t *lex, mp_uint_t num_digits, mp_uint_t *result) {
    mp_uint_t num = 0;
275
276
277
278
279
280
281
282
283
284
285
286
    while (num_digits-- != 0) {
        next_char(lex);
        unichar c = CUR_CHAR(lex);
        if (!unichar_isxdigit(c)) {
            return false;
        }
        num = (num << 4) + hex_digit(c);
    }
    *result = num;
    return true;
}

287
288
289
290
STATIC void mp_lexer_next_token_into(mp_lexer_t *lex, bool first_token) {
    // start new token text
    vstr_reset(&lex->vstr);

291
    // skip white space and comments
Damien's avatar
Damien committed
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
    bool had_physical_newline = false;
    while (!is_end(lex)) {
        if (is_physical_newline(lex)) {
            had_physical_newline = true;
            next_char(lex);
        } else if (is_whitespace(lex)) {
            next_char(lex);
        } else if (is_char(lex, '#')) {
            next_char(lex);
            while (!is_end(lex) && !is_physical_newline(lex)) {
                next_char(lex);
            }
            // had_physical_newline will be set on next loop
        } else if (is_char(lex, '\\')) {
            // backslash (outside string literals) must appear just before a physical newline
            next_char(lex);
            if (!is_physical_newline(lex)) {
309
                // SyntaxError: unexpected character after line continuation character
310
311
312
                lex->tok_line = lex->line;
                lex->tok_column = lex->column;
                lex->tok_kind = MP_TOKEN_BAD_LINE_CONTINUATION;
313
                return;
Damien's avatar
Damien committed
314
315
316
317
318
319
320
321
            } else {
                next_char(lex);
            }
        } else {
            break;
        }
    }

322
    // set token source information
323
324
    lex->tok_line = lex->line;
    lex->tok_column = lex->column;
325
326
327
328
329
330

    if (first_token && lex->line == 1 && lex->column != 1) {
        // check that the first token is in the first column
        // if first token is not on first line, we get a physical newline and
        // this check is done as part of normal indent/dedent checking below
        // (done to get equivalence with CPython)
331
        lex->tok_kind = MP_TOKEN_INDENT;
332
333

    } else if (lex->emit_dent < 0) {
334
        lex->tok_kind = MP_TOKEN_DEDENT;
Damien's avatar
Damien committed
335
336
337
        lex->emit_dent += 1;

    } else if (lex->emit_dent > 0) {
338
        lex->tok_kind = MP_TOKEN_INDENT;
Damien's avatar
Damien committed
339
340
        lex->emit_dent -= 1;

341
    } else if (had_physical_newline && lex->nested_bracket_level == 0) {
342
        lex->tok_kind = MP_TOKEN_NEWLINE;
Damien's avatar
Damien committed
343

344
        mp_uint_t num_spaces = lex->column - 1;
Damien's avatar
Damien committed
345
346
347
348
349
350
351
352
353
354
355
        lex->emit_dent = 0;
        if (num_spaces == indent_top(lex)) {
        } else if (num_spaces > indent_top(lex)) {
            indent_push(lex, num_spaces);
            lex->emit_dent += 1;
        } else {
            while (num_spaces < indent_top(lex)) {
                indent_pop(lex);
                lex->emit_dent -= 1;
            }
            if (num_spaces != indent_top(lex)) {
356
                lex->tok_kind = MP_TOKEN_DEDENT_MISMATCH;
Damien's avatar
Damien committed
357
358
359
360
361
            }
        }

    } else if (is_end(lex)) {
        if (indent_top(lex) > 0) {
362
            lex->tok_kind = MP_TOKEN_NEWLINE;
Damien's avatar
Damien committed
363
364
365
366
367
368
            lex->emit_dent = 0;
            while (indent_top(lex) > 0) {
                indent_pop(lex);
                lex->emit_dent -= 1;
            }
        } else {
369
            lex->tok_kind = MP_TOKEN_END;
Damien's avatar
Damien committed
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
        }

    } else if (is_char_or(lex, '\'', '\"')
               || (is_char_or3(lex, 'r', 'u', 'b') && is_char_following_or(lex, '\'', '\"'))
               || ((is_char_and(lex, 'r', 'b') || is_char_and(lex, 'b', 'r')) && is_char_following_following_or(lex, '\'', '\"'))) {
        // a string or bytes literal

