fig 0.1.81 → 0.2.1

data/lib/fig/statement.rb

@@ -8,8 +8,6 @@ module Fig; end
 class Fig::Statement
   ENVIRONMENT_VARIABLE_NAME_REGEX = %r< \A \w+ \z >x
 
-  attr_reader :line, :column, :source_description
-
   def self.position_description(line, column, source_description)
     if not line or not column
       return '' if not source_description
@@ -26,6 +24,8 @@ class Fig::Statement
     return description
   end
 
+  attr_reader :line, :column, :source_description
+
   # This mess of getting these as a single array necessary is due to
   # limitations of the "*" array splat operator in ruby v1.8.
   def initialize(line_column, source_description)
@@ -39,7 +39,7 @@ class Fig::Statement
   # A name for this kind of Statement, usually a keyword for this statement as
   # it appears in package definition files.
   def statement_type()
-    raise NotImplementedError
+    raise NotImplementedError.new self.class.name
   end
 
   # Block will receive a Statement.
@@ -48,19 +48,19 @@ class Fig::Statement
   end
 
   def unparse_as_version(unparser)
-    raise NotImplementedError
+    raise NotImplementedError.new self.class.name
   end
 
   # Returns a two element array containing the version and an explanation of
   # why the version is necessary if the version is greater than 0.
   def minimum_grammar_for_emitting_input()
-    raise NotImplementedError
+    raise NotImplementedError.new self.class.name
   end
 
   # Returns a two element array containing the version and an explanation of
   # why the version is necessary if the version is greater than 0.
   def minimum_grammar_for_publishing()
-    raise NotImplementedError
+    raise NotImplementedError.new self.class.name
   end
 
   def urls()

data/lib/fig/statement/asset.rb

@@ -1,6 +1,3 @@
-require 'cgi'
-
-require 'fig/parser'
 require 'fig/statement'
 require 'fig/string_tokenizer'
 require 'fig/url'
@@ -84,16 +81,7 @@ module Fig::Statement::Asset
 
   module ClassMethods
     def validate_and_process_escapes_in_location(location, &block)
-      tokenizer = Fig::StringTokenizer.new
-      tokenized_string = tokenizer.tokenize(location, &block)
-      return if ! tokenized_string
-
-      # "config" is a reasonable asset name, so we let that pass.
-      if Fig::Parser.strict_keyword?(tokenized_string.to_expanded_string)
-        yield 'is a keyword.'
-      end
-
-      return tokenized_string
+      return Fig::StringTokenizer.new.tokenize(location, &block)
     end
   end
 end

data/lib/fig/statement/command.rb

@@ -7,6 +7,15 @@ module Fig; end
 class Fig::Statement::Command < Fig::Statement
   attr_reader :command
 
+
+  def self.validate_and_process_escapes_in_argument(
+    command_line_argument, &block
+  )
+    tokenizer = Fig::StringTokenizer.new TOKENIZING_SUBEXPRESSION_MATCHER, '@'
+
+    return tokenizer.tokenize command_line_argument, &block
+  end
+
   def initialize(line_column, source_description, command)
     super(line_column, source_description)
 
@@ -21,7 +30,45 @@ class Fig::Statement::Command < Fig::Statement
     return unparser.command(self)
   end
 
+  def minimum_grammar_for_emitting_input()
+    return minimum_grammar()
+  end
+
   def minimum_grammar_for_publishing()
+    return minimum_grammar()
+  end
+
+  private
+
+  def minimum_grammar()
+    if command.size > 1
+      return [1, 'contains multiple components']
+    end
+
+    argument = command.first.to_escaped_string
+
+    # Can't have octothorpes anywhere in v0 due to comment stripping via
+    # regex.
+    if argument =~ /#/
+      return [1, 'contains a comment ("#") character']
+    end
+
+    if argument =~ /"/
+      return [1, %Q<contains a double quote>]
+    end
+
     return [0]
   end
+
+  TOKENIZING_SUBEXPRESSION_MATCHER = [
+    {
+      :pattern => %r< \@ [a-zA-Z0-9_.-]* >x,
+      :action =>
+        lambda {
+          |subexpression, error_block|
+
+          Fig::TokenizedString::Token.new :package_path, subexpression[1..-1]
+        }
+    }
+  ]
 end

data/lib/fig/statement/environment_variable.rb

@@ -1,9 +1,19 @@
+require 'fig/statement'
+require 'fig/string_tokenizer'
+require 'fig/tokenized_string/token'
+
 module Fig; end
 class Fig::Statement; end
 
