forked from whitequark/parser
/
lexer.rl
2543 lines (2107 loc) · 67.8 KB
/
lexer.rl
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
%%machine lex; # % fix highlighting
#
# === BEFORE YOU START ===
#
# Read the Ruby Hacking Guide chapter 11, available in English at
# http://whitequark.org/blog/2013/04/01/ruby-hacking-guide-ch-11-finite-state-lexer/
#
# Remember two things about Ragel scanners:
#
# 1) Longest match wins.
#
# 2) If two matches have the same length, the first
# in source code wins.
#
# General rules of making Ragel and Bison happy:
#
# * `p` (position) and `@te` contain the index of the character
# they're pointing to ("current"), plus one. `@ts` contains the index
# of the corresponding character. The code for extracting matched token is:
#
# @source_buffer.slice(@ts...@te)
#
# * If your input is `foooooooobar` and the rule is:
#
# 'f' 'o'+
#
# the result will be:
#
# foooooooobar
# ^ ts=0 ^ p=te=9
#
# * A Ragel lexer action should not emit more than one token, unless
# you know what you are doing.
#
# * All Ragel commands (fnext, fgoto, ...) end with a semicolon.
#
# * If an action emits the token and transitions to another state, use
# these Ragel commands:
#
# emit($whatever)
# fnext $next_state; fbreak;
#
# If you perform `fgoto` in an action which does not emit a token nor
# rewinds the stream pointer, the parser's side-effectful,
# context-sensitive lookahead actions will break in a hard to detect
# and debug way.
#
# * If an action does not emit a token:
#
# fgoto $next_state;
#
# * If an action features lookbehind, i.e. matches characters with the
# intent of passing them to another action:
#
# p = @ts - 1
# fgoto $next_state;
#
# or, if the lookbehind consists of a single character:
#
# fhold; fgoto $next_state;
#
# * Ragel merges actions. So, if you have `e_lparen = '(' %act` and
# `c_lparen = '('` and a lexer action `e_lparen | c_lparen`, the result
# _will_ invoke the action `act`.
#
# e_something stands for "something with **e**mbedded action".
#
# * EOF is explicit and is matched by `c_eof`. If you want to introspect
# the state of the lexer, add this rule to the state:
#
# c_eof => do_eof;
#
# * If you proceed past EOF, the lexer will complain:
#
# NoMethodError: undefined method `ord' for nil:NilClass
#
class Parser::Lexer
%% write data nofinal;
# %
ESCAPES = {
?a.ord => "\a", ?b.ord => "\b", ?e.ord => "\e", ?f.ord => "\f",
?n.ord => "\n", ?r.ord => "\r", ?s.ord => "\s", ?t.ord => "\t",
?v.ord => "\v", ?\\.ord => "\\"
}.freeze
REGEXP_META_CHARACTERS = Regexp.union(*"\\$()*+.<>?[]^{|}".chars).freeze
attr_reader :source_buffer
attr_accessor :diagnostics
attr_accessor :static_env
attr_accessor :force_utf32
attr_accessor :cond, :cmdarg, :in_kwarg, :context, :command_start
attr_accessor :tokens, :comments
def initialize(version)
@version = version
@static_env = nil
@context = nil
@tokens = nil
@comments = nil
reset
end
def reset(reset_state=true)
# Ragel state:
if reset_state
# Unit tests set state prior to resetting lexer.
@cs = self.class.lex_en_line_begin
@cond = StackState.new('cond')
@cmdarg = StackState.new('cmdarg')
@cond_stack = []
@cmdarg_stack = []
end
@force_utf32 = false # Set to true by some tests
@source_pts = nil # @source as a codepoint array
@p = 0 # stream position (saved manually in #advance)
@ts = nil # token start
@te = nil # token end
@act = 0 # next action
@stack = [] # state stack
@top = 0 # state stack top pointer
# Lexer state:
@token_queue = []
@literal_stack = []
