[ruby/psych] Add strict_integer option to parse numbers with commas as strings

Authored-by: Seth Boyles <sethboyles@gmail.com>

https://github.com/ruby/psych/commit/75bebb37b8
This commit is contained in:
Seth Boyles 2022-01-14 19:58:20 +00:00 committed by git
parent b815a0bd75
commit c1a6ff046d
6 changed files with 62 additions and 17 deletions

View File

@ -268,10 +268,10 @@ module Psych
# YAML documents that are supplied via user input. Instead, please use the # YAML documents that are supplied via user input. Instead, please use the
# load method or the safe_load method. # load method or the safe_load method.
# #
def self.unsafe_load yaml, filename: nil, fallback: false, symbolize_names: false, freeze: false def self.unsafe_load yaml, filename: nil, fallback: false, symbolize_names: false, freeze: false, strict_integer: false
result = parse(yaml, filename: filename) result = parse(yaml, filename: filename)
return fallback unless result return fallback unless result
result.to_ruby(symbolize_names: symbolize_names, freeze: freeze) result.to_ruby(symbolize_names: symbolize_names, freeze: freeze, strict_integer: strict_integer)
end end
### ###
@ -319,13 +319,13 @@ module Psych
# Psych.safe_load("---\n foo: bar") # => {"foo"=>"bar"} # Psych.safe_load("---\n foo: bar") # => {"foo"=>"bar"}
# Psych.safe_load("---\n foo: bar", symbolize_names: true) # => {:foo=>"bar"} # Psych.safe_load("---\n foo: bar", symbolize_names: true) # => {:foo=>"bar"}
# #
def self.safe_load yaml, permitted_classes: [], permitted_symbols: [], aliases: false, filename: nil, fallback: nil, symbolize_names: false, freeze: false def self.safe_load yaml, permitted_classes: [], permitted_symbols: [], aliases: false, filename: nil, fallback: nil, symbolize_names: false, freeze: false, strict_integer: false
result = parse(yaml, filename: filename) result = parse(yaml, filename: filename)
return fallback unless result return fallback unless result
class_loader = ClassLoader::Restricted.new(permitted_classes.map(&:to_s), class_loader = ClassLoader::Restricted.new(permitted_classes.map(&:to_s),
permitted_symbols.map(&:to_s)) permitted_symbols.map(&:to_s))
scanner = ScalarScanner.new class_loader scanner = ScalarScanner.new class_loader, strict_integer: strict_integer
visitor = if aliases visitor = if aliases
Visitors::ToRuby.new scanner, class_loader, symbolize_names: symbolize_names, freeze: freeze Visitors::ToRuby.new scanner, class_loader, symbolize_names: symbolize_names, freeze: freeze
else else
@ -365,14 +365,15 @@ module Psych
# Raises a TypeError when `yaml` parameter is NilClass. This method is # Raises a TypeError when `yaml` parameter is NilClass. This method is
# similar to `safe_load` except that `Symbol` objects are allowed by default. # similar to `safe_load` except that `Symbol` objects are allowed by default.
# #
def self.load yaml, permitted_classes: [Symbol], permitted_symbols: [], aliases: false, filename: nil, fallback: nil, symbolize_names: false, freeze: false def self.load yaml, permitted_classes: [Symbol], permitted_symbols: [], aliases: false, filename: nil, fallback: nil, symbolize_names: false, freeze: false, strict_integer: false
safe_load yaml, permitted_classes: permitted_classes, safe_load yaml, permitted_classes: permitted_classes,
permitted_symbols: permitted_symbols, permitted_symbols: permitted_symbols,
aliases: aliases, aliases: aliases,
filename: filename, filename: filename,
fallback: fallback, fallback: fallback,
symbolize_names: symbolize_names, symbolize_names: symbolize_names,
freeze: freeze freeze: freeze,
strict_integer: strict_integer
end end
### ###

View File

@ -46,8 +46,8 @@ module Psych
# Convert this node to Ruby. # Convert this node to Ruby.
# #
# See also Psych::Visitors::ToRuby # See also Psych::Visitors::ToRuby
def to_ruby(symbolize_names: false, freeze: false) def to_ruby(symbolize_names: false, freeze: false, strict_integer: false)
Visitors::ToRuby.create(symbolize_names: symbolize_names, freeze: freeze).accept(self) Visitors::ToRuby.create(symbolize_names: symbolize_names, freeze: freeze, strict_integer: strict_integer).accept(self)
end end
alias :transform :to_ruby alias :transform :to_ruby

