ruby-next-parser 3.1.1.3 → 3.2.2.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/lib/parser/ruby-next/lexer-F0.rb +12957 -0
- data/lib/parser/ruby-next/lexer-F1.rb +14968 -0
- data/lib/parser/ruby-next/lexer.rl +201 -739
- data/lib/parser/ruby-next/parser_ext.rb +6 -1
- data/lib/parser/ruby-next/version.rb +1 -1
- data/lib/parser/rubynext.rb +8151 -3791
- data/lib/parser/rubynext.y +36 -15
- metadata +5 -4
- data/lib/parser/ruby-next/lexer.rb +0 -25177
@@ -82,14 +82,6 @@ class Next
|
|
82
82
|
%% write data nofinal;
|
83
83
|
# %
|
84
84
|
|
85
|
-
ESCAPES = {
|
86
|
-
?a.ord => "\a", ?b.ord => "\b", ?e.ord => "\e", ?f.ord => "\f",
|
87
|
-
?n.ord => "\n", ?r.ord => "\r", ?s.ord => "\s", ?t.ord => "\t",
|
88
|
-
?v.ord => "\v", ?\\.ord => "\\"
|
89
|
-
}.freeze
|
90
|
-
|
91
|
-
REGEXP_META_CHARACTERS = Regexp.union(*"\\$()*+.<>?[]^{|}".chars).freeze
|
92
|
-
|
93
85
|
attr_reader :source_buffer
|
94
86
|
|
95
87
|
attr_accessor :diagnostics
|
@@ -100,7 +92,7 @@ class Next
|
|
100
92
|
|
101
93
|
attr_accessor :tokens, :comments
|
102
94
|
|
103
|
-
attr_reader :paren_nest, :cmdarg_stack, :cond_stack, :lambda_stack
|
95
|
+
attr_reader :paren_nest, :cmdarg_stack, :cond_stack, :lambda_stack, :version
|
104
96
|
|
105
97
|
def initialize(version)
|
106
98
|
@version = version
|
@@ -110,6 +102,26 @@ class Next
|
|
110
102
|
@tokens = nil
|
111
103
|
@comments = nil
|
112
104
|
|
105
|
+
@_lex_actions =
|
106
|
+
if self.class.respond_to?(:_lex_actions, true)
|
107
|
+
self.class.send :_lex_actions
|
108
|
+
else
|
109
|
+
[]
|
110
|
+
end
|
111
|
+
|
112
|
+
@emit_integer = lambda { |chars, p| emit(:tINTEGER, chars); p }
|
113
|
+
@emit_rational = lambda { |chars, p| emit(:tRATIONAL, Rational(chars)); p }
|
114
|
+
@emit_imaginary = lambda { |chars, p| emit(:tIMAGINARY, Complex(0, chars)); p }
|
115
|
+
@emit_imaginary_rational = lambda { |chars, p| emit(:tIMAGINARY, Complex(0, Rational(chars))); p }
|
116
|
+
@emit_integer_re = lambda { |chars, p| emit(:tINTEGER, chars, @ts, @te - 2); p - 2 }
|
117
|
+
@emit_integer_if = lambda { |chars, p| emit(:tINTEGER, chars, @ts, @te - 2); p - 2 }
|
118
|
+
@emit_integer_rescue = lambda { |chars, p| emit(:tINTEGER, chars, @ts, @te - 6); p - 6 }
|
119
|
+
|
120
|
+
@emit_float = lambda { |chars, p| emit(:tFLOAT, Float(chars)); p }
|
121
|
+
@emit_imaginary_float = lambda { |chars, p| emit(:tIMAGINARY, Complex(0, Float(chars))); p }
|
122
|
+
@emit_float_if = lambda { |chars, p| emit(:tFLOAT, Float(chars), @ts, @te - 2); p - 2 }
|
123
|
+
@emit_float_rescue = lambda { |chars, p| emit(:tFLOAT, Float(chars), @ts, @te - 6); p - 6 }
|
124
|
+
|
113
125
|
reset
|
114
126
|
end
|
115
127
|
|
@@ -139,7 +151,6 @@ class Next
|
|
139
151
|
|
140
152
|
# Lexer state:
|
141
153
|
@token_queue = []
|
142
|
-
@literal_stack = []
|
143
154
|
|
144
155
|
@eq_begin_s = nil # location of last encountered =begin
|
145
156
|
@sharp_s = nil # location of last encountered #
|
@@ -151,34 +162,20 @@ class Next
|
|
151
162
|
@num_suffix_s = nil # starting position of numeric suffix
|
152
163
|
@num_xfrm = nil # numeric suffix-induced transformation
|
153
164
|
|
154
|
-
@escape_s = nil # starting position of current sequence
|
155
|
-
@escape = nil # last escaped sequence, as string
|
156
|
-
|
157
|
-
@herebody_s = nil # starting position of current heredoc line
|
158
|
-
|
159
165
|
# Ruby 1.9 ->() lambdas emit a distinct token if do/{ is
|
160
166
|
# encountered after a matching closing parenthesis.
|
161
167
|
@paren_nest = 0
|
162
168
|
@lambda_stack = []
|
163
169
|
|
164
|
-
# After encountering the closing line of <<~SQUIGGLY_HEREDOC,
|
165
|
-
# we store the indentation level and give it out to the parser
|
166
|
-
# on request. It is not possible to infer indentation level just
|
167
|
-
# from the AST because escape sequences such as `\ ` or `\t` are
|
168
|
-
# expanded inside the lexer, but count as non-whitespace for
|
169
|
-
# indentation purposes.
|
170
|
-
@dedent_level = nil
|
171
|
-
|
172
170
|
# If the lexer is in `command state' (aka expr_value)
|
173
171
|
# at the entry to #advance, it will transition to expr_cmdarg
|
174
172
|
# instead of expr_arg at certain points.
|
175
173
|
@command_start = true
|
176
174
|
|
177
|
-
# True at the end of "def foo a:"
|
178
|
-
@in_kwarg = false
|
179
|
-
|
180
175
|
# State before =begin / =end block comment
|
181
176
|
@cs_before_block_comment = self.class.lex_en_line_begin
|
177
|
+
|
178
|
+
@strings = Parser::LexerStrings.new(self, @version)
|
182
179
|
end
|
183
180
|
|
184
181
|
def source_buffer=(source_buffer)
|
@@ -200,6 +197,9 @@ class Next
|
|
200
197
|
else
|
201
198
|
@source_pts = nil
|
202
199
|
end
|
200
|
+
|
201
|
+
@strings.source_buffer = @source_buffer
|
202
|
+
@strings.source_pts = @source_pts
|
203
203
|
end
|
204
204
|
|
205
205
|
def encoding
|
@@ -220,10 +220,7 @@ class Next
|
|
220
220
|
:expr_endfn => lex_en_expr_endfn,
|
221
221
|
:expr_labelarg => lex_en_expr_labelarg,
|
222
222
|
|
223
|
-
:
|
224
|
-
:interp_words => lex_en_interp_words,
|
225
|
-
:plain_string => lex_en_plain_string,
|
226
|
-
:plain_words => lex_en_plain_string,
|
223
|
+
:inside_string => lex_en_inside_string
|
227
224
|
}
|
228
225
|
|
229
226
|
def state
|
@@ -253,15 +250,12 @@ class Next
|
|
253
250
|
end
|
254
251
|
|
255
252
|
def dedent_level
|
256
|
-
|
257
|
-
# using a stale value.
|
258
|
-
dedent_level, @dedent_level = @dedent_level, nil
|
259
|
-
dedent_level
|
253
|
+
@strings.dedent_level
|
260
254
|
end
|
261
255
|
|
262
256
|
# Return next token: [type, value].
|
263
257
|
def advance
|
264
|
-
|
258
|
+
unless @token_queue.empty?