@eq_begin_s = nil # location of last encountered =begin
@sharp_s = nil # location of last encountered #
@newline_s = nil # location of last encountered newline
@num_base = nil # last numeric base
@num_digits_s = nil # starting position of numeric digits
@num_suffix_s = nil # starting position of numeric suffix
@num_xfrm = nil # numeric suffix-induced transformation
@escape_s = nil # starting position of current sequence
@escape = nil # last escaped sequence, as string
@herebody_s = nil # starting position of current heredoc line
# Ruby 1.9 ->() lambdas emit a distinct token if do/{ is
# encountered after a matching closing parenthesis.
@paren_nest = 0
@lambda_stack = []
# After encountering the closing line of <<~SQUIGGLY_HEREDOC,
# we store the indentation level and give it out to the parser
# on request. It is not possible to infer indentation level just
# from the AST because escape sequences such as `\ ` or `\t` are
# expanded inside the lexer, but count as non-whitespace for
# indentation purposes.
@dedent_level = nil
# If the lexer is in `command state' (aka expr_value)
# at the entry to #advance, it will transition to expr_cmdarg
# instead of expr_arg at certain points.
@command_start = true
# True at the end of "def foo a:"
@in_kwarg = false
# State before =begin / =end block comment
@cs_before_block_comment = self.class.lex_en_line_begin
end
def source_buffer=(source_buffer)
@source_buffer = source_buffer
if @source_buffer
source = @source_buffer.source
if source.encoding == Encoding::UTF_8
@source_pts = source.unpack('U*')
else
@source_pts = source.unpack('C*')
end
if @source_pts[0] == 0xfeff
# Skip byte order mark.
@p = 1
end
else
@source_pts = nil
end
end
def encoding
@source_buffer.source.encoding
end
LEX_STATES = {
:line_begin => lex_en_line_begin,
:expr_dot => lex_en_expr_dot,
:expr_fname => lex_en_expr_fname,
:expr_value => lex_en_expr_value,
:expr_beg => lex_en_expr_beg,
:expr_mid => lex_en_expr_mid,
:expr_arg => lex_en_expr_arg,
:expr_cmdarg => lex_en_expr_cmdarg,
:expr_end => lex_en_expr_end,
:expr_endarg => lex_en_expr_endarg,
:expr_endfn => lex_en_expr_endfn,
:expr_labelarg => lex_en_expr_labelarg,
:interp_string => lex_en_interp_string,
:interp_words => lex_en_interp_words,
:plain_string => lex_en_plain_string,
:plain_words => lex_en_plain_string,
}
def state
LEX_STATES.invert.fetch(@cs, @cs)
end
def state=(state)
@cs = LEX_STATES.fetch(state)
end
def push_cmdarg
@cmdarg_stack.push(@cmdarg)
@cmdarg = StackState.new("cmdarg.#{@cmdarg_stack.count}")
end
def pop_cmdarg
@cmdarg = @cmdarg_stack.pop
end
def push_cond
@cond_stack.push(@cond)
@cond = StackState.new("cond.#{@cond_stack.count}")
end
def pop_cond
@cond = @cond_stack.pop
end
def dedent_level
# We erase @dedent_level as a precaution to avoid accidentally
# using a stale value.
dedent_level, @dedent_level = @dedent_level, nil
dedent_level
end
# Return next token: [type, value].
def advance
if @token_queue.any?
return @token_queue.shift
end
# Ugly, but dependent on Ragel output. Consider refactoring it somehow.
klass = self.class
_lex_trans_keys = klass.send :_lex_trans_keys
_lex_key_spans = klass.send :_lex_key_spans
_lex_index_offsets = klass.send :_lex_index_offsets
_lex_indicies = klass.send :_lex_indicies
_lex_trans_targs = klass.send :_lex_trans_targs
_lex_trans_actions = klass.send :_lex_trans_actions
_lex_to_state_actions = klass.send :_lex_to_state_actions
_lex_from_state_actions = klass.send :_lex_from_state_actions
_lex_eof_trans = klass.send :_lex_eof_trans
pe = @source_pts.size + 2
p, eof = @p, pe
cmd_state = @command_start
@command_start = false
%% write exec;
# %
# Ragel creates a local variable called `testEof` but it doesn't use
# it any assignment. This dead code is here to swallow the warning.
# It has no runtime cost because Ruby doesn't produce any instructions from it.
if false
testEof
end
@p = p
if @token_queue.any?
