Notice to GitKraken users: A vulnerability has been found in the SSH key generation of GitKraken versions 7.6.0 to 8.0.0 (https://www.gitkraken.com/blog/weak-ssh-key-fix). If you use GitKraken and have generated a SSH key using one of these versions, please remove it both from your local workstation and from your LRZ GitLab profile.

21.10.2021, 9:00 - 11:00: Due to updates GitLab may be unavailable for some minutes between 09:00 and 11:00.

ebnf.py 24.9 KB
Newer Older
1
2
#!/usr/bin/python3

3
"""ebnf.py - EBNF -> Python-Parser compilation for DHParser
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

Copyright 2016  by Eckhart Arnold (arnold@badw.de)
                Bavarian Academy of Sciences an Humanities (badw.de)

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied.  See the License for the specific language governing
permissions and limitations under the License.
"""

from functools import partial
22
23
import keyword
import os
24
25
26
27
28
try:
    import regex as re
except ImportError:
    import re

29
30
from .__init__ import __version__
from .toolkit import load_if_file, escape_re, md5, sane_parser_name
31
from .parsers import GrammarBase, mixin_comment, Forward, RE, NegativeLookahead, \
32
33
    Alternative, Sequence, Optional, Required, OneOrMore, ZeroOrMore, Token, CompilerBase, \
    Capture, Retrieve
34
from .syntaxtree import Node, traverse, remove_enclosing_delimiters, reduce_single_child, \
di68kap's avatar
di68kap committed
35
    replace_by_single_child, TOKEN_KEYWORD, remove_expendables, remove_tokens, flatten, \
36
    forbid, assert_content, WHITESPACE_KEYWORD
37
38


Eckhart Arnold's avatar
Eckhart Arnold committed
39
40
41
42
43
__all__ = ['EBNFGrammar',
           'EBNFTransTable',
           'EBNFCompilerError',
           # 'Scanner',
           'EBNFCompiler']
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64


class EBNFGrammar(GrammarBase):
    r"""Parser for an EBNF source file, with this grammar:

    # EBNF-Grammar in EBNF

    @ comment    =  /#.*(?:\n|$)/                    # comments start with '#' and eat all chars up to and including '\n'
    @ whitespace =  /\s*/                            # whitespace includes linefeed
    @ literalws  =  right                            # trailing whitespace of literals will be ignored tacitly

    syntax     =  [~//] { definition | directive } §EOF
    definition =  symbol §"=" expression
    directive  =  "@" §symbol §"=" ( regexp | literal | list_ )

    expression =  term { "|" term }
    term       =  { factor }+
    factor     =  [flowmarker] [retrieveop] symbol !"="   # negative lookahead to be sure it's not a definition
                | [flowmarker] literal
                | [flowmarker] regexp
                | [flowmarker] group
65
                | [flowmarker] regexchain
66
67
68
69
70
71
72
73
74
                | [flowmarker] oneormore
                | repetition
                | option

    flowmarker =  "!"  | "&"  | "§" |                # '!' negative lookahead, '&' positive lookahead, '§' required
                  "-!" | "-&"                        # '-' negative lookbehind, '-&' positive lookbehind
    retrieveop =  "::" | ":"                         # '::' pop, ':' retrieve

    group      =  "(" expression §")"
75
    regexchain =  "<" expression §">"                # compiles "expression" into a singular regular expression
76
77
    oneormore  =  "{" expression "}+"
    repetition =  "{" expression §"}"
78
79
80
    option     =  "[" expression §"]"

    link       = regexp | symbol | literal           # semantic restriction: symbol must evaluate to a regexp or chain
81
82
83
84
85
86
87

