ruby/test/prism/regexp_test.rb
Peter Zhu eca3680c27 [ruby/prism] Fix use of uninitialized value when parsing regexp
Parsing the regexp /\A{/ causes uses an uninitialized value because it
tries to parse it as a range quantifier, so it reads the character after
the closing curly bracket. This is using uninitialized values because
prism strings are not null terminated. This can be seen in the Valgrind
output:

    ==834710== Conditional jump or move depends on uninitialised value(s)
    ==834710==    at 0x5DA010: pm_regexp_parse_range_quantifier (regexp.c:163)
    ==834710==    by 0x5DA010: pm_regexp_parse_quantifier (regexp.c:243)
    ==834710==    by 0x5DAD69: pm_regexp_parse_expression (regexp.c:738)
    ==834710==    by 0x5DAD69: pm_regexp_parse_pattern (regexp.c:761)
    ==834710==    by 0x5DAD69: pm_regexp_parse (regexp.c:773)
    ==834710==    by 0x5A2EE7: parse_regular_expression_named_captures (prism.c:20886)
    ==834710==    by 0x5A2EE7: parse_expression_infix (prism.c:21388)
    ==834710==    by 0x5A5FA5: parse_expression (prism.c:21804)
    ==834710==    by 0x5A64F3: parse_statements (prism.c:13858)
    ==834710==    by 0x5A9730: parse_program (prism.c:22011)
    ==834710==    by 0x576F0D: parse_input_success_p (extension.c:1062)
    ==834710==    by 0x576F0D: parse_success_p (extension.c:1084)

This commit adds checks for the end of the string to
pm_regexp_parse_range_quantifier.