        // parse type codes
        bool is_raw = false;
        bool is_bytes = false;
        if (is_char(lex, 'u')) {
            next_char(lex);
        } else if (is_char(lex, 'b')) {
            is_bytes = true;
            next_char(lex);
            if (is_char(lex, 'r')) {
                is_raw = true;
                next_char(lex);
            }
        } else if (is_char(lex, 'r')) {
            is_raw = true;
            next_char(lex);
            if (is_char(lex, 'b')) {
                is_bytes = true;
                next_char(lex);
            }
        }

        // set token kind
        if (is_bytes) {
400
            lex->tok_kind = MP_TOKEN_BYTES;
Damien's avatar
Damien committed
401
        } else {
402
            lex->tok_kind = MP_TOKEN_STRING;
Damien's avatar
Damien committed
403
404
405
406
407
408
409
410
411
412
        }

        // get first quoting character
        char quote_char = '\'';
        if (is_char(lex, '\"')) {
            quote_char = '\"';
        }
        next_char(lex);

        // work out if it's a single or triple quoted literal
413
        mp_uint_t num_quotes;
Damien's avatar
Damien committed
414
415
416
417
418
419
420
421
422
423
424
        if (is_char_and(lex, quote_char, quote_char)) {
            // triple quotes
            next_char(lex);
            next_char(lex);
            num_quotes = 3;
        } else {
            // single quotes
            num_quotes = 1;
        }

        // parse the literal
425
        mp_uint_t n_closing = 0;
Damien's avatar
Damien committed
426
427
428
        while (!is_end(lex) && (num_quotes > 1 || !is_char(lex, '\n')) && n_closing < num_quotes) {
            if (is_char(lex, quote_char)) {
                n_closing += 1;
429
                vstr_add_char(&lex->vstr, CUR_CHAR(lex));
Damien's avatar
Damien committed
430
431
            } else {
                n_closing = 0;
432
                if (is_char(lex, '\\')) {
Damien's avatar
Damien committed
433
                    next_char(lex);
434
                    unichar c = CUR_CHAR(lex);
435
436
437
438
439
                    if (is_raw) {
                        // raw strings allow escaping of quotes, but the backslash is also emitted
                        vstr_add_char(&lex->vstr, '\\');
                    } else {
                        switch (c) {
440
441
                            case MP_LEXER_EOF: break; // TODO a proper error message?
                            case '\n': c = MP_LEXER_EOF; break; // TODO check this works correctly (we are supposed to ignore it
442
443
444
445
446
447
448
449
450
451
                            case '\\': break;
                            case '\'': break;
                            case '"': break;
                            case 'a': c = 0x07; break;
                            case 'b': c = 0x08; break;
                            case 't': c = 0x09; break;
                            case 'n': c = 0x0a; break;
                            case 'v': c = 0x0b; break;
                            case 'f': c = 0x0c; break;
                            case 'r': c = 0x0d; break;
452
453
454
455
456
457
458
459
                            case 'u':
                            case 'U':
                                if (is_bytes) {
                                    // b'\u1234' == b'\\u1234'
                                    vstr_add_char(&lex->vstr, '\\');
                                    break;
                                }
                                // Otherwise fall through.
460
461
                            case 'x':
                            {
462
                                mp_uint_t num = 0;
463
                                if (!get_hex(lex, (c == 'x' ? 2 : c == 'u' ? 4 : 8), &num)) {
464
465
                                    // TODO error message
                                    assert(0);
466
467
                                }
                                c = num;
468
                                break;
469
                            }
470
471
472
473
474
475
476
477
                            case 'N':
                                // Supporting '\N{LATIN SMALL LETTER A}' == 'a' would require keeping the
                                // entire Unicode name table in the core. As of Unicode 6.3.0, that's nearly
                                // 3MB of text; even gzip-compressed and with minimal structure, it'll take
                                // roughly half a meg of storage. This form of Unicode escape may be added
                                // later on, but it's definitely not a priority right now. -- CJA 20140607
                                assert(!"Unicode name escapes not supported");
                                break;
478
479
480
                            default:
                                if (c >= '0' && c <= '7') {
                                    // Octal sequence, 1-3 chars
481
482
                                    mp_uint_t digits = 3;
                                    mp_uint_t num = c - '0';
483
484
485
486
487
488
489
490
491
492
493
                                    while (is_following_odigit(lex) && --digits != 0) {
                                        next_char(lex);
                                        num = num * 8 + (CUR_CHAR(lex) - '0');
                                    }
                                    c = num;
                                } else {
                                    // unrecognised escape character; CPython lets this through verbatim as '\' and then the character
                                    vstr_add_char(&lex->vstr, '\\');
                                }
                                break;
                        }
494
                    }
495
                    if (c != MP_LEXER_EOF) {
496
497
498
499
500
501
502
                        if (c < 0x110000 && !is_bytes) {
                            vstr_add_char(&lex->vstr, c);
                        } else if (c < 0x100 && is_bytes) {
                            vstr_add_byte(&lex->vstr, c);
                        } else {
                            assert(!"TODO: Throw an error, invalid escape code probably");
                        }
503
504
                    }
                } else {
505
506
507
                    // Add the "character" as a byte so that we remain 8-bit clean.
                    // This way, strings are parsed correctly whether or not they contain utf-8 chars.
                    vstr_add_byte(&lex->vstr, CUR_CHAR(lex));
Damien's avatar
Damien committed
508
509
510
511
512
513
514
                }
            }
            next_char(lex);
        }

