faraday_json 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.gitignore +38 -0
- data/.travis.yml +25 -0
- data/Gemfile +18 -0
- data/Gemfile-0.7.rb +3 -0
- data/Gemfile-0.7.rb.lock +57 -0
- data/Gemfile-0.8.rb +3 -0
- data/Gemfile-0.8.rb.lock +53 -0
- data/Gemfile.lock +52 -0
- data/LICENSE +30 -0
- data/README.md +30 -0
- data/Rakefile +28 -0
- data/faraday_json.gemspec +25 -0
- data/lib/faraday_json.rb +22 -0
- data/lib/faraday_json/encode_json.rb +94 -0
- data/lib/faraday_json/encoding.rb +219 -0
- data/lib/faraday_json/parse_json.rb +165 -0
- data/lib/faraday_json/version.rb +10 -0
- data/spec/data/iso8859-15_file.json +1 -0
- data/spec/data/utf16be_file.json +0 -0
- data/spec/data/utf16le_file.json +0 -0
- data/spec/data/utf8_file.json +1 -0
- data/spec/encode_json_spec.rb +376 -0
- data/spec/helper.rb +74 -0
- data/spec/parse_json_spec.rb +217 -0
- metadata +123 -0
@@ -0,0 +1,219 @@
|
|
1
|
+
#
|
2
|
+
# FaradayJSON
|
3
|
+
# https://github.com/spriteCloud/faraday_json
|
4
|
+
#
|
5
|
+
# Copyright (c) 2015 spriteCloud B.V. and other FaradayJSON contributors.
|
6
|
+
# All rights reserved.
|
7
|
+
#
|
8
|
+
|
9
|
+
module FaradayJSON
|
10
|
+
|
11
|
+
# Character encoding helper functions
|
12
|
+
module Encoding
|
13
|
+
|
14
|
+
# Two versions of transcode, one for Ruby 1.8 and one for greater versions.
|
15
|
+
|
16
|
+
if RUBY_VERSION.start_with?("1.8")
|
17
|
+
|
18
|
+
def transcode(data, input_charset, output_charset, opts = {})
|
19
|
+
# In Ruby 1.8, we pretty much have to believe the given charsets; there's
|
20
|
+
# not a lot of choice.
|
21
|
+
|
22
|
+
# If we don't have an input charset, we can't do better than US-ASCII.
|
23
|
+
if input_charset.nil? or input_charset.empty?
|
24
|
+
input_charset = opts.fetch('default_input_charset', 'us-ascii')
|
25
|
+
end
|
26
|
+
|
27
|
+
# The default output charset, on the other hand, should be UTF-8.
|
28
|
+
if output_charset.nil? or output_charset.empty?
|
29
|
+
output_charset = opts.fetch('default_output_charset', 'UTF-8//IGNORE')
|
30
|
+
end
|
31
|
+
|
32
|
+
# Transcode using iconv
|
33
|
+
require 'iconv'
|
34
|
+
return ::Iconv.conv(output_charset, input_charset, data)
|
35
|
+
end
|
36
|
+
|
37
|
+
else # end ruby 1.8/start ruby > 1.8
|
38
|
+
|
39
|
+
def transcode(data, input_charset, output_charset, opts = {})
|
40
|
+
# Strings have an encode function in Ruby > 1.8
|
41
|
+
if not data.respond_to?(:encode)
|
42
|
+
return data
|
43
|
+
end
|
44
|
+
|
45
|
+
# If we don't have a charset, just use whatever is in the string
|
46
|
+
# currently. If we do have a charset, we'll have to run some extra
|
47
|
+
# checks.
|
48
|
+
if not (input_charset.nil? or input_charset.empty?)
|
49
|
+
# Check passed charset is *understood* by finding it. If this fails,
|
50
|
+
# an exception is raised, which it also should be.
|
51
|
+
canonical = ::Encoding.find(input_charset)
|
52
|
+
|
53
|
+
# Second, ensure the canonical charset and the actual string encoding
|
54
|
+
# are identical. If not, we'll have to do a little more than just
|
55
|
+
# transcode to UTF-8.
|
56
|
+
if canonical != data.encoding
|
57
|
+
if opts.fetch('force_input_charset', false)
|
58
|
+
data.force_encoding(canonical)
|
59
|
+
else
|
60
|
+
raise "Provided charset was #{canonical}, but data was #{data.encoding}"
|
61
|
+
end
|
62
|
+
end
|
63
|
+
end
|
64
|
+
|
65
|
+
# If there's no output charset, we should default to UTF-8.
|
66
|
+
if output_charset.nil? or output_charset.empty?
