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
|
# frozen_string_literal: true
require 'spec_helper'
RSpec.describe('Escape scanning') do
include_examples 'scan', /c\at/, 1 => [:escape, :bell, '\a', 1, 3]
# not an escape outside a character set
include_examples 'scan', /c\bt/, 1 => [:anchor, :word_boundary, '\b', 1, 3]
include_examples 'scan', /c\ft/, 1 => [:escape, :form_feed, '\f', 1, 3]
include_examples 'scan', /c\nt/, 1 => [:escape, :newline, '\n', 1, 3]
include_examples 'scan', /c\tt/, 1 => [:escape, :tab, '\t', 1, 3]
include_examples 'scan', /c\vt/, 1 => [:escape, :vertical_tab, '\v', 1, 3]
# ineffectual literal escapes
# these cause "Unknown escape" warnings in Ruby for ascii chars,
# and simply drop the backslash for non-ascii chars (/\ü/.inspect == '/ü/').
# In terms of matching, Ruby treats them both like non-escaped literals.
include_examples 'scan', 'c\qt', 1 => [:escape, :literal, '\q', 1, 3]
include_examples 'scan', 'a\üc', 1 => [:escape, :literal, '\ü', 1, 3]
include_examples 'scan', 'a\😋c', 1 => [:escape, :literal, '\😋', 1, 3]
# these incomplete ref/call sequences are treated as literal escapes by Ruby
include_examples 'scan', 'c\gt', 1 => [:escape, :literal, '\g', 1, 3]
include_examples 'scan', 'c\kt', 1 => [:escape, :literal, '\k', 1, 3]
include_examples 'scan', 'a\012c', 1 => [:escape, :octal, '\012', 1, 5]
include_examples 'scan', 'a\0124', 1 => [:escape, :octal, '\012', 1, 5]
include_examples 'scan', '\712+7', 0 => [:escape, :octal, '\712', 0, 4]
# Multi-digit escaped numbers that occur before sufficient capturing groups
# have been opened are treated as octal or literal.
# "\10"[/\10()()()()()()()()()()/] # => "\b" # treated as octal
# "\70"[/\70()()()()()()()()()()/] # => "8" # treated as octal
# "90"[/\90()()()()()()()()()()/] # => "90" # treated as literal
# For cases treated as backrefs, see ./refcalls_spec.rb
include_examples 'scan', "\\10#{'()' * 10}",0 => [:escape, :octal, '\10', 0, 3]
include_examples 'scan', "\\90#{'()' * 90}",0 => [:escape, :literal, '\9', 0, 2],
1 => [:literal, :literal, '0', 2, 3]
# special case: "out-of-bound octal escapes" (digits > 7) are not treated as backrefs
include_examples 'scan', '\80', 0 => [:escape, :literal, '\8', 0, 2]
include_examples 'scan', '\80', 1 => [:literal, :literal, '0', 2, 3]
include_examples 'scan', 'a\xA', 1 => [:escape, :hex, '\xA', 1, 4]
include_examples 'scan', 'a\x24c', 1 => [:escape, :hex, '\x24', 1, 5]
include_examples 'scan', 'a\x0640c', 1 => [:escape, :hex, '\x06', 1, 5]
include_examples 'scan', 'a\xE2\x82\xAC', 1 => [:escape, :utf8_hex, '\xE2\x82\xAC', 1, 13]
include_examples 'scan', /a\xE2\x82\xAC/n, 1 => [:escape, :hex, '\xE2', 1, 5]
include_examples 'scan', /a\xE2\x82\xAC/n, 2 => [:escape, :hex, '\x82', 5, 9]
include_examples 'scan', /a\xE2\x82\xAC/n, 3 => [:escape, :hex, '\xAC', 9, 13]