@token_queue.shift
elsif @cs == klass.lex_error
[ false, [ '$error'.freeze, range(p - 1, p) ] ]
else
eof = @source_pts.size
[ false, [ '$eof'.freeze, range(eof, eof) ] ]
end
end
protected
def eof_codepoint?(point)
[0x04, 0x1a, 0x00].include? point
end
def version?(*versions)
versions.include?(@version)
end
def stack_pop
@top -= 1
@stack[@top]
end
def encode_escape(ord)
ord.chr.force_encoding(@source_buffer.source.encoding)
end
def tok(s = @ts, e = @te)
@source_buffer.slice(s...e)
end
def range(s = @ts, e = @te)
Parser::Source::Range.new(@source_buffer, s, e)
end
def emit(type, value = tok, s = @ts, e = @te)
token = [ type, [ value, range(s, e) ] ]
@token_queue.push(token)
@tokens.push(token) if @tokens
token
end
def emit_table(table, s = @ts, e = @te)
value = tok(s, e)
emit(table[value], value, s, e)
end
def emit_do(do_block=false)
if @cond.active?
emit(:kDO_COND, 'do'.freeze)
elsif @cmdarg.active? || do_block
emit(:kDO_BLOCK, 'do'.freeze)
else
emit(:kDO, 'do'.freeze)
end
end
def arg_or_cmdarg(cmd_state)
if cmd_state
self.class.lex_en_expr_cmdarg
else
self.class.lex_en_expr_arg
end
end
def emit_comment(s = @ts, e = @te)
if @comments
@comments.push(Parser::Source::Comment.new(range(s, e)))
end
if @tokens
@tokens.push([ :tCOMMENT, [ tok(s, e), range(s, e) ] ])
end
nil
end
def diagnostic(type, reason, arguments=nil, location=range, highlights=[])
@diagnostics.process(
Parser::Diagnostic.new(type, reason, arguments, location, highlights))
end
#
# === LITERAL STACK ===
#
def push_literal(*args)
new_literal = Literal.new(self, *args)
@literal_stack.push(new_literal)
next_state_for_literal(new_literal)
end
def next_state_for_literal(literal)
if literal.words? && literal.backslash_delimited?
if literal.interpolate?
self.class.lex_en_interp_backslash_delimited_words
else
self.class.lex_en_plain_backslash_delimited_words
end
elsif literal.words? && !literal.backslash_delimited?
if literal.interpolate?
self.class.lex_en_interp_words
else
self.class.lex_en_plain_words
end
elsif !literal.words? && literal.backslash_delimited?
if literal.interpolate?
self.class.lex_en_interp_backslash_delimited
else
self.class.lex_en_plain_backslash_delimited
end
else
if literal.interpolate?
