Class: HtmlTokenizer::Parser

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

Instance Method Summary collapse

Constructor Details

#initialize ⇒ Object 



464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
 
# File 'ext/html_tokenizer_ext/parser.c', line 464

static VALUE parser_initialize_method(VALUE self)
{
  struct parser_t *parser = NULL;

  Parser_Get_Struct(self, parser);
  DBG_PRINT("parser=%p initialize", parser);

  memset(parser, 0, sizeof(struct parser_t));

  parser->context = PARSER_NONE;

  tokenizer_init(&parser->tk);
  parser->tk.callback_data = parser;
  parser->tk.f_callback = parser_tokenize_callback;

  parser->doc.length = 0;
  parser->doc.data = NULL;
  parser->doc.enc_index = 0;
  parser->doc.mb_length = 0;

  parser->doc.line_number = 1;
  parser->doc.column_number = 0;

  parser->errors_count = 0;
  parser->errors = NULL;

  return Qnil;
}

Instance Method Details

#append_placeholder(source) ⇒ Object 



562
563
564
565
 
# File 'ext/html_tokenizer_ext/parser.c', line 562

static VALUE parser_append_placeholder_method(VALUE self, VALUE source)
{
  return parser_append_data(self, source, 1);
}

#attribute_name ⇒ Object 



663
664
665
666
667
668
 
# File 'ext/html_tokenizer_ext/parser.c', line 663

static VALUE parser_attribute_name_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return ref_to_str(parser, &parser->attribute.name);
}

#attribute_quoted? ⇒ Boolean

Returns:

  • (Boolean) 


686
687
688
689
690
691
 
# File 'ext/html_tokenizer_ext/parser.c', line 686

static VALUE parser_attribute_is_quoted_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return parser->attribute.is_quoted ? Qtrue : Qfalse;
}

#attribute_value ⇒ Object 



670
671
672
673
674
675
 
# File 'ext/html_tokenizer_ext/parser.c', line 670

static VALUE parser_attribute_value_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return ref_to_str(parser, &parser->attribute.value);
}

#cdata_text ⇒ Object 



700
701
702
703
704
705
 
# File 'ext/html_tokenizer_ext/parser.c', line 700

static VALUE parser_cdata_text_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return ref_to_str(parser, &parser->cdata.text);
}

#closing_tag? ⇒ Boolean

Returns:

  • (Boolean) 


649
650
651
652
653
654
 
# File 'ext/html_tokenizer_ext/parser.c', line 649

static VALUE parser_closing_tag_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return parser->tk.is_closing_tag ? Qtrue : Qfalse;
}

#column_number ⇒ Object 



757
758
759
760
761
762
 
# File 'ext/html_tokenizer_ext/parser.c', line 757

static VALUE parser_column_number_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return ULONG2NUM(parser->doc.column_number);
}

#comment_text ⇒ Object 



693
694
695
696
697
698
 
# File 'ext/html_tokenizer_ext/parser.c', line 693

static VALUE parser_comment_text_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return ref_to_str(parser, &parser->comment.text);
}

#context ⇒ Object 



596
597
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
627
628
629
630
631
632
 
# File 'ext/html_tokenizer_ext/parser.c', line 596

static VALUE parser_context_method(VALUE self)
{
  struct parser_t *parser = NULL;

  Parser_Get_Struct(self, parser);

  switch(parser->context) {
  case PARSER_NONE:
    return rawtext_context(parser) ? ID2SYM(rb_intern("rawtext")) : ID2SYM(rb_intern("none"));
  case PARSER_SOLIDUS_OR_TAG_NAME:
    return ID2SYM(rb_intern("solidus_or_tag_name"));
  case PARSER_TAG_NAME:
    return ID2SYM(rb_intern("tag_name"));
  case PARSER_TAG:
    return ID2SYM(rb_intern("tag"));
  case PARSER_ATTRIBUTE_NAME:
    return ID2SYM(rb_intern("attribute_name"));
  case PARSER_ATTRIBUTE_WHITESPACE_OR_EQUAL:
    return ID2SYM(rb_intern("after_attribute_name"));
  case PARSER_ATTRIBUTE_WHITESPACE_OR_VALUE:
    return ID2SYM(rb_intern("after_equal"));
  case PARSER_ATTRIBUTE_QUOTED_VALUE:
    return ID2SYM(rb_intern("quoted_value"));
  case PARSER_SPACE_AFTER_ATTRIBUTE:
    return ID2SYM(rb_intern("space_after_attribute"));
  case PARSER_ATTRIBUTE_UNQUOTED_VALUE:
    return ID2SYM(rb_intern("unquoted_value"));
  case PARSER_TAG_END:
    return ID2SYM(rb_intern("tag_end"));
  case PARSER_COMMENT:
    return ID2SYM(rb_intern("comment"));
  case PARSER_CDATA:
    return ID2SYM(rb_intern("cdata"));
  }

