Module: Cataract

Defined in:
lib/cataract.rb,
lib/cataract/pure.rb,
lib/cataract/pure.rb,
lib/cataract/rule.rb,
lib/cataract/error.rb,
lib/cataract/at_rule.rb,
lib/cataract/version.rb,
lib/cataract/constants.rb,
lib/cataract/stylesheet.rb,
lib/cataract/declaration.rb,
lib/cataract/media_query.rb,
lib/cataract/pure/parser.rb,
lib/cataract/declarations.rb,
lib/cataract/pure/flatten.rb,
lib/cataract/pure/helpers.rb,
lib/cataract/import_resolver.rb,
lib/cataract/pure/serializer.rb,
lib/cataract/import_statement.rb,
lib/cataract/pure/specificity.rb,
lib/cataract/stylesheet_scope.rb,
lib/cataract/pure/byte_constants.rb,
ext/cataract/cataract.c

Overview

Pure Ruby CSS parser - Byte constants for fast parsing Using getbyte() instead of String#[] to avoid allocating millions of string objects

Defined Under Namespace

Modules: Flatten, ImportResolver Classes: AtRule, ColorConversionError, Declaration, Declarations, DepthError, Error, ImportError, ImportStatement, MediaQuery, ParseError, Parser, Rule, SizeError, Stylesheet, StylesheetScope

Constant Summary collapse

PURE_RUBY_LOADED =

Flag to indicate pure Ruby version is loaded

true
IMPLEMENTATION =

Implementation type constant

ID2SYM(rb_intern("native"))
COMPILE_FLAGS =

Compile flags (mimic C version)

compile_flags
VERSION = 
'0.3.0'
DEFAULT_URI_RESOLVER =

Default URI resolver proc for converting relative URLs to absolute Uses Ruby's URI stdlib to merge base and relative URIs

lambda { |base, relative|
  require 'uri'
  URI.parse(base).merge(relative).to_s
}.freeze
PSEUDO_ELEMENT_KEYWORDS =

Legacy CSS2 pseudo-elements written with a single colon (e.g. :before) that must still be counted as pseudo-elements, not pseudo-classes.

%w[before after first-line first-letter selection].freeze
BYTE_SPACE =

Whitespace bytes

32
BYTE_TAB =

' '

9
BYTE_NEWLINE =

'\t'

10
BYTE_CR =

'\n'

13
BYTE_AT =

CSS structural characters

64
BYTE_LBRACE =

'@'

123
BYTE_RBRACE =

'{'

125
BYTE_LPAREN =

'}'

40
BYTE_RPAREN =

'('

41
BYTE_LBRACKET =

')'

91
BYTE_RBRACKET =

'['

93
BYTE_SEMICOLON =

']'

59
BYTE_COLON =

';'

58
BYTE_COMMA =

':'

44
BYTE_SLASH =

Comment characters

47
BYTE_STAR =

'/'

42
BYTE_SQUOTE =

Quote characters

39
BYTE_DQUOTE =

"'"

34
BYTE_HASH =

Selector characters

35
BYTE_DOT =

'#'

46
BYTE_GT =

'.'

62
BYTE_PLUS =

'>'

43
BYTE_TILDE =

'+'

126
BYTE_ASTERISK =

'~'

42
BYTE_AMPERSAND =

'*'

38
BYTE_HYPHEN =

Other

45
BYTE_UNDERSCORE =

'-'

95
BYTE_BACKSLASH =

'_'

92
BYTE_BANG =

'\'

33
BYTE_PERCENT =

'!'