self.class.lex_en_interp_string
else
self.class.lex_en_plain_string
end
end
end
def literal
@literal_stack.last
end
def pop_literal
old_literal = @literal_stack.pop
@dedent_level = old_literal.dedent_level
if old_literal.type == :tREGEXP_BEG
# Fetch modifiers.
self.class.lex_en_regexp_modifiers
else
self.class.lex_en_expr_end
end
end
# Mapping of strings to parser tokens.
PUNCTUATION = {
'=' => :tEQL, '&' => :tAMPER2, '|' => :tPIPE,
'!' => :tBANG, '^' => :tCARET, '+' => :tPLUS,
'-' => :tMINUS, '*' => :tSTAR2, '/' => :tDIVIDE,
'%' => :tPERCENT, '~' => :tTILDE, ',' => :tCOMMA,
';' => :tSEMI, '.' => :tDOT, '..' => :tDOT2,
'...' => :tDOT3, '[' => :tLBRACK2, ']' => :tRBRACK,
'(' => :tLPAREN2, ')' => :tRPAREN, '?' => :tEH,
':' => :tCOLON, '&&' => :tANDOP, '||' => :tOROP,
'-@' => :tUMINUS, '+@' => :tUPLUS, '~@' => :tTILDE,
'**' => :tPOW, '->' => :tLAMBDA, '=~' => :tMATCH,
'!~' => :tNMATCH, '==' => :tEQ, '!=' => :tNEQ,
'>' => :tGT, '>>' => :tRSHFT, '>=' => :tGEQ,
'<' => :tLT, '<<' => :tLSHFT, '<=' => :tLEQ,
'=>' => :tASSOC, '::' => :tCOLON2, '===' => :tEQQ,
'<=>' => :tCMP, '[]' => :tAREF, '[]=' => :tASET,
'{' => :tLCURLY, '}' => :tRCURLY, '`' => :tBACK_REF2,
'!@' => :tBANG, '&.' => :tANDDOT,
}
PUNCTUATION_BEGIN = {
'&' => :tAMPER, '*' => :tSTAR, '**' => :tDSTAR,
'+' => :tUPLUS, '-' => :tUMINUS, '::' => :tCOLON3,
'(' => :tLPAREN, '{' => :tLBRACE, '[' => :tLBRACK,
}
KEYWORDS = {
'if' => :kIF_MOD, 'unless' => :kUNLESS_MOD,
'while' => :kWHILE_MOD, 'until' => :kUNTIL_MOD,
'rescue' => :kRESCUE_MOD, 'defined?' => :kDEFINED,
'BEGIN' => :klBEGIN, 'END' => :klEND,
}
KEYWORDS_BEGIN = {
'if' => :kIF, 'unless' => :kUNLESS,
'while' => :kWHILE, 'until' => :kUNTIL,
'rescue' => :kRESCUE, 'defined?' => :kDEFINED,
'BEGIN' => :klBEGIN, 'END' => :klEND,
}
%w(class module def undef begin end then elsif else ensure case when
for break next redo retry in do return yield super self nil true
false and or not alias __FILE__ __LINE__ __ENCODING__).each do |keyword|
KEYWORDS_BEGIN[keyword] = KEYWORDS[keyword] = :"k#{keyword.upcase}"
end
%%{
# %
access @;
getkey (@source_pts[p] || 0);
# === CHARACTER CLASSES ===
#
# Pay close attention to the differences between c_any and any.
# c_any does not include EOF and so will cause incorrect behavior
# for machine subtraction (any-except rules) and default transitions
# for scanners.
action do_nl {
# Record position of a newline for precise location reporting on tNL
# tokens.
#
# This action is embedded directly into c_nl, as it is idempotent and
# there are no cases when we need to skip it.
@newline_s = p
}
c_nl = '\n' $ do_nl;
c_space = [ \t\r\f\v];
c_space_nl = c_space | c_nl;
c_eof = 0x04 | 0x1a | 0 | zlen; # ^D, ^Z, \0, EOF
c_eol = c_nl | c_eof;
c_any = any - c_eof;
c_nl_zlen = c_nl | zlen;
c_line = any - c_nl_zlen;
c_unicode = c_any - 0x00..0x7f;
c_upper = [A-Z];
c_lower = [a-z_] | c_unicode;
c_alpha = c_lower | c_upper;
c_alnum = c_alpha | [0-9];
action do_eof {
# Sit at EOF indefinitely. #advance would return $eof each time.
# This allows to feed the lexer more data if needed; this is only used
# in tests.
#
# Note that this action is not embedded into e_eof like e_heredoc_nl and e_bs
# below. This is due to the fact that scanner state at EOF is observed
# by tests, and encapsulating it in a rule would break the introspection.
fhold; fbreak;
