Class: HamlLint::RubyExtraction::ChunkExtractor

Inherits:
Object
  • Object
show all
Includes:
HamlVisitor
Defined in:
lib/haml_lint/ruby_extraction/chunk_extractor.rb

Overview

Extracts “chunks” of the haml file into instances of subclasses of HamlLint::RubyExtraction::BaseChunk.

This is the first step of generating Ruby code from a HAML file to then be processed by RuboCop. See HamlLint::RubyExtraction::BaseChunk for more details.

Constant Summary collapse

HAML_PARSER_INSTANCE = 
if Haml::VERSION >= '5.0.0'
  ::Haml::Parser.new({})
else
  ::Haml::Parser.new('', {})
end
BLOCK_KEYWORD_REGEX =

HAML strips newlines when handling multi-line statements (using pipes or trailing comma) We don’t. So the regex must be fixed to correctly detect the start of the string.

Regexp.new(Haml::Parser::BLOCK_KEYWORD_REGEX.source.sub('^', '\A'))
START_BLOCK_KEYWORDS = 
%w[if unless case begin for until while].freeze
LOOP_KEYWORDS = 
%w[for until while].freeze

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Methods included from HamlVisitor

#visit, #visit_children

Constructor Details

#initialize(document, script_output_prefix:, autocorrect: nil) ⇒ ChunkExtractor

Returns a new instance of ChunkExtractor.



24
25
26
27
28
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 24

def initialize(document, script_output_prefix:, autocorrect: nil)
  @document = document
  @script_output_prefix = script_output_prefix
  @autocorrect = autocorrect
end

Instance Attribute Details

#autocorrect ⇒ Object (readonly)

Returns the value of attribute autocorrect.



12
13
14
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 12

def autocorrect
  @autocorrect
end

#script_output_prefix ⇒ Object (readonly)

Returns the value of attribute script_output_prefix.



12
13
14
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 12

def script_output_prefix
  @script_output_prefix
end

Class Method Details

.anonymous_block?(code) ⇒ Boolean

Returns:

  • (Boolean) 


715
716
717
718
719
720
721
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 715

def self.anonymous_block?(code)
  # Don't start with a comment and end with a `do`
  # Definitely not perfect for the comment handling, but otherwise a more advanced parsing system is needed.
  # Move the comment to its own line if it's annoying.
  code !~ /\A\s*#/ &&
    code =~ /\bdo\s*(\|[^|]*\|\s*)?(#.*)?\z/
end

.block_keyword(code) ⇒ Object 



729
730
731
732
733
734
735
736
737
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 729

def self.block_keyword(code)
  # Need to handle 'for'/'while' since regex stolen from HAML parser doesn't
  if (keyword = code[/\A\s*([^\s]+)\s+/, 1]) && LOOP_KEYWORDS.include?(keyword)
    return keyword
  end

  return unless keyword = code.scan(BLOCK_KEYWORD_REGEX)[0]
  keyword[0] || keyword[1]
end

.start_block_keyword?(code) ⇒ Boolean

Returns:

  • (Boolean) 


724
725
726
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 724

def self.start_block_keyword?(code)
  START_BLOCK_KEYWORDS.include?(block_keyword(code))
end

.start_nesting_after?(code) ⇒ Boolean

Returns:

  • (Boolean) 


711
712
713
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 711

def self.start_nesting_after?(code)
  anonymous_block?(code) || start_block_keyword?(code)
end

Instance Method Details

#add_following_empty_lines(node, lines) ⇒ Object

Adds empty lines that follow the lines (Used for scripts), so that RuboCop can receive them too. Some cops are sensitive to empty lines.



683
684
685
686
687
688
689
690
691
692
693
694
695
696
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 683

def add_following_empty_lines(node, lines)
  first_line_index = node.line - 1 + lines.size
  extra_lines = []

  extra_lines << '' while HamlLint::Utils.is_blank_line?(@original_haml_lines[first_line_index + extra_lines.size])

  if @original_haml_lines[first_line_index + extra_lines.size].nil?
    # Since we reached the end of the document without finding content,
    # then we don't add those lines.
    return lines
  end

  lines + extra_lines
end

#add_html_attributes_value_usage(node, attributes_code, indent:) ⇒ Object

For HTML-style (parens) attributes, adds the attribute value expressions as a non-autocorrectable chunk purely so RuboCop counts the variables/methods used there.



