antelope 0.3.2 → 0.4.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.gitignore +25 -25
- data/.rspec +3 -3
- data/.travis.yml +10 -10
- data/.yardopts +7 -7
- data/CONTRIBUTING.md +50 -38
- data/GENERATORS.md +180 -124
- data/Gemfile +7 -7
- data/LICENSE.txt +22 -22
- data/README.md +240 -104
- data/Rakefile +2 -2
- data/TODO.md +58 -58
- data/antelope.gemspec +29 -28
- data/bin/antelope +7 -7
- data/examples/deterministic.ace +35 -35
- data/examples/example.ace +52 -51
- data/examples/example.ace.err +192 -192
- data/examples/example.ace.inf +432 -432
- data/examples/example.ate +70 -70
- data/examples/example.ate.err +192 -192
- data/examples/example.ate.inf +432 -432
- data/examples/liquidscript.ace +233 -233
- data/examples/simple.ace +22 -22
- data/lib/antelope/ace/compiler.rb +334 -334
- data/lib/antelope/ace/errors.rb +30 -30
- data/lib/antelope/ace/scanner/argument.rb +57 -57
- data/lib/antelope/ace/scanner/first.rb +89 -89
- data/lib/antelope/ace/scanner/second.rb +178 -178
- data/lib/antelope/ace/scanner/third.rb +27 -27
- data/lib/antelope/ace/scanner.rb +144 -144
- data/lib/antelope/ace.rb +47 -47
- data/lib/antelope/cli.rb +60 -60
- data/lib/antelope/errors.rb +25 -25
- data/lib/antelope/generation/constructor/first.rb +86 -86
- data/lib/antelope/generation/constructor/follow.rb +105 -105
- data/lib/antelope/generation/constructor/nullable.rb +64 -64
- data/lib/antelope/generation/constructor.rb +127 -127
- data/lib/antelope/generation/errors.rb +17 -17
- data/lib/antelope/generation/null.rb +13 -13
- data/lib/antelope/generation/recognizer/rule.rb +216 -216
- data/lib/antelope/generation/recognizer/state.rb +129 -129
- data/lib/antelope/generation/recognizer.rb +177 -177
- data/lib/antelope/generation/tableizer.rb +176 -176
- data/lib/antelope/generation.rb +15 -15
- data/lib/antelope/generator/base/coerce.rb +115 -0
- data/lib/antelope/generator/base/extra.rb +50 -0
- data/lib/antelope/generator/base.rb +134 -264
- data/lib/antelope/generator/c.rb +11 -11
- data/lib/antelope/generator/c_header.rb +105 -105
- data/lib/antelope/generator/c_source.rb +39 -39
- data/lib/antelope/generator/error.rb +34 -34
- data/lib/antelope/generator/group.rb +60 -57
- data/lib/antelope/generator/html.rb +51 -51
- data/lib/antelope/generator/info.rb +47 -47
- data/lib/antelope/generator/null.rb +18 -18
- data/lib/antelope/generator/output.rb +17 -17
- data/lib/antelope/generator/ruby.rb +112 -79
- data/lib/antelope/generator/templates/c_header.ant +36 -36
- data/lib/antelope/generator/templates/c_source.ant +202 -202
- data/lib/antelope/generator/templates/error.erb +40 -0
- data/lib/antelope/generator/templates/html/antelope.css +53 -1
- data/lib/antelope/generator/templates/html/antelope.html +82 -1
- data/lib/antelope/generator/templates/html/antelope.js +9 -1
- data/lib/antelope/generator/templates/html/css.ant +53 -53
- data/lib/antelope/generator/templates/html/html.ant +82 -82
- data/lib/antelope/generator/templates/html/js.ant +9 -9
- data/lib/antelope/generator/templates/info.erb +61 -0