|
265
259
|
return @token_queue.shift
|
266
260
|
end
|
267
261
|
|
@@ -276,6 +270,7 @@ class Next
|
|
276
270
|
_lex_to_state_actions = klass.send :_lex_to_state_actions
|
277
271
|
_lex_from_state_actions = klass.send :_lex_from_state_actions
|
278
272
|
_lex_eof_trans = klass.send :_lex_eof_trans
|
273
|
+
_lex_actions = @_lex_actions
|
279
274
|
|
280
275
|
pe = @source_pts.size + 2
|
281
276
|
p, eof = @p, pe
|
@@ -307,10 +302,6 @@ class Next
|
|
307
302
|
|
308
303
|
protected
|
309
304
|
|
310
|
-
def eof_codepoint?(point)
|
311
|
-
[0x04, 0x1a, 0x00].include? point
|
312
|
-
end
|
313
|
-
|
314
305
|
def version?(*versions)
|
315
306
|
versions.include?(@version)
|
316
307
|
end
|
@@ -320,12 +311,8 @@ class Next
|
|
320
311
|
@stack[@top]
|
321
312
|
end
|
322
313
|
|
323
|
-
def encode_escape(ord)
|
324
|
-
ord.chr.force_encoding(@source_buffer.source.encoding)
|
325
|
-
end
|
326
|
-
|
327
314
|
def tok(s = @ts, e = @te)
|
328
|
-
@source_buffer.slice(s
|
315
|
+
@source_buffer.slice(s, e - s)
|
329
316
|
end
|
330
317
|
|
331
318
|
def range(s = @ts, e = @te)
|
@@ -378,64 +365,110 @@ class Next
|
|
378
365
|
nil
|
379
366
|
end
|
380
367
|
|
368
|
+
def emit_comment_from_range(p, pe)
|
369
|
+
emit_comment(@sharp_s, p == pe ? p - 2 : p)
|
370
|
+
end
|
371
|
+
|
381
372
|
def diagnostic(type, reason, arguments=nil, location=range, highlights=[])
|
382
373
|
@diagnostics.process(
|
383
374
|
Parser::Diagnostic.new(type, reason, arguments, location, highlights))
|
384
375
|
end
|
385
376
|
|
386
|
-
#
|
387
|
-
# === LITERAL STACK ===
|
388
|
-
#
|
389
377
|
|
390
|
-
def
|
391
|
-
|
392
|
-
|
393
|
-
|
378
|
+
def e_lbrace
|
379
|
+
@cond.push(false); @cmdarg.push(false)
|
380
|
+
|
381
|
+
current_literal = @strings.literal
|
382
|
+
if current_literal
|
383
|
+
current_literal.start_interp_brace
|
384
|
+
end
|
394
385
|
end
|
395
386
|
|
396
|
-
def
|
397
|
-
|
398
|
-
|
399
|
-
|
400
|
-
|
401
|
-
|
402
|
-
|
403
|
-
|
404
|
-
|
405
|
-
|
406
|
-
|
407
|
-
|
408
|
-
|
409
|
-
|
410
|
-
|
411
|
-
|
387
|
+
def numeric_literal_int
|
388
|
+
digits = tok(@num_digits_s, @num_suffix_s)
|
389
|
+
|
390
|
+
if digits.end_with? '_'.freeze
|
391
|
+
diagnostic :error, :trailing_in_number, { :character => '_'.freeze },
|
392
|
+
range(@te - 1, @te)
|
393
|
+
elsif digits.empty? && @num_base == 8 && version?(18)
|
394
|
+
# 1.8 did not raise an error on 0o.
|
395
|
+
digits = '0'.freeze
|
396
|
+
elsif digits.empty?
|
397
|
+
diagnostic :error, :empty_numeric
|
398
|
+
elsif @num_base == 8 && (invalid_idx = digits.index(/[89]/))
|
399
|
+
invalid_s = @num_digits_s + invalid_idx
|
400
|
+
diagnostic :error, :invalid_octal, nil,
|
401
|
+
range(invalid_s, invalid_s + 1)
|
402
|
+
end
|
403
|
+
digits
|
404
|
+
end
|
405
|
+
|
406
|
+
def on_newline(p)
|
407
|
+
@strings.on_newline(p)
|
408
|
+
end
|
409
|
+
|
410
|
+
def check_ambiguous_slash(tm)
|
411
|
+
if tok(tm, tm + 1) == '/'.freeze
|
412
|
+
# Ambiguous regexp literal.
|
413
|
+
if @version < 30
|
414
|
+
diagnostic :warning, :ambiguous_literal, nil, range(tm, tm + 1)
|
412
415
|
else
|
413
|
-
|
416
|
+
diagnostic :warning, :ambiguous_regexp, nil, range(tm, tm + 1)
|
414
417
|
end
|
418
|
+
end
|
419
|
+
end
|
420
|
+
|
421
|
+
def emit_global_var(ts = @ts, te = @te)
|
422
|
+
if tok(ts, te) =~ /^\$([1-9][0-9]*)$/
|
423
|
+
emit(:tNTH_REF, tok(ts + 1, te).to_i, ts, te)
|
424
|
+
elsif tok =~ /^\$([&`'+])$/
|
425
|
+
emit(:tBACK_REF, tok(ts, te), ts, te)
|
415
426
|
else
|
416
|
-
|
417
|
-
self.class.lex_en_interp_string
|
418
|
-
else
|
419
|
-
self.class.lex_en_plain_string
|
420
|
-
end
|
427
|
+
emit(:tGVAR, tok(ts, te), ts, te)
|
421
428
|
end
|
422
429
|
end
|
423
430
|
|
424
|
-
def
|
425
|
-
|
431
|
+
def emit_class_var(ts = @ts, te = @te)
|
432
|
+
if tok(ts, te) =~ /^@@[0-9]/
|
433
|
+
diagnostic :error, :cvar_name, { :name => tok(ts, te) }
|
434
|
+
end
|
435
|
+
|
436
|
+
emit(:tCVAR, tok(ts, te), ts, te)
|
426
437
|
end
|
427
438
|
|
428
|
-
def
|
429
|
-
|
439
|
+
def emit_instance_var(ts = @ts, te = @te)
|
440
|
+
if tok(ts, te) =~ /^@[0-9]/
|
441
|
+
diagnostic :error, :ivar_name, { :name => tok(ts, te) }
|
442
|
+
end
|
443
|
+
|
444
|
+
emit(:tIVAR, tok(ts, te), ts, te)
|
445
|
+
end
|
446
|
+
|
447
|
+
def emit_rbrace_rparen_rbrack
|
448
|
+
emit_table(PUNCTUATION)
|
430
449
|
|
431
|
-
@
|
450
|
+
if @version < 24
|
451
|
+
@cond.lexpop
|
452
|
+
@cmdarg.lexpop
|
453
|
+
else
|
454
|
+
@cond.pop
|
455
|
+
@cmdarg.pop
|
456
|
+
end
|
457
|
+
end
|
432
458
|
|
433
|
-
|
434
|
-
|
435
|
-
|
459
|
+
def emit_colon_with_digits(p, tm, diag_msg)
|
460
|
+
if @version >= 27
|
461
|
+
diagnostic :error, diag_msg, { name: tok(tm, @te) }, range(tm, @te)
|
436
462
|
else
|
437
|
-
|
463
|
+
emit(:tCOLON, tok(@ts, @ts + 1), @ts, @ts + 1)
|
464
|
+
p = @ts
|
438
465
|
end
|
466
|
+
p
|
467
|
+
end
|
468
|
+
|
469
|
+
def emit_singleton_class
|
470
|
+
emit(:kCLASS, 'class'.freeze, @ts, @ts + 5)
|
471
|
+
emit(:tLSHFT, '<<'.freeze, @te - 2, @te)