 # A statement that manipulates an environment variable.
 module Fig::Statement::EnvironmentVariable
-  attr_reader :name, :value
+  attr_reader :name, :tokenized_value
+
+  def self.included(class_included_into)
+    class_included_into.extend(ClassMethods)
+
+    return
+  end
 
   def minimum_grammar_for_emitting_input()
     return minimum_grammar()
@@ -15,7 +25,8 @@ module Fig::Statement::EnvironmentVariable
 
   private
 
-  def minimum_grammar()
+  def standard_minimum_grammar()
+    value = tokenized_value.to_escaped_string
     if value =~ /\s/
       return [1, 'contains whitespace']
     end
@@ -23,7 +34,7 @@ module Fig::Statement::EnvironmentVariable
     # Can't have octothorpes anywhere in v0 due to comment stripping via
     # regex.
     if value =~ /#/
-      return [1, 'contains a "#" character']
+      return [1, 'contains a comment ("#") character']
     end
 
     if value =~ / ( ["'] ) /x
@@ -32,4 +43,54 @@ module Fig::Statement::EnvironmentVariable
 
     return [0]
   end
+
+  module ClassMethods
+    def seperate_name_and_value(combined, &error_block)
+      variable, raw_value = combined.split '=', 2
+      if variable !~ Fig::Statement::ENVIRONMENT_VARIABLE_NAME_REGEX
+        yield \
+          %Q<"#{variable}" does not consist solely of alphanumerics and underscores.>
+        return
+      end
+
+      return [variable, raw_value || '']
+    end
+
+    def tokenize_value(value, &error_block)
+      tokenizer = Fig::StringTokenizer.new TOKENIZING_SUBEXPRESSION_MATCHER, '@'
+      return tokenizer.tokenize value, &error_block
+    end
+
+    # TODO: Test coverage doesn't appear to be running this.
+    def base_v0_value_validation(variable, raw_value)
+      if raw_value =~ /\s/
+        yield %Q<The value of #{variable} (#{raw_value}) contains whitespace.>
+        return
+      end
+      if raw_value =~ /'/
+        yield %Q<The value of #{variable} (#{raw_value}) contains a single quote.>
+        return
+      end
+      if raw_value =~ /"/
+        yield %Q<The value of #{variable} (#{raw_value}) contains a double quote.>
+        return
+      end
+
+      return
+    end
+
+    private
+
+    TOKENIZING_SUBEXPRESSION_MATCHER = [
+      {
+        :pattern => %r<\@>,
+        :action =>
+          lambda {
+            |subexpression, error_block|
+
+            Fig::TokenizedString::Token.new :package_path, '@'
+          }
+      }
+    ]
+  end
 end

data/lib/fig/statement/grammar_version.rb

@@ -20,6 +20,10 @@ class Fig::Statement::GrammarVersion < Fig::Statement
     return unparser.grammar_version(self)
   end
 
+  def minimum_grammar_for_emitting_input()
+    return [version]
+  end
+
   def minimum_grammar_for_publishing()
     return [version]
   end

data/lib/fig/statement/include.rb

@@ -70,6 +70,10 @@ class Fig::Statement::Include < Fig::Statement
     return unparser.include(self)
   end
 
+  def minimum_grammar_for_emitting_input()
+    return [0]
+  end
+
   def minimum_grammar_for_publishing()
     return [0]
   end

data/lib/fig/statement/override.rb

@@ -41,6 +41,10 @@ class Fig::Statement::Override < Fig::Statement
     return unparser.override(self)
   end
 
+  def minimum_grammar_for_emitting_input()
+    return [0]
+  end
+
   def minimum_grammar_for_publishing()
     return [0]
   end

data/lib/fig/statement/path.rb

@@ -8,33 +8,43 @@ module Fig; end
 class Fig::Statement::Path < Fig::Statement
   include Fig::Statement::EnvironmentVariable
 
-  # We block single-quotes right now in order to allow for using them for
-  # quoting later.
-  VALUE_REGEX          = %r< \A [^;:'"<>|\s]+ \z >x
-  ARGUMENT_DESCRIPTION =
-    %q[The value must look like "NAME=VALUE". VALUE cannot contain any of ";:<>|", double quotes, or whitespace.]
-
   # Yields on error.
-  def self.parse_name_value(combined)
-    variable, value = combined.split('=')
+  def self.parse_name_value(combined, &error_block)
+    variable, raw_value = seperate_name_and_value combined, &error_block
+
+    tokenized_value = tokenize_value(raw_value, &error_block)
+
+    if tokenized_value.to_escaped_string.length < 1
+      yield %Q<The value of path variable #{variable} is empty.>
+      return
+    end
+
+    return [variable, tokenized_value]
+  end
+
+  def self.parse_v0_name_value(combined, &error_block)
+    variable, raw_value = seperate_name_and_value combined, &error_block
 
-    if variable !~ ENVIRONMENT_VARIABLE_NAME_REGEX
-      yield
+    if raw_value.length < 1
+      yield %Q<The value of path variable #{variable} is empty.>
+      return
     end
 
-    value = '' if value.nil?
-    if value !~ VALUE_REGEX
-      yield
+    base_v0_value_validation(variable, raw_value, &error_block)
+
+    if raw_value =~ /([;:<>|])/
+      yield %Q<The value of path variable #{variable} (#{raw_value}) contains a "#{raw_value}" character.>
+      return
     end
 
-    return [variable, value]
+    return [variable, tokenize_value(raw_value, &error_block)]
   end
 
-  def initialize(line_column, source_description, name, value)
+  def initialize(line_column, source_description, name, tokenized_value)
     super(line_column, source_description)
 