- data/lib/antelope/generator/templates/{ruby.ant → ruby.erb} +171 -178
- data/lib/antelope/generator.rb +62 -66
- data/lib/antelope/grammar/generation.rb +76 -76
- data/lib/antelope/grammar/loading.rb +84 -84
- data/lib/antelope/grammar/precedence.rb +59 -59
- data/lib/antelope/grammar/precedences.rb +64 -64
- data/lib/antelope/grammar/production.rb +56 -56
- data/lib/antelope/grammar/productions.rb +154 -154
- data/lib/antelope/grammar/symbols.rb +64 -64
- data/lib/antelope/grammar/token/epsilon.rb +23 -23
- data/lib/antelope/grammar/token/error.rb +24 -24
- data/lib/antelope/grammar/token/nonterminal.rb +15 -15
- data/lib/antelope/grammar/token/terminal.rb +15 -15
- data/lib/antelope/grammar/token.rb +231 -231
- data/lib/antelope/grammar.rb +68 -68
- data/lib/antelope/version.rb +6 -6
- data/lib/antelope.rb +18 -19
- data/optimizations.txt +42 -42
- data/spec/antelope/ace/compiler_spec.rb +60 -60
- data/spec/antelope/ace/scanner_spec.rb +27 -27
- data/spec/antelope/generation/constructor_spec.rb +131 -131
- data/spec/fixtures/simple.ace +22 -22
- data/spec/spec_helper.rb +39 -39
- data/spec/support/benchmark_helper.rb +5 -5
- data/spec/support/grammar_helper.rb +14 -14
- data/subl/Ace (Ruby).JSON-tmLanguage +94 -94
- data/subl/Ace (Ruby).tmLanguage +153 -153
- metadata +22 -11
- data/lib/antelope/generator/templates/error.ant +0 -34
- data/lib/antelope/generator/templates/info.ant +0 -53
- data/lib/antelope/template/compiler.rb +0 -78
- data/lib/antelope/template/errors.rb +0 -9
- data/lib/antelope/template/scanner.rb +0 -109
- data/lib/antelope/template.rb +0 -64
- data/spec/antelope/template_spec.rb +0 -50
@@ -1,129 +1,129 @@
|
|
1
|
-
# encoding: utf-8
|
2
|
-
|
3
|
-
require 'forwardable'
|
4
|
-
require 'securerandom'
|
5
|
-
|
6
|
-
module Antelope
|
7
|
-
module Generation
|
8
|
-
class Recognizer
|
9
|
-
|
10
|
-
# A state within the parser. A state has a set of rules, as
|
11
|
-
# well as transitions on those rules.
|
12
|
-
class State
|
13
|
-
|
14
|
-
# All of the rules in this state.
|
15
|
-
#
|
16
|
-
# @return [Set<Rule>]
|
17
|
-
attr_reader :rules
|
18
|
-
|
19
|
-
# All of the transitions that can be made on this state.
|
20
|
-
#
|
21
|
-
# @return [Hash<(Symbol, State)>]
|
22
|
-
attr_reader :transitions
|
23
|
-
|
24
|
-
# The id of this state. This starts off as a string of
|
25
|
-
# hexadecmial characters, but after all of the states are
|
26
|
-
# finalized, this becomes a numeric.
|
27
|
-
#
|
28
|
-
# @return [String, Numeric]
|
29
|
-
attr_accessor :id
|
30
|
-
|
31
|
-
include Enumerable
|
32
|
-
extend Forwardable
|
33
|
-
|
34
|
-
def_delegator :@rules, :each
|
35
|
-
|
36
|
-
# Initialize the state.
|
37
|
-
def initialize
|
38
|
-
@rules = Set.new
|
39
|
-
@transitions = {}
|
40
|
-
@id = format('%10x', object_id)
|
41
|
-
end
|
42
|
-
|
43
|
-
# Gives a nice string representation of the state.
|
44
|
-
#
|
45
|
-
# @return [String]
|
46
|
-
def inspect
|
47
|
-
"#<#{self.class} id=#{id} " \
|
48
|
-
"transitions=[#{transitions.keys.join(', ')}] " \
|
49
|
-
"rules=[{#{rules.to_a.join('} {')}}]>"