|
439
472
|
end
|
440
473
|
|
441
474
|
# Mapping of strings to parser tokens.
|
@@ -457,7 +490,7 @@ class Next
|
|
457
490
|
'=>' => :tASSOC, '::' => :tCOLON2, '===' => :tEQQ,
|
458
491
|
'<=>' => :tCMP, '[]' => :tAREF, '[]=' => :tASET,
|
459
492
|
'{' => :tLCURLY, '}' => :tRCURLY, '`' => :tBACK_REF2,
|
460
|
-
'!@' => :tBANG, '&.' => :tANDDOT,
|
493
|
+
'!@' => :tBANG, '&.' => :tANDDOT, '.:' => :tMETHREF,
|
461
494
|
}
|
462
495
|
|
463
496
|
PUNCTUATION_BEGIN = {
|
@@ -480,6 +513,11 @@ class Next
|
|
480
513
|
'BEGIN' => :klBEGIN, 'END' => :klEND,
|
481
514
|
}
|
482
515
|
|
516
|
+
ESCAPE_WHITESPACE = {
|
517
|
+
" " => '\s', "\r" => '\r', "\n" => '\n', "\t" => '\t',
|
518
|
+
"\v" => '\v', "\f" => '\f'
|
519
|
+
}
|
520
|
+
|
483
521
|
%w(class module def undef begin end then elsif else ensure case when
|
484
522
|
for break next redo retry in do return yield super self nil true
|
485
523
|
false and or not alias __FILE__ __LINE__ __ENCODING__).each do |keyword|
|
@@ -531,7 +569,7 @@ class Next
|
|
531
569
|
# This allows to feed the lexer more data if needed; this is only used
|
532
570
|
# in tests.
|
533
571
|
#
|
534
|
-
# Note that this action is not embedded into e_eof like
|
572
|
+
# Note that this action is not embedded into e_eof like e_nl and e_bs
|
535
573
|
# below. This is due to the fact that scanner state at EOF is observed
|
536
574
|
# by tests, and encapsulating it in a rule would break the introspection.
|
537
575
|
fhold; fbreak;
|
@@ -633,581 +671,43 @@ class Next
|
|
633
671
|
flo_pow = [eE] [+\-]? ( digit+ '_' )* digit+;
|
634
672
|
|
635
673
|
int_suffix =
|
636
|
-
'' % { @num_xfrm =
|
637
|
-
| 'r' % { @num_xfrm =
|
638
|
-
| 'i' % { @num_xfrm =
|
639
|
-
| 'ri' % { @num_xfrm =
|
640
|
-
| 're' % { @num_xfrm =
|
641
|
-
| 'if' % { @num_xfrm =
|
642
|
-
| 'rescue' % { @num_xfrm =
|
674
|
+
'' % { @num_xfrm = @emit_integer }
|
675
|
+
| 'r' % { @num_xfrm = @emit_rational }
|
676
|
+
| 'i' % { @num_xfrm = @emit_imaginary }
|
677
|
+
| 'ri' % { @num_xfrm = @emit_imaginary_rational }
|
678
|
+
| 're' % { @num_xfrm = @emit_integer_re }
|
679
|
+
| 'if' % { @num_xfrm = @emit_integer_if }
|
680
|
+
| 'rescue' % { @num_xfrm = @emit_integer_rescue };
|
643
681
|
|
644
682
|
flo_pow_suffix =
|
645
|
-
'' % { @num_xfrm =
|
646
|
-
| 'i' % { @num_xfrm =
|
647
|
-
| 'if' % { @num_xfrm =
|
683
|
+
'' % { @num_xfrm = @emit_float }
|
684
|
+
| 'i' % { @num_xfrm = @emit_imaginary_float }
|
685
|
+
| 'if' % { @num_xfrm = @emit_float_if };
|
648
686
|
|
649
687
|
flo_suffix =
|
650
688
|
flo_pow_suffix
|
651
|
-
| 'r' % { @num_xfrm =
|
652
|
-
| 'ri' % { @num_xfrm =
|
653
|
-
| 'rescue' % { @num_xfrm =
|
654
|
-
|
655
|
-
#
|
656
|
-
# === ESCAPE SEQUENCE PARSING ===
|
657
|
-
#
|
658
|
-
|
659
|
-
# Escape parsing code is a Ragel pattern, not a scanner, and therefore
|
660
|
-
# it shouldn't directly raise errors or perform other actions with side effects.
|
661
|
-
# In reality this would probably just mess up error reporting in pathological
|
662
|
-
# cases, through.
|
663
|
-
|
664
|
-
# The amount of code required to parse \M\C stuff correctly is ridiculous.
|
665
|
-
|
666
|
-
escaped_nl = "\\" c_nl;
|
667
|
-
|
668
|
-
action unicode_points {
|
669
|
-
@escape = ""
|
670
|
-
|
671
|
-
codepoints = tok(@escape_s + 2, p - 1)
|
672
|
-
codepoint_s = @escape_s + 2
|
673
|
-
|
674
|
-
if @version < 24
|
675
|
-
if codepoints.start_with?(" ") || codepoints.start_with?("\t")
|
676
|
-
diagnostic :fatal, :invalid_unicode_escape, nil,
|
677
|
-
range(@escape_s + 2, @escape_s + 3)
|
678
|
-
end
|
679
|
-
|
680
|
-
if spaces_p = codepoints.index(/[ \t]{2}/)
|
681
|
-
diagnostic :fatal, :invalid_unicode_escape, nil,
|
682
|
-
range(codepoint_s + spaces_p + 1, codepoint_s + spaces_p + 2)
|
683
|
-
end
|
684
|
-
|
685
|
-
if codepoints.end_with?(" ") || codepoints.end_with?("\t")
|
686
|
-
diagnostic :fatal, :invalid_unicode_escape, nil, range(p - 1, p)
|
687
|
-
end
|
688
|
-
end
|
689
|
-
|
690
|
-
codepoints.scan(/([0-9a-fA-F]+)|([ \t]+)/).each do |(codepoint_str, spaces)|
|
691
|
-
if spaces
|
692
|
-
codepoint_s += spaces.length
|
693
|
-
else
|
694
|
-
codepoint = codepoint_str.to_i(16)
|
695
|
-
|
696
|
-
if codepoint >= 0x110000
|
697
|
-
diagnostic :error, :unicode_point_too_large, nil,
|
698
|
-
range(codepoint_s, codepoint_s + codepoint_str.length)
|
699
|
-
break
|
700
|
-
end
|
701
|
-
|
702
|
-
@escape += codepoint.chr(Encoding::UTF_8)
|
703
|
-
codepoint_s += codepoint_str.length
|
704
|
-
end
|
705
|
-
end
|
706
|
-
}
|
707
|
-
|
708
|
-
action unescape_char {
|
709
|
-
codepoint = @source_pts[p - 1]
|
710
|
-
|
711
|
-
if @version >= 30 && (codepoint == 117 || codepoint == 85) # 'u' or 'U'
|
712
|
-
diagnostic :fatal, :invalid_escape
|
713
|
-
end
|
714
|
-
|
715
|
-
if (@escape = ESCAPES[codepoint]).nil?
|
716
|
-
@escape = encode_escape(@source_buffer.slice(p - 1))
|
717
|
-
end
|
718
|
-
}
|
719
|
-
|
720
|
-
action invalid_complex_escape {
|
721
|
-
diagnostic :fatal, :invalid_escape
|
722
|
-
}
|
723
|
-
|
724
|
-
action read_post_meta_or_ctrl_char {
|
725
|
-
@escape = @source_buffer.slice(p - 1).chr
|
726
|
-
|
727
|
-
if @version >= 27 && ((0..8).include?(@escape.ord) || (14..31).include?(@escape.ord))
|
728
|
-
diagnostic :fatal, :invalid_escape
|
729
|
-
end
|
730
|
-
}
|
731
|
-
|
732
|
-
action slash_c_char {
|
733
|
-
@escape = encode_escape(@escape[0].ord & 0x9f)
|
734
|
-
}
|
735
|
-
|
736
|
-
action slash_m_char {
|
737
|
-
@escape = encode_escape(@escape[0].ord | 0x80)
|
738
|
-
}
|
739
|
-
|
740
|
-
maybe_escaped_char = (
|
741
|
-
'\\' c_any %unescape_char
|
742
|
-
| '\\x' xdigit{1,2} % { @escape = encode_escape(tok(p - 2, p).to_i(16)) } %slash_c_char
|
743
|
-
| ( c_any - [\\] ) %read_post_meta_or_ctrl_char
|
744
|
-
);
|
745
|
-
|
746
|
-
maybe_escaped_ctrl_char = ( # why?!