View File

@ -12,24 +12,32 @@ module Psych
FLOAT = /^(?:[-+]?([0-9][0-9_,]*)?\.[0-9]*([eE][-+][0-9]+)?(?# base 10))$/x FLOAT = /^(?:[-+]?([0-9][0-9_,]*)?\.[0-9]*([eE][-+][0-9]+)?(?# base 10))$/x
# Taken from http://yaml.org/type/int.html # Taken from http://yaml.org/type/int.html
INTEGER = /^(?:[-+]?0b[0-1_,]+ (?# base 2) INTEGER_STRICT = /^(?:[-+]?0b[0-1_]+ (?# base 2)
|[-+]?0[0-7_,]+ (?# base 8) |[-+]?0[0-7_]+ (?# base 8)
|[-+]?(?:0|[1-9](?:[0-9]|,[0-9]|_[0-9])*) (?# base 10) |[-+]?(0|[1-9][0-9_]*) (?# base 10)
|[-+]?0x[0-9a-fA-F_,]+ (?# base 16))$/x |[-+]?0x[0-9a-fA-F_]+ (?# base 16))$/x
# Same as above, but allows commas.
# Not to YML spec, but kept for backwards compatibility
INTEGER_LEGACY = /^(?:[-+]?0b[0-1_,]+ (?# base 2)
|[-+]?0[0-7_,]+ (?# base 8)
|[-+]?(?:0|[1-9](?:[0-9]|,[0-9]|_[0-9])*) (?# base 10)
|[-+]?0x[0-9a-fA-F_,]+ (?# base 16))$/x
attr_reader :class_loader attr_reader :class_loader
# Create a new scanner # Create a new scanner
def initialize class_loader def initialize class_loader, strict_integer: false
@symbol_cache = {} @symbol_cache = {}
@class_loader = class_loader @class_loader = class_loader
@strict_integer = strict_integer
end end
# Tokenize +string+ returning the Ruby object # Tokenize +string+ returning the Ruby object
def tokenize string def tokenize string
return nil if string.empty? return nil if string.empty?
return @symbol_cache[string] if @symbol_cache.key?(string) return @symbol_cache[string] if @symbol_cache.key?(string)
integer_regex = @strict_integer ? INTEGER_STRICT : INTEGER_LEGACY
# Check for a String type, being careful not to get caught by hash keys, hex values, and # Check for a String type, being careful not to get caught by hash keys, hex values, and
# special floats (e.g., -.inf). # special floats (e.g., -.inf).
if string.match?(%r{^[^\d.:-]?[[:alpha:]_\s!@#$%\^&*(){}<>|/\\~;=]+}) || string.match?(/\n/) if string.match?(%r{^[^\d.:-]?[[:alpha:]_\s!@#$%\^&*(){}<>|/\\~;=]+}) || string.match?(/\n/)
@ -89,7 +97,7 @@ module Psych
else else
Float(string.gsub(/[,_]|\.([Ee]|$)/, '\1')) Float(string.gsub(/[,_]|\.([Ee]|$)/, '\1'))
end end
elsif string.match?(INTEGER) elsif string.match?(integer_regex)
parse_int string parse_int string
else else
string string

View File

@ -12,9 +12,9 @@ module Psych
### ###
# This class walks a YAML AST, converting each node to Ruby # This class walks a YAML AST, converting each node to Ruby
class ToRuby < Psych::Visitors::Visitor class ToRuby < Psych::Visitors::Visitor
def self.create(symbolize_names: false, freeze: false) def self.create(symbolize_names: false, freeze: false, strict_integer: false)
class_loader = ClassLoader.new class_loader = ClassLoader.new
scanner = ScalarScanner.new class_loader scanner = ScalarScanner.new class_loader, strict_integer: strict_integer
new(scanner, class_loader, symbolize_names: symbolize_names, freeze: freeze) new(scanner, class_loader, symbolize_names: symbolize_names, freeze: freeze)
end end

View File

@ -43,5 +43,16 @@ module Psych
str = Psych.load('--- 1.1.1') str = Psych.load('--- 1.1.1')
assert_equal '1.1.1', str assert_equal '1.1.1', str
end end
# This behavior is not to YML spec, but is kept for backwards compatibility
def test_string_with_commas
number = Psych.load('--- 12,34,56')
assert_equal 123456, number
end
def test_string_with_commas_with_strict_integer
str = Psych.load('--- 12,34,56', strict_integer: true)
assert_equal '12,34,56', str
end
end end
end end

View File

@ -149,6 +149,31 @@ module Psych
assert_equal 0x123456789abcdef, ss.tokenize('0x12_,34,_56,_789abcdef__') assert_equal 0x123456789abcdef, ss.tokenize('0x12_,34,_56,_789abcdef__')
end end
def test_scan_strict_int_commas_and_underscores
# this test is to ensure adherance to YML spec using the 'strict_integer' option
scanner = Psych::ScalarScanner.new ClassLoader.new, strict_integer: true
assert_equal 123_456_789, scanner.tokenize('123_456_789')
assert_equal '123,456,789', scanner.tokenize('123,456,789')
assert_equal '1_2,3,4_5,6_789', scanner.tokenize('1_2,3,4_5,6_789')
assert_equal 1, scanner.tokenize('1')
assert_equal 1, scanner.tokenize('+1')
assert_equal(-1, scanner.tokenize('-1'))
assert_equal 0b010101010, scanner.tokenize('0b010101010')
assert_equal 0b010101010, scanner.tokenize('0b01_01_01_010')
assert_equal '0b0,1_0,1_,0,1_01,0', scanner.tokenize('0b0,1_0,1_,0,1_01,0')
assert_equal 01234567, scanner.tokenize('01234567')
assert_equal '0_,,,1_2,_34567', scanner.tokenize('0_,,,1_2,_34567')
assert_equal 0x123456789abcdef, scanner.tokenize('0x123456789abcdef')
assert_equal 0x123456789abcdef, scanner.tokenize('0x12_34_56_789abcdef')
assert_equal '0x12_,34,_56,_789abcdef', scanner.tokenize('0x12_,34,_56,_789abcdef')
assert_equal '0x_12_,34,_56,_789abcdef', scanner.tokenize('0x_12_,34,_56,_789abcdef')
assert_equal '0x12_,34,_56,_789abcdef__', scanner.tokenize('0x12_,34,_56,_789abcdef__')
end
def test_scan_dot def test_scan_dot
assert_equal '.', ss.tokenize('.') assert_equal '.', ss.tokenize('.')
end end