|
50
|
-
end
|
51
|
-
|
52
|
-
# Merges another state with this state. It copies all of the
|
53
|
-
# rules into this state, and then merges the transitions on
|
54
|
-
# the given state to this state. It then returns self.
|
55
|
-
#
|
56
|
-
# @raise [ArgumentError] if the given argument is not a state.
|
57
|
-
# @param other [State] the state to merge.
|
58
|
-
# @return [self]
|
59
|
-
def merge!(other)
|
60
|
-
raise ArgumentError, "Expected #{self.class}, " \
|
61
|
-
"got #{other.class}" unless other.is_a? State
|
62
|
-
|
63
|
-
self << other
|
64
|
-
transitions.merge! other.transitions
|
65
|
-
|
66
|
-
self
|
67
|
-
end
|
68
|
-
|
69
|
-
# Finds the rule that match the given production. It
|
70
|
-
# uses fuzzy equality checking. It returns the first rule
|
71
|
-
# that matches.
|
72
|
-
#
|
73
|
-
# @param production [Rule] the rule to compare.
|
74
|
-
# @return [Rule?]
|
75
|
-
def rule_for(production)
|
76
|
-
rules.find { |rule| production === rule }
|
77
|
-
end
|
78
|
-
|
79
|
-
# Appends the given object to this state. The given object
|
80
|
-
# must be a state or a rule. If it's a state, it appends all
|
81
|
-
# of the rules in the state to this state. If it's a rule, it
|
82
|
-
# adds the rule to our rules.
|
83
|
-
#
|
84
|
-
# @raise [ArgumentError] if the argument isn't a {State} or a
|
85
|
-
# {Rule}.
|
86
|
-
# @param rule [State, Rule] the object to append.
|
87
|
-
# @return [self]
|
88
|
-
def <<(rule)
|
89
|
-
case rule
|
90
|
-
when State
|
91
|
-
rule.rules.map(&:clone).each { |r| self << r }
|
92
|
-
when Rule
|
93
|
-
rules << rule
|
94
|
-
when Array, Set
|
95
|
-
rule.each do |part|
|
96
|
-
self << part
|
97
|
-
end
|
98
|
-
else
|
99
|
-
raise ArgumentError, "Expected State or Rule, " \
|
100
|
-
"got #{rule.class}"
|
101
|
-
end
|
102
|
-
|
103
|
-
self
|
104
|
-
end
|
105
|
-
|
106
|
-
alias_method :push, :<<
|
107
|
-
|
108
|
-
# Check to see if this state is fuzzily equivalent to another
|
109
|
-
# state. It does this by checking if the transitions are
|
110
|
-
# equivalent, and then that the rules are fuzzily equivalent.
|
111
|
-
# Ideally, the method is commutative; that is,
|
112
|
-
# `(a === b) == (b === a)`.
|
113
|
-
#
|
114
|
-
# @param other [State] the state to check.
|
115
|
-
# @return [Boolean]
|
116
|
-
# @see Rule#===
|
117
|
-
def ===(other)
|
118
|
-
return super unless other.is_a? State
|
119
|
-
|
120
|
-
other_rules = other.rules.to_a
|
121
|
-
other.transitions == transitions &&
|
122
|
-
rules.size == other_rules.size &&
|
123
|
-
rules.each_with_index
|
124
|
-
.all? { |rule, i| rule === other_rules[i] }
|
125
|
-
end
|
126
|
-
end
|
127
|
-
end
|
128
|
-
end
|
129
|
-
end
|
1
|
+
# encoding: utf-8
|
2
|
+
|
3
|
+
require 'forwardable'
|
4
|
+
require 'securerandom'
