1
|
|
|
<?php |
2
|
|
|
/****************************************************************************** |
3
|
|
|
* An implementation of dicto (scg.unibe.ch/dicto) in and for PHP. |
4
|
|
|
* |
5
|
|
|
* Copyright (c) 2016 Richard Klees <[email protected]> |
6
|
|
|
* |
7
|
|
|
* This software is licensed under The MIT License. You should have received |
8
|
|
|
* a copy of the license along with the code. |
9
|
|
|
*/ |
10
|
|
|
|
11
|
|
|
namespace Lechimp\Dicto\Definition; |
12
|
|
|
|
13
|
|
|
/** |
14
|
|
|
* Parser for Rulesets. |
15
|
|
|
* |
16
|
|
|
* The grammar looks like this: |
17
|
|
|
* |
18
|
|
|
* explanation = '/ ** ... * /' |
19
|
|
|
* comment = '//..' | '/ * ... * /' |
20
|
|
|
* assignment = name "=" def |
21
|
|
|
* string = '"..."' |
22
|
|
|
* atom = [a-z ]... |
23
|
|
|
* name = [A-Z] [a-z]... |
24
|
|
|
* def = name | "{" def,... "}" | def property | def "except" def |
25
|
|
|
* property = atom ((name|string|atom)...)? |
26
|
|
|
* statement = def qualifier rule |
27
|
|
|
* rule = atom ((name|string|atom)...)? |
28
|
|
|
*/ |
29
|
|
|
class ASTParser extends Parser { |
30
|
|
|
const EXPLANATION_RE = "[/][*][*](([^*]|([*][^/]))*)[*][/]"; |
31
|
|
|
const SINGLE_LINE_COMMENT_RE = "[/][/]([^\n]*)"; |
32
|
|
|
const MULTI_LINE_COMMENT_RE = "[/][*](([^*]|([*][^/]))*)[*][/]"; |
33
|
|
|
const RULE_MODE_RE = "must|can(not)?"; |
34
|
|
|
const STRING_RE = "[\"]((([\\\\][\"])|[^\"])+)[\"]"; |
35
|
|
|
const NAME_RE = "[A-Z][A-Za-z_]*"; |
36
|
|
|
const ATOM_RE = "[a-z ]*"; |
37
|
|
|
const ASSIGNMENT_RE = "(".self::NAME_RE.")\s*=\s*"; |
38
|
|
|
const ATOM_HEAD_RE = "(".self::ATOM_RE.")\s*:\s*"; |
39
|
|
|
|
40
|
|
|
/** |
41
|
|
|
* @var AST\Factory |
42
|
|
|
*/ |
43
|
|
|
protected $ast_factory; |
44
|
|
|
|
45
|
|
|
public function __construct(AST\Factory $ast_factory) { |
46
|
|
|
$this->ast_factory = $ast_factory; |
47
|
|
|
parent::__construct(); |
48
|
|
|
} |
49
|
|
|
|
50
|
|
|
// Definition of symbols in the parser |
51
|
|
|
|
52
|
|
|
/** |
53
|
|
|
* @inheritdocs |
54
|
|
|
*/ |
55
|
|
|
protected function add_symbols_to(SymbolTable $table) { |
56
|
|
|
$this->add_symbols_for_comments($table); |
57
|
|
|
|
58
|
|
|
$this->add_symbols_for_variables_to($table); |
59
|
|
|
|
60
|
|
|
$this->add_symbols_for_rules_to($table); |
61
|
|
|
|
62
|
|
|
// Assignment |
63
|
|
|
$table->symbol(self::ASSIGNMENT_RE); |
64
|
|
|
|
65
|
|
|
// Strings |
66
|
|
|
$table->symbol(self::STRING_RE); |
67
|
|
|
|
68
|
|
|
// Names |
69
|
|
|
$table->literal(self::NAME_RE, function (array &$matches) { |
70
|
|
|
return $this->ast_factory->name($matches[0]); |
71
|
|
|
}); |
72
|
|
|
|
73
|
|
|
// Head of a property or rule |
74
|
|
|
$table->symbol(self::ATOM_HEAD_RE, 20) |
75
|
|
|
->left_denotation_is(function ($left, &$matches) { |
76
|
|
|
if (!($left instanceof AST\Definition)) { |
77
|
|
|
throw new ParserException |
78
|
|
|
("Expected a variable at the left of \"{$matches[0]}\"."); |
79
|
|
|
} |
80
|
|
|
$id = $this->ast_factory->atom($matches[1]); |
81
|
|
|
$arguments = $this->arguments(); |
82
|
|
|
return $this->ast_factory->property($left, $id, $arguments); |
83
|
|
|
}) |
84
|
|
|
->null_denotation_is(function (&$matches) { |
85
|
|
|
return $this->ast_factory->atom($matches[1]); |
86
|
|
|
}); |