}
#
# === TOKEN DEFINITIONS ===
#
# All operators are punctuation. There is more to punctuation
# than just operators. Operators can be overridden by user;
# punctuation can not.
# A list of operators which are valid in the function name context, but
# have different semantics in others.
operator_fname = '[]' | '[]=' | '`' | '-@' | '+@' | '~@' | '!@' ;
# A list of operators which can occur within an assignment shortcut (+ → +=).
operator_arithmetic = '&' | '|' | '&&' | '||' | '^' | '+' | '-' |
'*' | '/' | '**' | '~' | '<<' | '>>' | '%' ;
# A list of all user-definable operators not covered by groups above.
operator_rest = '=~' | '!~' | '==' | '!=' | '!' | '===' |
'<' | '<=' | '>' | '>=' | '<=>' | '=>' ;
# Note that `{` and `}` need to be referred to as e_lbrace and e_rbrace,
# as they are ambiguous with interpolation `#{}` and should be counted.
# These braces are not present in punctuation lists.
# A list of punctuation which has different meaning when used at the
# beginning of expression.
punctuation_begin = '-' | '+' | '::' | '(' | '[' |
'*' | '**' | '&' ;
# A list of all punctuation except punctuation_begin.
punctuation_end = ',' | '=' | '->' | '(' | '[' | ']' |
'::' | '?' | ':' | '.' | '..' | '...' ;
# A list of keywords which have different meaning at the beginning of expression.
keyword_modifier = 'if' | 'unless' | 'while' | 'until' | 'rescue' ;
# A list of keywords which accept an argument-like expression, i.e. have the
# same post-processing as method calls or commands. Example: `yield 1`,
# `yield (1)`, `yield(1)`, are interpreted as if `yield` was a function.
keyword_with_arg = 'yield' | 'super' | 'not' | 'defined?' ;
# A list of keywords which accept a literal function name as an argument.
keyword_with_fname = 'def' | 'undef' | 'alias' ;
# A list of keywords which accept an expression after them.
keyword_with_value = 'else' | 'case' | 'ensure' | 'module' | 'elsif' | 'then' |
'for' | 'in' | 'do' | 'when' | 'begin' | 'class' |
'and' | 'or' ;
# A list of keywords which accept a value, and treat the keywords from
# `keyword_modifier` list as modifiers.
keyword_with_mid = 'rescue' | 'return' | 'break' | 'next' ;
# A list of keywords which do not accept an expression after them.
keyword_with_end = 'end' | 'self' | 'true' | 'false' | 'retry' |
'redo' | 'nil' | 'BEGIN' | 'END' | '__FILE__' |
'__LINE__' | '__ENCODING__';
# All keywords.
keyword = keyword_with_value | keyword_with_mid |
keyword_with_end | keyword_with_arg |
keyword_with_fname | keyword_modifier ;
constant = c_upper c_alnum*;
bareword = c_alpha c_alnum*;
call_or_var = c_lower c_alnum*;
class_var = '@@' bareword;
instance_var = '@' bareword;
global_var = '$'
( bareword | digit+
| [`'+~*$&?!@/\\;,.=:<>"] # `
| '-' c_alnum
)
;
# Ruby accepts (and fails on) variables with leading digit
# in literal context, but not in unquoted symbol body.
class_var_v = '@@' c_alnum+;
instance_var_v = '@' c_alnum+;
label = bareword [?!]? ':';
#
# === NUMERIC PARSING ===
#
int_hex = ( xdigit+ '_' )* xdigit* '_'? ;
int_dec = ( digit+ '_' )* digit* '_'? ;
int_bin = ( [01]+ '_' )* [01]* '_'? ;
flo_int = [1-9] [0-9]* ( '_' digit+ )* | '0';
flo_frac = '.' ( digit+ '_' )* digit+;
flo_pow = [eE] [+\-]? ( digit+ '_' )* digit+;