|
5
|
+
|
6
|
+
module Antelope
|
7
|
+
module Generation
|
8
|
+
class Recognizer
|
9
|
+
|
10
|
+
# A state within the parser. A state has a set of rules, as
|
11
|
+
# well as transitions on those rules.
|
12
|
+
class State
|
13
|
+
|
14
|
+
# All of the rules in this state.
|
15
|
+
#
|
16
|
+
# @return [Set<Rule>]
|
17
|
+
attr_reader :rules
|
18
|
+
|
19
|
+
# All of the transitions that can be made on this state.
|
20
|
+
#
|
21
|
+
# @return [Hash<(Symbol, State)>]
|
22
|
+
attr_reader :transitions
|
23
|
+
|
24
|
+
# The id of this state. This starts off as a string of
|
25
|
+
# hexadecmial characters, but after all of the states are
|
26
|
+
# finalized, this becomes a numeric.
|
27
|
+
#
|
28
|
+
# @return [String, Numeric]
|
29
|
+
attr_accessor :id
|
30
|
+
|
31
|
+
include Enumerable
|
32
|
+
extend Forwardable
|
33
|
+
|
34
|
+
def_delegator :@rules, :each
|
35
|
+
|
36
|
+
# Initialize the state.
|
37
|
+
def initialize
|
38
|
+
@rules = Set.new
|
39
|
+
@transitions = {}
|
40
|
+
@id = format('%10x', object_id)
|
41
|
+
end
|
42
|
+
|
43
|
+
# Gives a nice string representation of the state.
|
44
|
+
#
|
45
|
+
# @return [String]
|
46
|
+
def inspect
|
47
|
+
"#<#{self.class} id=#{id} " \
|
48
|
+
"transitions=[#{transitions.keys.join(', ')}] " \
|
49
|
+
"rules=[{#{rules.to_a.join('} {')}}]>"
|
50
|
+
end
|
51
|
+
|
52
|
+
# Merges another state with this state. It copies all of the
|
53
|
+
# rules into this state, and then merges the transitions on
|
54
|
+
# the given state to this state. It then returns self.
|
55
|
+
#
|
56
|
+
# @raise [ArgumentError] if the given argument is not a state.
|
57
|
+
# @param other [State] the state to merge.
|
58
|
+
# @return [self]
|
59
|
+
def merge!(other)
|
60
|
+
raise ArgumentError, "Expected #{self.class}, " \
|
61
|
+
"got #{other.class}" unless other.is_a? State
|
62
|
+
|
63
|
+
self << other
|
64
|
+
transitions.merge! other.transitions
|
65
|
+
|
66
|
+
self
|
67
|
+
end
|
68
|
+
|
69
|
+
# Finds the rule that match the given production. It
|
70
|
+
# uses fuzzy equality checking. It returns the first rule
|
71
|
+
# that matches.
|
72
|
+
#
|
73
|
+
# @param production [Rule] the rule to compare.
|
74
|
+
# @return [Rule?]
|
75
|
+
def rule_for(production)
|
76
|
+
rules.find { |rule| production === rule }
|
77
|
+
end
|
78
|
+
|
79
|
+
# Appends the given object to this state. The given object
|
80
|
+
# must be a state or a rule. If it's a state, it appends all
|
81
|
+
# of the rules in the state to this state. If it's a rule, it
|
82
|
+
# adds the rule to our rules.
|
83
|
+
#
|
84
|
+
# @raise [ArgumentError] if the argument isn't a {State} or a
|
85
|
+
# {Rule}.
|
86
|
+
# @param rule [State, Rule] the object to append.
|
87
|
+
# @return [self]
|
88
|
+
def <<(rule)
|
89
|
+
case rule
|
90
|
+
when State
|
91
|
+
rule.rules.map(&:clone).each { |r| self << r }
|
92
|
+
when Rule
|
93
|
+
rules << rule
|
94
|
+
when Array, Set
|
95
|
+
rule.each do |part|
|
96
|
+
self << part
|
97
|
+
end
|
98
|
+
else
|
99
|
+
raise ArgumentError, "Expected State or Rule, " \
|
100
|
+
"got #{rule.class}"