|
747
|
-
'\\' c_any %unescape_char %slash_c_char
|
748
|
-
| '?' % { @escape = "\x7f" }
|
749
|
-
| '\\x' xdigit{1,2} % { @escape = encode_escape(tok(p - 2, p).to_i(16)) } %slash_c_char
|
750
|
-
| ( c_any - [\\?] ) %read_post_meta_or_ctrl_char %slash_c_char
|
751
|
-
);
|
752
|
-
|
753
|
-
escape = (
|
754
|
-
# \377
|
755
|
-
[0-7]{1,3}
|
756
|
-
% { @escape = encode_escape(tok(@escape_s, p).to_i(8) % 0x100) }
|
757
|
-
|
758
|
-
# \xff
|
759
|
-
| 'x' xdigit{1,2}
|
760
|
-
% { @escape = encode_escape(tok(@escape_s + 1, p).to_i(16)) }
|
761
|
-
|
762
|
-
# %q[\x]
|
763
|
-
| 'x' ( c_any - xdigit )
|
764
|
-
% {
|
765
|
-
diagnostic :fatal, :invalid_hex_escape, nil, range(@escape_s - 1, p + 2)
|
766
|
-
}
|
767
|
-
|
768
|
-
# \u263a
|
769
|
-
| 'u' xdigit{4}
|
770
|
-
% { @escape = tok(@escape_s + 1, p).to_i(16).chr(Encoding::UTF_8) }
|
771
|
-
|
772
|
-
# \u123
|
773
|
-
| 'u' xdigit{0,3}
|
774
|
-
% {
|
775
|
-
diagnostic :fatal, :invalid_unicode_escape, nil, range(@escape_s - 1, p)
|
776
|
-
}
|
777
|
-
|
778
|
-
# u{not hex} or u{}
|
779
|
-
| 'u{' ( c_any - xdigit - [ \t}] )* '}'
|
780
|
-
% {
|
781
|
-
diagnostic :fatal, :invalid_unicode_escape, nil, range(@escape_s - 1, p)
|
782
|
-
}
|
783
|
-
|
784
|
-
# \u{ \t 123 \t 456 \t\t }
|
785
|
-
| 'u{' [ \t]* ( xdigit{1,6} [ \t]+ )*
|
786
|
-
(
|
787
|
-
( xdigit{1,6} [ \t]* '}'
|
788
|
-
%unicode_points
|
789
|
-
)
|
790
|
-
|
|
791
|
-
( xdigit* ( c_any - xdigit - [ \t}] )+ '}'
|
792
|
-
| ( c_any - [ \t}] )* c_eof
|
793
|
-
| xdigit{7,}
|
794
|
-
) % {
|
795
|
-
diagnostic :fatal, :unterminated_unicode, nil, range(p - 1, p)
|
796
|
-
}
|
797
|
-
)
|
798
|
-
|
799
|
-
# \C-\a \cx
|
800
|
-
| ( 'C-' | 'c' ) escaped_nl?
|
801
|
-
maybe_escaped_ctrl_char
|
802
|
-
|
803
|
-
# \M-a
|
804
|
-
| 'M-' escaped_nl?
|
805
|
-
maybe_escaped_char
|
806
|
-
%slash_m_char
|
807
|
-
|
808
|
-
# \C-\M-f \M-\cf \c\M-f
|
809
|
-
| ( ( 'C-' | 'c' ) escaped_nl? '\\M-'
|
810
|
-
| 'M-\\' escaped_nl? ( 'C-' | 'c' ) ) escaped_nl?
|
811
|
-
maybe_escaped_ctrl_char
|
812
|
-
%slash_m_char
|
813
|
-
|
814
|
-
| 'C' c_any %invalid_complex_escape
|
815
|
-
| 'M' c_any %invalid_complex_escape
|
816
|
-
| ( 'M-\\C' | 'C-\\M' ) c_any %invalid_complex_escape
|
817
|
-
|
818
|
-
| ( c_any - [0-7xuCMc] ) %unescape_char
|
819
|
-
|
820
|
-
| c_eof % {
|
821
|
-
diagnostic :fatal, :escape_eof, nil, range(p - 1, p)
|
822
|
-
}
|
823
|
-
);
|
824
|
-
|
825
|
-
# Use rules in form of `e_bs escape' when you need to parse a sequence.
|
826
|
-
e_bs = '\\' % {
|
827
|
-
@escape_s = p
|
828
|
-
@escape = nil
|
829
|
-
};
|
830
|
-
|
831
|
-
#
|
832
|
-
# === STRING AND HEREDOC PARSING ===
|
833
|
-
#
|
834
|
-
|
835
|
-
# Heredoc parsing is quite a complex topic. First, consider that heredocs
|
836
|
-
# can be arbitrarily nested. For example:
|
837
|
-
#
|
838
|
-
# puts <<CODE
|
839
|
-
# the result is: #{<<RESULT.inspect
|
840
|
-
# i am a heredoc
|
841
|
-
# RESULT
|
842
|
-
# }
|
843
|
-
# CODE
|
844
|
-
#
|
845
|
-
# which, incidentally, evaluates to:
|
846
|
-
#
|
847
|
-
# the result is: " i am a heredoc\n"
|
848
|
-
#
|
849
|
-
# To parse them, lexer refers to two kinds (remember, nested heredocs)
|
850
|
-
# of positions in the input stream, namely heredoc_e
|
851
|
-
# (HEREDOC declaration End) and @herebody_s (HEREdoc BODY line Start).
|
852
|
-
#
|
853
|
-
# heredoc_e is simply contained inside the corresponding Literal, and
|
854
|
-
# when the heredoc is closed, the lexing is restarted from that position.
|
855
|
-
#
|
856
|
-
# @herebody_s is quite more complex. First, @herebody_s changes after each
|
857
|
-
# heredoc line is lexed. This way, at '\n' tok(@herebody_s, @te) always
|
858
|
-
# contains the current line, and also when a heredoc is started, @herebody_s
|
859
|
-
# contains the position from which the heredoc will be lexed.
|
860
|
-
#
|
861
|
-
# Second, as (insanity) there are nested heredocs, we need to maintain a
|
862
|
-
# stack of these positions. Each time #push_literal is called, it saves current
|
863
|
-
# @heredoc_s to literal.saved_herebody_s, and after an interpolation (possibly
|
864
|
-
# containing another heredocs) is closed, the previous value is restored.
|
865
|
-
|
866
|
-
e_heredoc_nl = c_nl % {
|
867
|
-
# After every heredoc was parsed, @herebody_s contains the
|
868
|
-
# position of next token after all heredocs.
|
869
|
-
if @herebody_s
|
870
|
-
p = @herebody_s
|
871
|
-
@herebody_s = nil
|
872
|
-
end
|
873
|
-
};
|
874
|
-
|
875
|
-
action extend_string {
|
876
|
-
string = tok
|
877
|
-
|
878
|
-
# tLABEL_END is only possible in non-cond context on >= 2.2
|
879
|
-
if @version >= 22 && !@cond.active?