int_suffix =
'' % { @num_xfrm = lambda { |chars| emit(:tINTEGER, chars) } }
| 'r' % { @num_xfrm = lambda { |chars| emit(:tRATIONAL, Rational(chars)) } }
| 'i' % { @num_xfrm = lambda { |chars| emit(:tIMAGINARY, Complex(0, chars)) } }
| 'ri' % { @num_xfrm = lambda { |chars| emit(:tIMAGINARY, Complex(0, Rational(chars))) } }
| 're' % { @num_xfrm = lambda { |chars| emit(:tINTEGER, chars, @ts, @te - 2); p -= 2 } }
| 'if' % { @num_xfrm = lambda { |chars| emit(:tINTEGER, chars, @ts, @te - 2); p -= 2 } }
| 'rescue' % { @num_xfrm = lambda { |chars| emit(:tINTEGER, chars, @ts, @te - 6); p -= 6 } };
flo_pow_suffix =
'' % { @num_xfrm = lambda { |chars| emit(:tFLOAT, Float(chars)) } }
| 'i' % { @num_xfrm = lambda { |chars| emit(:tIMAGINARY, Complex(0, Float(chars))) } }
| 'if' % { @num_xfrm = lambda { |chars| emit(:tFLOAT, Float(chars), @ts, @te - 2); p -= 2 } };
flo_suffix =
flo_pow_suffix
| 'r' % { @num_xfrm = lambda { |chars| emit(:tRATIONAL, Rational(chars)) } }
| 'ri' % { @num_xfrm = lambda { |chars| emit(:tIMAGINARY, Complex(0, Rational(chars))) } }
| 'rescue' % { @num_xfrm = lambda { |chars| emit(:tFLOAT, Float(chars), @ts, @te - 6); p -= 6 } };
#
# === ESCAPE SEQUENCE PARSING ===
#
# Escape parsing code is a Ragel pattern, not a scanner, and therefore
# it shouldn't directly raise errors or perform other actions with side effects.
# In reality this would probably just mess up error reporting in pathological
# cases, through.
# The amount of code required to parse \M\C stuff correctly is ridiculous.
escaped_nl = "\\" c_nl;
action unicode_points {
@escape = ""
codepoints = tok(@escape_s + 2, p - 1)
codepoint_s = @escape_s + 2
if @version < 24
if codepoints.start_with?(" ") || codepoints.start_with?("\t")
diagnostic :fatal, :invalid_unicode_escape, nil,
range(@escape_s + 2, @escape_s + 3)
end
if spaces_p = codepoints.index(/[ \t]{2}/)
diagnostic :fatal, :invalid_unicode_escape, nil,
range(codepoint_s + spaces_p + 1, codepoint_s + spaces_p + 2)
end
if codepoints.end_with?(" ") || codepoints.end_with?("\t")
diagnostic :fatal, :invalid_unicode_escape, nil, range(p - 1, p)
end
end
codepoints.scan(/([0-9a-fA-F]+)|([ \t]+)/).each do |(codepoint_str, spaces)|
if spaces
codepoint_s += spaces.length
else
codepoint = codepoint_str.to_i(16)
if codepoint >= 0x110000
diagnostic :error, :unicode_point_too_large, nil,
range(codepoint_s, codepoint_s + codepoint_str.length)
break
end
@escape += codepoint.chr(Encoding::UTF_8)
codepoint_s += codepoint_str.length
end
end
}
action unescape_char {
codepoint = @source_pts[p - 1]
if (@escape = ESCAPES[codepoint]).nil?
@escape = encode_escape(@source_buffer.slice(p - 1))
end
}
action invalid_complex_escape {
diagnostic :fatal, :invalid_escape
}
action read_post_meta_or_ctrl_char {
@escape = @source_buffer.slice(p - 1).chr
if @version >= 27 && ((0..8).include?(@escape.ord) || (14..31).include?(@escape.ord))
diagnostic :fatal, :invalid_escape
end
}
action slash_c_char {
@escape = encode_escape(@escape[0].ord & 0x9f)
}
action slash_m_char {
@escape = encode_escape(@escape[0].ord | 0x80)
}
maybe_escaped_char = (
'\\' c_any %unescape_char
| ( c_any - [\\] ) %read_post_meta_or_ctrl_char
);
maybe_escaped_ctrl_char = ( # why?!