|
67
|
+
output_charset = opts.fetch('default_output_charset', 'UTF-8')
|
68
|
+
end
|
69
|
+
|
70
|
+
# Transcode!
|
71
|
+
return data.encode(output_charset)
|
72
|
+
end
|
73
|
+
|
74
|
+
end # ruby > 1.8
|
75
|
+
|
76
|
+
# Convenient helper. Output is UTF-8. Input is either a string, or some data
|
77
|
+
# data. There's a Ruby 1.8 version mostly because it has to iteratively convert
|
78
|
+
# included strings.
|
79
|
+
if RUBY_VERSION.start_with?("1.8")
|
80
|
+
|
81
|
+
def to_utf8(data, charset, opts = {})
|
82
|
+
if data.is_a? Hash
|
83
|
+
transcoded = {}
|
84
|
+
data.each do |key, value|
|
85
|
+
transcoded[to_utf8(key, charset, opts)] = to_utf8(value, charset, opts)
|
86
|
+
end
|
87
|
+
return transcoded
|
88
|
+
elsif data.is_a? Array
|
89
|
+
transcoded = []
|
90
|
+
data.each do |value|
|
91
|
+
transcoded << to_utf8(value, charest, opts)
|
92
|
+
end
|
93
|
+
return transcoded
|
94
|
+
elsif data.is_a? String
|
95
|
+
return transcode(data, charset, 'UTF-8//IGNORE', opts)
|
96
|
+
else
|
97
|
+
return data
|
98
|
+
end
|
99
|
+
end
|
100
|
+
|
101
|
+
else # end ruby 1.8/start ruby > 1.8
|
102
|
+
|
103
|
+
def to_utf8(data, charset, opts = {})
|
104
|
+
return transcode(data, charset, 'UTF-8', opts)
|
105
|
+
end
|
106
|
+
|
107
|
+
end # ruby > 1.8
|
108
|
+
|
109
|
+
|
110
|
+
# Helper function; strips a BOM for UTF-16 encodings
|
111
|
+
def strip_bom(data, charset, opts = {})
|
112
|
+
# Only need to do this on Strings
|
113
|
+
if not data.is_a? String
|
114
|
+
return data
|
115
|
+
end
|
116
|
+
|
117
|
+
# If the charset is given, it overrides string internal encoding.
|
118
|
+
enc = get_dominant_encoding(data, charset, opts)
|
119
|
+
|
120
|
+
# Make the encoding canonical (if we can find out about that).
|
121
|
+
canonical = get_canonical_encoding(enc)
|
122
|
+
|
123
|
+
# Determine what a BOM would look like.
|
124
|
+
bom = get_bom(canonical)
|
125
|
+
|
126
|
+
# We can't operate on data, we need a byte array.
|
127
|
+
arr = data.each_byte.to_a
|
128
|
+
|
129
|
+
# Match BOM
|
130
|
+
found = true
|
131
|
+
bom.each_index do |i|
|
132
|
+
if bom[i] != arr[i]
|
133
|
+
found = false
|
134
|
+
break
|
135
|
+
end
|
136
|
+
end
|
137
|
+
|
138
|
+
# So we may have found a BOM! Strip it.
|
139
|
+
if found
|
140
|
+
ret = arr[bom.length..-1].pack('c*')
|
141
|
+
if ret.respond_to? :force_encoding
|
142
|
+
ret.force_encoding(canonical)
|
143
|
+
end
|
144
|
+
return ret
|
145
|
+
end
|
146
|
+
|
147
|
+
# No BOM
|
148
|
+
return data
|
149
|
+
end
|
150
|
+
|
151
|
+
# Given a String with (potentially, this depends on Ruby version) an encoding,
|
152
|
+
# and a charset from a content-type header (which may be nil), determines the
|
153
|
+
# dominant encoding. (Charset, if given, overrides internal encoding,
|
154
|
+
# if present).
|
155
|
+
def get_dominant_encoding(str, charset, opts = {})
|
156
|
+
enc = nil
|
157
|
+
if str.respond_to? :encoding
|
158
|
+
enc = str.encoding
|
159
|
+
end
|
160
|
+
|
161
|
+
if charset.nil? or charset.empty?
