Cannot deserialize object from a JSON string (but only to Hash)? - ruby

I wrote the dictation gem on my Mac, and deserialization works fine. When I installed it on another Mac it would not work because it "fails" to deserialize object, because it can only deserialize to a Hash.
Private Mac Ruby version: ruby-1.9.3-p0, json v1.8.0
Another Mac Ruby version: ruby-1.9.3-p448, json v1.8.0
I also tried different Ruby versions and Gem versions on both, but none of them works, only the initial one where I first wrote it.
When I try this code in the working environment:
require 'json'
class Word
attr_accessor :value, :translation
def initialize(value, translation)
#value = value
#translation = translation
end
def to_json(*args)
{
'json_class' => self.class.name,
'data' => [ #value, #translation ]
}.to_json(*args)
end
class << self
def json_create(object)
new(*object['data'])
end
end
end
str = '{"json_class":"Word","data":["Morgen","Tomorrow"]}'
p JSON.parse(str)
It prints a Word object, which is expected:
#<Word:0x007fcce22c9c58 #translation="Tomorrow", #value="Morgen">
With the other environment, it always prints a Hash:
{"json_class"=>"Word", "data"=>["Morgen", "Tomorrow"]}
I also tried to pass :object_class key, it throws another exception:
p JSON.parse(str, :object_class => Word)
# => ArgumentError: wrong number of arguments (0 for 2)
I could not figure out the require 'json' version during runtime using:
puts Gem.loaded_specs['json'].version
because Gem.loaded_specs.keys doesn't contain it.
Thanks for any hint.

Replied from the author of JSON lib - on newer version, due to security reason, to deserialize custom object, either you can:
JSON.parse(str, :create_additions => true)
or you can:
JSON.load(str)
So, I overlooked the JSON#load part in ruby-doc:
load(source, proc = nil, options = {})
Load a ruby data structure from a JSON source and return it. A source
can either be a string-like object, an IO-like object, or an object
responding to the read method. If proc was given, it will be called
with any nested Ruby object as an argument recursively in depth first
order. To modify the default options pass in the optional options
argument as well.
BEWARE: This method is meant to serialise data from trusted user
input, like from your own database server or clients under your
control, it could be dangerous to allow untrusted users to pass JSON
sources into it. The default options for the parser can be changed via
the ::load_default_options method.
This method is part of the implementation of the load/dump interface
of Marshal and YAML.

Deserializing directly into a rich object (especially if your JSON comes from an unknown source) can be a pretty serious attack vector (recent Rails vulnerabilities are related to that).
I would guess that this ability was disabled between Ruby versions, or, at least changed to a whitelist-based approach. I wasn't able to find any links to support this claim though, so I might be wrong.
Anyway, you might find it simpler and more compatible to initialize your class from the deserialized hash instead:
class Word
def self.from_json(json)
args = JSON.parse(json)["data"];
new(*args)
end
end

Here is another workaround, because my code is not used in web communication, vulnerability is not a problem here.
Before I was doing:
JSON.parse(str)
Now just need to add few lines:
obj = JSON.parse(str)
if obj.is_a?(Hash)
class_name = obj['json_class'].split('::').inject(Kernel) { |namespace, const_name| namespace.const_get(const_name) }
args = obj['data']
word = class_name.new(*args)
else
word = obj
end

Related

How to inspect the body of a method? [duplicate]