        // check we got the required end quotes
        if (n_closing < num_quotes) {
515
            lex->tok_kind = MP_TOKEN_LONELY_STRING_OPEN;
Damien's avatar
Damien committed
516
517
        }

518
        // cut off the end quotes from the token text
519
        vstr_cut_tail_bytes(&lex->vstr, n_closing);
Damien's avatar
Damien committed
520
521

    } else if (is_head_of_identifier(lex)) {
522
        lex->tok_kind = MP_TOKEN_NAME;
Damien's avatar
Damien committed
523

524
525
        // get first char
        vstr_add_char(&lex->vstr, CUR_CHAR(lex));
Damien's avatar
Damien committed
526
527
        next_char(lex);

528
        // get tail chars
Damien's avatar
Damien committed
529
        while (!is_end(lex) && is_tail_of_identifier(lex)) {
530
            vstr_add_char(&lex->vstr, CUR_CHAR(lex));
Damien's avatar
Damien committed
531
532
533
534
            next_char(lex);
        }

    } else if (is_digit(lex) || (is_char(lex, '.') && is_following_digit(lex))) {
535
        lex->tok_kind = MP_TOKEN_NUMBER;
Damien's avatar
Damien committed
536

537
538
        // get first char
        vstr_add_char(&lex->vstr, CUR_CHAR(lex));
Damien's avatar
Damien committed
539
540
        next_char(lex);

541
        // get tail chars
Damien's avatar
Damien committed
542
543
        while (!is_end(lex)) {
            if (is_char_or(lex, 'e', 'E')) {
544
                vstr_add_char(&lex->vstr, 'e');
Damien's avatar
Damien committed
545
546
                next_char(lex);
                if (is_char(lex, '+') || is_char(lex, '-')) {
547
                    vstr_add_char(&lex->vstr, CUR_CHAR(lex));
Damien's avatar
Damien committed
548
549
550
                    next_char(lex);
                }
            } else if (is_letter(lex) || is_digit(lex) || is_char_or(lex, '_', '.')) {
551
                vstr_add_char(&lex->vstr, CUR_CHAR(lex));
Damien's avatar
Damien committed
552
553
554
555
556
557
                next_char(lex);
            } else {
                break;
            }
        }

558
559
560
561
562
563
564
565
566
567
568
569
    } else if (is_char(lex, '.')) {
        // special handling for . and ... operators, because .. is not a valid operator

        // get first char
        vstr_add_char(&lex->vstr, '.');
        next_char(lex);

        if (is_char_and(lex, '.', '.')) {
            vstr_add_char(&lex->vstr, '.');
            vstr_add_char(&lex->vstr, '.');
            next_char(lex);
            next_char(lex);
570
            lex->tok_kind = MP_TOKEN_ELLIPSIS;
571
        } else {
572
            lex->tok_kind = MP_TOKEN_DEL_PERIOD;
573
574
        }

