chinwagsocial/spec/services/fetch_link_card_service_spec.rb

# frozen_string_literal: true

require 'rails_helper'

RSpec.describe FetchLinkCardService do
  subject { described_class.new }

  let(:html) { '<!doctype html><title>Hello world</title>' }
  let(:oembed_cache) { nil }

  before do
    stub_request(:get, 'http://example.com/html').to_return(headers: { 'Content-Type' => 'text/html' }, body: html)
    stub_request(:get, 'http://example.com/not-found').to_return(status: 404, headers: { 'Content-Type' => 'text/html' }, body: html)
    stub_request(:get, 'http://example.com/text').to_return(status: 404, headers: { 'Content-Type' => 'text/plain' }, body: 'Hello')
    stub_request(:get, 'http://example.com/redirect').to_return(status: 302, headers: { 'Location' => 'http://example.com/html' })
    stub_request(:get, 'http://example.com/redirect-to-404').to_return(status: 302, headers: { 'Location' => 'http://example.com/not-found' })
    stub_request(:get, 'http://example.com/oembed?url=http://example.com/html').to_return(headers: { 'Content-Type' => 'application/json' }, body: '{ "version": "1.0", "type": "link", "title": "oEmbed title" }')
    stub_request(:get, 'http://example.com/oembed?format=json&url=http://example.com/html').to_return(headers: { 'Content-Type' => 'application/json' }, body: '{ "version": "1.0", "type": "link", "title": "oEmbed title" }')

    stub_request(:get, 'http://example.xn--fiqs8s')
    stub_request(:get, 'http://example.com/日本語')
    stub_request(:get, 'http://example.com/test?data=file.gpx%5E1')
    stub_request(:get, 'http://example.com/test-')

    stub_request(:get, 'http://example.com/sjis').to_return(request_fixture('sjis.txt'))
    stub_request(:get, 'http://example.com/sjis_with_wrong_charset').to_return(request_fixture('sjis_with_wrong_charset.txt'))
    stub_request(:get, 'http://example.com/koi8-r').to_return(request_fixture('koi8-r.txt'))
    stub_request(:get, 'http://example.com/windows-1251').to_return(request_fixture('windows-1251.txt'))
    stub_request(:get, 'http://example.com/low_confidence_latin1').to_return(request_fixture('low_confidence_latin1.txt'))
    stub_request(:get, 'http://example.com/latin1_posing_as_utf8_broken').to_return(request_fixture('latin1_posing_as_utf8_broken.txt'))
    stub_request(:get, 'http://example.com/latin1_posing_as_utf8_recoverable').to_return(request_fixture('latin1_posing_as_utf8_recoverable.txt'))
    stub_request(:get, 'http://example.com/aergerliche-umlaute').to_return(request_fixture('redirect_with_utf8_url.txt'))
    stub_request(:get, 'http://example.com/page_without_title').to_return(request_fixture('page_without_title.txt'))
    stub_request(:get, 'http://example.com/long_canonical_url').to_return(request_fixture('long_canonical_url.txt'))
    stub_request(:get, 'http://example.com/alternative_utf8_spelling_in_header').to_return(request_fixture('alternative_utf8_spelling_in_header.txt'))

    Rails.cache.write('oembed_endpoint:example.com', oembed_cache) if oembed_cache

    subject.call(status)
  end

  context 'with a local status' do
    context 'with URL of a regular HTML page' do
      let(:status) { Fabricate(:status, text: 'http://example.com/html') }

      it 'creates preview card' do
        expect(status.preview_card).to_not be_nil
        expect(status.preview_card.url).to eq 'http://example.com/html'
        expect(status.preview_card.title).to eq 'Hello world'
      end
    end

    context 'with URL of a page with no title' do
      let(:status) { Fabricate(:status, text: 'http://example.com/html') }
      let(:html) { '<!doctype html><title></title>' }

      it 'does not create a preview card' do
        expect(status.preview_card).to be_nil
      end
    end

    context 'with a URL of a plain-text page' do
      let(:status) { Fabricate(:status, text: 'http://example.com/text') }

      it 'does not create a preview card' do
        expect(status.preview_card).to be_nil
      end
    end

    context 'with multiple URLs' do
      let(:status) { Fabricate(:status, text: 'ftp://example.com http://example.com/html http://example.com/text') }

      it 'fetches the first valid URL' do
        expect(a_request(:get, 'http://example.com/html')).to have_been_made
      end

      it 'does not fetch the second valid URL' do
        expect(a_request(:get, 'http://example.com/text/')).to_not have_been_made
      end
    end

    context 'with a redirect URL' do
      let(:status) { Fabricate(:status, text: 'http://example.com/redirect') }

      it 'follows redirect' do
        expect(a_request(:get, 'http://example.com/redirect')).to have_been_made.once
        expect(a_request(:get, 'http://example.com/html')).to have_been_made.once
      end

      it 'creates preview card' do
        expect(status.preview_card).to_not be_nil
        expect(status.preview_card.url).to eq 'http://example.com/html'
        expect(status.preview_card.title).to eq 'Hello world'
      end
    end

    context 'with a broken redirect URL' do
      let(:status) { Fabricate(:status, text: 'http://example.com/redirect-to-404') }