'\\' c_any %unescape_char %slash_c_char
| '?' % { @escape = "\x7f" }
| ( c_any - [\\?] ) %read_post_meta_or_ctrl_char %slash_c_char
);
escape = (
# \377
[0-7]{1,3}
% { @escape = encode_escape(tok(@escape_s, p).to_i(8) % 0x100) }
# \xff
| 'x' xdigit{1,2}
% { @escape = encode_escape(tok(@escape_s + 1, p).to_i(16)) }
# %q[\x]
| 'x' ( c_any - xdigit )
% {
diagnostic :fatal, :invalid_hex_escape, nil, range(@escape_s - 1, p + 2)
}
# \u263a
| 'u' xdigit{4}
% { @escape = tok(@escape_s + 1, p).to_i(16).chr(Encoding::UTF_8) }
# \u123
| 'u' xdigit{0,3}
% {
diagnostic :fatal, :invalid_unicode_escape, nil, range(@escape_s - 1, p)
}
# u{not hex} or u{}
| 'u{' ( c_any - xdigit - [ \t}] )* '}'
% {
diagnostic :fatal, :invalid_unicode_escape, nil, range(@escape_s - 1, p)
}
# \u{ \t 123 \t 456 \t\t }
| 'u{' [ \t]* ( xdigit{1,6} [ \t]+ )*
(
( xdigit{1,6} [ \t]* '}'
%unicode_points
)
|
( xdigit* ( c_any - xdigit - [ \t}] )+ '}'
| ( c_any - [ \t}] )* c_eof
| xdigit{7,}
) % {
diagnostic :fatal, :unterminated_unicode, nil, range(p - 1, p)
}
)
# \C-\a \cx
| ( 'C-' | 'c' ) escaped_nl?
maybe_escaped_ctrl_char
# \M-a
| 'M-' escaped_nl?
maybe_escaped_char
%slash_m_char
# \C-\M-f \M-\cf \c\M-f
| ( ( 'C-' | 'c' ) escaped_nl? '\\M-'
| 'M-\\' escaped_nl? ( 'C-' | 'c' ) ) escaped_nl?
maybe_escaped_ctrl_char
%slash_m_char
| 'C' c_any %invalid_complex_escape
| 'M' c_any %invalid_complex_escape
| ( 'M-\\C' | 'C-\\M' ) c_any %invalid_complex_escape
| ( c_any - [0-7xuCMc] ) %unescape_char
| c_eof % {
diagnostic :fatal, :escape_eof, nil, range(p - 1, p)
}
);
# Use rules in form of `e_bs escape' when you need to parse a sequence.
e_bs = '\\' % {
@escape_s = p
@escape = nil
};
#
# === STRING AND HEREDOC PARSING ===
#
# Heredoc parsing is quite a complex topic. First, consider that heredocs
# can be arbitrarily nested. For example:
#
# puts <<CODE
# the result is: #{<<RESULT.inspect
# i am a heredoc
# RESULT
# }
# CODE
#
# which, incidentally, evaluates to:
#
# the result is: " i am a heredoc\n"
#
# To parse them, lexer refers to two kinds (remember, nested heredocs)
# of positions in the input stream, namely heredoc_e
# (HEREDOC declaration End) and @herebody_s (HEREdoc BODY line Start).
#
# heredoc_e is simply contained inside the corresponding Literal, and
# when the heredoc is closed, the lexing is restarted from that position.
#
# @herebody_s is quite more complex. First, @herebody_s changes after each
# heredoc line is lexed. This way, at '\n' tok(@herebody_s, @te) always
# contains the current line, and also when a heredoc is started, @herebody_s
# contains the position from which the heredoc will be lexed.
#
# Second, as (insanity) there are nested heredocs, we need to maintain a
# stack of these positions. Each time #push_literal is called, it saves current
# @heredoc_s to literal.saved_herebody_s, and after an interpolation (possibly
# containing another heredocs) is closed, the previous value is restored.
e_heredoc_nl = c_nl % {
# After every heredoc was parsed, @herebody_s contains the
# position of next token after all heredocs.
if @herebody_s
p = @herebody_s
@herebody_s = nil
end
};
action extend_string {
string = tok
# tLABEL_END is only possible in non-cond context on >= 2.2
if @version >= 22 && !@cond.active?