356
357
358
359
360
361
362
363
364
365
366
367
368
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 356

def add_html_attributes_value_usage(node, attributes_code, indent:)
  ast = parse_ruby(attributes_code)
  return unless ast&.type == :hash

  value_sources = ast.children.filter_map do |pair|
    next unless pair.type == :pair
    pair.children[1]&.source
  end
  return if value_sources.empty?

  lines = value_sources.map { |value| "#{' ' * indent}#{script_output_prefix}#{value}" }
  @ruby_chunks << AdHocChunk.new(node, lines)
end

#add_interpolation_chunks(node, code, haml_line_index, indent:, line_start_index: 0) ⇒ Object

Adds chunks for the interpolation within the given code



451
452
453
454
455
456
457
458
459
460
461
462
463
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
492
493
494
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 451

def add_interpolation_chunks(node, code, haml_line_index, indent:, line_start_index: 0)
  HamlLint::Utils.handle_interpolation_with_indexes(code) do |scanner, line_index, line_char_index|
    escapes = scanner[2].size
    next if escapes.odd?
    char = scanner[3] # '{', '@' or '$'
    if Gem::Version.new(Haml::VERSION) >= Gem::Version.new('5') && (char != '{')
      # Before Haml 5, scanner didn't have a scanner[3], it only handled `#{}`
      next
    end

    line_start_char_index = line_char_index
    line_start_char_index += line_start_index if line_index == 0
    code_start_char_index = scanner.charpos

    # This moves the scanner
    Haml::Util.balance(scanner, '{', '}', 1)

    # Need to manually get the code now that we have positions so that all whitespace is present,
    # because Haml::Util.balance does a strip...
    interpolated_code = code[code_start_char_index...scanner.charpos - 1]

    if interpolated_code.include?("\n")
      # We can't correct multiline interpolation.
      # Finding meaningful code to generate and then transfer back is pretty complex

      # Since we can't fix it, strip around the code to reduce RuboCop lints that we won't be able to fix.
      interpolated_code = interpolated_code.strip
      interpolated_code = "#{' ' * indent}#{script_output_prefix}#{interpolated_code}"

      placeholder_code = interpolated_code.gsub(/\s*\n\s*/, ' ').rstrip
      unless parse_ruby(placeholder_code)
        placeholder_code = interpolated_code.gsub(/\s*\n\s*/, '; ').rstrip
      end
      @ruby_chunks << AdHocChunk.new(node, [placeholder_code],
                                     haml_line_index: haml_line_index + line_index)
    else
      interpolated_code = "#{' ' * indent}#{script_output_prefix}#{interpolated_code}"
      @ruby_chunks << InterpolationChunk.new(node, [interpolated_code],
                                             haml_line_index: haml_line_index + line_index,
                                             start_char_index: line_start_char_index,
                                             end_marker_indent: indent)
    end
  end
end

#extract ⇒ Object 



30
31
32
33
34
35
36
37
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 30

def extract
  raise 'Already extracted' if @ruby_chunks

  prepare_extract

  visit(@document.tree)
  @ruby_chunks
end

#extract_piped_plain_multilines(first_line_index) ⇒ Object 



598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 598

def extract_piped_plain_multilines(first_line_index)
  lines = []

  cur_line = @original_haml_lines[first_line_index].rstrip
  cur_line_index = first_line_index

  # The pipes must also be on the last line of the multi-line section
  while cur_line && process_plain_multiline!(cur_line)
    lines << cur_line
    cur_line_index += 1
    cur_line = @original_haml_lines[cur_line_index].rstrip
  end

  if lines.empty?
    lines << cur_line
  end
  lines
end

#extract_raw_ruby_lines(haml_processed_ruby_code, first_line_index) ⇒ first_line_offset, ruby_lines

HAML transforms the ruby code in many ways as it parses a document. Often removing lines and/or indentation. This is quite annoying for us since we want the exact layout of the code to analyze it.