include_examples 'scan', 'a\u0640c', 1 => [:escape, :codepoint, '\u0640', 1, 7]
include_examples 'scan', 'a\u{640 0641}c', 1 => [:escape, :codepoint_list, '\u{640 0641}', 1, 13]
include_examples 'scan', 'a\u{10FFFF}c', 1 => [:escape, :codepoint_list, '\u{10FFFF}', 1, 11]
include_examples 'scan', 'ab\\\xcd', 1 => [:escape, :backslash, '\\\\', 2, 4]
include_examples 'scan', 'ab\\\0cd', 1 => [:escape, :backslash, '\\\\', 2, 4]
include_examples 'scan', 'ab\\\Kcd', 1 => [:escape, :backslash, '\\\\', 2, 4]
include_examples 'scan', 'ab\^cd', 1 => [:escape, :bol, '\^', 2, 4]
include_examples 'scan', 'ab\$cd', 1 => [:escape, :eol, '\$', 2, 4]
include_examples 'scan', 'ab\[cd', 1 => [:escape, :set_open, '\[', 2, 4]
# escaped whitespace in x-mode
include_examples 'scan', /a\ b/x, 0 => [:literal, :literal, 'a', 0, 1],
1 => [:escape, :literal, '\ ', 1, 3],
2 => [:literal, :literal, 'b', 3, 4]
# newline literals can't be escaped in x-mode, c.f. https://bugs.ruby-lang.org/issues/19639
include_examples 'scan', /a\
b/x, 0 => [:literal, :literal, 'ab', 0, 2]
# Meta/control escapes
#
# After the following fix in Ruby 3.1, a Regexp#source containing meta/control
# escapes can only be set with the Regexp::new constructor.
# In Regexp literals, these escapes are now pre-processed to hex escapes.
#
# https://github.com/ruby/ruby/commit/11ae581a4a7f5d5f5ec6378872eab8f25381b1b9
n = ->(regexp_body){ Regexp.new(regexp_body.dup.force_encoding('ascii-8bit')) }
include_examples 'scan', 'a\cBc', 1 => [:escape, :control, '\cB', 1, 4]
include_examples 'scan', 'a\c^c', 1 => [:escape, :control, '\c^', 1, 4]
include_examples 'scan', 'a\c\n', 1 => [:escape, :control, '\c\n', 1, 5]
include_examples 'scan', 'a\c\\\\b', 1 => [:escape, :control, '\c\\\\', 1, 5]
include_examples 'scan', 'a\C-bc', 1 => [:escape, :control, '\C-b', 1, 5]
include_examples 'scan', 'a\C-^b', 1 => [:escape, :control, '\C-^', 1, 5]
include_examples 'scan', 'a\C-\nb', 1 => [:escape, :control, '\C-\n', 1, 6]
include_examples 'scan', 'a\C-\\\\b', 1 => [:escape, :control, '\C-\\\\', 1, 6]
include_examples 'scan', n.('a\c\M-Bc'), 1 => [:escape, :control, '\c\M-B', 1, 7]
include_examples 'scan', n.('a\C-\M-Bc'), 1 => [:escape, :control, '\C-\M-B', 1, 8]
include_examples 'scan', n.('a\M-Bc'), 1 => [:escape, :meta_sequence, '\M-B', 1, 5]
include_examples 'scan', n.('a\M-\cBc'), 1 => [:escape, :meta_sequence, '\M-\cB', 1, 7]
include_examples 'scan', n.('a\M-\c^'), 1 => [:escape, :meta_sequence, '\M-\c^', 1, 7]
include_examples 'scan', n.('a\M-\c\n'), 1 => [:escape, :meta_sequence, '\M-\c\n', 1, 8]
include_examples 'scan', n.('a\M-\c\\\\'), 1 => [:escape, :meta_sequence, '\M-\c\\\\', 1, 8]
include_examples 'scan', n.('a\M-\C-Bc'), 1 => [:escape, :meta_sequence, '\M-\C-B', 1, 8]
include_examples 'scan', n.('a\M-\C-\\\\'), 1 => [:escape, :meta_sequence, '\M-\C-\\\\', 1, 9]
end
|