I would like to know whether I can get source code a method on the fly, and whether I can get which file is this method in.
like
A.new.method(:a).SOURCE_CODE
A.new.method(:a).FILE
Use source_location:
class A
def foo
end
end
file, line = A.instance_method(:foo).source_location
# or
file, line = A.new.method(:foo).source_location
puts "Method foo is defined in #{file}, line #{line}"
# => "Method foo is defined in temp.rb, line 2"
Note that for builtin methods, source_location returns nil. If want to check out the C source code (have fun!), you'll have to look for the right C file (they're more or less organized by class) and find the rb_define_method for the method (towards the end of the file).
In Ruby 1.8 this method does not exist, but you can use this gem.
None of the answers so far show how to display the source code of a method on the fly...
It's actually very easy if you use the awesome 'method_source' gem by John Mair (the maker of Pry):
The method has to be implemented in Ruby (not C), and has to be loaded from a file (not irb).
Here's an example displaying the method source code in the Rails console with method_source:
$ rails console
> require 'method_source'
> I18n::Backend::Simple.instance_method(:lookup).source.display
def lookup(locale, key, scope = [], options = {})
init_translations unless initialized?
keys = I18n.normalize_keys(locale, key, scope, options[:separator])
keys.inject(translations) do |result, _key|
_key = _key.to_sym
return nil unless result.is_a?(Hash) && result.has_key?(_key)
result = result[_key]
result = resolve(locale, _key, result, options.merge(:scope => nil)) if result.is_a?(Symbol)
result
end
end
=> nil
See also:
https://rubygems.org/gems/method_source
https://github.com/banister/method_source
http://banisterfiend.wordpress.com/
Here is how to print out the source code from ruby:
puts File.read(OBJECT_TO_GET.method(:METHOD_FROM).source_location[0])
Without dependencies
method = SomeConstant.method(:some_method_name)
file_path, line = method.source_location
# puts 10 lines start from the method define
IO.readlines(file_path)[line-1, 10]
If you want use this more conveniently, your can open the Method class:
# ~/.irbrc
class Method
def source(limit=10)
file, line = source_location
if file && line
IO.readlines(file)[line-1,limit]
else
nil
end
end
end
And then just call method.source
With Pry you can use the show-method to view a method source, and you can even see some ruby c source code with pry-doc installed, according pry's doc in codde-browing
Note that we can also view C methods (from Ruby Core) using the
pry-doc plugin; we also show off the alternate syntax for show-method:
pry(main)> show-method Array#select
From: array.c in Ruby Core (C Method):
Number of lines: 15
static VALUE
rb_ary_select(VALUE ary)
{
VALUE result;
long i;
RETURN_ENUMERATOR(ary, 0, 0);
result = rb_ary_new2(RARRAY_LEN(ary));
for (i = 0; i < RARRAY_LEN(ary); i++) {
if (RTEST(rb_yield(RARRAY_PTR(ary)[i]))) {
rb_ary_push(result, rb_ary_elt(ary, i));
}
}
return result;
}
I created the "ri_for" gem for this purpose
>> require 'ri_for'
>> A.ri_for :foo
... outputs the source (and location, if you're on 1.9).
GL.
-r
Internal methods don't have source or source location (e.g. Integer#to_s)
require 'method_source'
User.method(:last).source
User.method(:last).source_location
I had to implement a similar feature (grab the source of a block) as part of Wrong and you can see how (and maybe even reuse the code) in chunk.rb (which relies on Ryan Davis' RubyParser as well as some pretty funny source file glomming code). You'd have to modify it to use Method#source_location and maybe tweak some other things so it does or doesn't include the def.
BTW I think Rubinius has this feature built in. For some reason it's been left out of MRI (the standard Ruby implementation), hence this hack.
Oooh, I like some of the stuff in method_source! Like using eval to tell if an expression is valid (and keep glomming source lines until you stop getting parse errors, like Chunk does)...

Ruby syntax error unexpected tLABEL

I have the following configuration for using redis as a rails cache, adapted closely from the rails doc:
Rails.application.configure do
...
config.cache_store = :redis_cache_store, url: "redis://example.com:6379/0"
...
end
Using rails 5.2.0.beta2 and ruby 2.3.1p112, my IDE is telling me the following things about the config.cache_store = line:
unused literal ignored
syntax error, unexpected tLABEL
If I enclose the hash in {}, the errors go away, i.e.:
config.cache_store = :redis_cache_store, { url: "redis://example.com:6379/0" }
Is the documentation in error, or is it a ruby version thing (e.g. maybe the doc is assuming ruby >= 2.4)?
In any case, I don't understand what the assignment is actually doing - how can two things be assigned to one?
When the config block hits cache_store and a symbol is given, it assumes it is a cache store class and does things like:
:redis_cache_store.to_s.classify.constantize
# => RedisCacheStore
which is presumably a class. You can infer that Rails does this by looking at the docs and seeing that you can also pass a class.
RedisCacheStore is defined inside the ActiveSupport::Cache namespace, which inherits from ActiveSupport::Cache::Store, which is defined in cache.rb. This store class only initializes with a single argument of options = nil:
def initialize(options = nil)
#options = options ? options.dup : {}
end
which is indicative of a requirement to pass a hash. For one reason or another, it doesn't initialize with options = {}, which would make your initial code correct. Hence, passing a hash like in your second example resolved the error.

What are the equivalent to Lodash's get and set in Ruby?