This function receives the code as haml provides it and the line where it starts. It returns the actual code as it is in the haml file, keeping breaks and indentation for the following lines. In addition, the start position of the code in the first line.

The rules for handling multiline code in HAML are as follow:

  • if the line being processed ends with a space and a pipe, then append to the line (without newlines) every following lines that also end with a space and a pipe. This means the last line of the “block” also needs a pipe at the end.

  • after processing the pipes, when dealing with ruby code (and not in tag attributes’ hash), if the line (which maybe span across multiple lines) ends with a comma, add the next line to the current piece of code.

Returns:

  • (first_line_offset, ruby_lines) 


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
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
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 539

def extract_raw_ruby_lines(haml_processed_ruby_code, first_line_index)
  haml_processed_ruby_code = haml_processed_ruby_code.strip
  first_line = @original_haml_lines[first_line_index]

  char_index = first_line.index(haml_processed_ruby_code)

  if char_index
    return [char_index, [haml_processed_ruby_code]]
  end

  cur_line_index = first_line_index
  cur_line = first_line.rstrip
  lines = []

  # The pipes must also be on the last line of the multi-line section
  while cur_line && process_multiline!(cur_line)
    lines << cur_line
    cur_line_index += 1
    cur_line = @original_haml_lines[cur_line_index].rstrip
  end

  if lines.empty?
    lines << cur_line
  else
    # The pipes must also be on the last line of the multi-line section. So cur_line is not the next line.
    # We want to go back to check for commas
    cur_line_index -= 1
    cur_line = lines.last
  end

  while HAML_PARSER_INSTANCE.send(:is_ruby_multiline?, cur_line)
    cur_line_index += 1
    cur_line = @original_haml_lines[cur_line_index].rstrip
    lines << cur_line
  end

  joined_lines = lines.join("\n")

  if haml_processed_ruby_code.include?("\n")
    haml_processed_ruby_code = haml_processed_ruby_code.tr("\n", ' ')
  end

  haml_processed_ruby_code.split(/[, ]/)

  regexp = HamlLint::Utils.regexp_for_parts(haml_processed_ruby_code.split(/,\s*|\s+/), '(?:,\\s*|\\s+)')

  match = joined_lines.match(regexp)
  # This can happen when pipes are used as marker for multiline parts, and when tag attributes change lines
  # without ending by a comma. This is quite a can of worm and is probably not too frequent, so for now,
  # these cases are not supported.
  return if match.nil?

  raw_ruby = match[0]
  ruby_lines = raw_ruby.split("\n")
  first_line_offset = match.begin(0)

  [first_line_offset, ruby_lines]
end

#extract_raw_tag_attributes_ruby_lines(haml_processed_ruby_code, first_line_index) ⇒ Object

Tag attributes actually handle multiline differently than scripts. The basic system basically keeps considering more lines until it meets the closing braces, but still processes pipes too (same as extract_raw_ruby_lines).



620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 620

def extract_raw_tag_attributes_ruby_lines(haml_processed_ruby_code, first_line_index)
  haml_processed_ruby_code = haml_processed_ruby_code.strip
  first_line = @original_haml_lines[first_line_index]

  char_index = first_line.index(haml_processed_ruby_code)

  if char_index
    return [char_index, [haml_processed_ruby_code]]
  end

  # The +1 is for the closing brace, which we need
  min_non_white_chars_to_add = haml_processed_ruby_code.scan(/\S/).size + 1

  regexp = HamlLint::Utils.regexp_for_parts(
    haml_processed_ruby_code.split(/\s+/), '\\s+', prefix: '\s*', suffix: '\s*(?=[)}])'
  )

  joined_lines = first_line.rstrip
  process_multiline!(joined_lines)

  cur_line_index = first_line_index + 1
  while @original_haml_lines[cur_line_index] && min_non_white_chars_to_add > 0
    new_line = @original_haml_lines[cur_line_index].rstrip
    process_multiline!(new_line)

    min_non_white_chars_to_add -= new_line.scan(/\S/).size
    joined_lines << "\n"
    joined_lines << new_line
    cur_line_index += 1
  end

  match = joined_lines.match(regexp)

  return if match.nil?