87
|
|
|
|
88
|
|
|
|
89
|
|
|
|
90
|
|
|
// End of statement |
91
|
|
|
$table->symbol("\n"); |
92
|
|
|
} |
93
|
|
|
|
94
|
|
|
/** |
95
|
|
|
* @param SymbolTable |
96
|
|
|
* @return null |
97
|
|
|
*/ |
98
|
|
|
protected function add_symbols_for_comments(SymbolTable $table) { |
99
|
|
|
$table->symbol(self::EXPLANATION_RE); |
100
|
|
|
$table->symbol(self::SINGLE_LINE_COMMENT_RE); |
101
|
|
|
$table->symbol(self::MULTI_LINE_COMMENT_RE); |
102
|
|
|
} |
103
|
|
|
|
104
|
|
|
/** |
105
|
|
|
* @param SymbolTable |
106
|
|
|
* @return null |
107
|
|
|
*/ |
108
|
|
|
protected function add_symbols_for_variables_to(SymbolTable $table) { |
109
|
|
|
// Any |
110
|
|
|
$table->operator("{") |
111
|
|
|
->null_denotation_is(function() { |
112
|
|
|
$arr = array(); |
113
|
|
|
while(true) { |
114
|
|
|
$arr[] = $this->variable(0); |
115
|
|
|
if ($this->is_current_token_operator("}")) { |
116
|
|
|
$this->advance_operator("}"); |
117
|
|
|
return $this->ast_factory->any($arr); |
118
|
|
|
} |
119
|
|
|
$this->advance_operator(","); |
120
|
|
|
} |
121
|
|
|
}); |
122
|
|
|
$table->operator("}"); |
123
|
|
|
$table->operator(","); |
124
|
|
|
|
125
|
|
|
// Except |
126
|
|
|
$table->symbol("except", 10) |
127
|
|
|
->left_denotation_is(function($left) { |
128
|
|
|
if (!($left instanceof AST\Definition)) { |
129
|
|
|
throw new ParserException |
130
|
|
|
("Expected a variable at the left of except."); |
131
|
|
|
} |
132
|
|
|
$right = $this->variable(10); |
133
|
|
|
return $this->ast_factory->except($left, $right); |
134
|
|
|
}); |
135
|
|
|
} |
136
|
|
|
|
137
|
|
|
/** |
138
|
|
|
* @param SymbolTable |
139
|
|
|
* @return null |
140
|
|
|
*/ |
141
|
|
|
protected function add_symbols_for_rules_to(SymbolTable $table) { |
142
|
|
|
// Rules |
143
|
|
|
$table->symbol("only"); |
144
|
|
|
$table->symbol(self::RULE_MODE_RE, 0) |
145
|
|
|
->null_denotation_is(function (array &$matches) { |
146
|
|
|
if ($matches[0] == "can") { |
147
|
|
|
return $this->ast_factory->only_X_can(); |
148
|
|
|
} |
149
|
|
|
if ($matches[0] == "must") { |
150
|
|
|
return $this->ast_factory->must(); |
151
|
|
|
} |
152
|
|
|
if ($matches[0] == "cannot") { |
153
|
|
|
return $this->ast_factory->cannot(); |
154
|
|
|
} |
155
|
|
|
throw new \LogicException("Unexpected \"".$matches[0]."\"."); |
156
|
|
|
}); |
157
|
|
|
} |
158
|
|
|
|
159
|
|
|
// IMPLEMENTATION OF Parser |
160
|
|
|
|
161
|
|
|
/** |
162
|
|
|
* @return Ruleset |
163
|
|
|
*/ |
164
|
|
|
public function parse($source) { |
165
|
|
|
$this->variables = array(); |
|
|
|
|
166
|
|
|
$this->rules = array(); |
|
|
|
|
167
|
|
|
return parent::parse($source); |
168
|
|
|
} |
169
|
|
|
|
170
|
|
|
/** |
171
|
|
|
* Root expression for the parser is some whitespace or comment where a |
172
|
|
|
* top level statement is in the middle. |
173
|
|
|
* |
174
|
|
|
* @return Ruleset |
175
|
|
|
*/ |
176
|
|
|
protected function root() { |
177
|
|
|
$lines = []; |
178
|
|
|
while (true) { |
179
|
|
|
// drop empty lines |
180
|
|
|
while ($tok = $this->is_current_token_to_be_dropped()) { |
181
|
|
|
$this->advance($tok); |
182
|
|
|
} |
183
|
|
|
if ($this->is_end_of_file_reached()) { |
184
|
|
|
break; |
185
|
|
|
} |
186
|
|
|
|
187
|
|
|
$lines[] = $this->top_level_statement(); |
188
|
|
|