|
162
|
+
if enc.nil?
|
163
|
+
default_encoding = opts.fetch('default_encoding', nil)
|
164
|
+
if default_encoding.nil?
|
165
|
+
raise "No charset provided, don't know what to do!" # FIXME
|
166
|
+
end
|
167
|
+
enc = default_encoding
|
168
|
+
end
|
169
|
+
else
|
170
|
+
enc = charset
|
171
|
+
end
|
172
|
+
|
173
|
+
return enc
|
174
|
+
end
|
175
|
+
|
176
|
+
|
177
|
+
# Returns a canonical version of an encoding.
|
178
|
+
def get_canonical_encoding(enc)
|
179
|
+
if defined? ::Encoding and ::Encoding.respond_to? :find
|
180
|
+
# Oh... Ruby 1.9.2 doesn't like passing an Encoding to find()...
|
181
|
+
if not enc.is_a? ::Encoding
|
182
|
+
enc = ::Encoding.find(enc)
|
183
|
+
end
|
184
|
+
return enc.to_s.downcase
|
185
|
+
end
|
186
|
+
return enc.downcase
|
187
|
+
end
|
188
|
+
|
189
|
+
|
190
|
+
# Given a (canonical) encoding, returns a BOM as an array of byte values. If
|
191
|
+
# the given encoding does not have a BOM, an empty array is returned.
|
192
|
+
def get_bom(enc)
|
193
|
+
bom = []
|
194
|
+
if enc.start_with?('utf16be') or enc.start_with?('utf-16be')
|
195
|
+
bom = [0xfe, 0xff]
|
196
|
+
elsif enc.start_with?('utf16le') or enc.start_with?('utf-16le')
|
197
|
+
bom = [0xff, 0xfe]
|
198
|
+
elsif enc.start_with?('utf8') or enc.start_with?('utf-8')
|
199
|
+
bom = [0xef, 0xbb, 0xbf]
|
200
|
+
elsif enc.start_with?('utf32be') or enc.start_with?('utf-32be')
|
201
|
+
bom = [0x00, 0x00, 0xfe, 0xff]
|
202
|
+
elsif enc.start_with?('utf32le') or enc.start_with?('utf-32le')
|
203
|
+
bom = [0xff, 0xfe, 0x00, 0x00]
|
204
|
+
end
|
205
|
+
return bom
|
206
|
+
end
|
207
|
+
|
208
|
+
|
209
|
+
|
210
|
+
# Helper function for testing
|
211
|
+
def bin_to_hex(data)
|
212
|
+
if data.respond_to? :each_byte
|
213
|
+
return data.each_byte.map { |b| b.to_s(16) }.join
|
214
|
+
end
|
215
|
+
return data
|
216
|
+
end
|
217
|
+
|
218
|
+
end # module Encoding
|
219
|
+
end # module FaradayJSON
|
@@ -0,0 +1,165 @@
|
|
1
|
+
#
|
2
|
+
# FaradayJSON
|
3
|
+
# https://github.com/spriteCloud/faraday_json
|
4
|
+
#
|
5
|
+
# Copyright (c) 2015 spriteCloud B.V. and other FaradayJSON contributors.
|
6
|
+
# All rights reserved.
|
7
|
+
#
|
8
|
+
|
9
|
+
require 'faraday_json/encoding'
|
10
|
+
|
11
|
+
module FaradayJSON
|
12
|
+
# Public: Parse response bodies as JSON.
|
13
|
+
class ParseJson < Faraday::Middleware
|
14
|
+
CONTENT_TYPE = 'Content-Type'.freeze
|
15
|
+
|
16
|
+
include ::FaradayJSON::Encoding
|
17
|
+
|
18
|
+
dependency do
|
19
|
+
require 'json' unless defined?(::JSON)
|
20
|
+
end
|
21
|
+
|
22
|
+
def initialize(app = nil, options = {})
|
23
|
+
super(app)
|
24
|
+
@options = options
|
25
|
+
@content_types = Array(options[:content_type])
|
26
|
+
end
|
27
|
+
|
28
|
+
def call(environment)
|
29
|
+
@app.call(environment).on_complete do |env|
|
30
|
+
if process_response_type?(response_type(env)) and parse_response?(env)
|
31
|
+
process_response(env)
|
32
|
+
end
|
33
|
+
end
|
34
|
+
end
|
35
|
+
|
36
|
+
def process_response(env)
|
37
|
+
env[:raw_body] = env[:body] if preserve_raw?(env)
|
38
|
+
body = env[:body]