|
880
|
-
lookahead = @source_buffer.slice(@te...@te+2)
|
881
|
-
end
|
882
|
-
|
883
|
-
current_literal = literal
|
884
|
-
if !current_literal.heredoc? &&
|
885
|
-
(token = current_literal.nest_and_try_closing(string, @ts, @te, lookahead))
|
886
|
-
if token[0] == :tLABEL_END
|
887
|
-
p += 1
|
888
|
-
pop_literal
|
889
|
-
fnext expr_labelarg;
|
890
|
-
else
|
891
|
-
fnext *pop_literal;
|
892
|
-
end
|
893
|
-
fbreak;
|
894
|
-
else
|
895
|
-
current_literal.extend_string(string, @ts, @te)
|
896
|
-
end
|
897
|
-
}
|
898
|
-
|
899
|
-
action extend_string_escaped {
|
900
|
-
current_literal = literal
|
901
|
-
# Get the first character after the backslash.
|
902
|
-
escaped_char = @source_buffer.slice(@escape_s).chr
|
903
|
-
|
904
|
-
if current_literal.munge_escape? escaped_char
|
905
|
-
# If this particular literal uses this character as an opening
|
906
|
-
# or closing delimiter, it is an escape sequence for that
|
907
|
-
# particular character. Write it without the backslash.
|
908
|
-
|
909
|
-
if current_literal.regexp? && REGEXP_META_CHARACTERS.match(escaped_char)
|
910
|
-
# Regular expressions should include escaped delimiters in their
|
911
|
-
# escaped form, except when the escaped character is
|
912
|
-
# a closing delimiter but not a regexp metacharacter.
|
913
|
-
#
|
914
|
-
# The backslash itself cannot be used as a closing delimiter
|
915
|
-
# at the same time as an escape symbol, but it is always munged,
|
916
|
-
# so this branch also executes for the non-closing-delimiter case
|
917
|
-
# for the backslash.
|
918
|
-
current_literal.extend_string(tok, @ts, @te)
|
919
|
-
else
|
920
|
-
current_literal.extend_string(escaped_char, @ts, @te)
|
921
|
-
end
|
922
|
-
else
|
923
|
-
# It does not. So this is an actual escape sequence, yay!
|
924
|
-
if current_literal.squiggly_heredoc? && escaped_char == "\n".freeze
|
925
|
-
# Squiggly heredocs like
|
926
|
-
# <<~-HERE
|
927
|
-
# 1\
|
928
|
-
# 2
|
929
|
-
# HERE
|
930
|
-
# treat '\' as a line continuation, but still dedent the body, so the heredoc above becomes "12\n".
|
931
|
-
# This information is emitted as is, without escaping,
|
932
|
-
# later this escape sequence (\\\n) gets handled manually in the Lexer::Dedenter
|
933
|
-
current_literal.extend_string(tok, @ts, @te)
|
934
|
-
elsif current_literal.supports_line_continuation_via_slash? && escaped_char == "\n".freeze
|
935
|
-
# Heredocs, regexp and a few other types of literals support line
|
936
|
-
# continuation via \\\n sequence. The code like
|
937
|
-
# "a\
|
938
|
-
# b"
|
939
|
-
# must be parsed as "ab"
|
940
|
-
current_literal.extend_string(tok.gsub("\\\n".freeze, ''.freeze), @ts, @te)
|
941
|
-
elsif current_literal.regexp? && @version >= 31 && %w[c C m M].include?(escaped_char)
|
942
|
-
# Ruby >= 3.1 escapes \c- and \m chars, that's the only escape sequence
|
943
|
-
# supported by regexes so far, so it needs a separate branch.
|
944
|
-
current_literal.extend_string(@escape, @ts, @te)
|
945
|
-
elsif current_literal.regexp?
|
946
|
-
# Regular expressions should include escape sequences in their
|
947
|
-
# escaped form. On the other hand, escaped newlines are removed (in cases like "\\C-\\\n\\M-x")
|
948
|
-
current_literal.extend_string(tok.gsub("\\\n".freeze, ''.freeze), @ts, @te)
|
949
|
-
else
|
950
|
-
current_literal.extend_string(@escape || tok, @ts, @te)
|
951
|
-
end
|
952
|
-
end
|
953
|
-
}
|
954
|
-
|
955
|
-
# Extend a string with a newline or a EOF character.
|
956
|
-
# As heredoc closing line can immediately precede EOF, this action
|
957
|
-
# has to handle such case specially.
|
958
|
-
action extend_string_eol {
|
959
|
-
current_literal = literal
|
960
|
-
if @te == pe
|
961
|
-
diagnostic :fatal, :string_eof, nil,
|
962
|
-
range(current_literal.str_s, current_literal.str_s + 1)
|
963
|
-
end
|
964
|
-
|
965
|
-
if current_literal.heredoc?
|
966
|
-
line = tok(@herebody_s, @ts).gsub(/\r+$/, ''.freeze)
|
967
|
-
|
968
|
-
if version?(18, 19, 20)
|
969
|
-
# See ruby:c48b4209c
|
970
|
-
line = line.gsub(/\r.*$/, ''.freeze)
|
971
|
-
end
|
972
|
-
|
973
|
-
# Try ending the heredoc with the complete most recently
|
974
|
-
# scanned line. @herebody_s always refers to the start of such line.
|
975
|
-
if current_literal.nest_and_try_closing(line, @herebody_s, @ts)
|
976
|
-
# Adjust @herebody_s to point to the next line.
|
977
|
-
@herebody_s = @te
|
978
|
-
|
979
|
-
# Continue regular lexing after the heredoc reference (<<END).
|
980
|
-
p = current_literal.heredoc_e - 1
|
981
|
-
fnext *pop_literal; fbreak;
|
982
|
-
else
|
983
|
-
# Calculate indentation level for <<~HEREDOCs.
|
984
|
-
current_literal.infer_indent_level(line)
|
985
|
-
|
986
|
-
# Ditto.
|
987
|
-
@herebody_s = @te
|
988
|
-
end
|
989
|
-
else
|
990
|
-
# Try ending the literal with a newline.
|
991
|
-
if current_literal.nest_and_try_closing(tok, @ts, @te)
|
992
|
-
fnext *pop_literal; fbreak;
|
993
|
-
end
|
994
|
-
|
995
|
-
if @herebody_s
|
996
|
-
# This is a regular literal intertwined with a heredoc. Like:
|
997
|
-
#
|
998
|
-
# p <<-foo+"1
|
999
|
-
# bar
|
1000
|
-
# foo
|
1001
|
-
# 2"
|
1002
|
-
#
|
1003
|
-
# which, incidentally, evaluates to "bar\n1\n2".
|
1004
|
-
p = @herebody_s - 1
|
1005
|
-
@herebody_s = nil
|
1006
|
-
end
|
1007
|
-
end
|
1008
|
-
|
1009
|
-
if current_literal.words? && !eof_codepoint?(@source_pts[p])
|
1010
|
-
current_literal.extend_space @ts, @te
|
1011
|
-
else
|
1012
|
-
# A literal newline is appended if the heredoc was _not_ closed
|
1013
|
-
# this time (see fbreak above). See also Literal#nest_and_try_closing
|
1014
|
-
# for rationale of calling #flush_string here.
|
1015
|
-
current_literal.extend_string tok, @ts, @te
|
1016
|
-
current_literal.flush_string
|
1017
|
-
end
|
1018
|
-
}
|
1019
|
-
|
1020
|
-
action extend_string_space {
|
1021
|
-
literal.extend_space @ts, @te
|
1022
|
-
}
|
689
|
+
| 'r' % { @num_xfrm = @emit_rational }
|
690
|
+
| 'ri' % { @num_xfrm = @emit_imaginary_rational }
|
691
|
+
| 'rescue' % { @num_xfrm = @emit_float_rescue };
|
1023
692
|
|
1024
693
|
#
|
1025
694
|
# === INTERPOLATION PARSING ===
|
1026
695
|
#
|
1027
696
|
|
1028
|
-
# Interpolations with immediate variable names simply call into
|
1029
|
-
# the corresponding machine.
|
1030
|
-
|
1031
|
-
interp_var = '#' ( global_var | class_var_v | instance_var_v );
|
1032
|
-
|
1033
|
-
action extend_interp_var {
|
1034
|
-
current_literal = literal
|
1035
|
-
current_literal.flush_string
|
1036
|
-
current_literal.extend_content
|
1037
|
-
|
1038
|
-
emit(:tSTRING_DVAR, nil, @ts, @ts + 1)
|
1039
|
-
|
1040
|
-
p = @ts
|
1041
|
-
fcall expr_variable;
|
1042
|
-
}
|
1043
|
-
|
1044
|
-
# Special case for Ruby > 2.7
|
1045
|
-
# If interpolated instance/class variable starts with a digit we parse it as a plain substring
|
1046
|
-
# However, "#$1" is still a regular interpolation
|
1047
|
-
interp_digit_var = '#' ('@' | '@@') digit c_alpha*;
|
1048
|
-
|
1049
|
-
action extend_interp_digit_var {
|
1050
|
-
if @version >= 27
|
1051
|
-
literal.extend_string(tok, @ts, @te)
|
1052
|
-
else
|
1053
|
-
message = tok.start_with?('#@@') ? :cvar_name : :ivar_name
|
1054
|
-
diagnostic :error, message, { :name => tok(@ts + 1, @te) }, range(@ts + 1, @te)