    symbol     =  /(?!\d)\w+/~                       # e.g. expression, factor, parameter_list
    literal    =  /"(?:[^"]|\\")*?"/~                # e.g. "(", '+', 'while'
                | /'(?:[^']|\\')*?'/~                # whitespace following literals will be ignored tacitly.
    regexp     =  /~?\/(?:[^\/]|(?<=\\)\/)*\/~?/~    # e.g. /\w+/, ~/#.*(?:\n|$)/~
                                                     # '~' is a whitespace-marker, if present leading or trailing
                                                     # whitespace of a regular expression will be ignored tacitly.
88
    list_      =  /\w+/~ { "," /\w+/~ }              # comma separated list of symbols, e.g. BEGIN_LIST, END_LIST,
89
90
91
92
                                                     # BEGIN_QUOTE, END_QUOTE ; see CommonMark/markdown.py for an exmaple
    EOF =  !/./
    """
    expression = Forward()
93
    source_hash__ = "a410e1727fb7575e98ff8451dbf8f3bd"
94
    parser_initialization__ = "upon instatiation"
95
96
    COMMENT__ = r'#.*(?:\n|$)'
    WSP__ = mixin_comment(whitespace=r'\s*', comment=r'#.*(?:\n|$)')
97
    wspL__ = ''
98
    wspR__ = WSP__
99
    EOF = NegativeLookahead(RE('.', wR=''))
100
    list_ = Sequence(RE('\\w+'), ZeroOrMore(Sequence(Token(","), RE('\\w+'))))
101
102
103
    regexp = RE('~?/(?:[^/]|(?<=\\\\)/)*/~?')
    literal = Alternative(RE('"(?:[^"]|\\\\")*?"'), RE("'(?:[^']|\\\\')*?'"))
    symbol = RE('(?!\\d)\\w+')
104
105
    link = Alternative(regexp, symbol, literal)
    option = Sequence(Token("["), expression, Required(Token("]")))
106
107
    repetition = Sequence(Token("{"), expression, Required(Token("}")))
    oneormore = Sequence(Token("{"), expression, Token("}+"))
108
    regexchain = Sequence(Token("<"), expression, Required(Token(">")))
109
110
111
112
113
    group = Sequence(Token("("), expression, Required(Token(")")))
    retrieveop = Alternative(Token("::"), Token(":"))
    flowmarker = Alternative(Token("!"), Token("&"), Token("§"), Token("-!"), Token("-&"))
    factor = Alternative(Sequence(Optional(flowmarker), Optional(retrieveop), symbol, NegativeLookahead(Token("="))),
                         Sequence(Optional(flowmarker), literal), Sequence(Optional(flowmarker), regexp),
114
115
                         Sequence(Optional(flowmarker), group), Sequence(Optional(flowmarker), regexchain),
                         Sequence(Optional(flowmarker), oneormore), repetition, option)
116
117
118
119
    term = OneOrMore(factor)
    expression.set(Sequence(term, ZeroOrMore(Sequence(Token("|"), term))))
    directive = Sequence(Token("@"), Required(symbol), Required(Token("=")), Alternative(regexp, literal, list_))
    definition = Sequence(symbol, Required(Token("=")), expression)
120
    syntax = Sequence(Optional(RE('', wR='', wL=WSP__)), ZeroOrMore(Alternative(definition, directive)), Required(EOF))
121
122
123
    root__ = syntax


124
125
#TODO: Add Capture and Retrieve Validation: A variable mustn't be captured twice before retrival?!?

126
EBNF_transformation_table = {
127
128
129
130
131
132
133
134
135
136
137
138
139
    # AST Transformations for EBNF-grammar
    "syntax":
        remove_expendables,
    "directive, definition":
        partial(remove_tokens, tokens={'@', '='}),
    "expression, chain":
        [replace_by_single_child, flatten,
         partial(remove_tokens, tokens={'|', '--'})],
    "term":
        [replace_by_single_child, flatten],  # supports both idioms:  "{ factor }+" and "factor { factor }"
    "factor, flowmarker, retrieveop":
        replace_by_single_child,
    "group":
140
141
142
        [remove_enclosing_delimiters, replace_by_single_child],
    "oneormore, repetition, option, regexchain":
        [reduce_single_child, remove_enclosing_delimiters],
143
    "symbol, literal, regexp":
144
145
146
        [remove_expendables, reduce_single_child],
    (TOKEN_KEYWORD, WHITESPACE_KEYWORD):
        [remove_expendables, reduce_single_child],
147
148
    "list_":
        [partial(remove_tokens, tokens={','})],
149
150
151
152
    "":
        [remove_expendables, replace_by_single_child]
}

153

154
EBNF_validation_table = {
155
156
157
158
159
    # Semantic validation on the AST
    "repetition, option, oneormore":
        [partial(forbid, child_tags=['repetition', 'option', 'oneormore']),
         partial(assert_content, regex=r'(?!§)')],
}
160

161

162
163
164
def EBNFTransform(syntax_tree):
    for processing_table in [EBNF_transformation_table, EBNF_validation_table]:
        traverse(syntax_tree, processing_table)
di68kap's avatar
di68kap committed
165
166


167
168
169
170
171
172
173
174
175
176
177
178
179
180
class EBNFCompilerError(Exception):
    """Error raised by `EBNFCompiler` class. (Not compilation errors
    in the strict sense, see `CompilationError` below)"""
    pass


class EBNFCompiler(CompilerBase):
    """Generates a Parser from an abstract syntax tree of a grammar specified
    in EBNF-Notation.
    """
    COMMENT_KEYWORD = "COMMENT__"
    RESERVED_SYMBOLS = {TOKEN_KEYWORD, WHITESPACE_KEYWORD, COMMENT_KEYWORD}
    AST_ERROR = "Badly structured syntax tree. " \
                "Potentially due to erroneuos AST transformation."
181
182
183
184
    PREFIX_TABLE = {'§': 'Required',
                    '&': 'Lookahead', '!': 'NegativeLookahead',
                    '-&': 'Lookbehind', '-!': 'NegativeLookbehind',
                    '::': 'Pop', ':': 'Retrieve'}
185
186
187
    WHITESPACE = {'horizontal': r'[\t ]*',  # default: horizontal
                  'linefeed': r'[ \t]*\n?(?!\s*\n)[ \t]*',
                  'vertical': r'\s*'}
188

189
    def __init__(self, grammar_name="", grammar_source=""):
190
        super(EBNFCompiler, self).__init__()
191
        self.set_grammar_name(grammar_name, grammar_source)
192
193
194
195
196
197
198
199
200
        self._reset()

    def _reset(self):
        self.rules = set()
        self.symbols = set()
        self.variables = set()
        self.definition_names = []
        self.recursive = set()
        self.root = ""
201
        self.directives = {'whitespace': self.WHITESPACE['horizontal'],
202
                           'comment': '',
203
                           'literalws': ['right'],
204
                           'tokens': set(),     # alt. 'scanner_tokens'
205
                           'counterpart': set()}     # alt. 'retrieve_counterpart'
206

207
208
209
210
211
212
213
    def set_grammar_name(self, grammar_name, grammar_source):
        assert grammar_name == "" or re.match('\w+\Z', grammar_name)
        if not grammar_name and re.fullmatch(r'[\w/:\\]+', grammar_source):
            grammar_name = os.path.splitext(os.path.basename(grammar_source))[0]
        self.grammar_name = grammar_name
        self.grammar_source = load_if_file(grammar_source)

214
215
216
217
218
219
220
221
    def gen_scanner_skeleton(self):
        name = self.grammar_name + "Scanner"
        return "def %s(text):\n    return text\n" % name

    def gen_AST_skeleton(self):
        if not self.definition_names:
            raise EBNFCompilerError('Compiler has not been run before calling '
                                    '"gen_AST_Skeleton()"!')
222
223
        tt_name = self.grammar_name + '_AST_transformation_table'
        tf_name = self.grammar_name + 'Transform'
di68kap's avatar
di68kap committed
224
        transtable = [tt_name + ' = {',
225
226
227
                      '    # AST Transformations for the ' +
                      self.grammar_name + '-grammar']
        for name in self.definition_names:
di68kap's avatar
di68kap committed
228
            transtable.append('    "' + name + '": no_operation,')
229
230
        transtable += ['    "": no_operation', '}', '',  tf_name +
                       ' = partial(traverse, processing_table=%s)' % tt_name, '']
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
        return '\n'.join(transtable)

    def gen_compiler_skeleton(self):
        if not self.definition_names:
            raise EBNFCompilerError('Compiler has not been run before calling '
                                    '"gen_Compiler_Skeleton()"!')
        compiler = ['class ' + self.grammar_name + 'Compiler(CompilerBase):',
                    '    """Compiler for the abstract-syntax-tree of a ' +
                    self.grammar_name + ' source file.',
                    '    """', '',
                    '    def __init__(self, grammar_name="' +
                    self.grammar_name + '"):',
                    '        super(' + self.grammar_name +
                    'Compiler, self).__init__()',
                    "        assert re.match('\w+\Z', grammar_name)", '']
        for name in self.definition_names:
            if name == self.root:
248
                compiler += ['    def ' + name + '__(self, node):',
249
250
                             '        return node', '']
            else:
251
                compiler += ['    def ' + name + '__(self, node):',
252
                             '        pass', '']
253
        return '\n'.join(compiler)
254
255
256
257
258
259

    def gen_parser(self, definitions):
        # fix capture of variables that have been defined before usage [sic!]
        if self.variables:
            for i in range(len(definitions)):
                if definitions[i][0] in self.variables:
260
                    definitions[i] = (definitions[i][0], 'Capture(%s)' % definitions[1])
261
262

        self.definition_names = [defn[0] for defn in definitions]
Eckhart Arnold's avatar
Eckhart Arnold committed
263
264
265
266
        definitions.append(('wspR__', WHITESPACE_KEYWORD
                            if 'right' in self.directives['literalws'] else "''"))
        definitions.append(('wspL__', WHITESPACE_KEYWORD
                            if 'left' in self.directives['literalws'] else "''"))
267
268
269
270
271
272
273
274
        definitions.append((WHITESPACE_KEYWORD,
                            ("mixin_comment(whitespace="
                             "r'{whitespace}', comment=r'{comment}')").
                            format(**self.directives)))
        definitions.append((self.COMMENT_KEYWORD, "r'{comment}'".format(**self.directives)))

        # prepare parser class header and docstring and
        # add EBNF grammar to the doc string of the parser class
275
        article = 'an ' if self.grammar_name[0:1] in "AaEeIiOoUu" else 'a '  # what about 'hour', 'universe' etc.?
276
277
278
279
        declarations = ['class ' + self.grammar_name +
                        'Grammar(GrammarBase):',
                        'r"""Parser for ' + article + self.grammar_name +
                        ' source file' +
280
                        (', with this grammar:' if self.grammar_source else '.')]
281
        definitions.append(('parser_initialization__', '"upon instatiation"'))
282
        if self.grammar_source:
283
            definitions.append(('source_hash__',
284
                                '"%s"' % md5(self.grammar_source, __version__)))
285
            declarations.append('')
286
            declarations += [line for line in self.grammar_source.split('\n')]
287
288
289
290
            while declarations[-1].strip() == '':
                declarations = declarations[:-1]
        declarations.append('"""')

291
292
293
294
295
296
297
        # add default functions for counterpart filters of pop or retrieve operators
        for symbol in self.directives['counterpart']:
            # declarations.append('def %s_counterpart(value): \n' % symbol +
            #                     '        return value.replace("(", ")").replace("[", "]")'
            #                     '.replace("{", "}").replace(">", "<")\n')
            declarations.append(symbol + '_counterpart = lambda value: value.replace("(", ")")'
                                '.replace("[", "]").replace("{", "}").replace(">", "<")')
298

299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
        # turn definitions into declarations in reverse order
        self.root = definitions[0][0] if definitions else ""
        definitions.reverse()
        declarations += [symbol + ' = Forward()'
                         for symbol in sorted(list(self.recursive))]
        for symbol, statement in definitions:
            if symbol in self.recursive:
                declarations += [symbol + '.set(' + statement + ')']
            else:
                declarations += [symbol + ' = ' + statement]
        for nd in self.symbols:
            if nd.result not in self.rules:
                nd.add_error("Missing production for symbol '%s'" % nd.result)
        if self.root and 'root__' not in self.symbols:
            declarations.append('root__ = ' + self.root)
        declarations.append('')
        return '\n    '.join(declarations)

317
    def syntax__(self, node):
318
319
320
321
322
323
324
325
326
327
328
        self._reset()
        definitions = []

        # drop the wrapping sequence node
        if isinstance(node.parser, Sequence) and \
                isinstance(node.result[0].parser, ZeroOrMore):
            node = node.result[0]

        # compile definitions and directives and collect definitions
        for nd in node.result:
            if nd.parser.name == "definition":
329
                definitions.append(self._compile(nd))
330
331
            else:
                assert nd.parser.name == "directive", nd.as_sexpr()
332
                self._compile(nd)
333
334
335

        return self.gen_parser(definitions)

336
    def definition__(self, node):
337
        rule = node.result[0].result
338
339
340
        if rule in self.rules:
            node.add_error('A rule with name "%s" has already been defined.' % rule)
        elif rule in EBNFCompiler.RESERVED_SYMBOLS:
341
342
343
344
            node.add_error('Symbol "%s" is a reserved symbol.' % rule)
        elif not sane_parser_name(rule):
            node.add_error('Illegal symbol "%s". Symbols must not start or '
                           ' end with a doube underscore "__".' % rule)
345
        elif rule in self.directives['tokens']:
346
347
348
349
350
351
352
            node.add_error('Symbol "%s" has already been defined as '
                           'a scanner token.' % rule)
        elif keyword.iskeyword(rule):
            node.add_error('Python keyword "%s" may not be used as a symbol. '
                           % rule + '(This may change in the furute.)')
        try:
            self.rules.add(rule)
353
            defn = self._compile(node.result[1])
354
            if rule in self.variables:
355
                defn = 'Capture(%s)' % defn
356
357
358
359
360
                self.variables.remove(rule)
        except TypeError as error:
            errmsg = EBNFCompiler.AST_ERROR + " (" + str(error) + ")\n" + node.as_sexpr()
            node.add_error(errmsg)
            rule, defn = rule + ':error', '"' + errmsg + '"'
Eckhart Arnold's avatar
Eckhart Arnold committed
361
        return rule, defn
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376

    @staticmethod
    def _check_rx(node, rx):
        """Checks whether the string `rx` represents a valid regular
        expression. Makes sure that multiline regular expressions are
        prepended by the multiline-flag. Returns the regular expression string.
        """
        rx = rx if rx.find('\n') < 0 or rx[0:4] == '(?x)' else '(?x)' + rx
        try:
            re.compile(rx)
        except Exception as re_error:
            node.add_error("malformed regular expression %s: %s" %
                           (repr(rx), str(re_error)))
        return rx