|
39
|
+
|
40
|
+
# Body will be in an unknown encoding. Use charset field to coerce it to
|
41
|
+
# internal UTF-8.
|
42
|
+
charset = response_charset(env)
|
43
|
+
|
44
|
+
# We must ensure we're interpreting the body as the right charset. First,
|
45
|
+
# strip the BOM (if any).
|
46
|
+
body = strip_bom(body, charset, { 'default_encoding' => 'us-ascii' })
|
47
|
+
|
48
|
+
# Transcode to UTF-8
|
49
|
+
body = to_utf8(body, charset, { 'force_input_charset' => true })
|
50
|
+
|
51
|
+
# Now that's done, parse the JSON.
|
52
|
+
ret = nil
|
53
|
+
begin
|
54
|
+
ret = ::JSON.parse(body) unless body.strip.empty?
|
55
|
+
rescue StandardError, SyntaxError => err
|
56
|
+
raise err if err.is_a? SyntaxError and err.class.name != 'Psych::SyntaxError'
|
57
|
+
raise Faraday::Error::ParsingError, err
|
58
|
+
end
|
59
|
+
env[:body] = ret
|
60
|
+
end
|
61
|
+
|
62
|
+
def response_type(env)
|
63
|
+
type = env[:response_headers][CONTENT_TYPE].to_s
|
64
|
+
type = type.split(';', 2).first if type.index(';')
|
65
|
+
type
|
66
|
+
end
|
67
|
+
|
68
|
+
def response_charset(env)
|
69
|
+
header = env[:response_headers][CONTENT_TYPE].to_s
|
70
|
+
if header.index(';')
|
71
|
+
header.split(';').each do |part|
|
72
|
+
if part.index('charset=')
|
73
|
+
return part.split('charset=', 2).last
|
74
|
+
end
|
75
|
+
end
|
76
|
+
end
|
77
|
+
return nil
|
78
|
+
end
|
79
|
+
|
80
|
+
def process_response_type?(type)
|
81
|
+
@content_types.empty? or @content_types.any? { |pattern|
|
82
|
+
pattern.is_a?(Regexp) ? type =~ pattern : type == pattern
|
83
|
+
}
|
84
|
+
end
|
85
|
+
|
86
|
+
def parse_response?(env)
|
87
|
+
env[:body].respond_to? :to_str
|
88
|
+
end
|
89
|
+
|
90
|
+
def preserve_raw?(env)
|
91
|
+
env[:request].fetch(:preserve_raw, @options[:preserve_raw])
|
92
|
+
end
|
93
|
+
|
94
|
+
|
95
|
+
|
96
|
+
# DRAGONS
|
97
|
+
module OptionsExtension
|
98
|
+
attr_accessor :preserve_raw
|
99
|
+
|
100
|
+
def to_hash
|
101
|
+
super.update(:preserve_raw => preserve_raw)
|
102
|
+
end
|
103
|
+
|
104
|
+
def each
|
105
|
+
return to_enum(:each) unless block_given?
|
106
|
+
super
|
107
|
+
yield :preserve_raw, preserve_raw
|
108
|
+
end
|
109
|
+
|
110
|
+
def fetch(key, *args)
|
111
|
+
if :preserve_raw == key
|
112
|
+
value = __send__(key)
|
113
|
+
value.nil? ? args.fetch(0) : value
|
114
|
+
else
|
115
|
+
super
|
116
|
+
end
|
117
|
+
end
|
118
|
+
end
|
119
|
+
|
120
|
+
if defined?(Faraday::RequestOptions)
|
121
|
+
begin
|
122
|
+
Faraday::RequestOptions.from(:preserve_raw => true)
|
123
|
+
rescue NoMethodError
|
124
|
+
Faraday::RequestOptions.send(:include, OptionsExtension)