} |
189
|
|
|
return $this->ast_factory->root($lines); |
190
|
|
|
} |
191
|
|
|
|
192
|
|
|
/** |
193
|
|
|
* Parses the top level statements in the rules file. |
194
|
|
|
* |
195
|
|
|
* @return null |
196
|
|
|
*/ |
197
|
|
|
public function top_level_statement() { |
198
|
|
|
// A top level statements is either.. |
199
|
|
|
// ..an explanation |
200
|
|
|
if ($this->is_current_token_matched_by(self::EXPLANATION_RE)) { |
201
|
|
|
$m = $this->current_match(); |
202
|
|
|
$this->advance(self::EXPLANATION_RE); |
203
|
|
|
return $this->ast_factory->explanation($this->trim_explanation($m[1])); |
204
|
|
|
} |
205
|
|
|
// ..an assignment to a variable. |
206
|
|
|
elseif ($this->is_current_token_matched_by(self::ASSIGNMENT_RE)) { |
207
|
|
|
return $this->variable_assignment(); |
208
|
|
|
} |
209
|
|
|
// ..or a rule declaration |
210
|
|
|
else { |
211
|
|
|
return $this->rule_declaration(); |
212
|
|
|
} |
213
|
|
|
} |
214
|
|
|
|
215
|
|
|
/** |
216
|
|
|
* Returns currently matched whitespace or comment token if there is any. |
217
|
|
|
* |
218
|
|
|
* @return string|null |
219
|
|
|
*/ |
220
|
|
|
public function is_current_token_to_be_dropped() { |
221
|
|
|
if ($this->is_current_token_matched_by("\n")) { |
222
|
|
|
return "\n"; |
223
|
|
|
} |
224
|
|
|
if ($this->is_current_token_matched_by(self::SINGLE_LINE_COMMENT_RE)) { |
225
|
|
|
return self::SINGLE_LINE_COMMENT_RE; |
226
|
|
|
} |
227
|
|
|
if ($this->is_current_token_matched_by(self::MULTI_LINE_COMMENT_RE)) { |
228
|
|
|
return self::MULTI_LINE_COMMENT_RE; |
229
|
|
|
} |
230
|
|
|
return null; |
231
|
|
|
} |
232
|
|
|
|
233
|
|
|
/** |
234
|
|
|
* @param string |
235
|
|
|
* @return string |
236
|
|
|
*/ |
237
|
|
|
protected function trim_explanation($content) { |
238
|
|
|
return trim( |
239
|
|
|
preg_replace("%\s*\n\s*([*]\s*)?%s", "\n", $content) |
240
|
|
|
); |
241
|
|
|
} |
242
|
|
|
|
243
|
|
|
/** |
244
|
|
|
* Fetch a rule mode from the stream. |
245
|
|
|
* |
246
|
|
|
* @return mixed |
247
|
|
|
*/ |
248
|
|
|
protected function rule_mode() { |
249
|
|
|
$this->is_current_token_matched_by(self::RULE_MODE_RE); |
250
|
|
|
$t = $this->current_symbol(); |
251
|
|
|
$m = $this->current_match(); |
252
|
|
|
$this->fetch_next_token(); |
253
|
|
|
$mode = $t->null_denotation($m); |
254
|
|
|
return $mode; |
255
|
|
|
} |
256
|
|
|
|
257
|
|
|
/** |
258
|
|
|
* Fetch a string from the stream. |
259
|
|
|
* |
260
|
|
|
* @return string |
261
|
|
|
*/ |
262
|
|
|
protected function string() { |
263
|
|
|
$m = $this->current_match(); |
264
|
|
|
$this->fetch_next_token(); |
265
|
|
|
return $this->ast_factory->string_value |
266
|
|
|
(str_replace("\\\"", "\"", |
267
|
|
|
str_replace("\\n", "\n", |
268
|
|
|
$m[1]))); |
269
|
|
|
} |
270
|
|
|
|
271
|
|
|
/** |
272
|
|
|
* Fetch a variable from the stream. |
273
|
|
|
* |
274
|
|
|
* @return V\Variable |
275
|
|
|
*/ |
276
|
|
|
protected function variable($right_binding_power = 0) { |
277
|
|
|
$t = $this->current_symbol(); |
278
|
|
|
$m = $this->current_match(); |
279
|
|
|
$this->fetch_next_token(); |
280
|
|
|
$left = $t->null_denotation($m); |
281
|
|
|
|
282
|
|
|
while ($right_binding_power < $this->token[0]->binding_power()) { |
283
|
|
|
$t = $this->current_symbol(); |
284
|
|
|