|
101
|
+
end
|
102
|
+
|
103
|
+
self
|
104
|
+
end
|
105
|
+
|
106
|
+
alias_method :push, :<<
|
107
|
+
|
108
|
+
# Check to see if this state is fuzzily equivalent to another
|
109
|
+
# state. It does this by checking if the transitions are
|
110
|
+
# equivalent, and then that the rules are fuzzily equivalent.
|
111
|
+
# Ideally, the method is commutative; that is,
|
112
|
+
# `(a === b) == (b === a)`.
|
113
|
+
#
|
114
|
+
# @param other [State] the state to check.
|
115
|
+
# @return [Boolean]
|
116
|
+
# @see Rule#===
|
117
|
+
def ===(other)
|
118
|
+
return super unless other.is_a? State
|
119
|
+
|
120
|
+
other_rules = other.rules.to_a
|
121
|
+
other.transitions == transitions &&
|
122
|
+
rules.size == other_rules.size &&
|
123
|
+
rules.each_with_index
|
124
|
+
.all? { |rule, i| rule === other_rules[i] }
|
125
|
+
end
|
126
|
+
end
|
127
|
+
end
|
128
|
+
end
|
129
|
+
end
|
@@ -1,177 +1,177 @@
|
|
1
|
-
# encoding: utf-8
|
2
|
-
|
3
|
-
require 'antelope/generation/recognizer/rule'
|
4
|
-
require 'antelope/generation/recognizer/state'
|
5
|
-
|
6
|
-
module Antelope
|
7
|
-
module Generation
|
8
|
-
# Recognizes all of the states in the grammar.
|
9
|
-
#
|
10
|
-
# @see http://redjazz96.tumblr.com/post/88446352960
|
11
|
-
class Recognizer
|
12
|
-
# A list of all of the states in the grammar.
|
13
|
-
#
|
14
|
-
# @return [Set<State>]
|
15
|
-
attr_reader :states
|
16
|
-
|
17
|
-
# The initial state. This is the state that is constructed from
|
18
|
-
# the rule with the left-hand side being `$start`.
|
19
|
-
#
|
20
|
-
# @return [State]
|
21
|
-
attr_reader :start
|
22
|
-
|
23
|
-
# The grammar that the recognizer is running off of.
|
24
|
-
#
|
25
|
-
# @return [Grammar]
|
26
|
-
attr_reader :grammar
|
27
|
-
|
28
|
-
# Initialize the recognizer.
|
29
|
-
#
|
30
|
-
# @param grammar [Grammar]
|
31
|
-
def initialize(grammar)
|
32
|
-
@grammar = grammar
|
33
|
-
@states = Set.new
|
34
|
-
@map = {}
|
35
|
-
end
|
36
|
-
|
37
|
-
# Runs the recognizer. After all states have been created, it
|
38
|
-
# resets the state ids into a more friendly form (they were
|
39
|
-
# originally hexadecimal, see {State#initialize}), and then
|
40
|
-
# resets the rule ids in each state into a more friendly form
|
41
|
-
# (they were also originally hexadecmial, see {Rule#initialize}
|
42
|
-
# ).
|
43
|
-
#
|
44
|
-
# @see #compute_initial_state
|
45
|
-
# @return [void]
|
46
|
-
def call
|
47
|
-
@states = Set.new
|
48
|
-
@start = compute_initial_state
|
49
|
-
redefine_state_ids
|
50
|
-
redefine_rule_ids
|
51
|
-
grammar.states = states
|
52
|
-
end
|
53
|
-
|
54
|
-
# Computes the initial state. Starting with the default
|
55
|
-
# production of `$start`, it then generates the whole state
|
56
|
-
# and then the spawned states from it.
|
57
|
-
#
|
58
|
-
# @return [State]
|
59
|
-
def compute_initial_state
|
60
|
-
production = grammar.productions[:$start][0]
|
61
|
-
rule = Rule.new(production, 0)
|
62
|
-
compute_whole_state(rule)