|
1055
|
-
end
|
1056
|
-
}
|
1057
|
-
|
1058
|
-
# Interpolations with code blocks must match nested curly braces, as
|
1059
|
-
# interpolation ending is ambiguous with a block ending. So, every
|
1060
|
-
# opening and closing brace should be matched with e_[lr]brace rules,
|
1061
|
-
# which automatically perform the counting.
|
1062
|
-
#
|
1063
|
-
# Note that interpolations can themselves be nested, so brace balance
|
1064
|
-
# is tied to the innermost literal.
|
1065
|
-
#
|
1066
|
-
# Also note that literals themselves should not use e_[lr]brace rules
|
1067
|
-
# when matching their opening and closing delimiters, as the amount of
|
1068
|
-
# braces inside the characters of a string literal is independent.
|
1069
|
-
|
1070
|
-
interp_code = '#{';
|
1071
|
-
|
1072
697
|
e_lbrace = '{' % {
|
1073
|
-
|
1074
|
-
|
1075
|
-
current_literal = literal
|
1076
|
-
if current_literal
|
1077
|
-
current_literal.start_interp_brace
|
1078
|
-
end
|
698
|
+
e_lbrace
|
1079
699
|
};
|
1080
700
|
|
1081
701
|
e_rbrace = '}' % {
|
1082
|
-
|
1083
|
-
|
1084
|
-
|
1085
|
-
|
1086
|
-
emit(:tRCURLY, '}'.freeze, p - 1, p)
|
1087
|
-
@cond.lexpop
|
1088
|
-
@cmdarg.lexpop
|
1089
|
-
else
|
1090
|
-
emit(:tSTRING_DEND, '}'.freeze, p - 1, p)
|
1091
|
-
end
|
1092
|
-
|
1093
|
-
if current_literal.saved_herebody_s
|
1094
|
-
@herebody_s = current_literal.saved_herebody_s
|
1095
|
-
end
|
1096
|
-
|
1097
|
-
|
1098
|
-
fhold;
|
1099
|
-
fnext *next_state_for_literal(current_literal);
|
1100
|
-
fbreak;
|
1101
|
-
end
|
702
|
+
if @strings.close_interp_on_current_literal(p)
|
703
|
+
fhold;
|
704
|
+
fnext inside_string;
|
705
|
+
fbreak;
|
1102
706
|
end
|
1103
707
|
|
1104
708
|
@paren_nest -= 1
|
1105
709
|
};
|
1106
710
|
|
1107
|
-
action extend_interp_code {
|
1108
|
-
current_literal = literal
|
1109
|
-
current_literal.flush_string
|
1110
|
-
current_literal.extend_content
|
1111
|
-
|
1112
|
-
emit(:tSTRING_DBEG, '#{'.freeze)
|
1113
|
-
|
1114
|
-
if current_literal.heredoc?
|
1115
|
-
current_literal.saved_herebody_s = @herebody_s
|
1116
|
-
@herebody_s = nil
|
1117
|
-
end
|
1118
|
-
|
1119
|
-
current_literal.start_interp_brace
|
1120
|
-
@command_start = true
|
1121
|
-
fnext expr_value;
|
1122
|
-
fbreak;
|
1123
|
-
}
|
1124
|
-
|
1125
|
-
# Actual string parsers are simply combined from the primitives defined
|
1126
|
-
# above.
|
1127
|
-
|
1128
|
-
interp_words := |*
|
1129
|
-
interp_code => extend_interp_code;
|
1130
|
-
interp_digit_var => extend_interp_digit_var;
|
1131
|
-
interp_var => extend_interp_var;
|
1132
|
-
e_bs escape => extend_string_escaped;
|
1133
|
-
c_space+ => extend_string_space;
|
1134
|
-
c_eol => extend_string_eol;
|
1135
|
-
c_any => extend_string;
|
1136
|
-
*|;
|
1137
|
-
|
1138
|
-
interp_string := |*
|
1139
|
-
interp_code => extend_interp_code;
|
1140
|
-
interp_digit_var => extend_interp_digit_var;
|
1141
|
-
interp_var => extend_interp_var;
|
1142
|
-
e_bs escape => extend_string_escaped;
|
1143
|
-
c_eol => extend_string_eol;
|
1144
|
-
c_any => extend_string;
|
1145
|
-
*|;
|
1146
|
-
|
1147
|
-
plain_words := |*
|
1148
|
-
e_bs c_any => extend_string_escaped;
|
1149
|
-
c_space+ => extend_string_space;
|
1150
|
-
c_eol => extend_string_eol;
|
1151
|
-
c_any => extend_string;
|
1152
|
-
*|;
|
1153
|
-
|
1154
|
-
plain_string := |*
|
1155
|
-
'\\' c_nl => extend_string_eol;
|
1156
|
-
e_bs c_any => extend_string_escaped;
|
1157
|
-
c_eol => extend_string_eol;
|
1158
|
-
c_any => extend_string;
|
1159
|
-
*|;
|
1160
|
-
|
1161
|
-
interp_backslash_delimited := |*
|
1162
|
-
interp_code => extend_interp_code;
|
1163
|
-
interp_digit_var => extend_interp_digit_var;
|
1164
|
-
interp_var => extend_interp_var;
|
1165
|
-
c_eol => extend_string_eol;
|
1166
|
-
c_any => extend_string;
|
1167
|
-
*|;
|
1168
|
-
|
1169
|
-
plain_backslash_delimited := |*
|
1170
|
-
c_eol => extend_string_eol;
|
1171
|
-
c_any => extend_string;
|
1172
|
-
*|;
|
1173
|
-
|
1174
|
-
interp_backslash_delimited_words := |*
|
1175
|
-
interp_code => extend_interp_code;
|
1176
|
-
interp_digit_var => extend_interp_digit_var;
|
1177
|
-
interp_var => extend_interp_var;
|
1178
|
-
c_space+ => extend_string_space;
|
1179
|
-
c_eol => extend_string_eol;
|
1180
|
-
c_any => extend_string;
|
1181
|
-
*|;
|
1182
|
-
|
1183
|
-
plain_backslash_delimited_words := |*
|
1184
|
-
c_space+ => extend_string_space;
|
1185
|
-
c_eol => extend_string_eol;
|
1186
|
-
c_any => extend_string;
|
1187
|
-
*|;
|
1188
|
-
|
1189
|
-
regexp_modifiers := |*
|
1190
|
-
[A-Za-z]+
|
1191
|
-
=> {
|
1192
|
-
unknown_options = tok.scan(/[^imxouesn]/)
|
1193
|
-
if unknown_options.any?