|
125
|
+
end
|
126
|
+
end
|
127
|
+
end # class ParseJson
|
128
|
+
|
129
|
+
# Public: Override the content-type of the response with "application/json"
|
130
|
+
# if the response body looks like it might be JSON, i.e. starts with an
|
131
|
+
# open bracket.
|
132
|
+
#
|
133
|
+
# This is to fix responses from certain API providers that insist on serving
|
134
|
+
# JSON with wrong MIME-types such as "text/javascript".
|
135
|
+
class ParseJsonMimeTypeFix < ParseJson
|
136
|
+
MIME_TYPE = 'application/json'.freeze
|
137
|
+
|
138
|
+
def process_response(env)
|
139
|
+
old_type = env[:response_headers][CONTENT_TYPE].to_s
|
140
|
+
new_type = MIME_TYPE.dup
|
141
|
+
new_type << ';' << old_type.split(';', 2).last if old_type.index(';')
|
142
|
+
env[:response_headers][CONTENT_TYPE] = new_type
|
143
|
+
end
|
144
|
+
|
145
|
+
BRACKETS = %w- [ { -
|
146
|
+
WHITESPACE = [ " ", "\n", "\r", "\t" ]
|
147
|
+
|
148
|
+
def parse_response?(env)
|
149
|
+
super and BRACKETS.include? first_char(env[:body])
|
150
|
+
end
|
151
|
+
|
152
|
+
def first_char(body)
|
153
|
+
idx = -1
|
154
|
+
begin
|
155
|
+
char = body[idx += 1]
|
156
|
+
char = char.chr if char
|
157
|
+
end while char and WHITESPACE.include? char
|
158
|
+
char
|
159
|
+
end
|
160
|
+
end # class ParseJson
|
161
|
+
|
162
|
+
end
|
163
|
+
|
164
|
+
# deprecated alias
|
165
|
+
Faraday::Response::ParseJson = FaradayJSON::ParseJson
|
@@ -0,0 +1 @@
|
|
1
|
+
{"a":"Hell�, W�rld!"}
|
Binary file
|
Binary file
|
@@ -0,0 +1 @@
|
|
1
|
+
{"a":"Hellö, Wörld!"}
|
@@ -0,0 +1,376 @@
|
|
1
|
+
# encoding: utf-8
|
2
|
+
|
3
|
+
require 'helper'
|
4
|
+
require 'faraday_json/encode_json'
|
5
|
+
|
6
|
+
describe FaradayJSON::EncodeJson do
|
7
|
+
let(:middleware) { described_class.new(lambda{|env| env}) }
|
8
|
+
|
9
|
+
def process(body, content_type = nil)
|
10
|
+
env = {:body => body, :request_headers => Faraday::Utils::Headers.new}
|
11
|
+
env[:request_headers]['content-type'] = content_type if content_type
|
12
|
+
middleware.call(faraday_env(env))
|
13
|
+
end
|
14
|
+
|
15
|
+
def result_body() result[:body] end
|
16
|
+
def result_type() result[:request_headers]['content-type'] end
|
17
|
+
def result_length() result[:request_headers]['content-length'].to_i end
|
18
|
+
|
19
|
+
context "no body" do
|
20
|
+
let(:result) { process(nil) }
|
21
|
+
|
22
|
+
it "doesn't change body" do
|
23
|
+
expect(result_body).to be_nil
|
24
|
+
end
|
25
|
+
|
26
|
+
it "doesn't add content type" do
|
27
|
+
expect(result_type).to be_nil
|
28
|
+
end
|
29
|
+
end
|
30
|
+
|
31
|
+
context "empty body" do
|
32
|
+
let(:result) { process('') }
|
33
|
+
|
34
|
+
it "doesn't change body" do
|
35
|
+
expect(result_body).to be_empty
|
36
|
+
end
|
37
|
+
|
38
|
+
it "doesn't add content type" do
|
39
|
+
expect(result_type).to be_nil
|
40
|
+
end
|
41
|
+
end
|
42
|
+
|
43
|
+
context "string body" do
|
44
|
+
let(:result) { process('{"a":1}') }
|
45
|
+
|
46
|
+
it "doesn't change body" do
|
47
|
+
expect(result_body).to eq('{"a":1}')
|
48
|
+
end
|
49
|
+
|
50
|
+
it "adds content type" do
|
51
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
52
|
+
end
|
53
|
+
end
|
54
|
+
|
55
|
+
context "object body" do
|
56
|
+
let(:result) { process({:a => 1}) }
|
57