  first_line_offset = match.begin(0)
  raw_ruby = match[0]
  ruby_lines = raw_ruby.split("\n", -1)

  [first_line_offset, ruby_lines]
end

#finish_visit_any_script(node, lines, raw_code: nil, must_start_chunk: false, first_output_prefix: '=') ⇒ Object

Code common to both silent and outputting scripts

raw_code is the code before we do transformations, such as adding the ‘HL.out = `



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
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 215

def finish_visit_any_script(node, lines, raw_code: nil, must_start_chunk: false, first_output_prefix: '=')
  raw_code ||= lines.join("\n")
  start_nesting = self.class.start_nesting_after?(raw_code)

  lines = add_following_empty_lines(node, lines)

  my_indent = lines.first.index(/\S/)
  indent_after = indent_after_line_index(node.line - 1 + lines.size - 1) || 0
  indent_after = [my_indent, indent_after].max

  @ruby_chunks << ScriptChunk.new(node, lines,
                                  end_marker_indent: indent_after,
                                  must_start_chunk: must_start_chunk,
                                  previous_chunk: @ruby_chunks.last,
                                  first_output_haml_prefix: first_output_prefix)

  yield

  if start_nesting
    if node.children.empty?
      raise "Line #{node.line} should be followed by indentation. This might actually" \
            " work in Haml, but it's almost a bug that it does. haml-lint cannot process."
    end

    last_child = node.children.last
    if last_child.is_a?(HamlLint::Tree::SilentScriptNode) && last_child.keyword == 'end'
      # This is allowed in Haml 5, gotta handle it!
      # No need for the implicit end chunk since there is an explicit one
    else
      @ruby_chunks << ImplicitEndChunk.new(node, [' ' * my_indent + 'end'],
                                           haml_line_index: @ruby_chunks.last.haml_end_line_index,
                                           end_marker_indent: my_indent)
    end
  end
end

#indent_after_line_index(line_index) ⇒ Object 



703
704
705
706
707
708
709
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 703

def indent_after_line_index(line_index)
  (line_index + 1..@original_haml_lines.size - 1).each do |i|
    indent = @original_haml_lines[i].index(/\S/)
    return indent if indent
  end
  nil
end

#parse_ruby(source) ⇒ Object 



698
699
700
701
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 698

def parse_ruby(source)
  @ruby_parser ||= HamlLint::RubyParser.new
  @ruby_parser.parse(source)
end

#prepare_extract ⇒ Object

Useful for tests



40
41
42
43
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 40

def prepare_extract
  @ruby_chunks = []
  @original_haml_lines = @document.source_lines
end

#process_multiline!(line) ⇒ Object 



496
497
498
499
500
501
502
503
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 496

def process_multiline!(line)
  if HAML_PARSER_INSTANCE.send(:is_multiline?, line)
    line.chop!.rstrip!
    true
  else
    false
  end
end

#process_plain_multiline!(line) ⇒ Object 



505
506
507
508
509
510
511
512
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 505

def process_plain_multiline!(line)
  if line&.end_with?(' |')
    line[-2..] = ''
    true
  else
    false
  end
end

#visit_comment(node) ⇒ Object

Visiting comments which are output to HTML. Lines looking like

`  / This will be in the HTML source!`



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
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 89

def visit_comment(node)
  line = @original_haml_lines[node.line - 1]
  indent = line.index(/\S/)

  @ruby_chunks << PlaceholderMarkerChunk.new(node, 'comment', indent: indent)

