Class: Hocon::Impl::PathParser

Inherits:
Object
  • Object
show all
Defined in:
lib/hocon/impl/path_parser.rb

Defined Under Namespace

Classes: Element

Constant Summary collapse

ConfigSyntax = 
Hocon::ConfigSyntax
SimpleConfigOrigin = 
Hocon::Impl::SimpleConfigOrigin
Tokenizer = 
Hocon::Impl::Tokenizer
Tokens = 
Hocon::Impl::Tokens
ConfigNodePath = 
Hocon::Impl::ConfigNodePath
ConfigValueType = 
Hocon::ConfigValueType
ConfigBadPathError = 
Hocon::ConfigError::ConfigBadPathError

Class Method Summary collapse

Class Method Details

.add_path_text(buf, was_quoted, new_text) ⇒ Object 



188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
 
# File 'lib/hocon/impl/path_parser.rb', line 188

def self.add_path_text(buf, was_quoted, new_text)
  i = if was_quoted
        -1
      else
        new_text.index('.') || -1
      end
  current = buf.last
  if i < 0
    # add to current path element
    current.sb << new_text
    # any empty quoted string means this element can
    # now be empty.
    if was_quoted && (current.sb.length == 0)
      current.can_be_empty = true
    end
  else
    # "buf" plus up to the period is an element
    current.sb << new_text[0, i]
    # then start a new element
    buf.push(Element.new("", false))
    # recurse to consume remainder of new_text
    add_path_text(buf, false, new_text[i + 1, new_text.length - 1])
  end
end

.api_origin ⇒ Object 



36
37
38
 
# File 'lib/hocon/impl/path_parser.rb', line 36

def self.api_origin
  SimpleConfigOrigin.new_simple("path parameter")
end

.fast_path_build(tail, s, path_end) ⇒ Object 



255
256
257
258
259
260
261
262
263
264
265
266
267
 
# File 'lib/hocon/impl/path_parser.rb', line 255

def self.fast_path_build(tail, s, path_end)
  # rindex takes last index it should look at, end - 1 not end
  split_at = s.rindex(".", path_end - 1)
  tokens = []
  tokens << Tokens.new_unquoted_text(nil, s)
  # this works even if split_at is -1; then we start the substring at 0
  with_one_more_element = Path.new(s[split_at + 1..path_end], tail)
  if split_at < 0
    with_one_more_element
  else
    fast_path_build(with_one_more_element, s, split_at)
  end
end

.looks_unsafe_for_fast_parser(s) ⇒ Object

the idea is to see if the string has any chars or features that might require the full parser to deal with.



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
253
 
# File 'lib/hocon/impl/path_parser.rb', line 215

def self.looks_unsafe_for_fast_parser(s)
  last_was_dot = true # // start of path is also a "dot"
  len = s.length
  if s.empty?
    return true
  end
  if s[0] == "."
    return true
  end
  if s[len - 1] == "."
    return true
  end

  (0..len).each do |i|
    c = s[i]
    if c =~ /^\w$/
      last_was_dot = false
      next
    elsif c == '.'
      if last_was_dot
        return true # ".." means we need to throw an error
      end
      last_was_dot = true
    elsif c == '-'
      if last_was_dot
        return true
      end
      next
    else
      return true
    end
  end

  if last_was_dot
    return true
  end

  false
end

.parse_path(path) ⇒ Object 



53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
 
# File 'lib/hocon/impl/path_parser.rb', line 53

def self.parse_path(path)
  speculated = speculative_fast_parse_path(path)
  if not speculated.nil?
    return speculated
  end

  reader = StringIO.new(path)

  begin
    tokens = Tokenizer.tokenize(api_origin, reader, ConfigSyntax::CONF)
    tokens.next # drop START
    return parse_path_expression(tokens, api_origin, path)
  ensure
    reader.close
  end
end

.parse_path_expression(expression, origin, original_text = nil, path_tokens = nil, flavor = ConfigSyntax::CONF) ⇒ Object 



77
78
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
160
161
162
163
164
165
 
