CoCalc provides the best real-time collaborative environment for Jupyter Notebooks, LaTeX documents, and SageMath, scalable from individual users to large groups and classes!

1
# frozen_string_literal: true
2

3
module RuboCop
4
  module Cop
5
    module Lint
6
      # Looks for invalid pack/unpack directives with Ruby 3.3.0 some directives
7
      # that used to not raise errors, now will - context: https://bugs.ruby-lang.org/issues/19150:
8
      #   * Array#pack now raises ArgumentError for unknown directives
9
      #   * String#unpack now raises ArgumentError for unknown directives
10
      #
11
      # @example
12
      #   # bad
13
      #   ```
14
      #   3.3.0-preview1 :003 > [0x1].pack('<L')
15
      #   <internal:pack>:8:in `pack': unknown pack directive '<' in '<L' (ArgumentError)
16
      #   ```
17
      #
18
      #   # good
19
      #   ```
20
      #   3.3.0-preview1 :001 > [0x1].pack('L<')
21
      #     => "\x01\x00\x00\x00"
22
      #   ```
23
      class DetectInvalidPackDirectives < RuboCop::Cop::Base
24
        # https://github.com/ruby/ruby/blob/7cfabe1acc55b24fc2c479a87efa71cf74e9e8fc/pack.c#L38
25
        MODIFIABLE_DIRECTIVES = %w[s S i I l L q Q j J]
26

27
        # https://github.com/ruby/ruby/blob/7cfabe1acc55b24fc2c479a87efa71cf74e9e8fc/pack.c#L298
28
        ACCEPTABLE_DIRECTIVES = %w[U m A B H a A Z b B h H c C s S i I l L q Q j J n N v V f F e E d D g G x X @ % U u m M P p w]
29

30
        # @param [RuboCop::AST::SendNode] node Node for the ruby `send` method
31
        # @return [[RuboCop::AST::Node], raise] offense when an invalid directive is found, or raise if unexpected error found
32
        def on_send(node)
33
          _callee, method_name = *node
34

35
          return unless %i[pack unpack pack1 unpack1].include?(method_name)
36

37
          args = node.arguments
38
          return if args.empty?
39

40
          args.each do |arg|
41
            next unless string_arg?(arg)
42

43
            # if multiline arguments are passed
44
            if arg.type == :dstr
45
              idx = []
46

47
              pack_directive = arg.children.map do |child|
48
                if begin_arg?(child)
49
                  next
50
                else
51
                  idx << child.children.first.length
52
                  child.children.join
53
                end
54
              end.join
55

56
            # elsif single line arguments are passed
57
            elsif arg.type == :str
58
              pack_directive = arg.children.first
59
            end
60

61
            error = validate_directive(pack_directive)
62
            if error.nil?
63
              next
64
            else
65
              offense_range = get_error_range(arg, error[:index], idx)
66
              return if offense_range.nil?
67

68
              add_offense(offense_range, message: error[:message])
69
            end
70
          end
71
        end
72

73
        private
74

75
        # Check if the pack directives are valid. See link for pack docs https://apidock.com/ruby/Array/pack
76
        #
77
        # Code based on https://github.com/ruby/ruby/blob/6391132c03ac08da0483adb986ff9a54e41f9e14/pack.c#L196
78
        # adapted into Ruby
79
        #
80
        # @param [String] pack_directive The ruby pack/unpack directive to validate
81
        # @return [Hash,nil] A hash with the message and index that the invalidate directive was found at, or nil.
82
        def validate_directive(pack_directive)
83
          # current pointer value
84
          p = 0
85

86
          # end of pointer range
87
          pend = pack_directive.length
88

89
          while p < pend
90
            explicit_endian = 0
91
            type_index = p
92

93
            # get data type
94
            type = pack_directive[type_index]
95
            p += 1
96

97
            if type.blank?
98
              next
99
            end
100

101
            if type == '#'
102
              p += 1 while p < pend && pack_directive[p] != "\n"
103
              next
104
            end
105