  # Comment can have subnodes, such as plain. This happens for conditional comments, such as:
  # %head
  #   /[if mso]
  #     %div
  if node.children
    # We don't want to use a block because assignments in a block are local to that block,
    # so the semantics of the extracted ruby would be different from the one generated by
    # Haml. Those differences can make some cops, such as UselessAssignment, have false
    # positives
    begin_chunk = AdHocChunk.new(node, [' ' * indent + 'begin'])
    @ruby_chunks << begin_chunk
    indent += 2

    yield

    indent -= 2

    if @ruby_chunks.last.equal?(begin_chunk)
      # So there is nothing nesting, remove the wrapping "begin"
      @ruby_chunks.pop
    else
      @ruby_chunks << AdHocChunk.new(node,
                                     [' ' * indent + 'ensure', ' ' * indent + '  HL.noop', ' ' * indent + 'end'],
                                     haml_line_index: @ruby_chunks.last.haml_end_line_index)
    end
  end
end

#visit_filter(node) ⇒ Object

Visiting a HAML filter. Lines looking like ‘ :javascript` and the following lines that are nested.



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
438
439
440
441
442
443
444
445
446
447
448
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 411

def visit_filter(node)
  # For unknown reasons, haml doesn't escape interpolations in filters.
  # So we can rely on \n to split / get the number of lines.
  filter_name_indent = @original_haml_lines[node.line - 1].index(/\S/)
  if node.filter_type == 'ruby'
    # The indentation in node.text is normalized, so that at least one line
    # is indented by 0. Split("\n", -1) + pop keeps trailing blank lines (plain split drops them),
    # which the end marker needs so it doesn't land right after the last code line.
    lines = node.text.split("\n", -1)
    lines.pop
    lines.map! do |line|
      if !/\S/.match?(line)
        # whitespace or empty
        ''
      else
        ' ' * filter_name_indent + line
      end
    end

    @ruby_chunks << RubyFilterChunk.new(node, lines,
                                        haml_line_index: node.line, # it's one the next line, no need for -1
                                        start_marker_indent: filter_name_indent,
                                        end_marker_indent: filter_name_indent)
  elsif node.text.include?('#')
    name_indentation = ' ' * @original_haml_lines[node.line - 1].index(/\S/)
    # TODO: HAML_LINT_FILTER could be in the string and mess things up
    lines = ["#{name_indentation}#{script_output_prefix}<<~HAML_LINT_FILTER"]
    lines.concat @original_haml_lines[node.line..(node.line + node.text.count("\n") - 1)]
    lines << "#{name_indentation}HAML_LINT_FILTER"
    @ruby_chunks << NonRubyFilterChunk.new(node, lines,
                                           end_marker_indent: filter_name_indent)
  # Those could be interpolation. We treat them as a here-doc, which is nice since we can
  # keep the indentation as-is.
  else
    @ruby_chunks << PlaceholderMarkerChunk.new(node, 'filter', indent: filter_name_indent,
                                               nb_lines: 1 + node.text.count("\n"))
  end
end

#visit_haml_comment(node) ⇒ Object

Visiting lines like ‘ -# Some commenting!`



56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 56

def visit_haml_comment(node)
  # We want to preserve leading whitespace if it exists, but add a leading
  # whitespace if it doesn't exist so that RuboCop's LeadingCommentSpace
  # doesn't complain
  line_index = node.line - 1
  lines = @original_haml_lines[line_index..(line_index + node.text.count("\n"))].dup
  indent = lines.first.index(/\S/)
  # Remove only the -, the # will align with regular code
  #  -# comment
  #  - foo()
  # becomes
  #  # comment
  #  foo()
  lines[0] = lines[0].sub('-', '')

  # Adding a space before the comment if its missing
  # We can't fix those, so make sure not to generate warnings for them.
  lines[0] = lines[0].sub(/\A(\s*)#(\S)/, '\\1# \\2')

  HamlLint::Utils.map_after_first!(lines) do |line|
    # Since the indent/spaces of the extra line comments isn't exactly in the haml,
    # it's not RuboCop's job to fix indentation, so just make a reasonable indentation
    # to avoid offenses.
    ' ' * indent + line.sub(/^\s*/, '# ').rstrip
  end