37
BYTE_EQUALS =

'%'

61
BYTE_CARET =

'='

94
BYTE_DOLLAR =

'^'

36
BYTE_PIPE =

'$'

124
BYTE_LOWER_U =

Specific lowercase letters (for keyword matching)

117
BYTE_LOWER_R =

'u'

114
BYTE_LOWER_L =

'r'

108
BYTE_LOWER_D =

'l'

100
BYTE_LOWER_T =

'd'

116
BYTE_LOWER_N =

't'

110
BYTE_UPPER_U =

Specific uppercase letters (for case-insensitive matching)

85
BYTE_UPPER_R =

'U'

82
BYTE_UPPER_L =

'R'

76
BYTE_UPPER_D =

'L'

68
BYTE_UPPER_T =

'D'

84
BYTE_UPPER_N =

'T'

78
BYTE_LOWER_A =

Letter ranges (a-z, A-Z)

97
BYTE_LOWER_Z =

'a'

122
BYTE_UPPER_A =

'z'

65
BYTE_UPPER_Z =

'A'

90
BYTE_CASE_DIFF =

'Z'

32
BYTE_DIGIT_0 =

Digit range (0-9)

48
BYTE_DIGIT_9 =

'0'

57
NESTING_STYLE_IMPLICIT =

Nesting styles (for CSS nesting support)

0
NESTING_STYLE_EXPLICIT =

Implicit nesting: .parent { .child { ... } } => .parent .child

1
NATIVE_EXTENSION_LOADED =

Flag to indicate native extension is loaded (for pure Ruby fallback detection)

Qtrue

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

._parse_css(*args) ⇒ Hash

Parse CSS and return hash with parsed data This matches the old parse_css API

Parameters:

  • css_string (String) —

    CSS to parse

  • parser_options (Hash) —

    Parser options (optional, defaults to {})

Returns:

  • (Hash) —

    { rules: [...], media_index: ..., charset: "..." }



91
92
93
94
 
# File 'lib/cataract/pure.rb', line 91

def self._parse_css(css_string, parser_options = {})
  parser = Parser.new(css_string, parser_options: parser_options)
  parser.parse
end

._rule_parent_id(rule) ⇒ Integer?

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.

Helper: read a rule's parent_rule_id, guarding against AtRule (which has no such member - only Rule participates in CSS nesting, AtRule never is nested via '&' nesting). Calling .parent_rule_id directly on an AtRule raises NoMethodError.

Parameters:

Returns:

  • (Integer, nil) —

    parent_rule_id, or nil if the rule isn't a Rule



66
67
68
 
# File 'lib/cataract/pure/serializer.rb', line 66

def self._rule_parent_id(rule)
  rule.is_a?(Rule) ? rule.parent_rule_id : nil
end

._serialize_children(result, parent_selector, children, rule_children, media_queries, parent_has_declarations) ⇒ Object

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.

Helper: recursively serialize a rule's nested children. CSS nesting can go as deep as the parser allows (MAX_PARSE_DEPTH), so this recurses rather than hand-unrolling a fixed number of levels - a nested rule's own nested rules are found the same way its parent's were, via rule_children.



206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
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
 
# File 'lib/cataract/pure/serializer.rb', line 206

def self._serialize_children(result, parent_selector, children, rule_children, media_queries, parent_has_declarations)
  children.each_with_index do |child, index|
    # Add space before nested content
    # - Always add space if the parent had declarations
    # - Add space between nested rules (not before first if no declarations)
    if parent_has_declarations || index > 0
      result << ' '
    end

    # Check if this child has @media nesting (parent_rule_id present but nesting_style is nil)
    if child.nesting_style.nil? && child.media_query_id && media_queries[child.media_query_id]
      # This is a nested @media rule
      mq = media_queries[child.media_query_id]
      media_query_string = _build_media_query_string(mq, nil, nil, media_queries)
      result << "@media #{media_query_string} { "
      _serialize_declarations(result, child.declarations)

      # Serialize any children of this @media rule
      media_children = rule_children[child.id] || []
      media_children.each_with_index do |media_child, media_idx|
        result << ' ' if media_idx > 0 || !child.declarations.empty?