106
            # Modifiers
107
            loop do
108
              case pack_directive[p]
109
              when '_', '!'
110
                if MODIFIABLE_DIRECTIVES.include?(type)
111
                  p += 1
112
                else
113
                  return { message: "'#{pack_directive[p]}' allowed only after types #{MODIFIABLE_DIRECTIVES.join}", index: p }
114
                end
115
              when '<', '>'
116
                unless MODIFIABLE_DIRECTIVES.include?(type)
117
                  return { message: "'#{pack_directive[p]}' allowed only after types #{MODIFIABLE_DIRECTIVES.join}", index: p }
118
                end
119

120
                if explicit_endian != 0
121
                  return { message: "Can't use both '<' and '>'.", index: p }
122
                end
123

124
                explicit_endian = pack_directive[p]
125
                p += 1
126
              else
127
                break
128
              end
129
            end
130

131
            # Data length
132
            if pack_directive[p] == '*'
133
              p += 1
134
            elsif pack_directive[p]&.match?(/\d/)
135
              p += 1 while pack_directive[p]&.match?(/\d/)
136
            end
137

138
            # Check type
139
            unless ACCEPTABLE_DIRECTIVES.include?(type)
140
              return { message: "unknown pack directive '#{type}' in '#{pack_directive}'", index: type_index }
141
            end
142
          end
143

144
          nil
145
        end
146

147
        # Checks if the current node is of type `:str` or `:dstr` - `dstr` being multiline
148
        #
149
        # @param [RuboCop::AST::SendNode] node Node for the ruby `send` method
150
        # @return [TrueClass, FalseClass]
151
        def string_arg?(node)
152
          node&.type == :str || node&.type == :dstr
153
        end
154

155
        # Check if the node if of type `:begin`
156
        #
157
        # @param [RuboCop::AST::SendNode] node Node for the ruby `send` method
158
        # @return [TrueClass, FalseClass]
159
        def begin_arg?(node)
160
          node.type == :begin
161
        end
162

163
        # Get the range of the offense to more accurately raise offenses against specific directives
164
        #
165
        # @param [RuboCop::AST::DstrNode, RuboCop::AST::StrNode ] arg The node that need its range calculated
166
        # @param [Integer] p The current pointer value
167
        # @param [Array] idx An array holding to number of indexes for the node
168
        # @return [Parser::Source::Range] The range of the node value
169
        def get_error_range(arg, p, idx)
170
          # Logic for multiline strings
171
          if arg.type == :dstr
172
            total = 0
173
            index = 0
174

175
            idx.each_with_index do |idx_length, count|
176
              if total < p
177
                total += idx_length
178
                index = count
179
              end
180
            end
181
            adjusted_index = p - idx[0..(index - 1)].sum
182

183
            indexed_arg = arg.children[index]
184

185
            if begin_arg?(indexed_arg)
186
              return nil
187
            else
188
              newline_adjustment = indexed_arg.children.first[0..adjusted_index].scan(/[\n\t]/).count
189
            end
190

191
            # If there's opening quotes present, i.e. "a", instead of heredoc which doesn't have preceding opening quotes:
192
            if indexed_arg.loc.begin
193
              range_start = indexed_arg.loc.begin.end_pos + (p - adjusted_index)
194
            else
195
              expression = processed_source.raw_source[indexed_arg.loc.expression.begin.begin_pos...indexed_arg.loc.expression.end.end_pos]
196
              if expression[/^\s+/].nil?
197
                leading_whitespace_size = 0
198
              else
199
                leading_whitespace_size = expression[/^\s+/].length
200
              end
201
              adjusted_index += leading_whitespace_size
202
              range_start = indexed_arg.loc.expression.begin_pos + (adjusted_index + newline_adjustment)
203
            end
204
          # Logic for single line strings
205
          else
206
            newline_adjustment = arg.children.first[0..p].scan(/[\n\t]/).count
207
            range_start = arg.loc.begin.end_pos + (p + newline_adjustment)
208
          end
209

210
          range_end = range_start + 1
211

212
          Parser::Source::Range.new(arg.loc.expression.source_buffer, range_start, range_end)
213
        end
214
      end
215
    end
216
  end
217
end
218

219