      it 'follows redirect' do
        expect(a_request(:get, 'http://example.com/redirect-to-404')).to have_been_made.once
        expect(a_request(:get, 'http://example.com/not-found')).to have_been_made.once
      end

      it 'does not create a preview card' do
        expect(status.preview_card).to be_nil
      end
    end

    context 'with a redirect URL with faulty encoding' do
      let(:status) { Fabricate(:status, text: 'http://example.com/aergerliche-umlaute') }

      it 'does not create a preview card' do
        expect(status.preview_card).to be_nil
      end
    end

    context 'with a page that has no title' do
      let(:status) { Fabricate(:status, text: 'http://example.com/page_without_title') }

      it 'does not create a preview card' do
        expect(status.preview_card).to be_nil
      end
    end

    context 'with a 404 URL' do
      let(:status) { Fabricate(:status, text: 'http://example.com/not-found') }

      it 'does not create a preview card' do
        expect(status.preview_card).to be_nil
      end
    end

    context 'with an IDN URL' do
      let(:status) { Fabricate(:status, text: 'Check out http://example.中国') }

      it 'fetches the URL' do
        expect(a_request(:get, 'http://example.xn--fiqs8s/')).to have_been_made.once
      end
    end

    context 'with a URL of a page in Shift JIS encoding' do
      let(:status) { Fabricate(:status, text: 'Check out http://example.com/sjis') }

      it 'decodes the HTML' do
        expect(status.preview_card.title).to eq('SJISのページ')
      end
    end

    context 'with a URL of a page in Shift JIS encoding labeled as UTF-8' do
      let(:status) { Fabricate(:status, text: 'Check out http://example.com/sjis_with_wrong_charset') }

      it 'decodes the HTML despite the wrong charset header' do
        expect(status.preview_card.title).to eq('SJISのページ')
      end
    end

    context 'with a URL of a page in KOI8-R encoding' do
      let(:status) { Fabricate(:status, text: 'Check out http://example.com/koi8-r') }

      it 'decodes the HTML' do
        expect(status.preview_card.title).to eq('Московя начинаетъ только въ XVI ст. привлекать внимане иностранцевъ.')
      end
    end

    context 'with a URL of a page in Windows-1251 encoding' do
      let(:status) { Fabricate(:status, text: 'Check out http://example.com/windows-1251') }

      it 'decodes the HTML' do
        expect(status.preview_card.title).to eq('сэмпл текст')
      end
    end

    context 'with a URL of a page in ISO-8859-1 encoding, that charlock_holmes cannot detect' do
      context 'when encoding in http header is correct' do
        let(:status) { Fabricate(:status, text: 'Check out http://example.com/low_confidence_latin1') }

        it 'decodes the HTML' do
          expect(status.preview_card.title).to eq("Tofu á l'orange")
        end
      end

      context 'when encoding in http header is incorrect' do
        context 'when encoding problems appear in unrelated tags' do
          let(:status) { Fabricate(:status, text: 'Check out http://example.com/latin1_posing_as_utf8_recoverable') }

          it 'decodes the HTML' do
            expect(status.preview_card.title).to eq('Tofu with orange sauce')
          end
        end

        context 'when encoding problems appear in title tag' do
          let(:status) { Fabricate(:status, text: 'Check out http://example.com/latin1_posing_as_utf8_broken') }

          it 'creates a preview card anyway that replaces invalid bytes with U+FFFD (replacement char)' do
            expect(status.preview_card.title).to eq("Tofu <20> l'orange")
          end
        end
      end
    end

    context 'with a Japanese path URL' do
      let(:status) { Fabricate(:status, text: 'テストhttp://example.com/日本語') }

      it 'fetches the URL' do
        expect(a_request(:get, 'http://example.com/日本語')).to have_been_made.once
      end
    end

    context 'with a hyphen-suffixed URL' do
      let(:status) { Fabricate(:status, text: 'test http://example.com/test-') }

      it 'fetches the URL' do
        expect(a_request(:get, 'http://example.com/test-')).to have_been_made.once
      end
    end

    context 'with a caret-suffixed URL' do
      let(:status) { Fabricate(:status, text: 'test http://example.com/test?data=file.gpx^1') }

      it 'fetches the URL' do
        expect(a_request(:get, 'http://example.com/test?data=file.gpx%5E1')).to have_been_made.once
      end

      it 'does not strip the caret before fetching' do
        expect(a_request(:get, 'http://example.com/test?data=file.gpx')).to_not have_been_made
      end
    end

    context 'with a non-isolated URL' do
      let(:status) { Fabricate(:status, text: 'testhttp://example.com/sjis') }

      it 'does not fetch URLs not isolated from their surroundings' do
        expect(a_request(:get, 'http://example.com/sjis')).to_not have_been_made
      end
    end

    context 'with a URL of a page with oEmbed support' do
      let(:html) { '<!doctype html><title>Hello world</title><link rel="alternate" type="application/json+oembed" href="http://example.com/oembed?url=http://example.com/html">' }
      let(:status) { Fabricate(:status, text: 'http://example.com/html') }