        nested_media_selector = _reconstruct_nested_selector(
          child.selector, media_child.selector,
          media_child.nesting_style
        )

        result << "#{nested_media_selector} { "
        _serialize_declarations(result, media_child.declarations)
        result << ' }'
      end

      result << ' }'
    else
      # Regular nested selector - determine if we need to reconstruct it with &
      nested_selector = _reconstruct_nested_selector(parent_selector, child.selector, child.nesting_style)

      result << "#{nested_selector} { "
      _serialize_declarations(result, child.declarations)

      # Recurse into this child's own nested children, however deep they go
      _serialize_children(result, child.selector, rule_children[child.id] || [], rule_children, media_queries, !child.declarations.empty?)

      result << ' }'
    end
  end
end

._serialize_children_formatted(result, parent_selector, children, rule_children, media_queries, indent) ⇒ Object

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.

Helper: recursively serialize a rule's nested children with indentation. CSS nesting can go as deep as the parser allows (MAX_PARSE_DEPTH), so this recurses rather than hand-unrolling a fixed number of levels - a nested rule's own nested rules are found the same way its parent's were, via rule_children, with the indent growing by one level each call.



683
684
685
686
687
688
689
690
691
692
693
694
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
 
# File 'lib/cataract/pure/serializer.rb', line 683

def self._serialize_children_formatted(result, parent_selector, children, rule_children, media_queries, indent)
  children.each do |child|
    if child.nesting_style.nil? && child.media_query_id && media_queries[child.media_query_id]
      # Nested @media
      mq = media_queries[child.media_query_id]
      media_query_string = _build_media_query_string(mq, nil, nil, media_queries)
      result << indent
      result << "@media #{media_query_string} {\n"

      unless child.declarations.empty?
        _serialize_declarations_formatted(result, child.declarations, "#{indent}  ")
      end

      # Handle nested media children
      media_children = rule_children[child.id] || []
      media_children.each do |media_child|
        nested_media_selector = _reconstruct_nested_selector(
          child.selector,
          media_child.selector,
          media_child.nesting_style
        )

        result << indent
        result << "  #{nested_media_selector} {\n"
        unless media_child.declarations.empty?
          _serialize_declarations_formatted(result, media_child.declarations, "#{indent}    ")
        end
        result << indent
        result << "  }\n"
      end

      result << indent
      result << "}\n"
    else
      # Regular nested selector
      nested_selector = _reconstruct_nested_selector(parent_selector, child.selector, child.nesting_style)

      result << indent
      result << "#{nested_selector} {\n"

      unless child.declarations.empty?
        _serialize_declarations_formatted(result, child.declarations, "#{indent}  ")
      end

      # Recurse into this child's own nested children, however deep they go
      _serialize_children_formatted(result, child.selector, rule_children[child.id] || [], rule_children, media_queries, "#{indent}  ")

      result << indent
      result << "}\n"
    end
  end
end

.calculate_specificity(selector) ⇒ Object

Deprecated alias for backwards compatibility



20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
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
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
 
# File 'lib/cataract/pure/specificity.rb', line 20

def self.calculate_specificity(selector)
  return 0 if selector.nil? || selector.empty?

  # Counters for specificity components
  id_count = 0
  class_count = 0
  attr_count = 0
  pseudo_class_count = 0
  pseudo_element_count = 0
  element_count = 0

  i = 0
  len = selector.length

  while i < len
    byte = selector.getbyte(i)

    if byte == BYTE_HASH
      id_count += 1
      i = skip_identifier(selector, i + 1, len)
    elsif byte == BYTE_DOT
      class_count += 1
      i = skip_identifier(selector, i + 1, len)
    elsif byte == BYTE_LBRACKET
      attr_count += 1
      i = skip_attribute_selector(selector, i, len)
    elsif byte == BYTE_COLON
      i, is_not, not_content, counts_as_element = parse_pseudo(selector, i, len)
      if not_content
        # :not() doesn't count itself, but its content does
        not_specificity = calculate_specificity(not_content)
        id_count += not_specificity / 100
        class_count += (not_specificity % 100) / 10
        element_count += not_specificity % 10
      elsif !is_not
        if counts_as_element
          pseudo_element_count += 1
        else
          pseudo_class_count += 1
        end
      end
    elsif letter?(byte)
      element_count += 1
      i = skip_identifier(selector, i + 1, len)
    else
      # Whitespace, combinators, and the universal selector (*) all have
      # zero specificity - just skip a single byte.
      i += 1
    end
  end