|
1194
|
-
diagnostic :error, :regexp_options,
|
1195
|
-
{ :options => unknown_options.join }
|
1196
|
-
end
|
1197
|
-
|
1198
|
-
emit(:tREGEXP_OPT)
|
1199
|
-
fnext expr_end;
|
1200
|
-
fbreak;
|
1201
|
-
};
|
1202
|
-
|
1203
|
-
any
|
1204
|
-
=> {
|
1205
|
-
emit(:tREGEXP_OPT, tok(@ts, @te - 1), @ts, @te - 1)
|
1206
|
-
fhold;
|
1207
|
-
fgoto expr_end;
|
1208
|
-
};
|
1209
|
-
*|;
|
1210
|
-
|
1211
711
|
#
|
1212
712
|
# === WHITESPACE HANDLING ===
|
1213
713
|
#
|
@@ -1221,16 +721,20 @@ class Next
|
|
1221
721
|
# comment is deemed equivalent to non-newline whitespace
|
1222
722
|
# (c_space character class).
|
1223
723
|
|
724
|
+
e_nl = c_nl % {
|
725
|
+
p = on_newline(p)
|
726
|
+
};
|
727
|
+
|
1224
728
|
w_space =
|
1225
729
|
c_space+
|
1226
|
-
| '\\'
|
730
|
+
| '\\' e_nl
|
1227
731
|
;
|
1228
732
|
|
1229
733
|
w_comment =
|
1230
734
|
'#' %{ @sharp_s = p - 1 }
|
1231
735
|
# The (p == pe) condition compensates for added "\0" and
|
1232
736
|
# the way Ragel handles EOF.
|
1233
|
-
c_line* %{
|
737
|
+
c_line* %{ emit_comment_from_range(p, pe) }
|
1234
738
|
;
|
1235
739
|
|
1236
740
|
w_space_comment =
|
@@ -1253,7 +757,7 @@ class Next
|
|
1253
757
|
# is equivalent to `foo = "bar\n" + 2`.
|
1254
758
|
|
1255
759
|
w_newline =
|
1256
|
-
|
760
|
+
e_nl;
|
1257
761
|
|
1258
762
|
w_any =
|
1259
763
|
w_space
|
@@ -1341,34 +845,22 @@ class Next
|
|
1341
845
|
expr_variable := |*
|
1342
846
|
global_var
|
1343
847
|
=> {
|
1344
|
-
|
1345
|
-
emit(:tNTH_REF, tok(@ts + 1).to_i)
|
1346
|
-
elsif tok =~ /^\$([&`'+])$/
|
1347
|
-
emit(:tBACK_REF)
|
1348
|
-
else
|
1349
|
-
emit(:tGVAR)
|
1350
|
-
end
|
848
|
+
emit_global_var
|
1351
849
|
|
1352
850
|
fnext *stack_pop; fbreak;
|
1353
851
|
};
|
1354
852
|
|
1355
853
|
class_var_v
|
1356
854
|
=> {
|
1357
|
-
|
1358
|
-
diagnostic :error, :cvar_name, { :name => tok }
|
1359
|
-
end
|
855
|
+
emit_class_var
|
1360
856
|
|
1361
|
-
emit(:tCVAR)
|
1362
857
|
fnext *stack_pop; fbreak;
|
1363
858
|
};
|
1364
859
|
|
1365
860
|
instance_var_v
|
1366
861
|
=> {
|
1367
|
-
|
1368
|
-
diagnostic :error, :ivar_name, { :name => tok }
|
1369
|
-
end
|
862
|
+
emit_instance_var
|
1370
863
|
|
1371
|
-
emit(:tIVAR)
|
1372
864
|
fnext *stack_pop; fbreak;
|
1373
865
|
};
|
1374
866
|
*|;
|
@@ -1418,7 +910,8 @@ class Next
|
|
1418
910
|
=> {
|
1419
911
|
if version?(23)
|
1420
912
|
type, delimiter = tok[0..-2], tok[-1].chr
|
1421
|
-
|
913
|
+
@strings.push_literal(type, delimiter, @ts)
|
914
|
+
fgoto inside_string;
|
1422
915
|
else
|
1423
916
|
p = @ts - 1
|
1424
917
|
fgoto expr_end;
|
@@ -1569,14 +1062,7 @@ class Next
|
|
1569
1062
|
| '<<'
|
1570
1063
|
)
|
1571
1064
|
=> {
|
1572
|
-
|
1573
|
-
# Ambiguous regexp literal.
|
1574
|
-
if @version < 30
|
1575
|
-
diagnostic :warning, :ambiguous_literal, nil, range(tm, tm + 1)
|
1576
|
-
else
|
1577
|
-
diagnostic :warning, :ambiguous_regexp, nil, range(tm, tm + 1)
|
1578
|
-
end
|
1579
|
-
end
|
1065
|
+
check_ambiguous_slash(tm)
|
1580
1066
|
|
1581
1067
|
p = tm - 1
|
1582
1068
|
fgoto expr_beg;
|
@@ -1774,21 +1260,26 @@ class Next
|
|
1774
1260
|
'/' c_any
|
1775
1261
|
=> {
|
1776
1262
|
type = delimiter = tok[0].chr
|
1777
|
-
|
1263
|
+
@strings.push_literal(type, delimiter, @ts)
|
1264
|
+
|
1265
|
+
fhold;
|
1266
|
+
fgoto inside_string;
|
1778
1267
|
};
|
1779
1268
|
|
1780
1269
|
# %<string>
|
1781
1270
|
'%' ( c_ascii - [A-Za-z0-9] )
|
1782
1271
|
=> {
|
1783
|
-
type, delimiter = @source_buffer.slice(@ts).chr, tok[-1].chr
|
1784
|
-
|
1272
|
+
type, delimiter = @source_buffer.slice(@ts, 1).chr, tok[-1].chr
|
1273
|
+
@strings.push_literal(type, delimiter, @ts)
|
1274
|
+
fgoto inside_string;
|
1785
1275
|
};
|
1786
1276
|
|
1787
1277
|
# %w(we are the people)
|
1788
1278
|
'%' [A-Za-z] (c_ascii - [A-Za-z0-9])
|
1789
1279
|
=> {
|
1790
1280
|
type, delimiter = tok[0..-2], tok[-1].chr
|
1791
|
-
|
1281
|
+
@strings.push_literal(type, delimiter, @ts)
|
1282
|
+
fgoto inside_string;
|
1792
1283
|
};
|
1793
1284
|
|
1794
1285
|
'%' c_eof
|
@@ -1834,10 +1325,11 @@ class Next
|
|
1834
1325
|
p = @ts + 1
|
1835
1326
|
fnext expr_beg; fbreak;
|
1836
1327
|
else
|
1837
|
-
|
1328
|
+
@strings.push_literal(type, delimiter, @ts, heredoc_e, indent, dedent_body);
|
1329
|
+
@strings.herebody_s ||= new_herebody_s
|
1838
1330
|
|
1839
|
-
@herebody_s
|
1840
|
-
|
1331
|
+
p = @strings.herebody_s - 1
|
1332
|
+
fnext inside_string;
|
1841
1333
|
end
|
1842
1334
|
};
|
1843
1335
|
|
@@ -1871,7 +1363,9 @@ class Next
|
|
1871
1363
|
':' ['"] # '
|
1872
1364
|
=> {
|
1873
1365
|
type, delimiter = tok, tok[-1].chr
|
1874
|
-
|
1366
|
+
@strings.push_literal(type, delimiter, @ts);
|
1367
|
+
|
1368
|
+
fgoto inside_string;
|
1875
1369
|
};
|
1876
1370
|
|
1877
1371
|
# :!@ is :!
|
@@ -1900,12 +1394,7 @@ class Next
|
|
1900
1394
|
| '@@' %{ tm = p - 2; diag_msg = :cvar_name }
|
1901
1395
|
) [0-9]*
|
1902
1396
|
=> {
|
1903
|
-
|
1904
|
-
diagnostic :error, diag_msg, { name: tok(tm, @te) }, range(tm, @te)
|
1905
|
-
else
|
1906
|
-
emit(:tCOLON, tok(@ts, @ts + 1), @ts, @ts + 1)
|
1907
|
-
p = @ts
|
1908
|
-
end
|
1397
|
+
emit_colon_with_digits(p, tm, diag_msg)
|
1909
1398
|
|
1910
1399
|
fnext expr_end; fbreak;
|
1911
1400
|
};
|
@@ -1916,29 +1405,19 @@ class Next
|
|
1916
1405
|
|
1917
1406
|
# Character constant, like ?a, ?\n, ?\u1000, and so on
|
1918
1407
|
# Don't accept \u escape with multiple codepoints, like \u{1 2 3}
|
1919
|
-
'?'
|
1920
|
-
| (c_any - c_space_nl - e_bs) % { @escape = nil }
|
1921
|
-
)
|
1408
|
+
'?' c_any
|
1922
1409
|
=> {
|
1923
|
-
|
1410
|
+
p, next_state = @strings.read_character_constant(@ts)
|
1411
|
+
fhold; # Ragel will do `p += 1` to consume input, prevent it
|
1924
1412
|
|
1925
|
-
|
1926
|
-
|
1413
|
+
# If strings lexer founds a character constant (?a) emit it,
|
1414
|
+
# otherwise read ternary operator
|
1415
|
+
if @token_queue.empty?