lookahead = @source_buffer.slice(@te...@te+2)
end
current_literal = literal
if !current_literal.heredoc? &&
(token = current_literal.nest_and_try_closing(string, @ts, @te, lookahead))
if token[0] == :tLABEL_END
p += 1
pop_literal
fnext expr_labelarg;
else
fnext *pop_literal;
end
fbreak;
else
current_literal.extend_string(string, @ts, @te)
end
}
action extend_string_escaped {
current_literal = literal
# Get the first character after the backslash.
escaped_char = @source_buffer.slice(@escape_s).chr
if current_literal.munge_escape? escaped_char
# If this particular literal uses this character as an opening
# or closing delimiter, it is an escape sequence for that
# particular character. Write it without the backslash.
if current_literal.regexp? && REGEXP_META_CHARACTERS.match(escaped_char)
# Regular expressions should include escaped delimiters in their
# escaped form, except when the escaped character is
# a closing delimiter but not a regexp metacharacter.
#
# The backslash itself cannot be used as a closing delimiter
# at the same time as an escape symbol, but it is always munged,
# so this branch also executes for the non-closing-delimiter case
# for the backslash.
current_literal.extend_string(tok, @ts, @te)
else
current_literal.extend_string(escaped_char, @ts, @te)
end
else
# It does not. So this is an actual escape sequence, yay!
if current_literal.squiggly_heredoc? && escaped_char == "\n".freeze
# Squiggly heredocs like
# <<~-HERE
# 1\
# 2
# HERE
# treat '\' as a line continuation, but still dedent the body, so the heredoc above becomes "12\n".
# This information is emitted as is, without escaping,
# later this escape sequence (\\\n) gets handled manually in the Lexer::Dedenter
current_literal.extend_string(tok, @ts, @te)
elsif current_literal.supports_line_continuation_via_slash? && escaped_char == "\n".freeze
# Heredocs, regexp and a few other types of literals support line
# continuation via \\\n sequence. The code like
# "a\
# b"
# must be parsed as "ab"
current_literal.extend_string(tok.gsub("\\\n".freeze, ''.freeze), @ts, @te)
elsif current_literal.regexp?
# Regular expressions should include escape sequences in their
# escaped form. On the other hand, escaped newlines are removed (in cases like "\\C-\\\n\\M-x")
current_literal.extend_string(tok.gsub("\\\n".freeze, ''.freeze), @ts, @te)
else
current_literal.extend_string(@escape || tok, @ts, @te)
end
end
}
# Extend a string with a newline or a EOF character.
# As heredoc closing line can immediately precede EOF, this action
# has to handle such case specially.
action extend_string_eol {
current_literal = literal
if @te == pe
diagnostic :fatal, :string_eof, nil,
range(current_literal.str_s, current_literal.str_s + 1)
end
if current_literal.heredoc?
line = tok(@herebody_s, @ts).gsub(/\r+$/, ''.freeze)
if version?(18, 19, 20)
# See ruby:c48b4209c
line = line.gsub(/\r.*$/, ''.freeze)
end
# Try ending the heredoc with the complete most recently
# scanned line. @herebody_s always refers to the start of such line.
if current_literal.nest_and_try_closing(line, @herebody_s, @ts)
# Adjust @herebody_s to point to the next line.
@herebody_s = @te
# Continue regular lexing after the heredoc reference (<<END).
p = current_literal.heredoc_e - 1
fnext *pop_literal; fbreak;
else
# Calculate indentation level for <<~HEREDOCs.
current_literal.infer_indent_level(line)
# Ditto.
@herebody_s = @te
end
else
# Try ending the literal with a newline.
if current_literal.nest_and_try_closing(tok, @ts, @te)
fnext *pop_literal; fbreak;
end
if @herebody_s
# This is a regular literal intertwined with a heredoc. Like:
#
# p <<-foo+"1
# bar
# foo
# 2"
#
# which, incidentally, evaluates to "bar\n1\n2".
p = @herebody_s - 1
@herebody_s = nil
end
end
if current_literal.words? && !eof_codepoint?(@source_pts[p])
current_literal.extend_space @ts, @te
else
# A literal newline is appended if the heredoc was _not_ closed
# this time (see fbreak above). See also Literal#nest_and_try_closing
# for rationale of calling #flush_string here.
current_literal.extend_string tok, @ts, @te