  # Calculate specificity using W3C formula
  (id_count * 100) +
    ((class_count + attr_count + pseudo_class_count) * 10) +
    ((element_count + pseudo_element_count) * 1)
end

.digit?(byte) ⇒ Boolean

Check if byte is a digit (0-9)

Parameters:

  • byte (Integer) —

    Byte value from String#getbyte

Returns:

  • (Boolean) —

    true if digit



25
26
27
 
# File 'lib/cataract/pure/helpers.rb', line 25

def self.digit?(byte)
  byte >= BYTE_DIGIT_0 && byte <= BYTE_DIGIT_9
end

.expand_shorthand(decl) ⇒ Array<Declaration>

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.

Expand a single shorthand declaration into longhand declarations. Underscore prefix indicates semi-private API - use with caution.

Parameters:

Returns:

  • (Array<Declaration>) —

    Array of expanded longhand declarations



136
137
138
 
# File 'lib/cataract/pure.rb', line 136

def self.expand_shorthand(decl)
  Flatten.expand_shorthand(decl)
end

.flatten(input) ⇒ Object

Output: Stylesheet with flattened declarations (cascade applied)



106
107
108
 
# File 'lib/cataract/pure.rb', line 106

def self.flatten(stylesheet)
  Flatten.flatten(stylesheet, mutate: false)
end

.flatten!(stylesheet) ⇒ Stylesheet

Flatten stylesheet rules in-place (mutates receiver)

Parameters:

  • stylesheet (Stylesheet) —

    Stylesheet to flatten

Returns:



114
115
116
 
# File 'lib/cataract/pure.rb', line 114

def self.flatten!(stylesheet)
  Flatten.flatten(stylesheet, mutate: true)
end

.ident_char?(byte) ⇒ Boolean

Check if byte is alphanumeric, hyphen, or underscore (CSS identifier char)

Parameters:

  • byte (Integer) —

    Byte value from String#getbyte

Returns:

  • (Boolean) —

    true if valid identifier character



32
33
34
 
# File 'lib/cataract/pure/helpers.rb', line 32

def self.ident_char?(byte)
  letter?(byte) || digit?(byte) || byte == BYTE_HYPHEN || byte == BYTE_UNDERSCORE
end

.is_whitespace?(byte) ⇒ Boolean

Check if a byte is whitespace (space, tab, newline, CR)

Parameters:

  • byte (Integer) —

    Byte value from String#getbyte

Returns:

  • (Boolean) —

    true if whitespace



10
11
12
 
# File 'lib/cataract/pure/helpers.rb', line 10

def self.is_whitespace?(byte)
  byte == BYTE_SPACE || byte == BYTE_TAB || byte == BYTE_NEWLINE || byte == BYTE_CR
end

.letter?(byte) ⇒ Boolean

Check if byte is a letter (a-z, A-Z)

Parameters:

  • byte (Integer) —

    Byte value from String#getbyte

Returns:

  • (Boolean) —

    true if letter



17
18
19
20
 
# File 'lib/cataract/pure/helpers.rb', line 17

def self.letter?(byte)
  (byte >= BYTE_LOWER_A && byte <= BYTE_LOWER_Z) ||
  (byte >= BYTE_UPPER_A && byte <= BYTE_UPPER_Z)
end

.merge(input) ⇒ Object

Output: Stylesheet with flattened declarations (cascade applied)



119
120
121
122
 
# File 'lib/cataract/pure.rb', line 119

def self.merge(stylesheet)
  warn 'Cataract.merge is deprecated, use Cataract.flatten instead', uplevel: 1
  flatten(stylesheet)
end