be6cbc23ef
2024-11-12 14:19:46 +00:00

265 lines
6.7 KiB
Ruby

# frozen_string_literal: true
require_relative "test_helper"
module Prism
class RegexpTest < TestCase
############################################################################
# These tests test the actual use case of extracting named capture groups
############################################################################
def test_named_captures_with_arrows
assert_equal([:foo], named_captures("(?<foo>bar)"))
end
def test_named_captures_with_single_quotes
assert_equal([:foo], named_captures("(?'foo'bar)"))
end
def test_nested_named_captures_with_arrows
assert_equal([:foo, :bar], named_captures("(?<foo>(?<bar>baz))"))
end
def test_nested_named_captures_with_single_quotes
assert_equal([:foo, :bar], named_captures("(?'foo'(?'bar'baz))"))
end
def test_allows_duplicate_named_captures
assert_equal([:foo], named_captures("(?<foo>bar)(?<foo>baz)"))
end
def test_named_capture_inside_fake_range_quantifier
assert_equal([:foo], named_captures("foo{1, (?<foo>2)}"))
end
def test_fake_named_captures_inside_character_sets
assert_equal([], named_captures("[a-z(?<foo>)]"))
end
def test_fake_named_capture_inside_character_set_with_escaped_ending
assert_equal([], named_captures("[a-z\\](?<foo>)]"))
end
############################################################################
# These tests test the rest of the AST. They are not exhaustive, but they
# should cover the most common cases. We test these to make sure we don't
# accidentally regress and stop being able to extract named captures.
############################################################################
def test_alternation
assert_valid_regexp("foo|bar")
end
def test_anchors
assert_valid_regexp("^foo$")
end
def test_any
assert_valid_regexp(".")
end
def test_posix_character_classes
assert_valid_regexp("[[:digit:]]")
end
def test_negated_posix_character_classes
assert_valid_regexp("[[:^digit:]]")
end
def test_invalid_posix_character_classes_should_fall_back_to_regular_classes
assert_valid_regexp("[[:foo]]")
end
def test_character_sets
assert_valid_regexp("[abc]")
end
def test_nested_character_sets
assert_valid_regexp("[[abc]]")
end
def test_nested_character_sets_with_operators
assert_valid_regexp("[[abc] && [def]]")
end
def test_named_capture_inside_nested_character_set
assert_equal([], named_captures("[foo (?<foo>bar)]"))
end
def test_negated_character_sets
assert_valid_regexp("[^abc]")
end
def test_character_ranges
assert_valid_regexp("[a-z]")
end
def test_negated_character_ranges
assert_valid_regexp("[^a-z]")
end
def test_comments
assert_valid_regexp("(?#foo)")
end
def test_comments_with_escaped_parentheses
assert_valid_regexp("(?#foo\\)\\))")
end
def test_non_capturing_groups
assert_valid_regexp("(?:foo)")
end
def test_positive_lookaheads
assert_valid_regexp("(?=foo)")
end
def test_negative_lookaheads
assert_valid_regexp("(?!foo)")
end
def test_positive_lookbehinds
assert_valid_regexp("(?<=foo)")
end
def test_negative_lookbehinds
assert_valid_regexp("(?<!foo)")
end
def test_atomic_groups
assert_valid_regexp("(?>foo)")
end
def test_absence_operator
assert_valid_regexp("(?~foo)")
end
def test_conditional_expression_with_index
assert_valid_regexp("(?(1)foo)")
end
def test_conditional_expression_with_name
assert_valid_regexp("(?(foo)bar)")
end
def test_conditional_expression_with_group
assert_valid_regexp("(?(<foo>)bar)")
end
def test_options_on_groups
assert_valid_regexp("(?imxdau:foo)")
end
def test_options_on_groups_getting_turned_off
assert_valid_regexp("(?-imx:foo)")
end
def test_options_on_groups_some_getting_turned_on_some_getting_turned_off
assert_valid_regexp("(?im-x:foo)")
end
def test_star_quantifier
assert_valid_regexp("foo*")
end
def test_plus_quantifier
assert_valid_regexp("foo+")
end
def test_question_mark_quantifier
assert_valid_regexp("foo?")
end
def test_endless_range_quantifier
assert_valid_regexp("foo{1,}")
end
def test_beginless_range_quantifier
assert_valid_regexp("foo{,1}")
end
def test_range_quantifier
assert_valid_regexp("foo{1,2}")
end
def test_fake_range_quantifier_because_of_spaces
assert_valid_regexp("foo{1, 2}")
end
def test_fake_range_quantifier_because_unclosed
assert_valid_regexp("\\A{")
end
############################################################################
# These test that flag values are correct.
############################################################################
def test_flag_ignorecase
assert_equal(Regexp::IGNORECASE, options("i"))
end
def test_flag_extended
assert_equal(Regexp::EXTENDED, options("x"))
end
def test_flag_multiline
assert_equal(Regexp::MULTILINE, options("m"))
end
def test_flag_fixedencoding
assert_equal(Regexp::FIXEDENCODING, options("e"))
assert_equal(Regexp::FIXEDENCODING, options("u"))
assert_equal(Regexp::FIXEDENCODING, options("s"))
end
def test_flag_noencoding
assert_equal(Regexp::NOENCODING, options("n"))
end
def test_flag_once
assert_equal(0, options("o"))
end
def test_flag_combined
value = Regexp::IGNORECASE | Regexp::MULTILINE | Regexp::EXTENDED
assert_equal(value, options("mix"))
end
def test_last_encoding_option_wins
regex = "/foo/nu"
option = Prism.parse_statement(regex).options
assert_equal Regexp::FIXEDENCODING, option
regex = "/foo/un"
option = Prism.parse_statement(regex).options
assert_equal Regexp::NOENCODING, option
end
private
def assert_valid_regexp(source)
assert Prism.parse_success?("/#{source}/ =~ \"\"")
end
def named_captures(source)
Prism.parse("/#{source}/ =~ \"\"").value.locals
end
def options(flags)
options =
["/foo/#{flags}", "/foo\#{1}/#{flags}"].map do |source|
Prism.parse_statement(source).options
end
# Check that we get the same set of options from both regular expressions
# and interpolated regular expressions.
assert_equal(1, options.uniq.length)
# Return the options from the first regular expression since we know they
# are the same.
options.first
end
end
end