# File 'lib/hocon/impl/path_parser.rb', line 77

def self.parse_path_expression(expression, origin, original_text = nil,
                                path_tokens = nil, flavor = ConfigSyntax::CONF)
  # each builder in "buf" is an element in the path
  buf = []
  buf.push(Element.new("", false))

  if !expression.has_next?
    raise ConfigBadPathError.new(
              origin,
              original_text,
              "Expecting a field name or path here, but got nothing")
  end

  while expression.has_next?
    t = expression.next

    if ! path_tokens.nil?
      path_tokens << t
    end

    # Ignore all IgnoredWhitespace tokens
    next if Tokens.ignored_whitespace?(t)

    if Tokens.value_with_type?(t, ConfigValueType::STRING)
      v = Tokens.value(t)
      # this is a quoted string; so any periods
      # in here don't count as path separators
      s = v.transform_to_string
      add_path_text(buf, true, s)
    elsif t == Tokens::EOF
      # ignore this; when parsing a file, it should not happen
      # since we're parsing a token list rather than the main
      # token iterator, and when parsing a path expression from the
      # API, it's expected to have an EOF.
    else
      # any periods outside of a quoted string count as
      # separators
      text = nil
      if Tokens.value?(t)
        # appending a number here may add
        # a period, but we _do_ count those as path
        # separators, because we basically want
        # "foo 3.0bar" to parse as a string even
        # though there's a number in it. The fact that
        # we tokenize non-string values is largely an
        # implementation detail.
        v = Tokens.value(t)

        # We need to split the tokens on a . so that we can get sub-paths but still preserve
        # the original path text when doing an insertion
        if ! path_tokens.nil?
          path_tokens.delete_at(path_tokens.size - 1)
          path_tokens.concat(split_token_on_period(t, flavor))
        end
        text = v.transform_to_string
      elsif Tokens.unquoted_text?(t)
        # We need to split the tokens on a . so that we can get sub-paths but still preserve
        # the original path text when doing an insertion on ConfigNodeObjects
        if ! path_tokens.nil?
          path_tokens.delete_at(path_tokens.size - 1)
          path_tokens.concat(split_token_on_period(t, flavor))
        end
        text = Tokens.unquoted_text(t)
      else
        raise ConfigBadPathError.new(
                  origin,
                  original_text,
                  "Token not allowed in path expression: #{t}" +
                      " (you can double-quote this token if you really want it here)")
      end

      add_path_text(buf, false, text)
    end
  end

  pb = Hocon::Impl::PathBuilder.new
  buf.each do |e|
    if (e.sb.length == 0) && !e.can_be_empty
      raise Hocon::ConfigError::ConfigBadPathError.new(
                origin,
                original_text,
                "path has a leading, trailing, or two adjacent period '.' (use quoted \"\" empty string if you want an empty element)")
    else
      pb.append_key(e.sb.string)
    end
  end

  pb.result
end

.parse_path_node(path, flavor = ConfigSyntax::CONF) ⇒ Object 



40
41
42
43
44
45
46
47
48
49
50
51
 
# File 'lib/hocon/impl/path_parser.rb', line 40

def self.parse_path_node(path, flavor = ConfigSyntax::CONF)
  reader = StringIO.new(path)

  begin
    tokens = Tokenizer.tokenize(api_origin, reader,
                                flavor)
    tokens.next # drop START
    parse_path_node_expression(tokens, api_origin, path, flavor)
  ensure
    reader.close
  end
end

.parse_path_node_expression(expression, origin, original_text = nil, flavor = ConfigSyntax::CONF) ⇒ Object 



70
71
72
73
74
75
 
# File 'lib/hocon/impl/path_parser.rb', line 70

def self.parse_path_node_expression(expression, origin, original_text = nil,
                                    flavor = ConfigSyntax::CONF)
  path_tokens = []
  path = parse_path_expression(expression, origin, original_text, path_tokens, flavor)
  ConfigNodePath.new(path, path_tokens);
end

.speculative_fast_parse_path(path) ⇒ Object

do something much faster than the full parser if we just have something like “foo” or “foo.bar”



271
272
273
274
275
276
277
278
 
# File 'lib/hocon/impl/path_parser.rb', line 271

def self.speculative_fast_parse_path(path)
  s = Hocon::Impl::ConfigImplUtil.unicode_trim(path)
  if looks_unsafe_for_fast_parser(s)
    return nil
  end

  fast_path_build(nil, s, s.length)
end

.split_token_on_period(t, flavor) ⇒ Object 



167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
 
# File 'lib/hocon/impl/path_parser.rb', line 167

def self.split_token_on_period(t, flavor)
  token_text = t.token_text
  if token_text == "."
    return [t]
  end
  split_token = token_text.split('.')
  split_tokens = []
  split_token.each do |s|
    if flavor == ConfigSyntax::CONF
      split_tokens << Tokens.new_unquoted_text(t.origin, s)
    else
      split_tokens << Tokens.new_string(t.origin, s, "\"#{s}\"")
    end
    split_tokens << Tokens.new_unquoted_text(t.origin, ".")
  end
  if token_text[-1] != "."
    split_tokens.delete_at(split_tokens.size - 1)
  end
  split_tokens
end