      it 'fetches the oEmbed URL' do
        expect(a_request(:get, 'http://example.com/oembed?url=http://example.com/html')).to have_been_made.once
      end

      it 'creates preview card' do
        expect(status.preview_card).to_not be_nil
        expect(status.preview_card.url).to eq 'http://example.com/html'
        expect(status.preview_card.title).to eq 'oEmbed title'
      end

      context 'when oEmbed endpoint cache populated' do
        let(:oembed_cache) { { endpoint: 'http://example.com/oembed?format=json&url={url}', format: :json } }

        it 'uses the cached oEmbed response' do
          expect(a_request(:get, 'http://example.com/oembed?url=http://example.com/html')).to_not have_been_made
          expect(a_request(:get, 'http://example.com/oembed?format=json&url=http://example.com/html')).to have_been_made
        end

        it 'creates preview card' do
          expect(status.preview_card).to_not be_nil
          expect(status.preview_card.url).to eq 'http://example.com/html'
          expect(status.preview_card.title).to eq 'oEmbed title'
        end
      end

      # If the original HTML URL for whatever reason (e.g. DOS protection) redirects to
      # an error page, we can still use the cached oEmbed but should not use the
      # redirect URL on the card.
      context 'when oEmbed endpoint cache populated but page returns 404' do
        let(:status) { Fabricate(:status, text: 'http://example.com/redirect-to-404') }
        let(:oembed_cache) { { endpoint: 'http://example.com/oembed?url=http://example.com/html', format: :json } }

        it 'uses the cached oEmbed response' do
          expect(a_request(:get, 'http://example.com/oembed?url=http://example.com/html')).to have_been_made
        end

        it 'creates preview card' do
          expect(status.preview_card).to_not be_nil
          expect(status.preview_card.title).to eq 'oEmbed title'
        end

        it 'uses the original URL' do
          expect(status.preview_card&.url).to eq 'http://example.com/redirect-to-404'
        end
      end
    end

    context 'with a URL of a page that includes a canonical URL too long for PostgreSQL unique indexes' do
      let(:status) { Fabricate(:status, text: 'test http://example.com/long_canonical_url') }

      it 'does not create a preview card' do
        expect(status.preview_card).to be_nil
      end
    end

    context 'with a URL where the `Content-Type` header uses `utf8` instead of `utf-8`' do
      let(:status) { Fabricate(:status, text: 'test http://example.com/alternative_utf8_spelling_in_header') }

      it 'does not create a preview card' do
        expect(status.preview_card.title).to eq 'Webserver Configs R Us'
      end
    end
  end

  context 'with a remote status' do
    let(:status) do
      Fabricate(:status, account: Fabricate(:account, domain: 'example.com'), text: <<-TEXT)
      Habt ihr ein paar gute Links zu <a>foo</a>
      #<span class="tag"><a href="https://quitter.se/tag/wannacry" target="_blank" rel="tag noopener noreferrer" title="https://quitter.se/tag/wannacry">Wannacry</a></span> herumfliegen?
      Ich will mal unter <br> <a href="http://example.com/not-found" target="_blank" rel="noopener noreferrer" title="http://example.com/not-found">http://example.com/not-found</a> was sammeln. !
      <a href="http://sn.jonkman.ca/group/416/id" target="_blank" rel="noopener noreferrer" title="http://sn.jonkman.ca/group/416/id">security</a>&nbsp;
      TEXT
    end

    it 'parses out URLs' do
      expect(a_request(:get, 'http://example.com/not-found')).to have_been_made.once
    end

    it 'ignores URLs to hashtags' do
      expect(a_request(:get, 'https://quitter.se/tag/wannacry')).to_not have_been_made
    end
  end
end
-												Enable Style/FrozenStringLiteralComment for specs (#23790)


											
										
										
											2023-02-22 11:55:31 +11:00
+								# frozen_string_literal: true
-												Punycode URI normalization (#2370)

* Fix #2119 - Whenever about to send a HTTP request, normalize the URI

* Add test for IDN request in FetchLinkCardService

* Perform IDN normalization on domains before they are stored in the DB

											
										
										
											2017-04-25 10:47:31 +10:00
+								require 'rails_helper'
-												Remove unneeded `type: :service` from spec/services files (#29304)


											
										
										
											2024-03-13 19:39:26 +11:00
+								RSpec.describe FetchLinkCardService do
-												Add support for structured data and more OpenGraph tags to link cards (#16938)

Save preview cards under their canonical URL

Increase max redirects to follow from 2 to 3
											
										
										
											2021-11-06 09:23:05 +11:00
+								  subject { described_class.new }
-												Fix #2572 - Resolve preview cards for remote statuses as well as local ones (#3088)


											
										
										
											2017-05-17 08:41:15 +10:00
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								  let(:html) { '<!doctype html><title>Hello world</title>' }
 								  let(:oembed_cache) { nil }
-												Punycode URI normalization (#2370)

* Fix #2119 - Whenever about to send a HTTP request, normalize the URI

* Add test for IDN request in FetchLinkCardService

* Perform IDN normalization on domains before they are stored in the DB

											
										
