14 changed files with 237 additions and 75 deletions
@ -0,0 +1,13 @@ |
|||
guard :rails do |
|||
watch('Gemfile.lock') |
|||
watch(%r{^(config|lib)/.*}) |
|||
end |
|||
|
|||
guard :livereload do |
|||
watch(%r{app/views/.+\.(erb|haml|slim)$}) |
|||
watch(%r{app/helpers/.+\.rb}) |
|||
watch(%r{public/.+\.(css|js|html)}) |
|||
watch(%r{config/locales/.+\.yml}) |
|||
# Rails Assets Pipeline |
|||
watch(%r{(app|vendor)(/assets/\w+/(.+\.(css|js|html|png|jpg|coffee|scss))).*}) { |m| "/assets/#{m[3]}" } |
|||
end |
|||
@ -0,0 +1,71 @@ |
|||
class Http |
|||
def initialize(url) |
|||
@url = url |
|||
@response = self.grab |
|||
end |
|||
|
|||
def code |
|||
@response.code |
|||
end |
|||
|
|||
def success? |
|||
@response.success? |
|||
end |
|||
|
|||
def html? |
|||
@response.content_type == 'text/html' |
|||
end |
|||
|
|||
def title |
|||
html = self.parse |
|||
return nil unless html |
|||
tag = html.at 'head title' |
|||
tag&.text |
|||
end |
|||
|
|||
def charset |
|||
return nil unless self.html? |
|||
|
|||
body = @response.body |
|||
html = Nokogiri::HTML.parse body |
|||
|
|||
# Content-Type charset seems already processed by HTTParty |
|||
# charset = @response.headers['content-type'] |
|||
# charset = /text\/html;\s*charset=(.*)/i.match charset |
|||
# return charset[1] if charset |
|||
|
|||
charset = html.at 'head meta[charset]' |
|||
return charset['charset'] if charset |
|||
|
|||
charset = html.at 'head meta[http-equiv="Content-Type"]' |
|||
if charset |
|||
charset = charset['content'] |
|||
charset = /text\/html;\s*charset=(.*)/i.match charset |
|||
return charset[1] if charset |
|||
end |
|||
|
|||
nil |
|||
end |
|||
|
|||
def body |
|||
charset = self.charset |
|||
body = @response.body |
|||
body = body.force_encoding charset if charset |
|||
body.encode! 'utf-8' unless body.encoding == Encoding::UTF_8 |
|||
body |
|||
end |
|||
|
|||
protected |
|||
|
|||
def grab |
|||
response = HTTParty.get @url, timeout: 10.seconds |
|||
raise "Receive #{response.code}" unless response.success? |
|||
response |
|||
end |
|||
|
|||
def parse |
|||
return nil unless self.html? |
|||
body = @response.body |
|||
Nokogiri::HTML.parse body |
|||
end |
|||
end |
|||
@ -0,0 +1,78 @@ |
|||
RSpec.describe Http do |
|||
let :site do |
|||
Http.new 'http://localhost/' |
|||
end |
|||
|
|||
it 'must encode to utf-8 if not HTML' do |
|||
stub_request(:any, 'localhost').to_return body: "\xC3\xA9", status: 200, |
|||
headers: { 'Content-Type': 'application/pdf' } |
|||
expect(site.body.encoding).to eq Encoding::UTF_8 |
|||
expect(site.body).to eq "\xC3\xA9" |
|||
end |
|||
|
|||
it 'must encode to utf-8 if HTML and nothing specified' do |
|||
body = <<-HEREDOC |
|||
<html> |
|||
<body>\xC3\xA9</body> |
|||
</html>' |
|||
HEREDOC |
|||
stub_request(:any, 'localhost').to_return body: body, status: 200, |
|||
headers: { 'Content-Type': 'text/html' } |
|||
body = site.body |
|||
expect(body.encoding).to eq Encoding::UTF_8 |
|||
body = Nokogiri::HTML.parse body |
|||
body = body.at('body').content |
|||
expect(body).to eq "\xC3\xA9" |
|||
end |
|||
|
|||
it 'must encode to given content-type charset if nothing specified' do |
|||
body = <<-HEREDOC |
|||
<html> |
|||
<body>\xE9</body> |
|||
</html>' |
|||
HEREDOC |
|||
stub_request(:any, 'localhost').to_return body: body, status: 200, |
|||
headers: { 'Content-Type': 'text/html; charset=iso-8859-1' } |
|||
body = site.body |
|||
expect(body.encoding).to eq Encoding::UTF_8 |
|||
body = Nokogiri::HTML.parse body |
|||
body = body.at('body').content |
|||
expect(body).to eq "\xC3\xA9" |
|||
end |
|||
|
|||
it 'must encode to given meta charset' do |
|||
body = <<-HEREDOC |
|||
<html> |
|||
<head> |
|||
<meta charset="ISO-8859-1"/> |
|||
</head> |
|||
<body>\xE9</body> |
|||
</html>' |
|||
HEREDOC |
|||
stub_request(:any, 'localhost').to_return body: body, status: 200, |
|||
headers: { 'Content-Type': 'text/html' } |
|||
body = site.body |
|||
expect(body.encoding).to eq Encoding::UTF_8 |
|||
body = Nokogiri::HTML.parse body |
|||
body = body.at('body').content |
|||
expect(body).to eq "\xC3\xA9" |
|||
end |
|||
|
|||
it 'must encode to given meta http-equiv' do |
|||
body = <<-HEREDOC |
|||
<html> |
|||
<head> |
|||
<meta http-equiv="Content-Type" content="text/html; CHARSET=iso-8859-1"> |
|||
</head> |
|||
<body>\xE9</body> |
|||
</html>' |
|||
HEREDOC |
|||
stub_request(:any, 'localhost').to_return body: body, status: 200, |
|||
headers: { 'Content-Type': 'text/html' } |
|||
body = site.body |
|||
expect(body.encoding).to eq Encoding::UTF_8 |
|||
body = Nokogiri::HTML.parse body |
|||
body = body.at('body').content |
|||
expect(body).to eq "\xC3\xA9" |
|||
end |
|||
end |
|||
Loading…
Reference in new issue