  # Using Placeholder instead of script because we can't revert back to the
  # exact original comment since multiple syntax lead to the exact same comment.
  @ruby_chunks << HamlCommentChunk.new(node, lines, end_marker_indent: indent)
end

#visit_plain(node) ⇒ Object

Visiting lines like ‘ Some raw text to output`



50
51
52
53
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 50

def visit_plain(node)
  indent = @original_haml_lines[node.line - 1].index(/\S/)
  @ruby_chunks << PlaceholderMarkerChunk.new(node, 'plain', indent: indent)
end

#visit_root(_node) ⇒ Object 



45
46
47
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 45

def visit_root(_node)
  yield # Collect lines of code from children
end

#visit_script(node, &block) ⇒ Object

Visit a script which outputs. Lines looking like ‘ = foo`



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
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 124

def visit_script(node, &block)
  raw_first_line = @original_haml_lines[node.line - 1]

  # ==, !, !==, &, &== means interpolation (was needed before HAML 2.2... it's still supported)
  # =, !=, &= mean actual ruby code is coming
  # Anything else is interpolation
  # The regex lists the case for Ruby Code. The 3 cases and making sure they are not followed by another = sign

  match = raw_first_line.match(/\A\s*(=|!=|&=)(?!=)/)
  unless match
    # The line doesn't start with a - or a =, this is actually a "plain"
    # that contains interpolation.
    indent = raw_first_line.index(/\S/)
    @ruby_chunks << PlaceholderMarkerChunk.new(node, 'interpolation', indent: indent)
    lines = extract_piped_plain_multilines(node.line - 1)
    add_interpolation_chunks(node, lines.join("\n"), node.line - 1, indent: indent)
    return
  end

  script_prefix = match[1]
  _first_line_offset, lines = extract_raw_ruby_lines(node.script, node.line - 1)
  # We want the actual indentation and prefix for the first line
  first_line = lines[0] = @original_haml_lines[node.line - 1].rstrip
  process_multiline!(first_line)

  lines[0] = lines[0].sub(/(#{script_prefix}[ \t]?)/, '')
  line_indentation = Regexp.last_match(1).size

  raw_code = lines.join("\n")

  if lines[0][/\S/] == '#'
    # a "=" script that only contains a comment... No need for the "HL.out = " prefix,
    # just treat it as comment which will turn into a "-" comment
  else
    lines[0] = HamlLint::Utils.insert_after_indentation(lines[0], script_output_prefix)
  end

  indent_delta = script_output_prefix.size - line_indentation
  HamlLint::Utils.map_after_first!(lines) do |line|
    HamlLint::Utils.indent(line, indent_delta)
  end

  prev_chunk = @ruby_chunks.last
  if prev_chunk.is_a?(ScriptChunk) &&
      prev_chunk.node.type == :script &&
      prev_chunk.node == node.parent
    # When an outputting script is nested under another outputting script,
    # we want to block them from being merged together by rubocop, because
    # this doesn't make sense in HAML.
    # Example:
    #   = if this_is_short
    #     = this_is_short_too
    # Could become (after RuboCop):
    #   HL.out = (HL.out = this_is_short_too if this_is_short)
    # Or in (broken) HAML style:
    #   = this_is_short_too = if this_is_short
    # By forcing this to start a chunk, there will be extra placeholders which
    # blocks rubocop from merging the lines.
    must_start_chunk = true
  elsif script_prefix != '='
    # In the few cases where &= and != are used to start the script,
    # We need to remember and put it back in the final HAML. Fusing scripts together
    # would make that basically impossible. Instead, a script has a "first_output_prefix"
    # field for this specific case
    must_start_chunk = true
  end

  finish_visit_any_script(node, lines, raw_code: raw_code, must_start_chunk: must_start_chunk,
                          first_output_prefix: script_prefix, &block)
end

#visit_silent_script(node, &block) ⇒ Object

Visit a script which doesn’t output. Lines looking like ‘ - foo`



196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 196

def visit_silent_script(node, &block)
  _first_line_offset, lines = extract_raw_ruby_lines(node.script, node.line - 1)
  # We want the actual indentation and prefix for the first line
  first_line = lines[0] = @original_haml_lines[node.line - 1].rstrip
  process_multiline!(first_line)

  lines[0] = lines[0].sub(/(-[ \t]?)/, '')
  nb_to_deindent = Regexp.last_match(1).size