|
63
|
-
end
|
64
|
-
|
65
|
-
# Computes the entire initial state from the initial rule.
|
66
|
-
# It starts with a blank state, adds the initial rule to it, and
|
67
|
-
# then generates the closure for that state; it then computes
|
68
|
-
# the rest of the states in the grammar.
|
69
|
-
#
|
70
|
-
# @param rule [Rule] the initial rule.
|
71
|
-
# @return [State]
|
72
|
-
def compute_whole_state(rule)
|
73
|
-
state = State.new
|
74
|
-
state << rule
|
75
|
-
compute_closure(state)
|
76
|
-
states << state
|
77
|
-
compute_states
|
78
|
-
state
|
79
|
-
end
|
80
|
-
|
81
|
-
# Computes all states. Uses a fix point iteration to determine
|
82
|
-
# when no states have been added. Loops through every state and
|
83
|
-
# every rule, looking for rules that have an active nonterminal
|
84
|
-
# and computing the closure for said rule.
|
85
|
-
#
|
86
|
-
# @return [void]
|
87
|
-
# @see #compute_closure
|
88
|
-
def compute_states
|
89
|
-
fixed_point(states) do
|
90
|
-
states.dup.each do |state|
|
91
|
-
compute_gotos(state)
|
92
|
-
end
|
93
|
-
end
|
94
|
-
end
|
95
|
-
|
96
|
-
# Given a state, it does a fixed point iteration on the rules of
|
97
|
-
# the state that have an active nonterminal, and add the
|
98
|
-
# corresponding production rules to the state.
|
99
|
-
#
|
100
|
-
# @return [void]
|
101
|
-
def compute_closure(state)
|
102
|
-
fixed_point(state.rules) do
|
103
|
-
state.rules.select { |r| r.active.nonterminal? }.each do |rule|
|
104
|
-
grammar.productions[rule.active.name].each do |prod|
|
105
|
-
state << Rule.new(prod, 0)
|
106
|
-
end
|
107
|
-
end
|
108
|
-
end
|
109
|
-
end
|
110
|
-
|
111
|
-
def compute_gotos(state)
|
112
|
-
actives = state.rules.map(&:active).select(&:name)
|
113
|
-
|
114
|
-
actives.each do |active|
|
115
|
-
next if state.transitions[active.name]
|
116
|
-
rules = state.rules
|
117
|
-
.select { |r| r.active == active && r.succ? }
|
118
|
-
.map(&:succ).to_set
|
119
|
-
s = states.find { |st| rules.subset? st.rules } || begin
|
120
|
-
s = State.new << rules
|
121
|
-
compute_closure(s)
|
122
|
-
states << s
|
123
|
-
s
|
124
|
-
end
|
125
|
-
|
126
|
-
state.transitions[active.name] = s
|
127
|
-
end
|
128
|
-
end
|
129
|
-
|
130
|
-
private
|
131
|
-
|
132
|
-
# Changes the IDs of the states into a more friendly format.
|
133
|
-
#
|
134
|
-
# @return [void]
|
135
|
-
def redefine_state_ids
|
136
|
-
states.each_with_index do |state, i|
|
137
|
-
state.id = i
|
138
|
-
end
|
139
|
-
end
|
140
|
-
|
141
|
-
# Redefines all of the rule ids to make them more friendly.
|
142
|
-
# Every rule in every state is given a unique ID, reguardless if
|
143
|
-
# the rules are equivalent.
|
144
|
-
#
|
145
|
-
# @return [void]
|
146
|
-
def redefine_rule_ids
|
147
|
-
start = 0
|
148
|
-
|
149
|
-
states.each do |state|
|
150
|
-
state.rules.each do |rule|
|
151
|
-
rule.id = start
|
152
|
-
start += 1
|
153
|
-
end
|
154
|
-
end
|
155
|
-
end
|
156
|
-
|
157
|
-
# Begins a fixed point iteration on the given enumerable. It
|
158
|
-
# initializes the added elements to one; then, while the number
|
159
|
-
# of added elements is not zero, it yields and checks for added
|
160
|
-
# elements.
|
161
|
-
#
|
162
|
-
# @param enum [Enumerable]
|
163
|
-
# @yield for every iteration. Guarenteed to do so at least
|
164
|
-
# once.
|
165
|
-
# @return [void]
|
166
|
-
def fixed_point(enum)
|
167
|
-
added = 1
|
168
|
-
|
169
|
-
until added.zero?