$m = $this->current_match(); |
285
|
|
|
$this->fetch_next_token(); |
286
|
|
|
$left = $t->left_denotation($left, $m); |
287
|
|
|
} |
288
|
|
|
|
289
|
|
|
if (!($left instanceof AST\Definition)) { |
290
|
|
|
throw new ParserException("Expected variable."); |
291
|
|
|
} |
292
|
|
|
|
293
|
|
|
return $left; |
294
|
|
|
} |
295
|
|
|
|
296
|
|
|
/** |
297
|
|
|
* Fetch some arguments from the stream. |
298
|
|
|
* |
299
|
|
|
* @return array of atoms, variables or strings |
300
|
|
|
*/ |
301
|
|
|
protected function arguments() { |
302
|
|
|
$args = []; |
303
|
|
|
while (true) { |
304
|
|
|
// An argument is either |
305
|
|
|
// ..a string |
306
|
|
|
if ($this->is_current_token_matched_by(self::STRING_RE)) { |
307
|
|
|
$m = $this->current_match(); |
308
|
|
|
$this->fetch_next_token(); |
309
|
|
|
$args[] = $this->ast_factory->string_value |
310
|
|
|
(str_replace("\\\"", "\"", |
311
|
|
|
str_replace("\\n", "\n", |
312
|
|
|
$m[1]))); |
313
|
|
|
} |
314
|
|
|
// ..a variable |
315
|
|
|
// TODO: this won't do with {..} |
316
|
|
|
if ($this->is_current_token_matched_by(self::NAME_RE)) { |
317
|
|
|
$args[] = $this->variable(0); |
318
|
|
|
} |
319
|
|
|
else { |
320
|
|
|
break; |
321
|
|
|
} |
322
|
|
|
} |
323
|
|
|
return $args; |
324
|
|
|
} |
325
|
|
|
|
326
|
|
|
/** |
327
|
|
|
* Fetch a rule schema and its arguments from the stream. |
328
|
|
|
* |
329
|
|
|
* @return array (R\Schema, array) |
330
|
|
|
*/ |
331
|
|
|
protected function schema() { |
332
|
|
|
$t = $this->current_symbol(); |
333
|
|
|
$m = $this->current_match(); |
334
|
|
|
$this->fetch_next_token(); |
335
|
|
|
$id = $t->null_denotation($m); |
336
|
|
|
if (!($id instanceof AST\Atom)) { |
337
|
|
|
throw new ParserException("Expected name of a rule schema."); |
338
|
|
|
} |
339
|
|
|
return $id; |
340
|
|
|
} |
341
|
|
|
|
342
|
|
|
// TOP LEVEL STATEMENTS |
343
|
|
|
|
344
|
|
|
/** |
345
|
|
|
* Process a variable assignment. |
346
|
|
|
* |
347
|
|
|
* @return null |
348
|
|
|
*/ |
349
|
|
|
protected function variable_assignment() { |
350
|
|
|
$m = $this->current_match(); |
351
|
|
|
$name = $this->ast_factory->name($m[1]); |
352
|
|
|
$this->fetch_next_token(); |
353
|
|
|
$def = $this->variable(); |
354
|
|
|
return $this->ast_factory->assignment($name, $def); |
355
|
|
|
} |
356
|
|
|
|
357
|
|
|
/** |
358
|
|
|
* Process a rule declaration. |
359
|
|
|
* |
360
|
|
|
* @return null |
361
|
|
|
*/ |
362
|
|
|
protected function rule_declaration() { |
363
|
|
|
if ($this->is_current_token_matched_by("only")) { |
364
|
|
|
$this->advance("only"); |
365
|
|
|
} |
366
|
|
|
$var = $this->variable(); |
367
|
|
|
$mode = $this->rule_mode(); |
368
|
|
|
$schema = $this->schema(); |
369
|
|
|
$arguments = $this->arguments(); |
370
|
|
|
assert('is_array($arguments)'); |
371
|
|
|
return $this->ast_factory->rule |
372
|
|
|
( $mode |
373
|
|
|
, $this->ast_factory->property |
374
|
|
|
( $var |
375
|
|
|
, $schema |
376
|
|
|
, $arguments |
377
|
|
|
) |
378
|
|
|
); |
379
|
|
|
} |
380
|
|
|
} |
381
|
|
|
|
In PHP it is possible to write to properties without declaring them. For example, the following is perfectly valid PHP code:
Generally, it is a good practice to explictly declare properties to avoid accidental typos and provide IDE auto-completion: