source: pyyaml/trunk/lib/yaml/scanner.py @ 149

Revision 149, 51.3 KB checked in by xi, 9 years ago (diff)

Raise an error for colons in the flow context.

RevLine 
[39]1
[55]2# Scanner produces tokens of the following types:
[118]3# STREAM-START
4# STREAM-END
[55]5# DIRECTIVE(name, value)
6# DOCUMENT-START
7# DOCUMENT-END
8# BLOCK-SEQUENCE-START
9# BLOCK-MAPPING-START
10# BLOCK-END
11# FLOW-SEQUENCE-START
12# FLOW-MAPPING-START
13# FLOW-SEQUENCE-END
14# FLOW-MAPPING-END
15# BLOCK-ENTRY
16# FLOW-ENTRY
17# KEY
18# VALUE
19# ALIAS(value)
20# ANCHOR(value)
21# TAG(value)
22# SCALAR(value, plain)
[57]23#
24# Read comments in the Scanner code for more details.
25#
[43]26
[46]27__all__ = ['Scanner', 'ScannerError']
[43]28
[52]29from error import MarkedYAMLError
[46]30from tokens import *
[39]31
[52]32class ScannerError(MarkedYAMLError):
33    pass
[51]34
[43]35class SimpleKey:
[51]36    # See below simple keys treatment.
37
[116]38    def __init__(self, token_number, required, index, line, column, mark):
[43]39        self.token_number = token_number
40        self.required = required
41        self.index = index
42        self.line = line
43        self.column = column
[116]44        self.mark = mark
[43]45
[39]46class Scanner:
47
[136]48    def __init__(self):
[39]49        """Initialize the scanner."""
[136]50        # It is assumed that Scanner and Reader will have a common descendant.
51        # Reader do the dirty work of checking for BOM and converting the
52        # input data to Unicode. It also adds NUL to the end.
[39]53        #
[46]54        # Reader supports the following methods
[136]55        #   self.peek(i=0)       # peek the next i-th character
56        #   self.prefix(l=1)     # peek the next l characters
57        #   self.forward(l=1)    # read the next l characters and move the pointer.
[39]58
59        # Had we reached the end of the stream?
60        self.done = False
61
62        # The number of unclosed '{' and '['. `flow_level == 0` means block
63        # context.
64        self.flow_level = 0
65
66        # List of processed tokens that are not yet emitted.
67        self.tokens = []
68
[118]69        # Add the STREAM-START token.
70        self.fetch_stream_start()
71
[39]72        # Number of tokens that were emitted through the `get_token` method.
73        self.tokens_taken = 0
74
75        # The current indentation level.
76        self.indent = -1
77
78        # Past indentation levels.
79        self.indents = []
80
[43]81        # Variables related to simple keys treatment.
[39]82
83        # A simple key is a key that is not denoted by the '?' indicator.
84        # Example of simple keys:
85        #   ---
86        #   block simple key: value
87        #   ? not a simple key:
88        #   : { flow simple key: value }
89        # We emit the KEY token before all keys, so when we find a potential
90        # simple key, we try to locate the corresponding ':' indicator.
91        # Simple keys should be limited to a single line and 1024 characters.
92
[43]93        # Can a simple key start at the current position? A simple key may
94        # start:
95        # - at the beginning of the line, not counting indentation spaces
96        #       (in block context),
97        # - after '{', '[', ',' (in the flow context),
98        # - after '?', ':', '-' (in the block context).
[60]99        # In the block context, this flag also signifies if a block collection
[43]100        # may start at the current position.
101        self.allow_simple_key = True
[39]102
103        # Keep track of possible simple keys. This is a dictionary. The key
104        # is `flow_level`; there can be no more that one possible simple key
[43]105        # for each level. The value is a SimpleKey record:
[116]106        #   (token_number, required, index, line, column, mark)
[43]107        # A simple key may start with ALIAS, ANCHOR, TAG, SCALAR(flow),
108        # '[', or '{' tokens.
[39]109        self.possible_simple_keys = {}
110
[51]111    # Public methods.
[39]112
[136]113    def check_token(self, *choices):
[51]114        # Check if the next token is one of the given types.
[43]115        while self.need_more_tokens():
[39]116            self.fetch_more_tokens()
117        if self.tokens:
[136]118            if not choices:
119                return True
[51]120            for choice in choices:
121                if isinstance(self.tokens[0], choice):
122                    return True
123        return False
124
[136]125    def peek_token(self):
[51]126        # Return the next token, but do not delete if from the queue.
127        while self.need_more_tokens():
128            self.fetch_more_tokens()
129        if self.tokens:
[39]130            return self.tokens[0]
131
[136]132    def get_token(self):
[51]133        # Return the next token.
[39]134        while self.need_more_tokens():
135            self.fetch_more_tokens()
136        if self.tokens:
137            self.tokens_taken += 1
138            return self.tokens.pop(0)
139
[51]140    def __iter__(self):
141        # Iterator protocol.
142        while self.need_more_tokens():
143            self.fetch_more_tokens()
144        while self.tokens:
145            self.tokens_taken += 1
146            yield self.tokens.pop(0)
147            while self.need_more_tokens():
148                self.fetch_more_tokens()
149
[43]150    # Private methods.
[39]151
152    def need_more_tokens(self):
153        if self.done:
154            return False
155        if not self.tokens:
156            return True
157        # The current token may be a potential simple key, so we
158        # need to look further.
[43]159        self.stale_possible_simple_keys()
[39]160        if self.next_possible_simple_key() == self.tokens_taken:
161            return True
162
163    def fetch_more_tokens(self):
164
165        # Eat whitespaces and comments until we reach the next token.
[43]166        self.scan_to_next_token()
[39]167
[43]168        # Remove obsolete possible simple keys.
169        self.stale_possible_simple_keys()
170
[39]171        # Compare the current indentation and column. It may add some tokens
[43]172        # and decrease the current indentation level.
[136]173        self.unwind_indent(self.column)
[39]174
175        # Peek the next character.
[136]176        ch = self.peek()
[39]177
[48]178        # Is it the end of stream?
[43]179        if ch == u'\0':
[48]180            return self.fetch_stream_end()
[39]181
182        # Is it a directive?
183        if ch == u'%' and self.check_directive():
184            return self.fetch_directive()
185
186        # Is it the document start?
187        if ch == u'-' and self.check_document_start():
188            return self.fetch_document_start()
189
190        # Is it the document end?
191        if ch == u'.' and self.check_document_end():
192            return self.fetch_document_end()
193
[52]194        # TODO: support for BOM within a stream.
195        #if ch == u'\uFEFF':
196        #    return self.fetch_bom()    <-- issue BOMToken
197
[39]198        # Note: the order of the following checks is NOT significant.
199
200        # Is it the flow sequence start indicator?
201        if ch == u'[':
202            return self.fetch_flow_sequence_start()
203
204        # Is it the flow mapping start indicator?
205        if ch == u'{':
206            return self.fetch_flow_mapping_start()
207
208        # Is it the flow sequence end indicator?
209        if ch == u']':
210            return self.fetch_flow_sequence_end()
211
212        # Is it the flow mapping end indicator?
213        if ch == u'}':
214            return self.fetch_flow_mapping_end()
215
[51]216        # Is it the flow entry indicator?
217        if ch in u',':
218            return self.fetch_flow_entry()
[43]219
[51]220        # Is it the block entry indicator?
221        if ch in u'-' and self.check_block_entry():
222            return self.fetch_block_entry()
223
[39]224        # Is it the key indicator?
225        if ch == u'?' and self.check_key():
226            return self.fetch_key()
227
228        # Is it the value indicator?
229        if ch == u':' and self.check_value():
230            return self.fetch_value()
231
232        # Is it an alias?
233        if ch == u'*':
234            return self.fetch_alias()
235
236        # Is it an anchor?
237        if ch == u'&':
238            return self.fetch_anchor()
239
[43]240        # Is it a tag?
[39]241        if ch == u'!':
242            return self.fetch_tag()
243
[43]244        # Is it a literal scalar?
245        if ch == u'|' and not self.flow_level:
[39]246            return self.fetch_literal()
247
248        # Is it a folded scalar?
[43]249        if ch == u'>' and not self.flow_level:
[39]250            return self.fetch_folded()
251
252        # Is it a single quoted scalar?
253        if ch == u'\'':
254            return self.fetch_single()
255
256        # Is it a double quoted scalar?
257        if ch == u'\"':
258            return self.fetch_double()
259
[43]260        # It must be a plain scalar then.
[39]261        if self.check_plain():
262            return self.fetch_plain()
263
[43]264        # No? It's an error. Let's produce a nice error message.
[48]265        raise ScannerError("while scanning for the next token", None,
266                "found character %r that cannot start any token"
[136]267                % ch.encode('utf-8'), self.get_mark())
[39]268
[43]269    # Simple keys treatment.
270
271    def next_possible_simple_key(self):
272        # Return the number of the nearest possible simple key. Actually we
273        # don't need to loop through the whole dictionary. We may replace it
274        # with the following code:
275        #   if not self.possible_simple_keys:
276        #       return None
277        #   return self.possible_simple_keys[
278        #           min(self.possible_simple_keys.keys())].token_number
279        min_token_number = None
280        for level in self.possible_simple_keys:
281            key = self.possible_simple_keys[level]
282            if min_token_number is None or key.token_number < min_token_number:
283                min_token_number = key.token_number
284        return min_token_number
285
286    def stale_possible_simple_keys(self):
287        # Remove entries that are no longer possible simple keys. According to
288        # the YAML specification, simple keys
289        # - should be limited to a single line,
290        # - should be no longer than 1024 characters.
291        # Disabling this procedure will allow simple keys of any length and
292        # height (may cause problems if indentation is broken though).
293        for level in self.possible_simple_keys.keys():
294            key = self.possible_simple_keys[level]
[136]295            if key.line != self.line  \
296                    or self.index-key.index > 1024:
[43]297                if key.required:
[116]298                    raise ScannerError("while scanning a simple key", key.mark,
[136]299                            "could not found expected ':'", self.get_mark())
[43]300                del self.possible_simple_keys[level]
301
302    def save_possible_simple_key(self):
303        # The next token may start a simple key. We check if it's possible
304        # and save its position. This function is called for
305        #   ALIAS, ANCHOR, TAG, SCALAR(flow), '[', and '{'.
306
307        # Check if a simple key is required at the current position.
[136]308        required = not self.flow_level and self.indent == self.column
[43]309
[47]310        # A simple key is required only if it is the first token in the current
311        # line. Therefore it is always allowed.
312        assert self.allow_simple_key or not required
313
[43]314        # The next token might be a simple key. Let's save it's number and
315        # position.
316        if self.allow_simple_key:
317            self.remove_possible_simple_key()
318            token_number = self.tokens_taken+len(self.tokens)
319            key = SimpleKey(token_number, required,
[136]320                    self.index, self.line, self.column, self.get_mark())
[43]321            self.possible_simple_keys[self.flow_level] = key
322
323    def remove_possible_simple_key(self):
324        # Remove the saved possible key position at the current flow level.
325        if self.flow_level in self.possible_simple_keys:
326            key = self.possible_simple_keys[self.flow_level]
[47]327           
328            # I don't think it's possible, but I could be wrong.
329            assert not key.required
330            #if key.required:
[116]331            #    raise ScannerError("while scanning a simple key", key.mark,
[136]332            #            "could not found expected ':'", self.get_mark())
[43]333
334    # Indentation functions.
335
336    def unwind_indent(self, column):
337
[117]338        ## In flow context, tokens should respect indentation.
339        ## Actually the condition should be `self.indent >= column` according to
340        ## the spec. But this condition will prohibit intuitively correct
341        ## constructions such as
342        ## key : {
343        ## }
344        #if self.flow_level and self.indent > column:
345        #    raise ScannerError(None, None,
346        #            "invalid intendation or unclosed '[' or '{'",
[136]347        #            self.get_mark())
[43]348
[117]349        # In the flow context, indentation is ignored. We make the scanner less
350        # restrictive then specification requires.
351        if self.flow_level:
352            return
353
[43]354        # In block context, we may need to issue the BLOCK-END tokens.
355        while self.indent > column:
[136]356            mark = self.get_mark()
[43]357            self.indent = self.indents.pop()
[116]358            self.tokens.append(BlockEndToken(mark, mark))
[43]359
360    def add_indent(self, column):
361        # Check if we need to increase indentation.
362        if self.indent < column:
363            self.indents.append(self.indent)
364            self.indent = column
365            return True
366        return False
367
368    # Fetchers.
369
[118]370    def fetch_stream_start(self):
371        # We always add STREAM-START as the first token and STREAM-END as the
372        # last token.
373
374        # Read the token.
[136]375        mark = self.get_mark()
[118]376       
[130]377        # Add STREAM-START.
378        self.tokens.append(StreamStartToken(mark, mark,
[136]379            encoding=self.encoding))
[118]380       
381
[48]382    def fetch_stream_end(self):
[39]383
384        # Set the current intendation to -1.
[43]385        self.unwind_indent(-1)
[39]386
387        # Reset everything (not really needed).
[43]388        self.allow_simple_key = False
[39]389        self.possible_simple_keys = {}
390
[43]391        # Read the token.
[136]392        mark = self.get_mark()
[43]393       
[118]394        # Add STREAM-END.
[116]395        self.tokens.append(StreamEndToken(mark, mark))
[39]396
[136]397        # The steam is finished.
[39]398        self.done = True
399
[43]400    def fetch_directive(self):
401       
402        # Set the current intendation to -1.
403        self.unwind_indent(-1)
[39]404
[43]405        # Reset simple keys.
406        self.remove_possible_simple_key()
407        self.allow_simple_key = False
[39]408
[43]409        # Scan and add DIRECTIVE.
[47]410        self.tokens.append(self.scan_directive())
[39]411
412    def fetch_document_start(self):
[44]413        self.fetch_document_indicator(DocumentStartToken)
[39]414
[43]415    def fetch_document_end(self):
[44]416        self.fetch_document_indicator(DocumentEndToken)
[43]417
418    def fetch_document_indicator(self, TokenClass):
419
[39]420        # Set the current intendation to -1.
[43]421        self.unwind_indent(-1)
[39]422
[43]423        # Reset simple keys. Note that there could not be a block collection
424        # after '---'.
425        self.remove_possible_simple_key()
426        self.allow_simple_key = False
[39]427
[43]428        # Add DOCUMENT-START or DOCUMENT-END.
[136]429        start_mark = self.get_mark()
430        self.forward(3)
431        end_mark = self.get_mark()
[116]432        self.tokens.append(TokenClass(start_mark, end_mark))
[39]433
[43]434    def fetch_flow_sequence_start(self):
[44]435        self.fetch_flow_collection_start(FlowSequenceStartToken)
[39]436
[43]437    def fetch_flow_mapping_start(self):
[44]438        self.fetch_flow_collection_start(FlowMappingStartToken)
[43]439
440    def fetch_flow_collection_start(self, TokenClass):
441
[44]442        # '[' and '{' may start a simple key.
443        self.save_possible_simple_key()
444
[43]445        # Increase the flow level.
446        self.flow_level += 1
447
448        # Simple keys are allowed after '[' and '{'.
449        self.allow_simple_key = True
450
451        # Add FLOW-SEQUENCE-START or FLOW-MAPPING-START.
[136]452        start_mark = self.get_mark()
453        self.forward()
454        end_mark = self.get_mark()
[116]455        self.tokens.append(TokenClass(start_mark, end_mark))
[39]456
[43]457    def fetch_flow_sequence_end(self):
[44]458        self.fetch_flow_collection_end(FlowSequenceEndToken)
[39]459
[43]460    def fetch_flow_mapping_end(self):
[44]461        self.fetch_flow_collection_end(FlowMappingEndToken)
[43]462
463    def fetch_flow_collection_end(self, TokenClass):
464
465        # Reset possible simple key on the current level.
466        self.remove_possible_simple_key()
467
468        # Decrease the flow level.
469        self.flow_level -= 1
470
471        # No simple keys after ']' or '}'.
472        self.allow_simple_key = False
473
474        # Add FLOW-SEQUENCE-END or FLOW-MAPPING-END.
[136]475        start_mark = self.get_mark()
476        self.forward()
477        end_mark = self.get_mark()
[116]478        self.tokens.append(TokenClass(start_mark, end_mark))
[39]479
[51]480    def fetch_flow_entry(self):
[39]481
[51]482        # Simple keys are allowed after ','.
483        self.allow_simple_key = True
484
485        # Reset possible simple key on the current level.
486        self.remove_possible_simple_key()
487
488        # Add FLOW-ENTRY.
[136]489        start_mark = self.get_mark()
490        self.forward()
491        end_mark = self.get_mark()
[116]492        self.tokens.append(FlowEntryToken(start_mark, end_mark))
[51]493
494    def fetch_block_entry(self):
495
[43]496        # Block context needs additional checks.
497        if not self.flow_level:
[39]498
[43]499            # Are we allowed to start a new entry?
500            if not self.allow_simple_key:
[47]501                raise ScannerError(None, None,
502                        "sequence entries are not allowed here",
[136]503                        self.get_mark())
[39]504
[43]505            # We may need to add BLOCK-SEQUENCE-START.
[136]506            if self.add_indent(self.column):
507                mark = self.get_mark()
[116]508                self.tokens.append(BlockSequenceStartToken(mark, mark))
[39]509
[51]510        # It's an error for the block entry to occur in the flow context,
511        # but we let the parser detect this.
512        else:
513            pass
514
515        # Simple keys are allowed after '-'.
[43]516        self.allow_simple_key = True
[39]517
[43]518        # Reset possible simple key on the current level.
519        self.remove_possible_simple_key()
[39]520
[51]521        # Add BLOCK-ENTRY.
[136]522        start_mark = self.get_mark()
523        self.forward()
524        end_mark = self.get_mark()
[116]525        self.tokens.append(BlockEntryToken(start_mark, end_mark))
[39]526
[43]527    def fetch_key(self):
528       
529        # Block context needs additional checks.
530        if not self.flow_level:
[39]531
[43]532            # Are we allowed to start a key (not nessesary a simple)?
533            if not self.allow_simple_key:
[47]534                raise ScannerError(None, None,
535                        "mapping keys are not allowed here",
[136]536                        self.get_mark())
[43]537
538            # We may need to add BLOCK-MAPPING-START.
[136]539            if self.add_indent(self.column):
540                mark = self.get_mark()
[116]541                self.tokens.append(BlockMappingStartToken(mark, mark))
[43]542
543        # Simple keys are allowed after '?' in the block context.
544        self.allow_simple_key = not self.flow_level
545
546        # Reset possible simple key on the current level.
547        self.remove_possible_simple_key()
548
549        # Add KEY.
[136]550        start_mark = self.get_mark()
551        self.forward()
552        end_mark = self.get_mark()
[116]553        self.tokens.append(KeyToken(start_mark, end_mark))
[39]554
[43]555    def fetch_value(self):
[39]556
[43]557        # Do we determine a simple key?
558        if self.flow_level in self.possible_simple_keys:
[39]559
[43]560            # Add KEY.
561            key = self.possible_simple_keys[self.flow_level]
562            del self.possible_simple_keys[self.flow_level]
563            self.tokens.insert(key.token_number-self.tokens_taken,
[116]564                    KeyToken(key.mark, key.mark))
[39]565
[43]566            # If this key starts a new block mapping, we need to add
567            # BLOCK-MAPPING-START.
568            if not self.flow_level:
569                if self.add_indent(key.column):
570                    self.tokens.insert(key.token_number-self.tokens_taken,
[116]571                            BlockMappingStartToken(key.mark, key.mark))
[37]572
[43]573            # There cannot be two simple keys one after another.
574            self.allow_simple_key = False
[37]575
[43]576        # It must be a part of a complex key.
577        else:
578           
[47]579            # Block context needs additional checks.
580            # (Do we really need them? They will be catched by the parser
581            # anyway.)
582            if not self.flow_level:
583
584                # We are allowed to start a complex value if and only if
585                # we can start a simple key.
586                if not self.allow_simple_key:
587                    raise ScannerError(None, None,
588                            "mapping values are not allowed here",
[136]589                            self.get_mark())
[47]590
[43]591            # Simple keys are allowed after ':' in the block context.
592            self.allow_simple_key = not self.flow_level
[37]593
[43]594            # Reset possible simple key on the current level.
595            self.remove_possible_simple_key()
[37]596
[43]597        # Add VALUE.
[136]598        start_mark = self.get_mark()
599        self.forward()
600        end_mark = self.get_mark()
[116]601        self.tokens.append(ValueToken(start_mark, end_mark))
[37]602
[43]603    def fetch_alias(self):
[37]604
[43]605        # ALIAS could be a simple key.
606        self.save_possible_simple_key()
[37]607
[43]608        # No simple keys after ALIAS.
609        self.allow_simple_key = False
[37]610
[43]611        # Scan and add ALIAS.
[47]612        self.tokens.append(self.scan_anchor(AliasToken))
[37]613
[43]614    def fetch_anchor(self):
[37]615
[43]616        # ANCHOR could start a simple key.
617        self.save_possible_simple_key()
[37]618
[43]619        # No simple keys after ANCHOR.
620        self.allow_simple_key = False
[37]621
[43]622        # Scan and add ANCHOR.
[47]623        self.tokens.append(self.scan_anchor(AnchorToken))
[37]624
[43]625    def fetch_tag(self):
[37]626
[43]627        # TAG could start a simple key.
628        self.save_possible_simple_key()
[37]629
[43]630        # No simple keys after TAG.
631        self.allow_simple_key = False
[37]632
[43]633        # Scan and add TAG.
[47]634        self.tokens.append(self.scan_tag())
[37]635
[43]636    def fetch_literal(self):
[130]637        self.fetch_block_scalar(style='|')
[37]638
[43]639    def fetch_folded(self):
[130]640        self.fetch_block_scalar(style='>')
[37]641
[130]642    def fetch_block_scalar(self, style):
[37]643
[43]644        # A simple key may follow a block scalar.
645        self.allow_simple_key = True
[37]646
[43]647        # Reset possible simple key on the current level.
648        self.remove_possible_simple_key()
[37]649
[43]650        # Scan and add SCALAR.
[130]651        self.tokens.append(self.scan_block_scalar(style))
[37]652
[43]653    def fetch_single(self):
[130]654        self.fetch_flow_scalar(style='\'')
[37]655
[43]656    def fetch_double(self):
[130]657        self.fetch_flow_scalar(style='"')
[37]658
[130]659    def fetch_flow_scalar(self, style):
[37]660
[43]661        # A flow scalar could be a simple key.
662        self.save_possible_simple_key()
[37]663
[43]664        # No simple keys after flow scalars.
665        self.allow_simple_key = False
[37]666
[43]667        # Scan and add SCALAR.
[130]668        self.tokens.append(self.scan_flow_scalar(style))
[37]669
[43]670    def fetch_plain(self):
[37]671
[43]672        # A plain scalar could be a simple key.
673        self.save_possible_simple_key()
[37]674
[43]675        # No simple keys after plain scalars. But note that `scan_plain` will
676        # change this flag if the scan is finished at the beginning of the
677        # line.
678        self.allow_simple_key = False
[37]679
[43]680        # Scan and add SCALAR. May change `allow_simple_key`.
[47]681        self.tokens.append(self.scan_plain())
[37]682
[43]683    # Checkers.
[37]684
[43]685    def check_directive(self):
[37]686
[43]687        # DIRECTIVE:        ^ '%' ...
688        # The '%' indicator is already checked.
[136]689        if self.column == 0:
[43]690            return True
[37]691
[43]692    def check_document_start(self):
[37]693
[43]694        # DOCUMENT-START:   ^ '---' (' '|'\n')
[136]695        if self.column == 0:
696            if self.prefix(3) == u'---'  \
697                    and self.peek(3) in u'\0 \t\r\n\x85\u2028\u2029':
[43]698                return True
[37]699
[43]700    def check_document_end(self):
[37]701
[43]702        # DOCUMENT-END:     ^ '...' (' '|'\n')
[136]703        if self.column == 0:
704            if self.prefix(3) == u'...'  \
705                    and self.peek(3) in u'\0 \t\r\n\x85\u2028\u2029':
[43]706                return True
[37]707
[51]708    def check_block_entry(self):
[43]709
[51]710        # BLOCK-ENTRY:      '-' (' '|'\n')
[136]711        return self.peek(1) in u'\0 \t\r\n\x85\u2028\u2029'
[43]712
713    def check_key(self):
714
715        # KEY(flow context):    '?'
716        if self.flow_level:
[37]717            return True
[43]718
719        # KEY(block context):   '?' (' '|'\n')
[37]720        else:
[136]721            return self.peek(1) in u'\0 \t\r\n\x85\u2028\u2029'
[37]722
[43]723    def check_value(self):
724
725        # VALUE(flow context):  ':'
726        if self.flow_level:
[37]727            return True
[43]728
729        # VALUE(block context): ':' (' '|'\n')
[37]730        else:
[136]731            return self.peek(1) in u'\0 \t\r\n\x85\u2028\u2029'
[37]732
[43]733    def check_plain(self):
[37]734
[48]735        # A plain scalar may start with any non-space character except:
736        #   '-', '?', ':', ',', '[', ']', '{', '}',
737        #   '#', '&', '*', '!', '|', '>', '\'', '\"',
738        #   '%', '@', '`'.
739        #
740        # It may also start with
741        #   '-', '?', ':'
742        # if it is followed by a non-space character.
743        #
744        # Note that we limit the last rule to the block context (except the
745        # '-' character) because we want the flow context to be space
746        # independent.
[136]747        ch = self.peek()
[48]748        return ch not in u'\0 \t\r\n\x85\u2028\u2029-?:,[]{}#&*!|>\'\"%@`'  \
[136]749                or (self.peek(1) not in u'\0 \t\r\n\x85\u2028\u2029'
[132]750                        and (ch == u'-' or (not self.flow_level and ch in u'?:')))
[48]751
[43]752    # Scanners.
753
754    def scan_to_next_token(self):
[47]755        # We ignore spaces, line breaks and comments.
756        # If we find a line break in the block context, we set the flag
757        # `allow_simple_key` on.
[51]758        # The byte order mark is stripped if it's the first character in the
759        # stream. We do not yet support BOM inside the stream as the
760        # specification requires. Any such mark will be considered as a part
761        # of the document.
[52]762        #
763        # TODO: We need to make tab handling rules more sane. A good rule is
764        #   Tabs cannot precede tokens
765        #   BLOCK-SEQUENCE-START, BLOCK-MAPPING-START, BLOCK-END,
766        #   KEY(block), VALUE(block), BLOCK-ENTRY
767        # So the checking code is
768        #   if <TAB>:
769        #       self.allow_simple_keys = False
770        # We also need to add the check for `allow_simple_keys == True` to
771        # `unwind_indent` before issuing BLOCK-END.
772        # Scanners for block, flow, and plain scalars need to be modified.
773
[136]774        if self.index == 0 and self.peek() == u'\uFEFF':
775            self.forward()
[43]776        found = False
777        while not found:
[136]778            while self.peek() == u' ':
779                self.forward()
780            if self.peek() == u'#':
781                while self.peek() not in u'\0\r\n\x85\u2028\u2029':
782                    self.forward()
[47]783            if self.scan_line_break():
[43]784                if not self.flow_level:
785                    self.allow_simple_key = True
[37]786            else:
[43]787                found = True
[37]788
[43]789    def scan_directive(self):
[48]790        # See the specification for details.
[136]791        start_mark = self.get_mark()
792        self.forward()
[116]793        name = self.scan_directive_name(start_mark)
[48]794        value = None
795        if name == u'YAML':
[116]796            value = self.scan_yaml_directive_value(start_mark)
[136]797            end_mark = self.get_mark()
[48]798        elif name == u'TAG':
[116]799            value = self.scan_tag_directive_value(start_mark)
[136]800            end_mark = self.get_mark()
[43]801        else:
[136]802            end_mark = self.get_mark()
803            while self.peek() not in u'\0\r\n\x85\u2028\u2029':
804                self.forward()
[116]805        self.scan_directive_ignored_line(start_mark)
806        return DirectiveToken(name, value, start_mark, end_mark)
[48]807
[116]808    def scan_directive_name(self, start_mark):
[48]809        # See the specification for details.
810        length = 0
[136]811        ch = self.peek(length)
[48]812        while u'0' <= ch <= u'9' or u'A' <= ch <= 'Z' or u'a' <= ch <= 'z'  \
813                or ch in u'-_':
814            length += 1
[136]815            ch = self.peek(length)
[48]816        if not length:
[116]817            raise ScannerError("while scanning a directive", start_mark,
[52]818                    "expected alphabetic or numeric character, but found %r"
[136]819                    % ch.encode('utf-8'), self.get_mark())
820        value = self.prefix(length)
821        self.forward(length)
822        ch = self.peek()
[48]823        if ch not in u'\0 \r\n\x85\u2028\u2029':
[116]824            raise ScannerError("while scanning a directive", start_mark,
[48]825                    "expected alphabetic or numeric character, but found %r"
[136]826                    % ch.encode('utf-8'), self.get_mark())
[48]827        return value
828
[116]829    def scan_yaml_directive_value(self, start_mark):
[48]830        # See the specification for details.
[136]831        while self.peek() == u' ':
832            self.forward()
[116]833        major = self.scan_yaml_directive_number(start_mark)
[136]834        if self.peek() != '.':
[116]835            raise ScannerError("while scanning a directive", start_mark,
[52]836                    "expected a digit or '.', but found %r"
[136]837                    % self.peek().encode('utf-8'),
838                    self.get_mark())
839        self.forward()
[116]840        minor = self.scan_yaml_directive_number(start_mark)
[136]841        if self.peek() not in u'\0 \r\n\x85\u2028\u2029':
[116]842            raise ScannerError("while scanning a directive", start_mark,
[52]843                    "expected a digit or ' ', but found %r"
[136]844                    % self.peek().encode('utf-8'),
845                    self.get_mark())
[48]846        return (major, minor)
[37]847
[116]848    def scan_yaml_directive_number(self, start_mark):
[48]849        # See the specification for details.
[136]850        ch = self.peek()
[48]851        if not (u'0' <= ch <= '9'):
[116]852            raise ScannerError("while scanning a directive", start_mark,
[48]853                    "expected a digit, but found %r" % ch.encode('utf-8'),
[136]854                    self.get_mark())
[48]855        length = 0
[136]856        while u'0' <= self.peek(length) <= u'9':
[48]857            length += 1
[136]858        value = int(self.prefix(length))
859        self.forward(length)
[48]860        return value
861
[116]862    def scan_tag_directive_value(self, start_mark):
[48]863        # See the specification for details.
[136]864        while self.peek() == u' ':
865            self.forward()
[116]866        handle = self.scan_tag_directive_handle(start_mark)
[136]867        while self.peek() == u' ':
868            self.forward()
[116]869        prefix = self.scan_tag_directive_prefix(start_mark)
[48]870        return (handle, prefix)
871
[116]872    def scan_tag_directive_handle(self, start_mark):
[48]873        # See the specification for details.
[116]874        value = self.scan_tag_handle('directive', start_mark)
[136]875        ch = self.peek()
[52]876        if ch != u' ':
[116]877            raise ScannerError("while scanning a directive", start_mark,
[48]878                    "expected ' ', but found %r" % ch.encode('utf-8'),
[136]879                    self.get_mark())
[48]880        return value
881
[116]882    def scan_tag_directive_prefix(self, start_mark):
[48]883        # See the specification for details.
[116]884        value = self.scan_tag_uri('directive', start_mark)
[136]885        ch = self.peek()
[48]886        if ch not in u'\0 \r\n\x85\u2028\u2029':
[116]887            raise ScannerError("while scanning a directive", start_mark,
[48]888                    "expected ' ', but found %r" % ch.encode('utf-8'),
[136]889                    self.get_mark())
[48]890        return value
891
[116]892    def scan_directive_ignored_line(self, start_mark):
[48]893        # See the specification for details.
[136]894        while self.peek() == u' ':
895            self.forward()
896        if self.peek() == u'#':
897            while self.peek() not in u'\0\r\n\x85\u2028\u2029':
898                self.forward()
899        ch = self.peek()
[48]900        if ch not in u'\0\r\n\x85\u2028\u2029':
[116]901            raise ScannerError("while scanning a directive", start_mark,
[48]902                    "expected a comment or a line break, but found %r"
[136]903                        % ch.encode('utf-8'), self.get_mark())
[48]904        self.scan_line_break()
905
[43]906    def scan_anchor(self, TokenClass):
[48]907        # The specification does not restrict characters for anchors and
908        # aliases. This may lead to problems, for instance, the document:
909        #   [ *alias, value ]
910        # can be interpteted in two ways, as
911        #   [ "value" ]
912        # and
913        #   [ *alias , "value" ]
914        # Therefore we restrict aliases to numbers and ASCII letters.
[136]915        start_mark = self.get_mark()
916        indicator = self.peek()
[48]917        if indicator == '*':
918            name = 'alias'
919        else:
920            name = 'anchor'
[136]921        self.forward()
[48]922        length = 0
[136]923        ch = self.peek(length)
[48]924        while u'0' <= ch <= u'9' or u'A' <= ch <= 'Z' or u'a' <= ch <= 'z'  \
925                or ch in u'-_':
926            length += 1
[136]927            ch = self.peek(length)
[48]928        if not length:
[116]929            raise ScannerError("while scanning an %s" % name, start_mark,
[52]930                    "expected alphabetic or numeric character, but found %r"
[136]931                    % ch.encode('utf-8'), self.get_mark())
932        value = self.prefix(length)
933        self.forward(length)
934        ch = self.peek()
[48]935        if ch not in u'\0 \t\r\n\x85\u2028\u2029?:,]}%@`':
[116]936            raise ScannerError("while scanning an %s" % name, start_mark,
[48]937                    "expected alphabetic or numeric character, but found %r"
[136]938                    % ch.encode('utf-8'), self.get_mark())
939        end_mark = self.get_mark()
[116]940        return TokenClass(value, start_mark, end_mark)
[37]941
[43]942    def scan_tag(self):
[48]943        # See the specification for details.
[136]944        start_mark = self.get_mark()
945        ch = self.peek(1)
[48]946        if ch == u'<':
947            handle = None
[136]948            self.forward(2)
[116]949            suffix = self.scan_tag_uri('tag', start_mark)
[136]950            if self.peek() != u'>':
[116]951                raise ScannerError("while parsing a tag", start_mark,
[136]952                        "expected '>', but found %r" % self.peek().encode('utf-8'),
953                        self.get_mark())
954            self.forward()
[48]955        elif ch in u'\0 \t\r\n\x85\u2028\u2029':
956            handle = None
957            suffix = u'!'
[136]958            self.forward()
[48]959        else:
960            length = 1
961            use_handle = False
962            while ch not in u'\0 \r\n\x85\u2028\u2029':
963                if ch == u'!':
964                    use_handle = True
965                    break
966                length += 1
[136]967                ch = self.peek(length)
[48]968            handle = u'!'
969            if use_handle:
[116]970                handle = self.scan_tag_handle('tag', start_mark)
[48]971            else:
972                handle = u'!'
[136]973                self.forward()
[116]974            suffix = self.scan_tag_uri('tag', start_mark)
[136]975        ch = self.peek()
[48]976        if ch not in u'\0 \r\n\x85\u2028\u2029':
[116]977            raise ScannerError("while scanning a tag", start_mark,
[48]978                    "expected ' ', but found %r" % ch.encode('utf-8'),
[136]979                    self.get_mark())
[48]980        value = (handle, suffix)
[136]981        end_mark = self.get_mark()
[116]982        return TagToken(value, start_mark, end_mark)
[43]983
[130]984    def scan_block_scalar(self, style):
[48]985        # See the specification for details.
986
[130]987        if style == '>':
988            folded = True
989        else:
990            folded = False
991
[48]992        chunks = []
[136]993        start_mark = self.get_mark()
[48]994
995        # Scan the header.
[136]996        self.forward()
[116]997        chomping, increment = self.scan_block_scalar_indicators(start_mark)
998        self.scan_block_scalar_ignored_line(start_mark)
[48]999
1000        # Determine the indentation level and go to the first non-empty line.
1001        min_indent = self.indent+1
1002        if min_indent < 1:
1003            min_indent = 1
1004        if increment is None:
[116]1005            breaks, max_indent, end_mark = self.scan_block_scalar_indentation()
[48]1006            indent = max(min_indent, max_indent)
1007        else:
1008            indent = min_indent+increment-1
[116]1009            breaks, end_mark = self.scan_block_scalar_breaks(indent)
[48]1010        line_break = u''
1011
1012        # Scan the inner part of the block scalar.
[136]1013        while self.column == indent and self.peek() != u'\0':
[48]1014            chunks.extend(breaks)
[136]1015            leading_non_space = self.peek() not in u' \t'
[48]1016            length = 0
[136]1017            while self.peek(length) not in u'\0\r\n\x85\u2028\u2029':
[48]1018                length += 1
[136]1019            chunks.append(self.prefix(length))
1020            self.forward(length)
[48]1021            line_break = self.scan_line_break()
[116]1022            breaks, end_mark = self.scan_block_scalar_breaks(indent)
[136]1023            if self.column == indent and self.peek() != u'\0':
[130]1024
[48]1025                # Unfortunately, folding rules are ambiguous.
1026                #
1027                # This is the folding according to the specification:
[51]1028               
1029                if folded and line_break == u'\n'   \
[136]1030                        and leading_non_space and self.peek() not in u' \t':
[51]1031                    if not breaks:
1032                        chunks.append(u' ')
1033                else:
1034                    chunks.append(line_break)
1035               
1036                # This is Clark Evans's interpretation (also in the spec
1037                # examples):
[48]1038                #
[51]1039                #if folded and line_break == u'\n':
[48]1040                #    if not breaks:
[136]1041                #        if self.peek() not in ' \t':
[51]1042                #            chunks.append(u' ')
1043                #        else:
1044                #            chunks.append(line_break)
[48]1045                #else:
1046                #    chunks.append(line_break)
1047            else:
1048                break
1049
1050        # Chomp the tail.
1051        if chomping is not False:
1052            chunks.append(line_break)
1053        if chomping is True:
1054            chunks.extend(breaks)
1055
1056        # We are done.
[130]1057        return ScalarToken(u''.join(chunks), False, start_mark, end_mark,
1058                style)
[48]1059
[116]1060    def scan_block_scalar_indicators(self, start_mark):
[48]1061        # See the specification for details.
1062        chomping = None
1063        increment = None
[136]1064        ch = self.peek()
[48]1065        if ch in u'+-':
1066            if ch == '+':
1067                chomping = True
1068            else:
1069                chomping = False
[136]1070            self.forward()
1071            ch = self.peek()
[48]1072            if ch in u'0123456789':
1073                increment = int(ch)
1074                if increment == 0:
[116]1075                    raise ScannerError("while scanning a block scalar", start_mark,
[48]1076                            "expected indentation indicator in the range 1-9, but found 0",
[136]1077                            self.get_mark())
1078                self.forward()
[48]1079        elif ch in u'0123456789':
1080            increment = int(ch)
1081            if increment == 0:
[116]1082                raise ScannerError("while scanning a block scalar", start_mark,
[48]1083                        "expected indentation indicator in the range 1-9, but found 0",
[136]1084                        self.get_mark())
1085            self.forward()
1086            ch = self.peek()
[48]1087            if ch in u'+-':
1088                if ch == '+':
1089                    chomping = True
1090                else:
1091                    chomping = False
[136]1092                self.forward()
1093        ch = self.peek()
[48]1094        if ch not in u'\0 \r\n\x85\u2028\u2029':
[116]1095            raise ScannerError("while scanning a block scalar", start_mark,
[48]1096                    "expected chomping or indentation indicators, but found %r"
[136]1097                        % ch.encode('utf-8'), self.get_mark())
[48]1098        return chomping, increment
1099
[116]1100    def scan_block_scalar_ignored_line(self, start_mark):
[48]1101        # See the specification for details.
[136]1102        while self.peek() == u' ':
1103            self.forward()
1104        if self.peek() == u'#':
1105            while self.peek() not in u'\0\r\n\x85\u2028\u2029':
1106                self.forward()
1107        ch = self.peek()
[48]1108        if ch not in u'\0\r\n\x85\u2028\u2029':
[116]1109            raise ScannerError("while scanning a block scalar", start_mark,
[48]1110                    "expected a comment or a line break, but found %r"
[136]1111                        % ch.encode('utf-8'), self.get_mark())
[48]1112        self.scan_line_break()
[43]1113
[48]1114    def scan_block_scalar_indentation(self):
1115        # See the specification for details.
1116        chunks = []
1117        max_indent = 0
[136]1118        end_mark = self.get_mark()
1119        while self.peek() in u' \r\n\x85\u2028\u2029':
1120            if self.peek() != u' ':
[48]1121                chunks.append(self.scan_line_break())
[136]1122                end_mark = self.get_mark()
[48]1123            else:
[136]1124                self.forward()
1125                if self.column > max_indent:
1126                    max_indent = self.column
[116]1127        return chunks, max_indent, end_mark
[48]1128
1129    def scan_block_scalar_breaks(self, indent):
1130        # See the specification for details.
1131        chunks = []
[136]1132        end_mark = self.get_mark()
1133        while self.column < indent and self.peek() == u' ':
1134            self.forward()
1135        while self.peek() in u'\r\n\x85\u2028\u2029':
[48]1136            chunks.append(self.scan_line_break())
[136]1137            end_mark = self.get_mark()
1138            while self.column < indent and self.peek() == u' ':
1139                self.forward()
[116]1140        return chunks, end_mark
[48]1141
[130]1142    def scan_flow_scalar(self, style):
[48]1143        # See the specification for details.
[117]1144        # Note that we loose indentation rules for quoted scalars. Quoted
1145        # scalars don't need to adhere indentation because " and ' clearly
1146        # mark the beginning and the end of them. Therefore we are less
1147        # restrictive then the specification requires. We only need to check
1148        # that document separators are not included in scalars.
[130]1149        if style == '"':
1150            double = True
1151        else:
1152            double = False
[48]1153        chunks = []
[136]1154        start_mark = self.get_mark()
1155        quote = self.peek()
1156        self.forward()
[117]1157        chunks.extend(self.scan_flow_scalar_non_spaces(double, start_mark))
[136]1158        while self.peek() != quote:
[117]1159            chunks.extend(self.scan_flow_scalar_spaces(double, start_mark))
1160            chunks.extend(self.scan_flow_scalar_non_spaces(double, start_mark))
[136]1161        self.forward()
1162        end_mark = self.get_mark()
[130]1163        return ScalarToken(u''.join(chunks), False, start_mark, end_mark,
1164                style)
[48]1165
1166    ESCAPE_REPLACEMENTS = {
1167        u'0':   u'\0',
1168        u'a':   u'\x07',
1169        u'b':   u'\x08',
1170        u't':   u'\x09',
1171        u'\t':  u'\x09',
1172        u'n':   u'\x0A',
1173        u'v':   u'\x0B',
1174        u'f':   u'\x0C',
1175        u'r':   u'\x0D',
1176        u'e':   u'\x1B',
1177        u' ':   u'\x20',
1178        u'\"':  u'\"',
1179        u'\\':  u'\\',
1180        u'N':   u'\x85',
1181        u'_':   u'\xA0',
1182        u'L':   u'\u2028',
1183        u'P':   u'\u2029',
1184    }
1185
1186    ESCAPE_CODES = {
1187        u'x':   2,
1188        u'u':   4,
1189        u'U':   8,
1190    }
1191
[117]1192    def scan_flow_scalar_non_spaces(self, double, start_mark):
[48]1193        # See the specification for details.
1194        chunks = []
1195        while True:
1196            length = 0
[136]1197            while self.peek(length) not in u'\'\"\\\0 \t\r\n\x85\u2028\u2029':
[48]1198                length += 1
1199            if length:
[136]1200                chunks.append(self.prefix(length))
1201                self.forward(length)
1202            ch = self.peek()
1203            if not double and ch == u'\'' and self.peek(1) == u'\'':
[48]1204                chunks.append(u'\'')
[136]1205                self.forward(2)
[48]1206            elif (double and ch == u'\'') or (not double and ch in u'\"\\'):
1207                chunks.append(ch)
[136]1208                self.forward()
[48]1209            elif double and ch == u'\\':
[136]1210                self.forward()
1211                ch = self.peek()
[48]1212                if ch in self.ESCAPE_REPLACEMENTS:
1213                    chunks.append(self.ESCAPE_REPLACEMENTS[ch])
[136]1214                    self.forward()
[48]1215                elif ch in self.ESCAPE_CODES:
1216                    length = self.ESCAPE_CODES[ch]
[136]1217                    self.forward()
[48]1218                    for k in range(length):
[136]1219                        if self.peek(k) not in u'0123456789ABCDEFabcdef':
[116]1220                            raise ScannerError("while scanning a double-quoted scalar", start_mark,
[48]1221                                    "expected escape sequence of %d hexdecimal numbers, but found %r" %
[136]1222                                        (length, self.peek(k).encode('utf-8')), self.get_mark())
1223                    code = int(self.prefix(length), 16)
[48]1224                    chunks.append(unichr(code))
[136]1225                    self.forward(length)
[48]1226                elif ch in u'\r\n\x85\u2028\u2029':
1227                    self.scan_line_break()
[117]1228                    chunks.extend(self.scan_flow_scalar_breaks(double, start_mark))
[48]1229                else:
[116]1230                    raise ScannerError("while scanning a double-quoted scalar", start_mark,
[136]1231                            "found unknown escape character %r" % ch.encode('utf-8'), self.get_mark())
[37]1232            else:
[48]1233                return chunks
[37]1234
[117]1235    def scan_flow_scalar_spaces(self, double, start_mark):
[48]1236        # See the specification for details.
1237        chunks = []
1238        length = 0
[136]1239        while self.peek(length) in u' \t':
[48]1240            length += 1
[136]1241        whitespaces = self.prefix(length)
1242        self.forward(length)
1243        ch = self.peek()
[48]1244        if ch == u'\0':
[116]1245            raise ScannerError("while scanning a quoted scalar", start_mark,
[136]1246                    "found unexpected end of stream", self.get_mark())
[48]1247        elif ch in u'\r\n\x85\u2028\u2029':
1248            line_break = self.scan_line_break()
[117]1249            breaks = self.scan_flow_scalar_breaks(double, start_mark)
[48]1250            if line_break != u'\n':
1251                chunks.append(line_break)
1252            elif not breaks:
1253                chunks.append(u' ')
1254            chunks.extend(breaks)
1255        else:
1256            chunks.append(whitespaces)
1257        return chunks
1258
[117]1259    def scan_flow_scalar_breaks(self, double, start_mark):
[48]1260        # See the specification for details.
1261        chunks = []
1262        while True:
[117]1263            # Instead of checking indentation, we check for document
1264            # separators.
[136]1265            prefix = self.prefix(3)
[117]1266            if (prefix == u'---' or prefix == u'...')   \
[136]1267                    and self.peek(3) in u'\0 \t\r\n\x85\u2028\u2029':
[116]1268                raise ScannerError("while scanning a quoted scalar", start_mark,
[136]1269                        "found unexpected document separator", self.get_mark())
1270            while self.peek() in u' \t':
1271                self.forward()
1272            if self.peek() in u'\r\n\x85\u2028\u2029':
[48]1273                chunks.append(self.scan_line_break())
1274            else:
1275                return chunks
1276
[43]1277    def scan_plain(self):
[48]1278        # See the specification for details.
1279        # We add an additional restriction for the flow context:
[117]1280        #   plain scalars in the flow context cannot contain ',', ':' and '?'.
[48]1281        # We also keep track of the `allow_simple_key` flag here.
[117]1282        # Indentation rules are loosed for the flow context.
[48]1283        chunks = []
[136]1284        start_mark = self.get_mark()
[116]1285        end_mark = start_mark
[43]1286        indent = self.indent+1
[117]1287        # We allow zero indentation for scalars, but then we need to check for
1288        # document separators at the beginning of the line.
1289        #if indent == 0:
1290        #    indent = 1
[48]1291        spaces = []
[43]1292        while True:
[48]1293            length = 0
[136]1294            if self.peek() == u'#':
[43]1295                break
[48]1296            while True:
[136]1297                ch = self.peek(length)
[48]1298                if ch in u'\0 \t\r\n\x85\u2028\u2029'   \
1299                        or (not self.flow_level and ch == u':' and
[136]1300                                self.peek(length+1) in u'\0 \t\r\n\x28\u2028\u2029') \
[48]1301                        or (self.flow_level and ch in u',:?[]{}'):
1302                    break
1303                length += 1
[149]1304            # It's not clear what we should do with ':' in the flow context.
