1 $ // This file is part of NIT ( http://www.nitlanguage.org ).
3 $ // Copyright 2008 Jean Privat <jean@pryen.org>
4 $ // Based on algorithms developped for ( http://www.sablecc.org/ ).
6 $ // Licensed under the Apache License, Version 2.0 (the "License");
7 $ // you may not use this file except in compliance with the License.
8 $ // You may obtain a copy of the License at
10 $ // http://www.apache.org/licenses/LICENSE-2.0
12 $ // Unless required by applicable law or agreed to in writing, software
13 $ // distributed under the License is distributed on an "AS IS" BASIS,
14 $ // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15 $ // See the License for the specific language governing permissions and
16 $ // limitations under the License.
18 $ template make_parser()
20 # State of the parser automata as stored in the parser stack.
22 # The internal state number
23 readable writable var _state: Int
25 # The node stored with the state in the stack
26 readable writable var _nodes: nullable Object
28 init(state: Int, nodes: nullable Object)
40 # Stack of pushed states and productions
41 var _stack: Array[State]
43 # Position in the stack
46 # Create a new parser based on a given lexer
50 _stack = new Array[State]
57 # Do a transition in the automata
58 private fun go_to(index: Int): Int
61 var table = _goto_table[index]
63 var high = table.length/2 - 1
66 var middle = (low + high) / 2
67 var subindex = middle * 2
69 if state < table[subindex] then
71 else if state > table[subindex] then
74 return table[subindex + 1]
78 return table[1] # Default value
81 # Push someting in the state stack
82 private fun push(numstate: Int, list_node: nullable Object)
84 var pos = _stack_pos + 1
86 if pos < _stack.length then
87 var state = _stack[pos]
88 state.state = numstate
89 state.nodes = list_node
91 _stack.push(new State(numstate, list_node))
96 private fun state: Int
98 return _stack[_stack_pos].state
101 # Pop something from the stack state
102 private fun pop: nullable Object
104 var res = _stack[_stack_pos].nodes
105 _stack_pos = _stack_pos -1
109 # Build and return a full AST.
116 var token = lexer.peek
117 if token isa PError then
118 return new Start(null, token)
121 var index = token.parser_index
122 var table = _action_table[state]
123 var action_type = table[1]
124 var action_value = table[2]
127 var high = table.length/3 - 1
130 var middle = (low + high) / 2
131 var subindex = middle * 3
133 if index < table[subindex] then
135 else if index > table[subindex] then
138 action_type = table[subindex + 1]
139 action_value = table[subindex + 2]
140 high = low -1 # break
144 if action_type == 0 then # SHIFT
145 push(action_value, lexer.next)
146 else if action_type == 1 then # REDUCE
147 _reduce_table[action_value].action(self)
148 else if action_type == 2 then # ACCEPT
149 var node2 = lexer.next
152 assert node1 isa ${/parser/prods/prod/@ename}
153 var node = new Start(node1, node2)
154 (new ComputeProdLocationVisitor).enter_visit(node)
156 else if action_type == 3 then # ERROR
157 var node2 = new PError.init_error("Syntax error: unexpected token.", token.location)
158 var node = new Start(null, node2)
161 if false then break # FIXME remove once unreach loop exits are in c_src
163 abort # FIXME remove once unreach loop exits are in c_src
166 var _reduce_table: Array[ReduceAction]
167 private fun build_reduce_table
169 _reduce_table = new Array[ReduceAction].with_items(
170 $ foreach {rules/rule}
171 new ReduceAction@index(@leftside)[-sep ','-]
178 # Location on the first token after the start of a production
179 # So outside the production for epilon production
180 var _first_location: nullable Location
182 # Location of the last token before the end of a production
183 # So outside the production for epilon production
184 var _last_location: nullable Location
187 # Find location of production nodes
188 # Uses existing token locations to infer location of productions.