.merge!(stylesheet) ⇒ Object

Deprecated: Use flatten! instead



125
126
127
128
 
# File 'lib/cataract/pure.rb', line 125

def self.merge!(stylesheet)
  warn 'Cataract.merge! is deprecated, use Cataract.flatten! instead', uplevel: 1
  flatten!(stylesheet)
end

.parse_css(css, **options) ⇒ Object

NOTE: Copied from cataract.rb Need to untangle this eventually



98
99
100
 
# File 'lib/cataract/pure.rb', line 98

def self.parse_css(css, **options)
  Stylesheet.parse(css, **options)
end

.parse_declarations(declarations_string) ⇒ Array<Declaration>

Ruby-facing wrapper for new_parse_declarations

Parameters:

  • declarations_string (String) —

    CSS declarations like "color: red; margin: 10px"

Returns:

  • (Array<Declaration>) —

    Array of parsed declaration structs



1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
 
# File 'ext/cataract/cataract.c', line 1158

static VALUE new_parse_declarations(VALUE self, VALUE declarations_string) {
    Check_Type(declarations_string, T_STRING);

    const char *input = RSTRING_PTR(declarations_string);
    long input_len = RSTRING_LEN(declarations_string);

    // Strip outer braces and whitespace (css_parser compatibility)
    const char *start = input;
    const char *end = input + input_len;

    while (start < end && (IS_WHITESPACE(*start) || *start == '{')) start++;
    while (end > start && (IS_WHITESPACE(*(end-1)) || *(end-1) == '}')) end--;

    VALUE result = new_parse_declarations_string(start, end);

    RB_GC_GUARD(result);
    return result;
}

.parse_media_types ⇒ Object

.stylesheet_to_formatted_s(rules_array, charset, has_nesting, selector_lists, media_queries, media_query_lists) ⇒ Object

Formatted version with indentation and newlines (with nesting support)



549
550
551
552
553
554
555
556
557
558
559
560
561
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
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 'lib/cataract/pure/serializer.rb', line 549

def self.stylesheet_to_formatted_s(rules, charset, has_nesting, selector_lists = {}, media_queries = [], media_query_lists = {})
  result = +''

  # Add @charset if present
  unless charset.nil?
    result << "@charset \"#{charset}\";\n"
  end

  # Fast path: no nesting - use simple algorithm
  unless has_nesting
    return _stylesheet_to_formatted_s_without_nesting(rules, result, selector_lists, media_queries, media_query_lists)
  end

  # Build parent-child relationships
  rule_children = {}
  rules.each do |rule|
    parent_rule_id = _rule_parent_id(rule)
    next unless parent_rule_id

    parent_id = parent_rule_id.is_a?(Integer) ? parent_rule_id : parent_rule_id.to_i
    rule_children[parent_id] ||= []
    rule_children[parent_id] << rule
  end

  # Build reverse map: media_query_id => list_id
  mq_id_to_list_id = {}
  media_query_lists.each do |list_id, mq_ids|
    mq_ids.each { |mq_id| mq_id_to_list_id[mq_id] = list_id }
  end

  # Serialize top-level rules only
  current_media_query_list_id = nil
  current_media_query = nil
  in_media_block = false

  rules.each do |rule|
    next if _rule_parent_id(rule)

    rule_media_query_id = rule.media_query_id
    rule_media_query = rule_media_query_id ? media_queries[rule_media_query_id] : nil
    rule_media_query_list_id = rule_media_query_id ? mq_id_to_list_id[rule_media_query_id] : nil

    if rule_media_query.nil?
      if in_media_block
        result << "}\n"
        in_media_block = false
        current_media_query = nil
        current_media_query_list_id = nil
      end

      # Add blank line before base rule if we just closed a media block (ends with "}\n")
      result << "\n" if result.length > 1 && result.getbyte(-1) == BYTE_NEWLINE && result.getbyte(-2) == BYTE_RBRACE