										
											2017-04-25 10:47:31 +10:00
+								  before do
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    stub_request(:get, 'http://example.com/html').to_return(headers: { 'Content-Type' => 'text/html' }, body: html)
 								    stub_request(:get, 'http://example.com/not-found').to_return(status: 404, headers: { 'Content-Type' => 'text/html' }, body: html)
 								    stub_request(:get, 'http://example.com/text').to_return(status: 404, headers: { 'Content-Type' => 'text/plain' }, body: 'Hello')
 								    stub_request(:get, 'http://example.com/redirect').to_return(status: 302, headers: { 'Location' => 'http://example.com/html' })
 								    stub_request(:get, 'http://example.com/redirect-to-404').to_return(status: 302, headers: { 'Location' => 'http://example.com/not-found' })
 								    stub_request(:get, 'http://example.com/oembed?url=http://example.com/html').to_return(headers: { 'Content-Type' => 'application/json' }, body: '{ "version": "1.0", "type": "link", "title": "oEmbed title" }')
 								    stub_request(:get, 'http://example.com/oembed?format=json&url=http://example.com/html').to_return(headers: { 'Content-Type' => 'application/json' }, body: '{ "version": "1.0", "type": "link", "title": "oEmbed title" }')
 								    stub_request(:get, 'http://example.xn--fiqs8s')
 								    stub_request(:get, 'http://example.com/日本語')
 								    stub_request(:get, 'http://example.com/test?data=file.gpx%5E1')
 								    stub_request(:get, 'http://example.com/test-')
-												Fix Nokogiri::HTML at FetchLinkCardService (#4072)


											
										
										
											2017-07-05 22:54:21 +10:00
+								    stub_request(:get, 'http://example.com/sjis').to_return(request_fixture('sjis.txt'))
-												Use charlock_holmes instead of nkf at FetchLinkCardService (#4080)

* Specs for language detection

* Use CharlockHolmes instead of NKF

* Correct mistakes

* Correct style

* Set hint_enc instead of falling back and strip_tags

* Improve specs

* Add dependencies

											
										
										
											2017-07-09 06:44:31 +10:00
+								    stub_request(:get, 'http://example.com/sjis_with_wrong_charset').to_return(request_fixture('sjis_with_wrong_charset.txt'))
 								    stub_request(:get, 'http://example.com/koi8-r').to_return(request_fixture('koi8-r.txt'))
-												 Ignore low-confidence CharlockHolmes guesses when parsing link cards (#9510)

* Add failing test for windows-1251 link cards

* Ignore low-confidence CharlockHolmes guesses

Fixes #9466

* Fix no method error when charlock holmes cannot detect charset

											
										
										
											2018-12-18 05:19:45 +11:00
+								    stub_request(:get, 'http://example.com/windows-1251').to_return(request_fixture('windows-1251.txt'))
-												Improve encoding detection for link cards (#30780)


											
										
										
											2024-06-21 22:51:10 +10:00
+								    stub_request(:get, 'http://example.com/low_confidence_latin1').to_return(request_fixture('low_confidence_latin1.txt'))
-												Improve handling of encoding problems when creating link previews (#30929)


											
										
										
											2024-07-05 21:54:38 +10:00
+								    stub_request(:get, 'http://example.com/latin1_posing_as_utf8_broken').to_return(request_fixture('latin1_posing_as_utf8_broken.txt'))
 								    stub_request(:get, 'http://example.com/latin1_posing_as_utf8_recoverable').to_return(request_fixture('latin1_posing_as_utf8_recoverable.txt'))
-												Catch encoding errors when creating link previews. (#30853)


											
										
										
											2024-06-28 00:41:03 +10:00
+								    stub_request(:get, 'http://example.com/aergerliche-umlaute').to_return(request_fixture('redirect_with_utf8_url.txt'))
-												Fix details extraction when no title exists. (#30933)


											
										
										
											2024-07-05 23:28:52 +10:00
+								    stub_request(:get, 'http://example.com/page_without_title').to_return(request_fixture('page_without_title.txt'))
-												Add size limit for all PreviewCard URLs (#30973)


											
										
										
											2024-07-09 23:11:34 +10:00
+								    stub_request(:get, 'http://example.com/long_canonical_url').to_return(request_fixture('long_canonical_url.txt'))
-												Do not pass unknown encoding names to nokogiri. (#30987)


											
										
										
											2024-07-11 00:25:39 +10:00
+								    stub_request(:get, 'http://example.com/alternative_utf8_spelling_in_header').to_return(request_fixture('alternative_utf8_spelling_in_header.txt'))
-												Fix #2572 - Resolve preview cards for remote statuses as well as local ones (#3088)


											
										
										
											2017-05-17 08:41:15 +10:00
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    Rails.cache.write('oembed_endpoint:example.com', oembed_cache) if oembed_cache
-												Fix #2572 - Resolve preview cards for remote statuses as well as local ones (#3088)


											
										
										
											2017-05-17 08:41:15 +10:00
+								    subject.call(status)
 								  end
-												Fix RSpec/ContextWording cop (#24739)


											
										