|
+
|
58
|
+
it "encodes body" do
|
59
|
+
expect(result_body).to eq('{"a":1}')
|
60
|
+
end
|
61
|
+
|
62
|
+
it "adds content type" do
|
63
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
64
|
+
end
|
65
|
+
end
|
66
|
+
|
67
|
+
context "empty object body" do
|
68
|
+
let(:result) { process({}) }
|
69
|
+
|
70
|
+
it "encodes body" do
|
71
|
+
expect(result_body).to eq('{}')
|
72
|
+
end
|
73
|
+
end
|
74
|
+
|
75
|
+
context "object body with json type" do
|
76
|
+
let(:result) { process({:a => 1}, 'application/json; charset=utf-8') }
|
77
|
+
|
78
|
+
it "encodes body" do
|
79
|
+
expect(result_body).to eq('{"a":1}')
|
80
|
+
end
|
81
|
+
|
82
|
+
it "doesn't change content type" do
|
83
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
84
|
+
end
|
85
|
+
end
|
86
|
+
|
87
|
+
context "object body with incompatible type" do
|
88
|
+
let(:result) { process({:a => 1}, 'application/xml; charset=utf-8') }
|
89
|
+
|
90
|
+
it "doesn't change body" do
|
91
|
+
expect(result_body).to eq({:a => 1})
|
92
|
+
end
|
93
|
+
|
94
|
+
it "doesn't change content type" do
|
95
|
+
expect(result_type).to eq('application/xml; charset=utf-8')
|
96
|
+
end
|
97
|
+
end
|
98
|
+
|
99
|
+
### Unicode test cases
|
100
|
+
# Ruby 1.8 will almost certainly fail if there is no charset given in a header.
|
101
|
+
# In Ruby >1.8, we have some more methods for guessing well.
|
102
|
+
|
103
|
+
### All Ruby versions should work with a charset given.
|
104
|
+
context "utf-8 in string body" do
|
105
|
+
let(:result) { process('{"a":"ä"}', 'application/json; charset=utf-8') }
|
106
|
+
|
107
|
+
it "doesn't change body" do
|
108
|
+
expect(result_body).to eq('{"a":"ä"}')
|
109
|
+
end
|
110
|
+
|
111
|
+
it "doesn't change content type" do
|
112
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
113
|
+
end
|
114
|
+
|
115
|
+
it "adds content length" do
|
116
|
+
expect(result_length).to eq(10)
|
117
|
+
end
|
118
|
+
end
|
119
|
+
|
120
|
+
context "utf-8 in object body" do
|
121
|
+
let(:result) { process({:a => "ä"}, 'application/json; charset=utf-8') }
|
122
|
+
|
123
|
+
it "encodes body" do
|
124
|
+
expect(result_body).to eq('{"a":"ä"}')
|
125
|
+
end
|
126
|
+
|
127
|
+
it "doesn't change content type" do
|
128
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
129
|
+
end
|
130
|
+
|
131
|
+
it "adds content length" do
|
132
|
+
expect(result_length).to eq(10)
|
133
|
+
end
|
134
|
+
end
|
135
|
+
|
136
|
+
context "non-unicode in string body" do
|
137
|
+
let(:result) {
|
138
|
+
process(test_encode('{"a":"ä"}', 'iso-8859-15'), 'application/json; charset=iso-8859-15')
|
139
|
+
}
|
140
|
+
|
141
|
+
it "changes body" do
|
142
|
+
expect(result_body).to eq('{"a":"ä"}')
|
143
|
+
end
|
144
|
+
|
145
|
+
it "changes content type" do
|
146
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
147
|
+
end
|
148
|
+
|
149
|
+
it "adds content length" do
|
150
|
+
expect(result_length).to eq(10)
|
151
|
+
end
|
152
|
+
end
|
153
|
+
|
154
|
+
context "non-unicode in object body" do
|
155
|
+
let(:result) {
|
156
|
+
process({:a => test_encode('ä', 'iso-8859-15')}, 'application/json; charset=iso-8859-15')
|
157
|
+
}
|
158
|
+
|
159
|
+
it "encodes body" do
|
160
|
+
expect(result_body).to eq('{"a":"ä"}')