Damien's avatar
Damien committed
575
576
577
578
    } else {
        // search for encoded delimiter or operator

        const char *t = tok_enc;
579
        mp_uint_t tok_enc_index = 0;
Damien's avatar
Damien committed
580
581
582
583
584
585
586
587
588
589
590
591
592
593
        for (; *t != 0 && !is_char(lex, *t); t += 1) {
            if (*t == 'e' || *t == 'c') {
                t += 1;
            } else if (*t == 'E') {
                tok_enc_index -= 1;
                t += 1;
            }
            tok_enc_index += 1;
        }

        next_char(lex);

        if (*t == 0) {
            // didn't match any delimiter or operator characters
594
            lex->tok_kind = MP_TOKEN_INVALID;
Damien's avatar
Damien committed
595
596
597
598
599
600

        } else {
            // matched a delimiter or operator character

            // get the maximum characters for a valid token
            t += 1;
601
            mp_uint_t t_index = tok_enc_index;
Damien's avatar
Damien committed
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
            for (;;) {
                for (; *t == 'e'; t += 1) {
                    t += 1;
                    t_index += 1;
                    if (is_char(lex, *t)) {
                        next_char(lex);
                        tok_enc_index = t_index;
                        break;
                    }
                }

                if (*t == 'E') {
                    t += 1;
                    if (is_char(lex, *t)) {
                        next_char(lex);
                        tok_enc_index = t_index;
                    } else {
619
                        lex->tok_kind = MP_TOKEN_INVALID;
620
                        goto tok_enc_no_match;
Damien's avatar
Damien committed
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
                    }
                    break;
                }

                if (*t == 'c') {
                    t += 1;
                    t_index += 1;
                    if (is_char(lex, *t)) {
                        next_char(lex);
                        tok_enc_index = t_index;
                        t += 1;
                    } else {
                        break;
                    }
                } else {
                    break;
                }
            }

            // set token kind
641
            lex->tok_kind = tok_enc_kind[tok_enc_index];
Damien's avatar
Damien committed
642

643
644
            tok_enc_no_match:

Damien's avatar
Damien committed
645
            // compute bracket level for implicit line joining
646
            if (lex->tok_kind == MP_TOKEN_DEL_PAREN_OPEN || lex->tok_kind == MP_TOKEN_DEL_BRACKET_OPEN || lex->tok_kind == MP_TOKEN_DEL_BRACE_OPEN) {
Damien's avatar
Damien committed
647
                lex->nested_bracket_level += 1;
648
            } else if (lex->tok_kind == MP_TOKEN_DEL_PAREN_CLOSE || lex->tok_kind == MP_TOKEN_DEL_BRACKET_CLOSE || lex->tok_kind == MP_TOKEN_DEL_BRACE_CLOSE) {
Damien's avatar
Damien committed
649
650
651
652
653
                lex->nested_bracket_level -= 1;
            }
        }
    }

654
    // check for keywords
655
    if (lex->tok_kind == MP_TOKEN_NAME) {
656
657
658
659
        // We check for __debug__ here and convert it to its value.  This is so
        // the parser gives a syntax error on, eg, x.__debug__.  Otherwise, we
        // need to check for this special token in many places in the compiler.
        // TODO improve speed of these string comparisons
660
661
        //for (mp_int_t i = 0; tok_kw[i] != NULL; i++) {
        for (mp_int_t i = 0; i < MP_ARRAY_SIZE(tok_kw); i++) {
662
            if (str_strn_equal(tok_kw[i], lex->vstr.buf, lex->vstr.len)) {
663
664
                if (i == MP_ARRAY_SIZE(tok_kw) - 1) {
                    // tok_kw[MP_ARRAY_SIZE(tok_kw) - 1] == "__debug__"
665
                    lex->tok_kind = (mp_optimise_value == 0 ? MP_TOKEN_KW_TRUE : MP_TOKEN_KW_FALSE);
666
                } else {
667
                    lex->tok_kind = MP_TOKEN_KW_FALSE + i;
668
                }
Damien's avatar
Damien committed
669
670
671
672
673
674
                break;
            }
        }
    }
}