										
											2023-05-04 13:49:08 +10:00
+								  context 'with a local status' do
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    context 'with URL of a regular HTML page' do
 								      let(:status) { Fabricate(:status, text: 'http://example.com/html') }
 								      it 'creates preview card' do
 								        expect(status.preview_card).to_not be_nil
 								        expect(status.preview_card.url).to eq 'http://example.com/html'
 								        expect(status.preview_card.title).to eq 'Hello world'
 								      end
 								    end
 								    context 'with URL of a page with no title' do
 								      let(:status) { Fabricate(:status, text: 'http://example.com/html') }
 								      let(:html) { '<!doctype html><title></title>' }
 								      it 'does not create a preview card' do
 								        expect(status.preview_card).to be_nil
 								      end
 								    end
 								    context 'with a URL of a plain-text page' do
 								      let(:status) { Fabricate(:status, text: 'http://example.com/text') }
 								      it 'does not create a preview card' do
 								        expect(status.preview_card).to be_nil
 								      end
 								    end
 								    context 'with multiple URLs' do
 								      let(:status) { Fabricate(:status, text: 'ftp://example.com http://example.com/html http://example.com/text') }
 								      it 'fetches the first valid URL' do
 								        expect(a_request(:get, 'http://example.com/html')).to have_been_made
 								      end
 								      it 'does not fetch the second valid URL' do
 								        expect(a_request(:get, 'http://example.com/text/')).to_not have_been_made
 								      end
 								    end
 								    context 'with a redirect URL' do
 								      let(:status) { Fabricate(:status, text: 'http://example.com/redirect') }
 								      it 'follows redirect' do
 								        expect(a_request(:get, 'http://example.com/redirect')).to have_been_made.once
 								        expect(a_request(:get, 'http://example.com/html')).to have_been_made.once
 								      end
 								      it 'creates preview card' do
 								        expect(status.preview_card).to_not be_nil
 								        expect(status.preview_card.url).to eq 'http://example.com/html'
 								        expect(status.preview_card.title).to eq 'Hello world'
 								      end
 								    end
 								    context 'with a broken redirect URL' do
 								      let(:status) { Fabricate(:status, text: 'http://example.com/redirect-to-404') }
 								      it 'follows redirect' do
 								        expect(a_request(:get, 'http://example.com/redirect-to-404')).to have_been_made.once
 								        expect(a_request(:get, 'http://example.com/not-found')).to have_been_made.once
 								      end
 								      it 'does not create a preview card' do
 								        expect(status.preview_card).to be_nil
 								      end
 								    end
-												Catch encoding errors when creating link previews. (#30853)


											
										
										
											2024-06-28 00:41:03 +10:00
+								    context 'with a redirect URL with faulty encoding' do
 								      let(:status) { Fabricate(:status, text: 'http://example.com/aergerliche-umlaute') }
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      it 'does not create a preview card' do
 								        expect(status.preview_card).to be_nil
 								      end
 								    end
-												Fix details extraction when no title exists. (#30933)


											
										
										
											2024-07-05 23:28:52 +10:00
+								    context 'with a page that has no title' do
 								      let(:status) { Fabricate(:status, text: 'http://example.com/page_without_title') }
 								      it 'does not create a preview card' do
 								        expect(status.preview_card).to be_nil
 								      end
 								    end
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    context 'with a 404 URL' do
 								      let(:status) { Fabricate(:status, text: 'http://example.com/not-found') }
 								      it 'does not create a preview card' do
 								        expect(status.preview_card).to be_nil
 								      end
 								    end
 								    context 'with an IDN URL' do
-												Fix #2572 - Resolve preview cards for remote statuses as well as local ones (#3088)


											
										
										
											2017-05-17 08:41:15 +10:00
+								      let(:status) { Fabricate(:status, text: 'Check out http://example.中国') }
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      it 'fetches the URL' do
 								        expect(a_request(:get, 'http://example.xn--fiqs8s/')).to have_been_made.once
-												Fix #2572 - Resolve preview cards for remote statuses as well as local ones (#3088)


											
										
										
											2017-05-17 08:41:15 +10:00
+								      end
 								    end
-												Fix Nokogiri::HTML at FetchLinkCardService (#4072)


											
										
										
											2017-07-05 22:54:21 +10:00
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    context 'with a URL of a page in Shift JIS encoding' do
-												Fix Nokogiri::HTML at FetchLinkCardService (#4072)


											
										
										
											2017-07-05 22:54:21 +10:00
+								      let(:status) { Fabricate(:status, text: 'Check out http://example.com/sjis') }
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      it 'decodes the HTML' do
-												Change link previews to keep original URL from the status (#27312)


											
										
										
											2023-11-13 20:58:28 +11:00
+								        expect(status.preview_card.title).to eq('SJISのページ')
-												Use charlock_holmes instead of nkf at FetchLinkCardService (#4080)

* Specs for language detection

* Use CharlockHolmes instead of NKF

* Correct mistakes

* Correct style

* Set hint_enc instead of falling back and strip_tags

* Improve specs

* Add dependencies

											
										
										
											2017-07-09 06:44:31 +10:00
+								      end
 								    end
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    context 'with a URL of a page in Shift JIS encoding labeled as UTF-8' do
-												Use charlock_holmes instead of nkf at FetchLinkCardService (#4080)