|
170
|
-
added = enum.size
|
171
|
-
yield
|
172
|
-
added = enum.size - added
|
173
|
-
end
|
174
|
-
end
|
175
|
-
end
|
176
|
-
end
|
177
|
-
end
|
1
|
+
# encoding: utf-8
|
2
|
+
|
3
|
+
require 'antelope/generation/recognizer/rule'
|
4
|
+
require 'antelope/generation/recognizer/state'
|
5
|
+
|
6
|
+
module Antelope
|
7
|
+
module Generation
|
8
|
+
# Recognizes all of the states in the grammar.
|
9
|
+
#
|
10
|
+
# @see http://redjazz96.tumblr.com/post/88446352960
|
11
|
+
class Recognizer
|
12
|
+
# A list of all of the states in the grammar.
|
13
|
+
#
|
14
|
+
# @return [Set<State>]
|
15
|
+
attr_reader :states
|
16
|
+
|
17
|
+
# The initial state. This is the state that is constructed from
|
18
|
+
# the rule with the left-hand side being `$start`.
|
19
|
+
#
|
20
|
+
# @return [State]
|
21
|
+
attr_reader :start
|
22
|
+
|
23
|
+
# The grammar that the recognizer is running off of.
|
24
|
+
#
|
25
|
+
# @return [Grammar]
|
26
|
+
attr_reader :grammar
|
27
|
+
|
28
|
+
# Initialize the recognizer.
|
29
|
+
#
|
30
|
+
# @param grammar [Grammar]
|
31
|
+
def initialize(grammar)
|
32
|
+
@grammar = grammar
|
33
|
+
@states = Set.new
|
34
|
+
@map = {}
|
35
|
+
end
|
36
|
+
|
37
|
+
# Runs the recognizer. After all states have been created, it
|
38
|
+
# resets the state ids into a more friendly form (they were
|
39
|
+
# originally hexadecimal, see {State#initialize}), and then
|
40
|
+
# resets the rule ids in each state into a more friendly form
|
41
|
+
# (they were also originally hexadecmial, see {Rule#initialize}
|
42
|
+
# ).
|
43
|
+
#
|
44
|
+
# @see #compute_initial_state
|
45
|
+
# @return [void]
|
46
|
+
def call
|
47
|
+
@states = Set.new
|
48
|
+
@start = compute_initial_state
|
49
|
+
redefine_state_ids
|
50
|
+
redefine_rule_ids
|
51
|
+
grammar.states = states
|
52
|
+
end
|
53
|
+
|
54
|
+
# Computes the initial state. Starting with the default
|
55
|
+
# production of `$start`, it then generates the whole state
|
56
|
+
# and then the spawned states from it.
|
57
|
+
#
|
58
|
+
# @return [State]
|
59
|
+
def compute_initial_state
|
60
|
+
production = grammar.productions[:$start][0]
|
61
|
+
rule = Rule.new(production, 0)
|
62
|
+
compute_whole_state(rule)
|
63
|
+
end
|
64
|
+
|
65
|
+
# Computes the entire initial state from the initial rule.
|
66
|
+
# It starts with a blank state, adds the initial rule to it, and
|
67
|
+
# then generates the closure for that state; it then computes
|
68
|
+
# the rest of the states in the grammar.
|
69
|
+
#
|
70
|
+
# @param rule [Rule] the initial rule.
|
71
|
+
# @return [State]
|
72
|
+
def compute_whole_state(rule)
|
73
|
+
state = State.new
|
74
|
+
state << rule
|
75
|
+
compute_closure(state)