1305            if (self.flow_level and ch == u':'
1306                    and self.peek(length+1) not in u'\0 \t\r\n\x28\u2028\u2029,[]{}'):
1307                self.forward(length)
1308                raise ScannerError("while scanning a plain scalar", start_mark,
1309                    "found unexpected ':'", self.get_mark(),
1310                    "Please check http://pyyaml.org/wiki/YAMLColonInFlowContext for details.")
[48]1311            if length == 0:
[43]1312                break
[48]1313            self.allow_simple_key = False
1314            chunks.extend(spaces)
[136]1315            chunks.append(self.prefix(length))
1316            self.forward(length)
1317            end_mark = self.get_mark()
[117]1318            spaces = self.scan_plain_spaces(indent, start_mark)
[136]1319            if not spaces or self.peek() == u'#' \
1320                    or (not self.flow_level and self.column < indent):
[48]1321                break
[116]1322        return ScalarToken(u''.join(chunks), True, start_mark, end_mark)
[37]1323
[117]1324    def scan_plain_spaces(self, indent, start_mark):
[48]1325        # See the specification for details.
1326        # The specification is really confusing about tabs in plain scalars.
1327        # We just forbid them completely. Do not use tabs in YAML!
1328        chunks = []
1329        length = 0
[136]1330        while self.peek(length) in u' ':
[48]1331            length += 1
[136]1332        whitespaces = self.prefix(length)
1333        self.forward(length)
1334        ch = self.peek()
[48]1335        if ch in u'\r\n\x85\u2028\u2029':
1336            line_break = self.scan_line_break()
1337            self.allow_simple_key = True
[136]1338            prefix = self.prefix(3)
[117]1339            if (prefix == u'---' or prefix == u'...')   \
[136]1340                    and self.peek(3) in u'\0 \t\r\n\x85\u2028\u2029':
[117]1341                return
[48]1342            breaks = []
[136]1343            while self.peek() in u' \r\n\x85\u2028\u2029':
1344                if self.peek() == ' ':
1345                    self.forward()
[48]1346                else:
1347                    breaks.append(self.scan_line_break())
[136]1348                    prefix = self.prefix(3)
[117]1349                    if (prefix == u'---' or prefix == u'...')   \
[136]1350                            and self.peek(3) in u'\0 \t\r\n\x85\u2028\u2029':
[117]1351                        return
[48]1352            if line_break != u'\n':
1353                chunks.append(line_break)
1354            elif not breaks:
1355                chunks.append(u' ')
1356            chunks.extend(breaks)
1357        elif whitespaces:
1358            chunks.append(whitespaces)
1359        return chunks
1360
[116]1361    def scan_tag_handle(self, name, start_mark):
[48]1362        # See the specification for details.
1363        # For some strange reasons, the specification does not allow '_' in
1364        # tag handles. I have allowed it anyway.
[136]1365        ch = self.peek()
[52]1366        if ch != u'!':
[116]1367            raise ScannerError("while scanning a %s" % name, start_mark,
[48]1368                    "expected '!', but found %r" % ch.encode('utf-8'),
[136]1369                    self.get_mark())
[48]1370        length = 1
[136]1371        ch = self.peek(length)
[48]1372        if ch != u' ':
1373            while u'0' <= ch <= u'9' or u'A' <= ch <= 'Z' or u'a' <= ch <= 'z'  \
1374                    or ch in u'-_':
1375                length += 1
[136]1376                ch = self.peek(length)
[48]1377            if ch != u'!':
[136]1378                self.forward(length)
[116]1379                raise ScannerError("while scanning a %s" % name, start_mark,
[48]1380                        "expected '!', but found %r" % ch.encode('utf-8'),
[136]1381                        self.get_mark())
[48]1382            length += 1
[136]1383        value = self.prefix(length)
1384        self.forward(length)
[48]1385        return value
1386
[116]1387    def scan_tag_uri(self, name, start_mark):
[48]1388        # See the specification for details.
1389        # Note: we do not check if URI is well-formed.
1390        chunks = []
1391        length = 0
[136]1392        ch = self.peek(length)
[48]1393        while u'0' <= ch <= u'9' or u'A' <= ch <= 'Z' or u'a' <= ch <= 'z'  \
1394                or ch in u'-;/?:@&=+$,_.!~*\'()[]%':
1395            if ch == u'%':
[136]1396                chunks.append(self.prefix(length))
1397                self.forward(length)
[48]1398                length = 0
[116]1399                chunks.append(self.scan_uri_escapes(name, start_mark))
[48]1400            else:
1401                length += 1
[136]1402            ch = self.peek(length)
[48]1403        if length:
[136]1404            chunks.append(self.prefix(length))
1405            self.forward(length)
[48]1406            length = 0
1407        if not chunks:
[116]1408            raise ScannerError("while parsing a %s" % name, start_mark,
[48]1409                    "expected URI, but found %r" % ch.encode('utf-8'),
[136]1410                    self.get_mark())
[48]1411        return u''.join(chunks)
1412
[116]1413    def scan_uri_escapes(self, name, start_mark):
[48]1414        # See the specification for details.
1415        bytes = []
[136]1416        mark = self.get_mark()
1417        while self.peek() == u'%':
1418            self.forward()
[48]1419            for k in range(2):
[136]1420                if self.peek(k) not in u'0123456789ABCDEFabcdef':
[116]1421                    raise ScannerError("while scanning a %s" % name, start_mark,
[48]1422                            "expected URI escape sequence of 2 hexdecimal numbers, but found %r" %
[136]1423                                (self.peek(k).encode('utf-8')), self.get_mark())
1424            bytes.append(chr(int(self.prefix(2), 16)))
1425            self.forward(2)
[48]1426        try:
1427            value = unicode(''.join(bytes), 'utf-8')
1428        except UnicodeDecodeError, exc:
[116]1429            raise ScannerError("while scanning a %s" % name, start_mark, str(exc), mark)
[48]1430        return value
1431
[47]1432    def scan_line_break(self):
1433        # Transforms:
1434        #   '\r\n'      :   '\n'
1435        #   '\r'        :   '\n'
1436        #   '\n'        :   '\n'
1437        #   '\x85'      :   '\n'
1438        #   '\u2028'    :   '\u2028'
1439        #   '\u2029     :   '\u2029'
1440        #   default     :   ''
[136]1441        ch = self.peek()
[47]1442        if ch in u'\r\n\x85':
[136]1443            if self.prefix(2) == u'\r\n':
1444                self.forward(2)
[47]1445            else:
[136]1446                self.forward()
[47]1447            return u'\n'
1448        elif ch in u'\u2028\u2029':
[136]1449            self.forward()
[47]1450            return ch
1451        return u''
1452
[45]1453#try:
1454#    import psyco
1455#    psyco.bind(Scanner)
1456#except ImportError:
1457#    pass
1458
Note: See TracBrowser for help on using the repository browser.