  return Qnil;
}

#document ⇒ Object 



567
568
569
570
571
572
573
574
575
576
 
# File 'ext/html_tokenizer_ext/parser.c', line 567

static VALUE parser_document_method(VALUE self)
{
  struct parser_t *parser = NULL;
  rb_encoding *enc;
  Parser_Get_Struct(self, parser);
  if(!parser->doc.data)
    return Qnil;
  enc = rb_enc_from_index(parser->doc.enc_index);
  return rb_enc_str_new(parser->doc.data, parser->doc.length, enc);
}

#document_length ⇒ Object 



578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
 
# File 'ext/html_tokenizer_ext/parser.c', line 578

static VALUE parser_document_length_method(VALUE self)
{
  struct parser_t *parser = NULL;
  rb_encoding *enc;
  const char *buf;

  Parser_Get_Struct(self, parser);

  if(parser->doc.data == NULL) {
    return ULONG2NUM(0);
  }
  else {
    buf = parser->doc.data;
    enc = rb_enc_from_index(parser->doc.enc_index);
    return ULONG2NUM(rb_enc_strlen(buf, buf + parser->doc.length, enc));
  }
}

#errors(error_p) ⇒ Object 



733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
 
# File 'ext/html_tokenizer_ext/parser.c', line 733

static VALUE parser_errors_method(VALUE self, VALUE error_p)
{
  struct parser_t *parser = NULL;
  VALUE list;
  size_t i;
  Parser_Get_Struct(self, parser);

  list = rb_ary_new();
  for(i=0; i<parser->errors_count; i++) {
    if(parser->errors[i].message) {
      rb_ary_push(list, create_parser_error(&parser->errors[i]));
    }
  }

  return list;
}

#errors_count ⇒ Object 



714
715
716
717
718
719
 
# File 'ext/html_tokenizer_ext/parser.c', line 714

static VALUE parser_errors_count_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return INT2NUM(parser->errors_count);
}

#line_number ⇒ Object 



750
751
752
753
754
755
 
# File 'ext/html_tokenizer_ext/parser.c', line 750

static VALUE parser_line_number_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return ULONG2NUM(parser->doc.line_number);
}

#parse(source) ⇒ Object 



557
558
559
560
 
# File 'ext/html_tokenizer_ext/parser.c', line 557

static VALUE parser_parse_method(VALUE self, VALUE source)
{
  return parser_append_data(self, source, 0);
}

#quote_character ⇒ Object 



677
678
679
680
681
682
683
684
 
# File 'ext/html_tokenizer_ext/parser.c', line 677

static VALUE parser_quote_character_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return parser->attribute.is_quoted ?
    rb_str_new(&parser->tk.attribute_value_start, 1) :
    Qnil;
}

#rawtext_text ⇒ Object 



707
708
709
710
711
712
 
# File 'ext/html_tokenizer_ext/parser.c', line 707

static VALUE parser_rawtext_text_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return ref_to_str(parser, &parser->rawtext.text);
}

#self_closing_tag? ⇒ Boolean

Returns:

  • (Boolean) 


656
657
658
659
660
661
 
# File 'ext/html_tokenizer_ext/parser.c', line 656

static VALUE parser_self_closing_tag_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return parser->tag.self_closing ? Qtrue : Qfalse;
}

#tag_name ⇒ Object 



642
643
644
645
646
647
 
# File 'ext/html_tokenizer_ext/parser.c', line 642

static VALUE parser_tag_name_method(VALUE self)
{
  struct parser_t *parser = NULL;
  Parser_Get_Struct(self, parser);
  return ref_to_str(parser, &parser->tag.name);
}