* Specs for language detection

* Use CharlockHolmes instead of NKF

* Correct mistakes

* Correct style

* Set hint_enc instead of falling back and strip_tags

* Improve specs

* Add dependencies

											
										
										
											2017-07-09 06:44:31 +10:00
+								      let(:status) { Fabricate(:status, text: 'Check out http://example.com/sjis_with_wrong_charset') }
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      it 'decodes the HTML despite the wrong charset header' do
-												Change link previews to keep original URL from the status (#27312)


											
										
										
											2023-11-13 20:58:28 +11:00
+								        expect(status.preview_card.title).to eq('SJISのページ')
-												Use charlock_holmes instead of nkf at FetchLinkCardService (#4080)

* Specs for language detection

* Use CharlockHolmes instead of NKF

* Correct mistakes

* Correct style

* Set hint_enc instead of falling back and strip_tags

* Improve specs

* Add dependencies

											
										
										
											2017-07-09 06:44:31 +10:00
+								      end
 								    end
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    context 'with a URL of a page in KOI8-R encoding' do
-												Use charlock_holmes instead of nkf at FetchLinkCardService (#4080)

* Specs for language detection

* Use CharlockHolmes instead of NKF

* Correct mistakes

* Correct style

* Set hint_enc instead of falling back and strip_tags

* Improve specs

* Add dependencies

											
										
										
											2017-07-09 06:44:31 +10:00
+								      let(:status) { Fabricate(:status, text: 'Check out http://example.com/koi8-r') }
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      it 'decodes the HTML' do
-												Change link previews to keep original URL from the status (#27312)


											
										
										
											2023-11-13 20:58:28 +11:00
+								        expect(status.preview_card.title).to eq('Московя начинаетъ только въ XVI ст. привлекать внимане иностранцевъ.')
-												Fix Nokogiri::HTML at FetchLinkCardService (#4072)


											
										
										
											2017-07-05 22:54:21 +10:00
+								      end
 								    end
-												Enable to recognize most kinds of characters as URL paths (#4941)


											
										
										
											2017-09-15 02:03:20 +10:00
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    context 'with a URL of a page in Windows-1251 encoding' do
-												 Ignore low-confidence CharlockHolmes guesses when parsing link cards (#9510)

* Add failing test for windows-1251 link cards

* Ignore low-confidence CharlockHolmes guesses

Fixes #9466

* Fix no method error when charlock holmes cannot detect charset

											
										
										
											2018-12-18 05:19:45 +11:00
+								      let(:status) { Fabricate(:status, text: 'Check out http://example.com/windows-1251') }
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      it 'decodes the HTML' do
-												Change link previews to keep original URL from the status (#27312)


											
										
										
											2023-11-13 20:58:28 +11:00
+								        expect(status.preview_card.title).to eq('сэмпл текст')
-												 Ignore low-confidence CharlockHolmes guesses when parsing link cards (#9510)

* Add failing test for windows-1251 link cards

* Ignore low-confidence CharlockHolmes guesses

Fixes #9466

* Fix no method error when charlock holmes cannot detect charset

											
										
										
											2018-12-18 05:19:45 +11:00
+								      end
 								    end
-												Improve encoding detection for link cards (#30780)


											
										
										
											2024-06-21 22:51:10 +10:00
+								    context 'with a URL of a page in ISO-8859-1 encoding, that charlock_holmes cannot detect' do
-												Improve handling of encoding problems when creating link previews (#30929)


											
										
										
											2024-07-05 21:54:38 +10:00
+								      context 'when encoding in http header is correct' do
 								        let(:status) { Fabricate(:status, text: 'Check out http://example.com/low_confidence_latin1') }
-												Improve encoding detection for link cards (#30780)


											
										
										
											2024-06-21 22:51:10 +10:00
-												Improve handling of encoding problems when creating link previews (#30929)


											
										
										
											2024-07-05 21:54:38 +10:00
+								        it 'decodes the HTML' do
 								          expect(status.preview_card.title).to eq("Tofu á l'orange")
 								        end
 								      end
 								      context 'when encoding in http header is incorrect' do
 								        context 'when encoding problems appear in unrelated tags' do
 								          let(:status) { Fabricate(:status, text: 'Check out http://example.com/latin1_posing_as_utf8_recoverable') }
 								          it 'decodes the HTML' do
 								            expect(status.preview_card.title).to eq('Tofu with orange sauce')
 								          end
 								        end
 								        context 'when encoding problems appear in title tag' do
 								          let(:status) { Fabricate(:status, text: 'Check out http://example.com/latin1_posing_as_utf8_broken') }
-												Change fetch link card service to parse as HTML5 (#31814)


											
										
										
											2024-09-09 20:59:42 +10:00
+								          it 'creates a preview card anyway that replaces invalid bytes with U+FFFD (replacement char)' do
 								            expect(status.preview_card.title).to eq("Tofu <20> l'orange")
-												Improve handling of encoding problems when creating link previews (#30929)


											
										
										
											2024-07-05 21:54:38 +10:00
+								          end
 								        end
-												Improve encoding detection for link cards (#30780)


											
										
										
											2024-06-21 22:51:10 +10:00
+								      end
 								    end
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    context 'with a Japanese path URL' do
-												Enable to recognize most kinds of characters as URL paths (#4941)