I would like to use something similar to Lodash's get and set, but in Ruby instead of JavaScript. I tried few searches but I can't find anything similar.
Lodash's documentation will probably explain it in a better way, but it's getting and setting a property from a string path ('x[0].y.z' for example). If the full path doesn't exist when setting a property, it is automatically created.
Lodash Set
Lodash Get
I eventually ported Lodash _.set and _.get from JavaScript to Ruby and made a Gem.
Ruby 2.3 introduces the new safe navigator operator for getting nested/chained values:
x[0]&.y&.z #=> result or nil
Otherwise, Rails monkey patches all objects with try(…), allowing you to:
x[0].try(:y).try(:z) #=> result or nil
Setting is a bit harder, and I'd recommend ensuring you have the final object before attempting to set a property, e.g.:
if obj = x[0]&.y&.z
z.name = "Dr Robot"
end
You can use the Rudash Gem that comes with most of the Lodash utilities, and not only the _.get and _.set.
Sometimes I have had the need to programmatically get the value for a property deep into an object, but the thing is that sometimes the property is really a method, and sometimes it needs parameters!
So I came up with this solution, hope it helps devising one for your problem:
(Needs Rails' #try)
def reduce_attributes_for( object, options )
options.reduce( {} ) do |hash, ( attribute, methods )|
hash[attribute] = methods.reduce( object ) { |a, e| a.try!(:send, *e) }
hash
end
end
# Usage example
o = Object.new
attribute_map = {
# same as o.object_id
id: [:object_id],
# same as o.object_id.to_s
id_as_string: [:object_id, :to_s],
# same as o.object_id.to_s.length
id_as_string_length: [:object_id, :to_s, :length],
# I know, this one is a contrived example, but its purpose is
# to illustrate how you would call methods with parameters
# same as o.object_id.to_s.scan(/\d/)[1].to_i
second_number_from_id: [:object_id, :to_s, [:scan, /\d/], [:[],1], :to_i]
}
reduce_attributes_for( o, attribute_map )
# {:id=>47295942175460,
# :id_as_string=>"47295942175460",
# :id_as_string_length=>14,
# :second_number_from_id=>7}

How to get to_json work on BSON::Code object as well in mongo-ruby-driver

I am trying 'mongo-ruby-driver' for some project. It's working fine except when I call to_json on mongo object. It gives well formed json but it's not converting BSON::Code into readable value i resulting JSON.
Instead of showing code text, it show something like
#<BSON::Code:0x00000100af6fa8>
Did anyone tried it. Any help id highly appreciated.
UPDATE
here is some code snippet:
#records is variable that contains Array of MongoDB documents in hash.
#records.to_json
When I call to_json on it it gives everything as expected. except for the key that contain BSON::Code (means javascript code). for example consider following doc:
{
"_id" : "contains",
"value" : function( obj, target ) { return obj.indexOf(target) != -1; };
}
Querying same doc from ruby gives output like:
{
"_id"=>"contains",
"value"=><BSON::Code:2160165280 #data="function( obj, target ) { return obj.indexOf(target) != -1; };" #scope="{}">
}
and calling to_json on this gives following:
{"_id":"contains","value":"#<BSON::Code:0x00000100b54658>"}
this is what the problem is. Instead of getting actual code for 'value' key I am getting ruby object as string.
Calling code method on BSON::Code we can get it converted into code. But for that I need to loop mongo docs, check values for each key, calling code on it if it is an object of BSON::Code and then assigning it back to key. And at last we can call to_json on it. But I dont want this much overhead. I need to_json itself should take care of it.
The serializer that is iterating over the attributes of the objects in the array is probably calling the to_s method which would cause the output you are seeing.
You can either monkey patch the BSON::Code class to include a to_s method which calls inspect (the method that produces the output you want) or modify the serializer to detect when it encounters a BSON::Code instance and call inspect on it rather than to_s.
The code to mokey patch the BSON::Code class would look like this:
module BSON
class Code
def to_s
inspect
end
end
end
This would have the same behavior as inspect. If you just wanted the code you could monkey patch this in:
module BSON
class Code
def to_s
#code
end
end
end
I might make this the default behavior for to_s in the driver but for now just include that in your code and it should work like a champ.
Which Mongo object, do you mean the module? If you could, please post the code and what you are trying to do.
The only to_json I see in the driver is BSON::ObjectId#to_json which (from the docs) is described as a method that does the following:
Convert to MongoDB extended JSON format. Since JSON includes type information, but lacks an ObjectId type, this JSON format encodes the type using an $oid key.
I get the same results when I use it:
1.9.3-p0 :001 > require 'mongo'
=> true
1.9.3-p0 :002 > BSON::ObjectId
=> BSON::ObjectId
1.9.3-p0 :003 > BSON::ObjectId.new()
=> BSON::ObjectId('4f17350eadd361e91d000001')
1.9.3-p0 :004 > BSON::ObjectId.new().to_json
=> "{\"$oid\": \"4f173512add361e91d000002\"}"
BSON ( http://bsonspec.org/ ) is a binary representation of JSON ( http://www.json.org/ ) . JSON is not meant to be used to describe functions, as it is meant to be portable. Functions in this manner are not portable to other systems. So there is no way to serialize it. There is some hacks defined here that may get you what you need, but ultimately, BSON/JSON may be a weird tool for the job if your trying to serialize functions.

Is there a way to print Ruby function source, like JavaScript's Function#toString / toSource? [duplicate]

I would like to know whether I can get source code a method on the fly, and whether I can get which file is this method in.
like
A.new.method(:a).SOURCE_CODE
A.new.method(:a).FILE
Use source_location:
class A
def foo
end
end
file, line = A.instance_method(:foo).source_location
# or
file, line = A.new.method(:foo).source_location
puts "Method foo is defined in #{file}, line #{line}"
# => "Method foo is defined in temp.rb, line 2"
Note that for builtin methods, source_location returns nil. If want to check out the C source code (have fun!), you'll have to look for the right C file (they're more or less organized by class) and find the rb_define_method for the method (towards the end of the file).
In Ruby 1.8 this method does not exist, but you can use this gem.
None of the answers so far show how to display the source code of a method on the fly...
It's actually very easy if you use the awesome 'method_source' gem by John Mair (the maker of Pry):
The method has to be implemented in Ruby (not C), and has to be loaded from a file (not irb).
Here's an example displaying the method source code in the Rails console with method_source:
$ rails console
> require 'method_source'
> I18n::Backend::Simple.instance_method(:lookup).source.display
def lookup(locale, key, scope = [], options = {})
init_translations unless initialized?
keys = I18n.normalize_keys(locale, key, scope, options[:separator])
keys.inject(translations) do |result, _key|
_key = _key.to_sym
return nil unless result.is_a?(Hash) && result.has_key?(_key)
result = result[_key]
result = resolve(locale, _key, result, options.merge(:scope => nil)) if result.is_a?(Symbol)
result
end
end
=> nil
See also:
https://rubygems.org/gems/method_source
https://github.com/banister/method_source
http://banisterfiend.wordpress.com/
Here is how to print out the source code from ruby:
puts File.read(OBJECT_TO_GET.method(:METHOD_FROM).source_location[0])
Without dependencies
method = SomeConstant.method(:some_method_name)
file_path, line = method.source_location
# puts 10 lines start from the method define
IO.readlines(file_path)[line-1, 10]
If you want use this more conveniently, your can open the Method class:
# ~/.irbrc
class Method
def source(limit=10)
file, line = source_location
if file && line
IO.readlines(file)[line-1,limit]
else
nil
end
end
end
And then just call method.source
With Pry you can use the show-method to view a method source, and you can even see some ruby c source code with pry-doc installed, according pry's doc in codde-browing
Note that we can also view C methods (from Ruby Core) using the
pry-doc plugin; we also show off the alternate syntax for show-method:
pry(main)> show-method Array#select
From: array.c in Ruby Core (C Method):
Number of lines: 15
static VALUE
rb_ary_select(VALUE ary)
{
VALUE result;
long i;
RETURN_ENUMERATOR(ary, 0, 0);
result = rb_ary_new2(RARRAY_LEN(ary));
for (i = 0; i < RARRAY_LEN(ary); i++) {
if (RTEST(rb_yield(RARRAY_PTR(ary)[i]))) {
rb_ary_push(result, rb_ary_elt(ary, i));
}
}
return result;
}
I created the "ri_for" gem for this purpose
>> require 'ri_for'
>> A.ri_for :foo
... outputs the source (and location, if you're on 1.9).
GL.
-r
Internal methods don't have source or source location (e.g. Integer#to_s)
require 'method_source'
User.method(:last).source
User.method(:last).source_location
I had to implement a similar feature (grab the source of a block) as part of Wrong and you can see how (and maybe even reuse the code) in chunk.rb (which relies on Ryan Davis' RubyParser as well as some pretty funny source file glomming code). You'd have to modify it to use Method#source_location and maybe tweak some other things so it does or doesn't include the def.
BTW I think Rubinius has this feature built in. For some reason it's been left out of MRI (the standard Ruby implementation), hence this hack.
Oooh, I like some of the stuff in method_source! Like using eval to tell if an expression is valid (and keep glomming source lines until you stop getting parse errors, like Chunk does)...

Resources