675
mp_lexer_t *mp_lexer_new(qstr src_name, void *stream_data, mp_lexer_stream_next_byte_t stream_next_byte, mp_lexer_stream_close_t stream_close) {
676
    mp_lexer_t *lex = m_new_obj_maybe(mp_lexer_t);
677
678
679
680
681
682
683
684

    // check for memory allocation error
    if (lex == NULL) {
        if (stream_close) {
            stream_close(stream_data);
        }
        return NULL;
    }
Damien's avatar
Damien committed
685

686
    lex->source_name = src_name;
687
    lex->stream_data = stream_data;
688
    lex->stream_next_byte = stream_next_byte;
689
    lex->stream_close = stream_close;
Damien's avatar
Damien committed
690
691
692
693
    lex->line = 1;
    lex->column = 1;
    lex->emit_dent = 0;
    lex->nested_bracket_level = 0;
694
    lex->alloc_indent_level = MICROPY_ALLOC_LEXER_INDENT_INIT;
Damien's avatar
Damien committed
695
    lex->num_indent_level = 1;
696
    lex->indent_level = m_new_maybe(uint16_t, lex->alloc_indent_level);
697
    vstr_init(&lex->vstr, 32);
Damien's avatar
Damien committed
698

699
700
701
702
703
704
705
706
707
    // check for memory allocation error
    if (lex->indent_level == NULL || vstr_had_error(&lex->vstr)) {
        mp_lexer_free(lex);
        return NULL;
    }

    // store sentinel for first indentation level
    lex->indent_level[0] = 0;

Damien's avatar
Damien committed
708
    // preload characters
709
710
711
    lex->chr0 = stream_next_byte(stream_data);
    lex->chr1 = stream_next_byte(stream_data);
    lex->chr2 = stream_next_byte(stream_data);
712
713

    // if input stream is 0, 1 or 2 characters long and doesn't end in a newline, then insert a newline at the end
714
    if (lex->chr0 == MP_LEXER_EOF) {
715
        lex->chr0 = '\n';
716
    } else if (lex->chr1 == MP_LEXER_EOF) {
Damien's avatar
Damien committed
717
        if (lex->chr0 != '\n' && lex->chr0 != '\r') {
718
            lex->chr1 = '\n';
Damien's avatar
Damien committed
719
        }
720
    } else if (lex->chr2 == MP_LEXER_EOF) {
Damien's avatar
Damien committed
721
        if (lex->chr1 != '\n' && lex->chr1 != '\r') {
722
            lex->chr2 = '\n';
Damien's avatar
Damien committed
723
724
725
        }
    }

726
    // preload first token
727
    mp_lexer_next_token_into(lex, true);
Damien's avatar
Damien committed
728
729
730
731

    return lex;
}

732
void mp_lexer_free(mp_lexer_t *lex) {
733
    if (lex) {
734
735
        if (lex->stream_close) {
            lex->stream_close(lex->stream_data);
736
        }
737
        vstr_clear(&lex->vstr);
738
        m_del(uint16_t, lex->indent_level, lex->alloc_indent_level);
739
        m_del_obj(mp_lexer_t, lex);
Damien's avatar
Damien committed
740
741
742
    }
}

743
void mp_lexer_to_next(mp_lexer_t *lex) {
744
    mp_lexer_next_token_into(lex, false);
Damien's avatar
Damien committed
745
746
}

747
748
#if MICROPY_DEBUG_PRINTERS
void mp_lexer_show_token(const mp_lexer_t *lex) {
749
    printf("(" UINT_FMT ":" UINT_FMT ") kind:%u str:%p len:%zu", lex->tok_line, lex->tok_column, lex->tok_kind, lex->vstr.buf, lex->vstr.len);
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
    if (lex->vstr.len > 0) {
        const byte *i = (const byte *)lex->vstr.buf;
        const byte *j = (const byte *)i + lex->vstr.len;
        printf(" ");
        while (i < j) {
            unichar c = utf8_get_char(i);
            i = utf8_next_char(i);
            if (unichar_isprint(c)) {
                printf("%c", c);
            } else {
                printf("?");
            }
        }
    }
    printf("\n");
Damien's avatar
Damien committed
765
}
766
#endif