|
161
|
+
end
|
162
|
+
|
163
|
+
it "changes content type" do
|
164
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
165
|
+
end
|
166
|
+
|
167
|
+
it "adds content length" do
|
168
|
+
expect(result_length).to eq(10)
|
169
|
+
end
|
170
|
+
end
|
171
|
+
|
172
|
+
context "non-utf-8 in string body" do
|
173
|
+
let(:result) {
|
174
|
+
process(test_encode('{"a":"ä"}', 'utf-16be'), 'application/json; charset=utf-16be')
|
175
|
+
}
|
176
|
+
|
177
|
+
it "changes body" do
|
178
|
+
expect(result_body).to eq('{"a":"ä"}')
|
179
|
+
end
|
180
|
+
|
181
|
+
it "changes content type" do
|
182
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
183
|
+
end
|
184
|
+
|
185
|
+
it "adds content length" do
|
186
|
+
expect(result_length).to eq(10)
|
187
|
+
end
|
188
|
+
end
|
189
|
+
|
190
|
+
context "non-utf-8 in object body" do
|
191
|
+
let(:result) {
|
192
|
+
process({:a => test_encode('ä', 'utf-16le')}, 'application/json; charset=utf-16le')
|
193
|
+
}
|
194
|
+
|
195
|
+
it "encodes body" do
|
196
|
+
expect(result_body).to eq('{"a":"ä"}')
|
197
|
+
end
|
198
|
+
|
199
|
+
it "changes content type" do
|
200
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
201
|
+
end
|
202
|
+
|
203
|
+
it "adds content length" do
|
204
|
+
expect(result_length).to eq(10)
|
205
|
+
end
|
206
|
+
end
|
207
|
+
|
208
|
+
|
209
|
+
### Ruby versions > 1.8 should be able to guess missing charsets at times.
|
210
|
+
if not RUBY_VERSION.start_with?("1.8")
|
211
|
+
context "utf-8 in string body without content type" do
|
212
|
+
let(:result) { process('{"a":"ä"}') }
|
213
|
+
|
214
|
+
it "doesn't change body" do
|
215
|
+
expect(result_body).to eq('{"a":"ä"}')
|
216
|
+
end
|
217
|
+
|
218
|
+
it "adds content type" do
|
219
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
220
|
+
end
|
221
|
+
|
222
|
+
it "adds content length" do
|
223
|
+
expect(result_length).to eq(10)
|
224
|
+
end
|
225
|
+
end
|
226
|
+
|
227
|
+
context "utf-8 in object body without content type" do
|
228
|
+
let(:result) { process({:a => "ä"}) }
|
229
|
+
|
230
|
+
it "encodes body" do
|
231
|
+
expect(result_body).to eq('{"a":"ä"}')
|
232
|
+
end
|
233
|
+
|
234
|
+
it "adds content type" do
|
235
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
236
|
+
end
|
237
|
+
|
238
|
+
it "adds content length" do
|
239
|
+
expect(result_length).to eq(10)
|
240
|
+
end
|
241
|
+
end
|
242
|
+
|
243
|
+
context "non-unicode in string body without content type" do
|
244
|
+
let(:result) {
|
245
|
+
process(test_encode('{"a":"ä"}', 'iso-8859-15'))
|
246
|
+
}
|
247
|
+
|
248
|
+
it "doesn't change body" do
|
249
|
+
expect(result_body).to eq('{"a":"ä"}')
|
250
|
+
end
|
251
|
+
|
252
|
+
it "adds content type" do
|
253
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
254
|
+
end
|
255
|
+
|
256
|
+
it "adds content length" do
|
257
|
+
expect(result_length).to eq(10)
|
258
|
+
end
|
259
|
+
end
|
260
|
+
|
261
|
+
context "non-unicode in object body without content type" do
|
262
|
+
let(:result) {
|
263
|
+
process({:a => test_encode('ä', 'iso-8859-15')})
|
264
|
+
}
|
265
|
+
|
266
|
+
it "encodes body" do
|
267
|
+
expect(result_body).to eq('{"a":"ä"}')
|
268
|
+
end
|
269
|
+
|
270
|
+
it "adds content type" do
|
271