      _serialize_rule_with_nesting_formatted(result, rule, rule_children, '', media_queries)
    else
      # For lists: compare list_id
      # For single queries: compare by content (type + conditions)
      needs_new_block = _needs_new_media_block?(current_media_query, current_media_query_list_id, rule_media_query,
                                                rule_media_query_list_id)

      if needs_new_block
        if in_media_block
          result << "}\n"
        elsif result.length > 0
          result << "\n"
        end
        current_media_query = rule_media_query
        current_media_query_list_id = rule_media_query_list_id
        # Serialize the media query (or list)
        media_query_string = _build_media_query_string(rule_media_query, rule_media_query_list_id, media_query_lists, media_queries)
        result << "@media #{media_query_string} {\n"
        in_media_block = true
      end

      _serialize_rule_with_nesting_formatted(result, rule, rule_children, '  ', media_queries)
    end
  end

  if in_media_block
    result << "}\n"
  end

  result
end

.stylesheet_to_s(rules_array, charset, has_nesting, selector_lists, media_queries, media_query_lists) ⇒ Object

New stylesheet serialization entry point - checks for nesting and delegates



79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
 
# File 'lib/cataract/pure/serializer.rb', line 79

def self.stylesheet_to_s(rules, charset, has_nesting, selector_lists = {}, media_queries = [], media_query_lists = {})
  result = +''

  # Add @charset if present
  unless charset.nil?
    result << "@charset \"#{charset}\";\n"
  end

  # Fast path: no nesting - use simple algorithm
  unless has_nesting
    return _stylesheet_to_s_without_nesting(rules, result, selector_lists, media_queries, media_query_lists)
  end

  # Build parent-child relationships
  rule_children = {}
  rules.each do |rule|
    parent_rule_id = _rule_parent_id(rule)
    next unless parent_rule_id

    parent_id = parent_rule_id.is_a?(Integer) ? parent_rule_id : parent_rule_id.to_i
    rule_children[parent_id] ||= []
    rule_children[parent_id] << rule
  end

  # Build reverse map: media_query_id => list_id
  mq_id_to_list_id = {}
  media_query_lists.each do |list_id, mq_ids|
    mq_ids.each { |mq_id| mq_id_to_list_id[mq_id] = list_id }
  end

  # Serialize top-level rules only (those without parent_rule_id)
  current_media_query_list_id = nil
  current_media_query = nil
  in_media_block = false

  rules.each do |rule|
    # Skip rules that have a parent (they'll be serialized as nested)
    next if _rule_parent_id(rule)

    rule_media_query_id = rule.media_query_id
    rule_media_query = rule_media_query_id ? media_queries[rule_media_query_id] : nil
    rule_media_query_list_id = rule_media_query_id ? mq_id_to_list_id[rule_media_query_id] : nil

    if rule_media_query.nil?
      # Close any open media block
      if in_media_block
        result << "}\n"
        in_media_block = false
        current_media_query = nil
        current_media_query_list_id = nil
      end
    else
      # Check if we need to open a new media block
      # For lists: compare list_id
      # For single queries: compare by content (type + conditions)
      needs_new_block = _needs_new_media_block?(current_media_query, current_media_query_list_id, rule_media_query,
                                                rule_media_query_list_id)

      if needs_new_block
        if in_media_block
          result << "}\n"
        end
        current_media_query = rule_media_query
        current_media_query_list_id = rule_media_query_list_id

        # Serialize the media query (or list)
        media_query_string = _build_media_query_string(rule_media_query, rule_media_query_list_id, media_query_lists, media_queries)
        result << "@media #{media_query_string} {\n"
        in_media_block = true
      end
    end

    _serialize_rule_with_nesting(result, rule, rule_children, media_queries)
  end

  if in_media_block
    result << "}\n"
  end

  result
end

Instance Method Details

#parse_css(css, **options) ⇒ Object 



79
80
81
 
# File 'lib/cataract.rb', line 79

def parse_css(css, **options)
  Stylesheet.parse(css, **options)
end