											
										
										
											2017-09-15 02:03:20 +10:00
+								      let(:status) { Fabricate(:status, text: 'テストhttp://example.com/日本語') }
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      it 'fetches the URL' do
 								        expect(a_request(:get, 'http://example.com/日本語')).to have_been_made.once
-												Enable to recognize most kinds of characters as URL paths (#4941)


											
										
										
											2017-09-15 02:03:20 +10:00
+								      end
 								    end
-												Fix URLs incorrectly having trailing hyphen removed (#6465)

In cases where a URL has a trailing hyphen the FetchLinkCardService incorrectly removes the hyphen when it is parsed

The hyphen is not a reserved character in the URI spec https://tools.ietf.org/html/rfc3986#section-2.2
											
										
										
											2018-02-12 09:49:18 +11:00
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    context 'with a hyphen-suffixed URL' do
-												Fix URLs incorrectly having trailing hyphen removed (#6465)

In cases where a URL has a trailing hyphen the FetchLinkCardService incorrectly removes the hyphen when it is parsed

The hyphen is not a reserved character in the URI spec https://tools.ietf.org/html/rfc3986#section-2.2
											
										
										
											2018-02-12 09:49:18 +11:00
+								      let(:status) { Fabricate(:status, text: 'test http://example.com/test-') }
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      it 'fetches the URL' do
 								        expect(a_request(:get, 'http://example.com/test-')).to have_been_made.once
-												Fix URLs incorrectly having trailing hyphen removed (#6465)

In cases where a URL has a trailing hyphen the FetchLinkCardService incorrectly removes the hyphen when it is parsed

The hyphen is not a reserved character in the URI spec https://tools.ietf.org/html/rfc3986#section-2.2
											
										
										
											2018-02-12 09:49:18 +11:00
+								      end
 								    end
-												Fix URL scanning in note length validator and preview card fetching (#15827)

* Add tests

* Fix URL scanning in note length validator and preview card fetching
											
										
										
											2021-03-04 10:12:26 +11:00
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    context 'with a caret-suffixed URL' do
 								      let(:status) { Fabricate(:status, text: 'test http://example.com/test?data=file.gpx^1') }
 								      it 'fetches the URL' do
 								        expect(a_request(:get, 'http://example.com/test?data=file.gpx%5E1')).to have_been_made.once
 								      end
 								      it 'does not strip the caret before fetching' do
 								        expect(a_request(:get, 'http://example.com/test?data=file.gpx')).to_not have_been_made
 								      end
 								    end
 								    context 'with a non-isolated URL' do
-												Fix URL scanning in note length validator and preview card fetching (#15827)

* Add tests

* Fix URL scanning in note length validator and preview card fetching
											
										
										
											2021-03-04 10:12:26 +11:00
+								      let(:status) { Fabricate(:status, text: 'testhttp://example.com/sjis') }
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      it 'does not fetch URLs not isolated from their surroundings' do
-												Fix URL scanning in note length validator and preview card fetching (#15827)

* Add tests

* Fix URL scanning in note length validator and preview card fetching
											
										
										
											2021-03-04 10:12:26 +11:00
+								        expect(a_request(:get, 'http://example.com/sjis')).to_not have_been_made
 								      end
 								    end
-												Allow carets in URL search params (#25216)


											
										
										
											2023-06-01 20:14:49 +10:00
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								    context 'with a URL of a page with oEmbed support' do
 								      let(:html) { '<!doctype html><title>Hello world</title><link rel="alternate" type="application/json+oembed" href="http://example.com/oembed?url=http://example.com/html">' }
 								      let(:status) { Fabricate(:status, text: 'http://example.com/html') }
-												Allow carets in URL search params (#25216)


											
										
										
											2023-06-01 20:14:49 +10:00
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      it 'fetches the oEmbed URL' do
 								        expect(a_request(:get, 'http://example.com/oembed?url=http://example.com/html')).to have_been_made.once
 								      end
 								      it 'creates preview card' do
 								        expect(status.preview_card).to_not be_nil
 								        expect(status.preview_card.url).to eq 'http://example.com/html'
 								        expect(status.preview_card.title).to eq 'oEmbed title'
 								      end
 								      context 'when oEmbed endpoint cache populated' do
 								        let(:oembed_cache) { { endpoint: 'http://example.com/oembed?format=json&url={url}', format: :json } }
 								        it 'uses the cached oEmbed response' do
 								          expect(a_request(:get, 'http://example.com/oembed?url=http://example.com/html')).to_not have_been_made
 								          expect(a_request(:get, 'http://example.com/oembed?format=json&url=http://example.com/html')).to have_been_made
 								        end
 								        it 'creates preview card' do
 								          expect(status.preview_card).to_not be_nil
 								          expect(status.preview_card.url).to eq 'http://example.com/html'
 								          expect(status.preview_card.title).to eq 'oEmbed title'
 								        end
 								      end
 								      # If the original HTML URL for whatever reason (e.g. DOS protection) redirects to
 								      # an error page, we can still use the cached oEmbed but should not use the
 								      # redirect URL on the card.
 								      context 'when oEmbed endpoint cache populated but page returns 404' do
 								        let(:status) { Fabricate(:status, text: 'http://example.com/redirect-to-404') }
 								        let(:oembed_cache) { { endpoint: 'http://example.com/oembed?url=http://example.com/html', format: :json } }
 								        it 'uses the cached oEmbed response' do
 								          expect(a_request(:get, 'http://example.com/oembed?url=http://example.com/html')).to have_been_made
 								        end
 								        it 'creates preview card' do
 								          expect(status.preview_card).to_not be_nil
 								          expect(status.preview_card.title).to eq 'oEmbed title'
 								        end
 								        it 'uses the original URL' do
 								          expect(status.preview_card&.url).to eq 'http://example.com/redirect-to-404'
 								        end
-												Allow carets in URL search params (#25216)