377
    def directive__(self, node):
378
        key = node.result[0].result.lower()
379
        assert key not in self.directives['tokens']
380
        if key in {'comment', 'whitespace'}:
381
382
            if node.result[1].parser.name == "list_":
                if len(node.result[1].result) != 1:
Eckhart Arnold's avatar
Eckhart Arnold committed
383
                    node.add_error('Directive "%s" must have one, but not %i values.' %
384
                                   (key, len(node.result[1])))
385
                value = self._compile(node.result[1]).pop()
386
387
                if key == 'whitespace' and value in EBNFCompiler.WHITESPACE:
                    value = EBNFCompiler.WHITESPACE[value]  # replace whitespace-name by regex
388
                else:
389
                    node.add_error('Value "%s" not allowed for directive "%s".' % (value, key))
390
            else:
391
392
393
394
395
396
397
398
                value = node.result[1].result.strip("~")
                if value != node.result[1].result:
                    node.add_error("Whitespace marker '~' not allowed in definition of "
                                   "%s regular expression." % key)
                if value[0] + value[-1] in {'""', "''"}:
                    value = escape_re(value[1:-1])
                elif value[0] + value[-1] == '//':
                    value = self._check_rx(node, value[1:-1])
399
400
401
                if key == 'whitespace' and not re.match(value, ''):
                    node.add_error("Implicit whitespace should always match the empty string, "
                                   "/%s/ does not." % value)
402
            self.directives[key] = value
403

404
        elif key == 'literalws':
405
            value = {item.lower() for item in self._compile(node.result[1])}
406
            if (len(value - {'left', 'right', 'both', 'none'}) > 0
Eckhart Arnold's avatar
Eckhart Arnold committed
407
                    or ('none' in value and len(value) > 1)):
408
409
410
411
412
413
414
                node.add_error('Directive "literalws" allows the values '
                               '`left`, `right`, `both` or `none`, '
                               'but not `%s`' % ", ".join(value))
            ws = {'left', 'right'} if 'both' in value \
                else {} if 'none' in value else value
            self.directives[key] = list(ws)

415
        elif key in {'tokens', 'scanner_tokens'}:
416
            self.directives['tokens'] |= self._compile(node.result[1])
417

418
        elif key in {'counterpart', 'retrieve_counterpart'}:
419
            self.directives['counterpart'] |= self._compile(node.result[1])
420

421
422
423
        else:
            node.add_error('Unknown directive %s ! (Known ones are %s .)' %
                           (key,
424
                            ', '.join(list(self.directives.keys()))))
425
426
        return ""

427
    def non_terminal(self, node, parser_class, custom_args=[]):
428
429
430
        """Compiles any non-terminal, where `parser_class` indicates the Parser class
        name for the particular non-terminal.
        """
431
        arguments = [self._compile(r) for r in node.result] + custom_args
432
433
        return parser_class + '(' + ', '.join(arguments) + ')'

434
    def expression__(self, node):
435
436
        return self.non_terminal(node, 'Alternative')

437
    def term__(self, node):
438
439
        return self.non_terminal(node, 'Sequence')

440
    def factor__(self, node):
441
442
443
444
        assert isinstance(node.parser, Sequence), node.as_sexpr()  # these assert statements can be removed
        assert node.children
        assert len(node.result) >= 2, node.as_sexpr()
        prefix = node.result[0].result
445
        custom_args = []
446
447
448

        if prefix in {'::', ':'}:
            assert len(node.result) == 2
449
450
            arg = node.result[-1]
            argstr = str(arg)
451
            if arg.parser.name != 'symbol':
Eckhart Arnold's avatar
Eckhart Arnold committed
452
                node.add_error(('Retrieve Operator "%s" requires a symbol, '
453
454
                                'and not a %s.') % (prefix, str(arg.parser)))
                return str(arg.result)
455
456
            if str(arg) in self.directives['counterpart']:
                custom_args = ['counterpart=%s_counterpart' % str(arg)]
457
458
            self.variables.add(arg.result)