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
272
|
+
end
|
273
|
+
|
274
|
+
it "adds content length" do
|
275
|
+
expect(result_length).to eq(10)
|
276
|
+
end
|
277
|
+
end
|
278
|
+
|
279
|
+
context "non-utf-8 in string body without content type" do
|
280
|
+
let(:result) {
|
281
|
+
process(test_encode('{"a":"ä"}', 'utf-16be'))
|
282
|
+
}
|
283
|
+
|
284
|
+
|
285
|
+
it "doesn't change body" do
|
286
|
+
expect(result_body).to eq('{"a":"ä"}')
|
287
|
+
end
|
288
|
+
|
289
|
+
it "adds content type" do
|
290
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
291
|
+
end
|
292
|
+
|
293
|
+
it "adds content length" do
|
294
|
+
expect(result_length).to eq(10)
|
295
|
+
end
|
296
|
+
end
|
297
|
+
|
298
|
+
context "non-utf-8 in object body without content type" do
|
299
|
+
let(:result) {
|
300
|
+
process({:a => test_encode('ä', 'utf-16le')})
|
301
|
+
}
|
302
|
+
|
303
|
+
it "encodes body" do
|
304
|
+
expect(result_body).to eq('{"a":"ä"}')
|
305
|
+
end
|
306
|
+
|
307
|
+
it "adds content type" do
|
308
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
309
|
+
end
|
310
|
+
|
311
|
+
it "adds content length" do
|
312
|
+
expect(result_length).to eq(10)
|
313
|
+
end
|
314
|
+
end
|
315
|
+
|
316
|
+
### Dealing with files in various encoding should ideally be easy
|
317
|
+
FILES = {
|
318
|
+
'spec/data/iso8859-15_file.json' => 'iso-8859-15',
|
319
|
+
'spec/data/utf16be_file.json' => 'utf-16be',
|
320
|
+
'spec/data/utf16le_file.json' => 'utf-16le',
|
321
|
+
'spec/data/utf8_file.json' => 'utf-8',
|
322
|
+
}
|
323
|
+
|
324
|
+
|
325
|
+
FILES.each do |fname, enc|
|
326
|
+
context "reading #{enc} encoded file '#{fname}'" do
|
327
|
+
# Read the string from file; read binary/with encoding. Ruby 1.8 will
|
328
|
+
# ignore this, but must still work.
|
329
|
+
data = File.new(fname, "rb:#{enc}").read
|
330
|
+
|
331
|
+
# Passing that data with a charset should do the right thing.
|
332
|
+
let(:result) {
|
333
|
+
process(data)
|
334
|
+
}
|
335
|
+
|
336
|
+
it "encodes body" do
|
337
|
+
expect(result_body).to eq("{\"a\":\"Hellö, Wörld!\"}\n")
|
338
|
+
end
|
339
|
+
|
340
|
+
it "adds content type" do
|
341
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
342
|
+
end
|
343
|
+
|
344
|
+
it "adds content length" do
|
345
|
+
expect(result_length).to eq(24)
|
346
|
+
end
|
347
|
+
end
|
348
|
+
end
|
349
|
+
|
350
|
+
FILES.each do |fname, enc|
|
351
|
+
context "reading #{enc} encoded file '#{fname}' as binary" do
|
352
|
+
# Read the string from file; read binary/with encoding. Ruby 1.8 will
|
353
|
+
# ignore this, but must still work.
|
354
|
+
data = File.new(fname, "rb").read
|
355
|
+
|
356
|
+
# Passing that data with a charset should do the right thing.
|
357
|
+
let(:result) {
|
358
|
+
process(data, "application/json; charset=#{enc}")
|
359
|
+
}
|
360
|
+
|
361
|
+
it "encodes body" do
|
362
|
+
expect(result_body).to eq("{\"a\":\"Hellö, Wörld!\"}\n")
|
363
|
+
end
|
364
|
+
|
365
|
+
it "adds content type" do
|
366
|
+
expect(result_type).to eq('application/json; charset=utf-8')
|
367
|
+
end
|
368
|
+
|
369
|
+
it "adds content length" do
|
370
|
+
expect(result_length).to eq(24)
|
371
|
+
end
|
372
|
+
end
|
373
|
+
end
|
374
|
+
|
375
|
+
end
|
376
|
+
end
|