     @name = name
-    @value = value
+    @tokenized_value = tokenized_value
   end
 
   def statement_type()
@@ -48,4 +58,18 @@ class Fig::Statement::Path < Fig::Statement
   def unparse_as_version(unparser)
     return unparser.path(self)
   end
+
+  private
+
+  def minimum_grammar()
+    base_grammar_version = standard_minimum_grammar
+    return base_grammar_version if base_grammar_version[0] != 0
+
+    value = tokenized_value.to_escaped_string
+    if value =~ / ( [;:<>|] ) /x
+      return [1, %Q<contains a "#{$1}" character>]
+    end
+
+    return [0]
+  end
 end

data/lib/fig/statement/retrieve.rb

@@ -7,14 +7,21 @@ module Fig; end
 # Specifies that files from a package should be copied into the current
 # directory when an environment variable has its value changed.
 class Fig::Statement::Retrieve < Fig::Statement
-  attr_reader   :var, :path
+  def self.tokenize_path(path, &error_block)
+    tokenizer = Fig::StringTokenizer.new TOKENIZING_SUBEXPRESSION_MATCHER, '\\['
+    return tokenizer.tokenize path, &error_block
+  end
+
+  attr_reader :variable
+  attr_reader :tokenized_path
 
-  def initialize(line_column, source_description, var, path)
+  def initialize(line_column, source_description, variable, tokenized_path)
     super(line_column, source_description)
 
-    @var = var
-    @path = path
+    @variable       = variable
+    @tokenized_path = tokenized_path
 
+    path = tokenized_path.to_escaped_string
     # Yeah, it's not cross-platform, but File doesn't have an #absolute? method
     # and this is better than nothing.
     if (
@@ -56,10 +63,59 @@ class Fig::Statement::Retrieve < Fig::Statement
   end
 
   def minimum_grammar_for_emitting_input()
-    return [0]
+    return minimum_grammar()
   end
 
   def minimum_grammar_for_publishing()
+    return minimum_grammar()
+  end
+
+  private
+
+  def minimum_grammar()
+    path = tokenized_path.to_escaped_string
+    if path =~ /\s/
+      return [1, 'contains whitespace']
+    end
+
+    # Can't have octothorpes anywhere in v0 due to comment stripping via
+    # regex.
+    if path =~ /#/
+      return [1, 'contains a comment ("#") character']
+    end
+
+    if path =~ %r< ' >x
+      return [1, %Q<contains a single quote character>]
+    end
+
+    if path =~ %r< " >x
+      return [1, %Q<contains a double quote character>]
+    end
+
+    if path =~ %r< ( [^a-zA-Z0-9_/.\[\]-] ) >x
+      return [1, %Q<contains a "#{$1}" character>]
+    end
+
     return [0]
   end
+
+  TOKENIZING_SUBEXPRESSION_MATCHER = [
+    {
+      # Shortest sequence of characters that starts with an open square bracket
+      # and ends with either a close square bracket or end of string.
+      :pattern => %r< \[ [^\]]* \]? >x,
+      :action  =>
+        lambda {
+          |subexpression, error_block|
+
+          if subexpression == '[package]'
+            return Fig::TokenizedString::Token.new :package_path, '[package]'
+          end
+
+          error_block.call \
+            %q<contains an unescaped "[" that is not followed by "ackage]".>
+          return
+        }
+    }
+  ]
 end

data/lib/fig/statement/set.rb

@@ -7,33 +7,24 @@ module Fig; end
 class Fig::Statement::Set < Fig::Statement
   include Fig::Statement::EnvironmentVariable
 
-  # We block quotes right now in order to allow for using them for
-  # quoting later.
-  VALUE_REGEX          = %r< \A [^\s\\'"]* \z >x
-  ARGUMENT_DESCRIPTION =
-    %q<The value must look like "NAME=VALUE"; VALUE cannot contain whitespace though it can be empty.>
-
   # Yields on error.
-  def self.parse_name_value(combined)
-    variable, value = combined.split('=')
-
-    if variable !~ ENVIRONMENT_VARIABLE_NAME_REGEX
-      yield
-    end
+  def self.parse_name_value(combined, &error_block)
+    variable, raw_value = seperate_name_and_value combined, &error_block
 
-    value = '' if value.nil?
-    if value !~ VALUE_REGEX
-      yield
-    end
+    return [variable, tokenize_value(raw_value, &error_block)]
+  end
 
-    return [variable, value]
+  def self.parse_v0_name_value(combined, &error_block)
+    variable, raw_value = seperate_name_and_value combined, &error_block
+    base_v0_value_validation(variable, raw_value, &error_block)
+    return [variable, tokenize_value(raw_value, &error_block)]
   end
 
-  def initialize(line_column, source_description, name, value)
+  def initialize(line_column, source_description, name, tokenized_value)
     super(line_column, source_description)
 
     @name = name
-    @value = value
+    @tokenized_value = tokenized_value
   end
 
   def statement_type()
@@ -47,4 +38,10 @@ class Fig::Statement::Set < Fig::Statement
   def unparse_as_version(unparser)
     return unparser.set(self)
   end
+
+  private
+
+  def minimum_grammar()
+    return standard_minimum_grammar
+  end
 end