|
1416
|
+
fgoto *next_state;
|
1927
1417
|
else
|
1928
|
-
|
1418
|
+
fnext *next_state;
|
1419
|
+
fbreak;
|
1929
1420
|
end
|
1930
|
-
|
1931
|
-
fnext expr_end; fbreak;
|
1932
|
-
};
|
1933
|
-
|
1934
|
-
'?' c_space_nl
|
1935
|
-
=> {
|
1936
|
-
escape = { " " => '\s', "\r" => '\r', "\n" => '\n', "\t" => '\t',
|
1937
|
-
"\v" => '\v', "\f" => '\f' }[@source_buffer.slice(@ts + 1)]
|
1938
|
-
diagnostic :warning, :invalid_escape_use, { :escape => escape }, range
|
1939
|
-
|
1940
|
-
p = @ts - 1
|
1941
|
-
fgoto expr_end;
|
1942
1421
|
};
|
1943
1422
|
|
1944
1423
|
'?' c_eof
|
@@ -1946,13 +1425,6 @@ class Next
|
|
1946
1425
|
diagnostic :fatal, :incomplete_escape, nil, range(@ts, @ts + 1)
|
1947
1426
|
};
|
1948
1427
|
|
1949
|
-
# f ?aa : b: Disambiguate with a character literal.
|
1950
|
-
'?' [A-Za-z_] bareword
|
1951
|
-
=> {
|
1952
|
-
p = @ts - 1
|
1953
|
-
fgoto expr_end;
|
1954
|
-
};
|
1955
|
-
|
1956
1428
|
#
|
1957
1429
|
# AMBIGUOUS EMPTY BLOCK ARGUMENTS
|
1958
1430
|
#
|
@@ -2028,7 +1500,7 @@ class Next
|
|
2028
1500
|
if version?(18)
|
2029
1501
|
ident = tok(@ts, @te - 2)
|
2030
1502
|
|
2031
|
-
emit((@source_buffer.slice(@ts) =~ /[A-Z]/) ? :tCONSTANT : :tIDENTIFIER,
|
1503
|
+
emit((@source_buffer.slice(@ts, 1) =~ /[A-Z]/) ? :tCONSTANT : :tIDENTIFIER,
|
2032
1504
|
ident, @ts, @te - 2)
|
2033
1505
|
fhold; # continue as a symbol
|
2034
1506
|
|
@@ -2133,7 +1605,7 @@ class Next
|
|
2133
1605
|
|
2134
1606
|
w_any;
|
2135
1607
|
|
2136
|
-
|
1608
|
+
e_nl '=begin' ( c_space | c_nl_zlen )
|
2137
1609
|
=> {
|
2138
1610
|
p = @ts - 1
|
2139
1611
|
@cs_before_block_comment = @cs
|
@@ -2186,7 +1658,8 @@ class Next
|
|
2186
1658
|
# "bar", 'baz'
|
2187
1659
|
['"] # '
|
2188
1660
|
=> {
|
2189
|
-
|
1661
|
+
@strings.push_literal(tok, tok, @ts)
|
1662
|
+
fgoto inside_string;
|
2190
1663
|
};
|
2191
1664
|
|
2192
1665
|
w_space_comment;
|
@@ -2247,8 +1720,7 @@ class Next
|
|
2247
1720
|
fnext expr_fname; fbreak; };
|
2248
1721
|
|
2249
1722
|
'class' w_any* '<<'
|
2250
|
-
=> {
|
2251
|
-
emit(:tLSHFT, '<<'.freeze, @te - 2, @te)
|
1723
|
+
=> { emit_singleton_class
|
2252
1724
|
fnext expr_value; fbreak; };
|
2253
1725
|
|
2254
1726
|
# a if b:c: Syntax error.
|
@@ -2307,27 +1779,13 @@ class Next
|
|
2307
1779
|
| '0' digit* '_'? %{ @num_base = 8; @num_digits_s = @ts } int_dec
|
2308
1780
|
) %{ @num_suffix_s = p } int_suffix
|
2309
1781
|
=> {
|
2310
|
-
digits =
|
2311
|
-
|
2312
|
-
if digits.end_with? '_'.freeze
|
2313
|
-
diagnostic :error, :trailing_in_number, { :character => '_'.freeze },
|
2314
|
-
range(@te - 1, @te)
|
2315
|
-
elsif digits.empty? && @num_base == 8 && version?(18)
|
2316
|
-
# 1.8 did not raise an error on 0o.
|
2317
|
-
digits = '0'.freeze
|
2318
|
-
elsif digits.empty?
|
2319
|
-
diagnostic :error, :empty_numeric
|
2320
|
-
elsif @num_base == 8 && (invalid_idx = digits.index(/[89]/))
|
2321
|
-
invalid_s = @num_digits_s + invalid_idx
|
2322
|
-
diagnostic :error, :invalid_octal, nil,
|
2323
|
-
range(invalid_s, invalid_s + 1)
|
2324
|
-
end
|
1782
|
+
digits = numeric_literal_int
|
2325
1783
|
|
2326
1784
|
if version?(18, 19, 20)
|
2327
1785
|
emit(:tINTEGER, digits.to_i(@num_base), @ts, @num_suffix_s)
|
2328
1786
|
p = @num_suffix_s - 1
|
2329
1787
|
else
|
2330
|
-
@num_xfrm.call(digits.to_i(@num_base))
|
1788
|
+
p = @num_xfrm.call(digits.to_i(@num_base), p)
|
2331
1789
|
end
|
2332
1790
|
fbreak;
|
2333
1791
|
};
|
@@ -2372,7 +1830,7 @@ class Next
|
|
2372
1830
|
emit(:tFLOAT, Float(digits), @ts, @num_suffix_s)
|
2373
1831
|
p = @num_suffix_s - 1
|
2374
1832
|
else
|
2375
|
-
@num_xfrm.call(digits)
|
1833
|
+
p = @num_xfrm.call(digits, p)
|
2376
1834
|
end
|
2377
1835
|
fbreak;
|
2378
1836
|
};
|
@@ -2385,7 +1843,8 @@ class Next
|
|
2385
1843
|
'`' | ['"] # '
|
2386
1844
|
=> {
|
2387
1845
|
type, delimiter = tok, tok[-1].chr
|
2388
|
-
|
1846
|
+
@strings.push_literal(type, delimiter, @ts, nil, false, false, true);
|
1847
|
+
fgoto inside_string;
|
2389
1848
|
};
|
2390
1849
|
|
2391
1850
|
#
|
@@ -2470,15 +1929,7 @@ class Next
|
|
2470
1929
|
|
2471
1930
|
e_rbrace | e_rparen | e_rbrack
|
2472
1931
|
=> {
|
2473
|
-
|
2474
|
-
|
2475
|
-
if @version < 24
|
2476
|
-
@cond.lexpop
|
2477
|
-
@cmdarg.lexpop
|
2478
|
-
else
|
2479
|
-
@cond.pop
|
2480
|
-
@cmdarg.pop
|
2481
|
-
end
|
1932
|
+
emit_rbrace_rparen_rbrack
|
2482
1933
|
|
2483
1934
|
if tok == '}'.freeze || tok == ']'.freeze
|
2484
1935
|
if @version >= 25
|
@@ -2635,6 +2086,17 @@ class Next
|
|
2635
2086
|
c_eof => do_eof;
|
2636
2087
|
*|;
|
2637
2088
|
|
2089
|
+
inside_string := |*
|
2090
|
+
any
|
2091
|
+
=> {
|
2092
|
+
p, next_state = @strings.advance(p)
|
2093
|
+
|
2094
|
+
fhold; # Ragel will do `p += 1` to consume input, prevent it
|
2095
|
+
fnext *next_state;
|
2096
|
+
fbreak;
|
2097
|
+
};
|
2098
|
+
*|;
|
2099
|
+
|
2638
2100
|
}%%
|
2639
2101
|
# %
|
2640
2102
|
end
|