189 private class ComputeProdLocationVisitor
191 # Currenlty visited productions that need a first token
192 var _need_first_prods: Array[Prod] = new Array[Prod]
194 # Already visited epsilon productions that waits something after them
195 var _need_after_epsilons: Array[Prod] = new Array[Prod]
197 # Already visited epsilon production that waits something before them
198 var _need_before_epsilons: Array[Prod] = new Array[Prod]
200 # Location of the last visited token in the current production
201 var _last_location: nullable Location = null
203 redef fun visit(n: nullable PNode)
207 else if n isa Token then
211 # Add a first token to productions that need one
212 for no in _need_first_prods do
213 no._first_location = loc
215 _need_first_prods.clear
217 # Find location for already visited epsilon production that need one
218 for no in _need_after_epsilons do
219 # Epsilon production that is in the middle of a non-epsilon production
220 # The epsilon production has both a token before and after it
222 var startl = no._last_location
223 no.location = new Location(endl.file, startl.line_end, endl.line_start, startl.column_end, endl.column_start)
225 _need_after_epsilons.clear
228 _need_first_prods.add(n)
230 var old_last = _last_location
231 _last_location = null
233 var endl = _last_location
234 if endl == null then _last_location = old_last
236 n._last_location = endl
237 var startl = n._first_location
238 if startl != null then
239 # Non-epsilon production
242 n.location = new Location(startl.file, startl.line_start, endl.line_end, startl.column_start, endl.column_end)
244 for no in _need_before_epsilons do
245 # Epsilon production that starts the current non-epsilon production
246 #var startl = n.location
247 no.location = new Location(startl.file, startl.line_start, startl.line_start, startl.column_start, startl.column_start)
249 _need_before_epsilons.clear
251 for no in _need_after_epsilons do
252 # Epsilon production that finishes the current non-epsilon production
253 #var endl = n.location
254 no.location = new Location(endl.file, endl.line_end, endl.line_end, endl.column_end, endl.column_end)
256 _need_after_epsilons.clear
258 # No first token means epsilon production (or "throw all my tokens" production)
259 # So, it must be located it later
261 # Epsilon production that starts a parent non-epsilon production
262 _need_before_epsilons.add(n)
264 # Epsilon production in the middle or that finishes a parent non-epsilon production
265 _need_after_epsilons.add(n)
274 # Each reduca action has its own class, this one is the root of the hierarchy.
275 private abstract class ReduceAction
276 fun action(p: Parser) is abstract
277 fun concat(l1, l2 : Array[Object]): Array[Object]
279 if l1.is_empty then return l2
285 $ foreach {rules/rule}
286 private class ReduceAction@index
288 redef fun action(p: Parser)
290 var node_list: nullable Object = null
294 var ${translate(@result,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")} = p.pop
296 $ when {@cmd='FETCHLIST'}
297 var ${translate(@result,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")} = ${translate(@from,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")}
298 assert ${translate(@result,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")} isa Array[Object]
300 $ when {@cmd='FETCHNODE'}
301 var ${translate(@result,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")} = ${translate(@from,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")}
302 assert ${translate(@result,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")} isa nullable @etype
304 $ when {@cmd='ADDNODE'}
305 if ${translate(@node,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")} != null then
306 ${translate(@tolist,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")}.add(${translate(@node,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")})
309 $ when {@cmd='ADDLIST'}
310 ${translate(@tolist,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")} = concat(${translate(@tolist,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")}, ${translate(@fromlist,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")})
312 $ when {@cmd='MAKELIST'}
313 var ${translate(@result,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")} = new Array[Object]
315 $ when {@cmd='MAKENODE'}
316 var ${translate(@result,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")}: nullable @etype = new @etype.init_${translate(@etype,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")}(
321 ${translate(.,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")}[-sep ','-]
326 $ when {@cmd='RETURNNODE'}
330 node_list = ${translate(@node,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")}
333 $ when {@cmd='RETURNLIST'}
334 node_list = ${translate(@list,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")}
338 p.push(p.go_to(_goto), node_list)
341 init(g: Int) do _goto = g
346 $ template make_parser_tables()
347 # Parser that build a full AST
348 abstract class ParserTable
349 var _action_table: Array[Array[Int]]
350 private fun build_action_table
352 _action_table = once [
353 $ foreach {parser_data/action_table/row}
354 action_table_row${position()}[-sep ','-]
359 $ foreach {parser_data/action_table/row}
360 private fun action_table_row${position()}: Array[Int]
364 @from, @action, @to[-sep ','-]
370 var _goto_table: Array[Array[Int]]
371 private fun build_goto_table
374 $ foreach {parser_data/goto_table/row}
377 @from, @to[-sep ','-]