											
										
										
											2023-06-01 20:14:49 +10:00
+								      end
 								    end
-												Add size limit for all PreviewCard URLs (#30973)


											
										
										
											2024-07-09 23:11:34 +10:00
 								    context 'with a URL of a page that includes a canonical URL too long for PostgreSQL unique indexes' do
 								      let(:status) { Fabricate(:status, text: 'test http://example.com/long_canonical_url') }
 								      it 'does not create a preview card' do
 								        expect(status.preview_card).to be_nil
 								      end
 								    end
-												Do not pass unknown encoding names to nokogiri. (#30987)


											
										
										
											2024-07-11 00:25:39 +10:00
 								    context 'with a URL where the `Content-Type` header uses `utf8` instead of `utf-8`' do
 								      let(:status) { Fabricate(:status, text: 'test http://example.com/alternative_utf8_spelling_in_header') }
 								      it 'does not create a preview card' do
 								        expect(status.preview_card.title).to eq 'Webserver Configs R Us'
 								      end
 								    end
-												Punycode URI normalization (#2370)

* Fix #2119 - Whenever about to send a HTTP request, normalize the URI

* Add test for IDN request in FetchLinkCardService

* Perform IDN normalization on domains before they are stored in the DB

											
										
										
											2017-04-25 10:47:31 +10:00
+								  end
-												Fix RSpec/ContextWording cop (#24739)


											
										
										
											2023-05-04 13:49:08 +10:00
+								  context 'with a remote status' do
-												Reformat large text arg in `FetchLinkCardService` spec (#26183)


											
										
										
											2023-07-26 17:44:51 +10:00
+								    let(:status) do
 								      Fabricate(:status, account: Fabricate(:account, domain: 'example.com'), text: <<-TEXT)
 								      Habt ihr ein paar gute Links zu <a>foo</a>
 								      #<span class="tag"><a href="https://quitter.se/tag/wannacry" target="_blank" rel="tag noopener noreferrer" title="https://quitter.se/tag/wannacry">Wannacry</a></span> herumfliegen?
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      Ich will mal unter <br> <a href="http://example.com/not-found" target="_blank" rel="noopener noreferrer" title="http://example.com/not-found">http://example.com/not-found</a> was sammeln. !
-												Reformat large text arg in `FetchLinkCardService` spec (#26183)


											
										
										
											2023-07-26 17:44:51 +10:00
+								      <a href="http://sn.jonkman.ca/group/416/id" target="_blank" rel="noopener noreferrer" title="http://sn.jonkman.ca/group/416/id">security</a>&nbsp;
 								      TEXT
 								    end
-												Fix #2572 - Resolve preview cards for remote statuses as well as local ones (#3088)


											
										
										
											2017-05-17 08:41:15 +10:00
 								    it 'parses out URLs' do
-												Use original URL in preview if it redirects to 4xx page (#26200)


											
										
										
											2023-07-29 07:02:08 +10:00
+								      expect(a_request(:get, 'http://example.com/not-found')).to have_been_made.once
-												Fix #2572 - Resolve preview cards for remote statuses as well as local ones (#3088)


											
										
										
											2017-05-17 08:41:15 +10:00
+								    end
-												Punycode URI normalization (#2370)

* Fix #2119 - Whenever about to send a HTTP request, normalize the URI

* Add test for IDN request in FetchLinkCardService

* Perform IDN normalization on domains before they are stored in the DB

											
										
										
											2017-04-25 10:47:31 +10:00
-												Fix #2572 - Resolve preview cards for remote statuses as well as local ones (#3088)


											
										
										
											2017-05-17 08:41:15 +10:00
+								    it 'ignores URLs to hashtags' do
-												Remove HEAD request from fetching link previews (#12028)

It is not really necessary and we need to reduce requests
											
										
										
											2019-10-01 12:54:10 +10:00
+								      expect(a_request(:get, 'https://quitter.se/tag/wannacry')).to_not have_been_made
-												Fix #2572 - Resolve preview cards for remote statuses as well as local ones (#3088)


											
										
										
											2017-05-17 08:41:15 +10:00
+								    end
-												Punycode URI normalization (#2370)

* Fix #2119 - Whenever about to send a HTTP request, normalize the URI

* Add test for IDN request in FetchLinkCardService

* Perform IDN normalization on domains before they are stored in the DB

											
										
										
											2017-04-25 10:47:31 +10:00
+								  end
 								end