Class: MiniHTML::Scanner

Inherits:
Object
  • Object
show all
Defined in:
ext/minihtml_scanner/minihtml_scanner.c

Instance Method Summary collapse

Constructor Details

#initialize(str) ⇒ Object 



141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
 
# File 'ext/minihtml_scanner/minihtml_scanner.c', line 141

static VALUE scanner_initialize(VALUE self, VALUE str) {
    Check_Type(str, T_STRING);
    scanner_t *t;
    TypedData_Get_Struct(self, scanner_t, &scanner_type, t);

    str = rb_str_dup(str);
    t->str = str;
    t->p = (const uint8_t *) RSTRING_PTR(str);
    t->end = t->p + RSTRING_LEN(str);
    t->idx_cp = 0;
    t->tokens = rb_ary_new();
    t->errors = rb_ary_new();
    t->line = 1;
    t->col = 1;

    // prime lookahead
    t->look[0] = next_utf8_cp(&t->p, t->end);
    t->look[1] = next_utf8_cp(&t->p, t->end);
    t->look[2] = next_utf8_cp(&t->p, t->end);
    t->look[3] = next_utf8_cp(&t->p, t->end);

    return self;
}

Instance Method Details

#eof? ⇒ Boolean 



488
489
490
491
492
 
# File 'ext/minihtml_scanner/minihtml_scanner.c', line 488

static VALUE scanner_at_eof(const VALUE self) {
    scanner_t *t;
    TypedData_Get_Struct(self, scanner_t, &scanner_type, t);
    return t->look[0] == EOF ? Qtrue : Qfalse;
}

#errors ⇒ Object 



472
473
474
475
476
 
# File 'ext/minihtml_scanner/minihtml_scanner.c', line 472

static VALUE scanner_errors(const VALUE self) {
    scanner_t *t;
    TypedData_Get_Struct(self, scanner_t, &scanner_type, t);
    return t->errors;
}

#stats ⇒ Object 



478
479
480
481
482
483
484
485
486
 
# File 'ext/minihtml_scanner/minihtml_scanner.c', line 478

static VALUE scanner_stats(const VALUE self) {
    scanner_t *t;
    TypedData_Get_Struct(self, scanner_t, &scanner_type, t);
    const VALUE h = rb_hash_new();
    rb_hash_aset(h, sym_line, LONG2NUM(t->line));
    rb_hash_aset(h, sym_column, LONG2NUM(t->col));
    rb_hash_aset(h, sym_offset, LONG2NUM(t->idx_cp));
    return h;
}

#tokenize ⇒ Object 



508
509
510
511
512
513
514
515
516
 
# File 'ext/minihtml_scanner/minihtml_scanner.c', line 508

static VALUE scanner_tokenize(const VALUE self) {
    scanner_t *t;
    TypedData_Get_Struct(self, scanner_t, &scanner_type, t);
    while (t->look[0] != EOF) {
        scanner_scan_token(t);
    }
    scanner_hydrate_tokens(t);
    return t->tokens;
}

#tokens ⇒ Object 



466
467
468
469
470
 
# File 'ext/minihtml_scanner/minihtml_scanner.c', line 466

static VALUE scanner_tokens(const VALUE self) {
    scanner_t *t;
    TypedData_Get_Struct(self, scanner_t, &scanner_type, t);
    return t->tokens;
}