459
        elif len(node.result) > 2:
460
461
            # shift = (Node(node.parser, node.result[1].result),)
            # node.result[1].result = shift + node.result[2:]
462
            node.result[1].result = (Node(node.result[1].parser, node.result[1].result),) \
463
464
465
466
467
                                    + node.result[2:]
            node.result[1].parser = node.parser
            node.result = (node.result[0], node.result[1])

        node.result = node.result[1:]
468
469
470
471
472
        try:
            parser_class = self.PREFIX_TABLE[prefix]
            return self.non_terminal(node, parser_class, custom_args)
        except KeyError:
            node.add_error('Unknown prefix "%s".' % prefix)
473

474
    def option__(self, node):
475
476
        return self.non_terminal(node, 'Optional')

477
    def repetition__(self, node):
478
479
        return self.non_terminal(node, 'ZeroOrMore')

480
    def oneormore__(self, node):
481
482
        return self.non_terminal(node, 'OneOrMore')

483
    def regexchain__(self, node):
484
485
        raise EBNFCompilerError("Not yet implemented!")

486
    def group__(self, node):
487
488
489
        raise EBNFCompilerError("Group nodes should have been eliminated by "
                                "AST transformation!")

490
    def symbol__(self, node):
491
        if node.result in self.directives['tokens']:
492
493
494
495
496
497
498
            return 'ScannerToken("' + node.result + '")'
        else:
            self.symbols.add(node)
            if node.result in self.rules:
                self.recursive.add(node.result)
            return node.result

499
    def literal__(self, node):
500
        return 'Token(' + node.result.replace('\\', r'\\') + ')'  # return 'Token(' + ', '.join([node.result]) + ')' ?
501

502
    def regexp__(self, node):
503
504
505
506
507
508
509
510
511
        rx = node.result
        name = []
        if rx[:2] == '~/':
            if not 'left' in self.directives['literalws']:
                name = ['wL=' + WHITESPACE_KEYWORD] + name
            rx = rx[1:]
        elif 'left' in self.directives['literalws']:
            name = ["wL=''"] + name
        if rx[-2:] == '/~':
Eckhart Arnold's avatar
Eckhart Arnold committed
512
            if 'right' not in self.directives['literalws']:
513
514
515
516
517
518
519
520
521
522
523
524
525
                name = ['wR=' + WHITESPACE_KEYWORD] + name
            rx = rx[:-1]
        elif 'right' in self.directives['literalws']:
            name = ["wR=''"] + name
        try:
            arg = repr(self._check_rx(node, rx[1:-1].replace(r'\/', '/')))
        except AttributeError as error:
            errmsg = EBNFCompiler.AST_ERROR + " (" + str(error) + ")\n" + \
                     node.as_sexpr()
            node.add_error(errmsg)
            return '"' + errmsg + '"'
        return 'RE(' + ', '.join([arg] + name) + ')'

526
    def list___(self, node):
527
528
        assert node.children
        return set(item.result.strip() for item in node.result)
529
530


531
532
533
def grammar_changed(grammar_class, grammar_source):
    """Returns ``True`` if ``grammar_class`` does not reflect the latest
    changes of ``grammar_source``
534
535
536
537

    Parameters:
        grammar_class:  the parser class representing the grammar
            or the file name of a compiler suite containing the grammar
538
539
540
        grammar_source:  File name or string representation of the
            EBNF code of the grammar
            
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
    Returns (bool):
        True, if the source text of the grammar is different from the
        source from which the grammar class was generated
    """
    grammar = load_if_file(grammar_source)
    chksum = md5(grammar, __version__)
    if isinstance(grammar_class, str):
        # grammar_class = load_compiler_suite(grammar_class)[1]
        with open(grammar_class, 'r', encoding='utf8') as f:
            pycode = f.read()
        m = re.search('class \w*\(GrammarBase\)', pycode)
        if m:
            m = re.search('    source_hash__ *= *"([a-z0-9]*)"',
                          pycode[m.span()[1]:])
            return not (m and m.groups() and m.groups()[-1] == chksum)
        else:
            return True
    else:
        return chksum != grammar_class.source_hash__