  HamlLint::Utils.map_after_first!(lines) do |line|
    line.sub(/^ {1,#{nb_to_deindent}}/, '')
  end

  finish_visit_any_script(node, lines, &block)
end

#visit_tag(node) ⇒ Object

Visiting a tag. Lines looking like ‘ %div`



252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 252

def visit_tag(node)
  indent = @original_haml_lines[node.line - 1].index(/\S/)

  @ruby_chunks << PlaceholderMarkerChunk.new(node, 'tag', indent: indent)

  current_line_index = visit_tag_attributes(node, indent: indent)
  visit_tag_script(node, line_index: current_line_index, indent: indent)

  # We don't want to use a block because assignments in a block are local to that block,
  # so the semantics of the extracted ruby would be different from the one generated by
  # Haml. Those differences can make some cops, such as UselessAssignment, have false
  # positives
  code = 'begin'
  begin_chunk = AdHocChunk.new(node, [' ' * indent + code])
  @ruby_chunks << begin_chunk

  indent += 2

  yield

  indent -= 2

  if @ruby_chunks.last.equal?(begin_chunk)
    # So there is nothing going "in" the tag, remove the wrapping "begin" and replace the PlaceholderMarkerChunk
    # by one less indented
    @ruby_chunks.pop
  else
    @ruby_chunks << AdHocChunk.new(node,
                                   [' ' * indent + 'ensure', ' ' * indent + '  HL.noop', ' ' * indent + 'end'],
                                   haml_line_index: @ruby_chunks.last.haml_end_line_index)
  end
end

#visit_tag_attributes(node, indent:) ⇒ Object

(Called manually form visit_tag) Visiting the attributes of a tag. Lots of different examples below in the code. A common syntax is: ‘%div’yes_please’‘

Returns the new line_index we reached, useful to handle the script that follows



290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 290

def visit_tag_attributes(node, indent:)
  final_line_index = node.line - 1
  additional_attributes = node.dynamic_attributes_sources

  attributes_code = additional_attributes.first
  if !attributes_code && node.hash_attributes? && node.dynamic_attributes_sources.empty?
    # No idea why .foo{:bar => 123} doesn't get here, but .foo{:bar => '123'} does...
    # The code we get for the latter is {:bar => '123'}.
    # We normalize it by removing the { } so that it matches wha we normally get
    attributes_code = node.dynamic_attributes_source[:hash][1...-1]
  end

  if attributes_code&.start_with?('{')
    # HTML-style (parens) attributes, e.g. %div(foo=foo), arrive as a synthesized hash string
    # like '{"foo" => foo,}'. That text isn't present verbatim in the HAML, so we can't map
    # RuboCop corrections back and intentionally don't autocorrect it. We still extract the
    # attribute *values* as a non-correctable chunk so RuboCop sees variables used here;
    # otherwise they look unused (false Lint/UselessAssignment, plus unsafe autocorrect that
    # removes the assignment). See issue #648.
    add_html_attributes_value_usage(node, attributes_code, indent: indent)
    attributes_code = nil
  end

  return final_line_index unless attributes_code
  # Attributes have different ways to be given to us:
  #   .foo{bar: 123} => "bar: 123"
  #   .foo{:bar => 123} => ":bar => 123"
  #   .foo{:bar => '123'} => "{:bar => '123'}" # No idea why this is different
  #   .foo(bar = 123) => '{"bar" => 123,}'
  #   .foo{html_attrs('fr-fr')} => html_attrs('fr-fr')
  #
  # The (bar = 123) case is extra painful to autocorrect (so is ignored up there).
  # #raw_ruby_from_haml  will "detect" this case by not finding the code.
  #
  # We wrap the result in a method to have a valid syntax for all 3 ways
  # without having to differentiate them.
  first_line_offset, raw_attributes_lines = extract_raw_tag_attributes_ruby_lines(attributes_code,
                                                                                  node.line - 1)
  return final_line_index unless raw_attributes_lines

  final_line_index += raw_attributes_lines.size - 1

  # Since .foo{bar: 123} => "bar: 123" needs wrapping (Or it would be a syntax error) and
  # .foo{html_attrs('fr-fr')} => html_attrs('fr-fr') doesn't care about being
  # wrapped, we always wrap to place them to a similar offset to how they are in the haml.
  wrap_by = first_line_offset - indent
  if wrap_by < 2
    # Need 2 minimum, for "W(". If we have less, we must indent everything for the difference
    extra_indent = 2 - wrap_by
    HamlLint::Utils.map_after_first!(raw_attributes_lines) do |line|
      HamlLint::Utils.indent(line, extra_indent)
    end
    wrap_by = 2
  end
  raw_attributes_lines = wrap_lines(raw_attributes_lines, wrap_by)
  raw_attributes_lines[0] = ' ' * indent + raw_attributes_lines[0]

  @ruby_chunks << TagAttributesChunk.new(node, raw_attributes_lines,
                                         end_marker_indent: indent,
                                         indent_to_remove: extra_indent)

  final_line_index
end

#visit_tag_script(node, line_index:, indent:) ⇒ Object

Visiting the script besides tag. The part to the right of the equal sign of lines looking like ‘ %div= foo(bar)`



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
400
401
402
403
404
405
406
407
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 372

def visit_tag_script(node, line_index:, indent:)
  return if node.script.nil? || node.script.empty?
  # We ignore scripts which are just a comment
  return if node.script[/\S/] == '#'

  first_line_offset, script_lines = extract_raw_ruby_lines(node.script, line_index)

  if script_lines.nil?
    # This is a string with interpolation after a tag
    # ex: %tag hello #{world}
    # Sadly, the text with interpolation is escaped from the original, but this code
    # needs the original.

    interpolation_original = @document.unescape_interpolation_to_original_cache[node.script]
    line_start_index = @original_haml_lines[node.line - 1].rindex(interpolation_original)
    if line_start_index.nil?
      raw_lines = extract_piped_plain_multilines(node.line - 1)
      equivalent_haml_code = "#{raw_lines.first} #{raw_lines[1..].map(&:lstrip).join(' ')}"
      line_start_index = equivalent_haml_code.rindex(interpolation_original)

      interpolation_original = raw_lines.join("\n")
    end
    add_interpolation_chunks(node, interpolation_original, node.line - 1,
                             line_start_index: line_start_index, indent: indent)
  else
    script_lines[0] = "#{' ' * indent}#{script_output_prefix}#{script_lines[0]}"
    indent_delta = script_output_prefix.size - first_line_offset + indent
    HamlLint::Utils.map_after_first!(script_lines) do |line|
      HamlLint::Utils.indent(line, indent_delta)
    end

    @ruby_chunks << TagScriptChunk.new(node, script_lines,
                                       haml_line_index: line_index,
                                       end_marker_indent: indent)
  end
end

#wrap_lines(lines, wrap_depth) ⇒ Object 



662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
 
# File 'lib/haml_lint/ruby_extraction/chunk_extractor.rb', line 662

def wrap_lines(lines, wrap_depth)
  lines = lines.dup
  wrapping_prefix = 'W' * (wrap_depth - 1) + '('

  # Strip leading/trailing space only when linting (not autocorrecting) to avoid SpaceInsideParens violations.
  # Preserve spaces for fully nested attribute hashes that need no adjustment.
  # Preserve spaces during autocorrect, so RuboCop can fix them and map corrections back correctly.
  unless lines[0] =~ /^\s*$/ || autocorrect
    leading_spaces = lines[0][/^\s*/].length
    lines = lines.map { |line| line.gsub(/^\s{,#{leading_spaces}}/, '') }
    lines[-1] = lines[-1].gsub(/\s+\z/, '')
  end

  lines[0] = wrapping_prefix + lines[0]
  lines[-1] = lines[-1] + ')'

  lines
end