|
76
|
+
states << state
|
77
|
+
compute_states
|
78
|
+
state
|
79
|
+
end
|
80
|
+
|
81
|
+
# Computes all states. Uses a fix point iteration to determine
|
82
|
+
# when no states have been added. Loops through every state and
|
83
|
+
# every rule, looking for rules that have an active nonterminal
|
84
|
+
# and computing the closure for said rule.
|
85
|
+
#
|
86
|
+
# @return [void]
|
87
|
+
# @see #compute_closure
|
88
|
+
def compute_states
|
89
|
+
fixed_point(states) do
|
90
|
+
states.dup.each do |state|
|
91
|
+
compute_gotos(state)
|
92
|
+
end
|
93
|
+
end
|
94
|
+
end
|
95
|
+
|
96
|
+
# Given a state, it does a fixed point iteration on the rules of
|
97
|
+
# the state that have an active nonterminal, and add the
|
98
|
+
# corresponding production rules to the state.
|
99
|
+
#
|
100
|
+
# @return [void]
|
101
|
+
def compute_closure(state)
|
102
|
+
fixed_point(state.rules) do
|
103
|
+
state.rules.select { |r| r.active.nonterminal? }.each do |rule|
|
104
|
+
grammar.productions[rule.active.name].each do |prod|
|
105
|
+
state << Rule.new(prod, 0)
|
106
|
+
end
|
107
|
+
end
|
108
|
+
end
|
109
|
+
end
|
110
|
+
|
111
|
+
def compute_gotos(state)
|
112
|
+
actives = state.rules.map(&:active).select(&:name)
|
113
|
+
|
114
|
+
actives.each do |active|
|
115
|
+
next if state.transitions[active.name]
|
116
|
+
rules = state.rules
|
117
|
+
.select { |r| r.active == active && r.succ? }
|
118
|
+
.map(&:succ).to_set
|
119
|
+
s = states.find { |st| rules.subset? st.rules } || begin
|
120
|
+
s = State.new << rules
|
121
|
+
compute_closure(s)
|
122
|
+
states << s
|
123
|
+
s
|
124
|
+
end
|
125
|
+
|
126
|
+
state.transitions[active.name] = s
|
127
|
+
end
|
128
|
+
end
|
129
|
+
|
130
|
+
private
|
131
|
+
|
132
|
+
# Changes the IDs of the states into a more friendly format.
|
133
|
+
#
|
134
|
+
# @return [void]
|
135
|
+
def redefine_state_ids
|
136
|
+
states.each_with_index do |state, i|
|
137
|
+
state.id = i
|
138
|
+
end
|
139
|
+
end
|
140
|
+
|
141
|
+
# Redefines all of the rule ids to make them more friendly.
|
142
|
+
# Every rule in every state is given a unique ID, reguardless if
|
143
|
+
# the rules are equivalent.
|
144
|
+
#
|
145
|
+
# @return [void]
|
146
|
+
def redefine_rule_ids
|
147
|
+
start = 0
|
148
|
+
|
149
|
+
states.each do |state|
|
150
|
+
state.rules.each do |rule|
|
151
|
+
rule.id = start
|
152
|
+
start += 1
|
153
|
+
end
|
154
|
+
end
|
155
|
+
end
|
156
|
+
|
157
|
+
# Begins a fixed point iteration on the given enumerable. It
|
158
|
+
# initializes the added elements to one; then, while the number
|
159
|
+
# of added elements is not zero, it yields and checks for added
|
160
|
+
# elements.
|
161
|
+
#
|
162
|
+
# @param enum [Enumerable]
|
163
|
+
# @yield for every iteration. Guarenteed to do so at least
|
164
|
+
# once.
|
165
|
+
# @return [void]
|
166
|
+
def fixed_point(enum)
|
167
|
+
added = 1
|
168
|
+
|
169
|
+
until added.zero?
|
170
|
+
added = enum.size
|
171
|
+
yield
|
172
|
+
added = enum.size - added
|
173
|
+
end
|
174
|
+
end
|
175
|
+
end
|
176
|
+
end
|
177
|
+
end
|