feat(deps): use html2rss in latest development status

Dockerfile

@@ -57,6 +57,8 @@ RUN apk add --no-cache \
   --no-create-home \
   --uid "$UID" "$USER" \
   && mkdir -p /app \
+  && mkdir -p /app/tmp/rack-cache-body \
+  && mkdir -p /app/tmp/rack-cache-meta \
   && chown "$USER":"$USER" -R /app
 
 WORKDIR /app

Gemfile

@@ -4,7 +4,9 @@ source 'https://rubygems.org'
 
 git_source(:github) { |repo_name| "https://github.com/#{repo_name}" }
 
-gem 'html2rss', '~> 0.14'
+# gem 'html2rss', '~> 0.14'
+gem 'html2rss', github: 'html2rss/html2rss'
+
 gem 'html2rss-configs', github: 'html2rss/html2rss-configs'
 
 # Use these instead of the two above (uncomment them) when developing locally:

app.rb

@@ -84,11 +84,38 @@ def self.development? = ENV['RACK_ENV'] == 'development'
         end
 
         r.on String, String do |folder_name, config_name_with_ext|
-          handle_html2rss_configs(request, folder_name, config_name_with_ext)
+          response['Content-Type'] = CONTENT_TYPE_RSS
+
+          name = "#{folder_name}/#{File.basename(config_name_with_ext, '.*')}"
+          config = Html2rss::Configs.find_by_name(name)
+
+          if (params = request.params).any?
+            config[:params] ||= {}
+            config[:params].merge!(params)
+          end
+
+          feed = Html2rss.feed(config)
+
+          HttpCache.expires(response, feed.channel.ttl.to_i * 60, cache_control: 'public')
+
+          feed.to_s
         end
 
         r.on String do |config_name_with_ext|
-          handle_local_config_feeds(request, config_name_with_ext)
+          response['Content-Type'] = CONTENT_TYPE_RSS
+
+          config = LocalConfig.find(File.basename(config_name_with_ext, '.*'))
+
+          if (params = request.params).any?
+            config[:params] ||= {}
+            config[:params].merge!(params)
+          end
+
+          feed = Html2rss.feed(config)
+
+          HttpCache.expires(response, feed.channel.ttl.to_i * 60, cache_control: 'public')
+
+          feed.to_s
         end
       end
 

app/http_cache.rb

@@ -15,6 +15,8 @@ module HttpCache
       # @param seconds [Integer]
       # @param cache_control [String, nil]
       def expires(response, seconds, cache_control: nil)
+        expires_now(response) and return if seconds <= 0
+
         response['Expires'] = (Time.now + seconds).httpdate
 
         cache_value = "max-age=#{seconds}"

app/ssrf_filter_strategy.rb

@@ -16,6 +16,7 @@ def execute
         response = SsrfFilter.get(ctx.url, headers:)
 
         Html2rss::RequestService::Response.new(body: response.body,
+                                               url: ctx.url,
                                                headers: response.to_hash.transform_values(&:first))
       end
     end

config/feeds.yml

@@ -16,6 +16,6 @@ feeds:
         selector: "li > div"
       title:
         selector: "h4"
-      link:
+      url:
         selector: "a"
         extractor: "href"

helpers/handle_error.rb

@@ -8,12 +8,10 @@ module Web
     class App
       def handle_error(error) # rubocop:disable Metrics/MethodLength
         case error
-        when Html2rss::Config::ParamsMissing,
+        when Html2rss::Config::DynamicParams::ParamsMissing,
              Roda::RodaPlugins::TypecastParams::Error
           set_error_response('Parameters missing or invalid', 422)
-        when Html2rss::AttributePostProcessors::UnknownPostProcessorName,
-             Html2rss::ItemExtractors::UnknownExtractorName,
-             Html2rss::Config::ChannelMissing
+        when Html2rss::Selectors::PostProcessors::UnknownPostProcessorName
           set_error_response('Invalid feed config', 422)
         when LocalConfig::NotFound,
              Html2rss::Configs::ConfigNotFound

routes/auto_source.rb

@@ -21,28 +21,19 @@ class App
 
             r.on String, method: :get do |encoded_url|
               strategy = (request.params['strategy'] || :ssrf_filter).to_sym
-              unless Html2rss::RequestService.strategy_registered?(strategy)
-                raise Html2rss::RequestService::UnknownStrategy
-              end
 
-              response['Content-Type'] = CONTENT_TYPE_RSS
-
-              url = Addressable::URI.parse Base64.urlsafe_decode64(encoded_url)
-              rss = Html2rss.auto_source(url, strategy:)
+              url = Addressable::URI.parse(Base64.urlsafe_decode64(encoded_url))
 
-              # Unfortunately, Ruby's rss gem does not provide a direct method to
-              # add an XML stylesheet to the RSS::RSS object itself.
-              stylesheet = Html2rss::RssBuilder::Stylesheet.new(href: '/rss.xsl', type: 'text/xsl').to_xml
+              feed = Html2rss.feed(stylesheets: [{ href: '/rss.xsl', type: 'text/xsl' }],
+                                   strategy:,
+                                   channel: { url: url.to_s },
+                                   auto_source: {})
 
-              xml_content = rss.to_xml
-              xml_content.sub!(/^<\?xml version="1.0" encoding="UTF-8"\?>/,
-                               "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n#{stylesheet}")
+              HttpCache.expires(response, AutoSource.ttl_in_seconds(feed), cache_control: 'private, must-revalidate')
 
-              HttpCache.expires response,
-                                AutoSource.ttl_in_seconds(rss),
-                                cache_control: 'private, must-revalidate'
-
-              xml_content
+              response['Content-Type'] = CONTENT_TYPE_RSS
+              response.status = 200
+              feed.to_xml
             end
           else
             # auto_source feature is disabled

spec/routes/auto_source_spec.rb

@@ -76,8 +76,9 @@ def app = described_class
 
       it 'responds successfully', :aggregate_failures do
         expect(response).to be_ok
+        expect(response.status).to eq 200
         expect(response.body).to start_with '<?xml version="1.0" encoding="UTF-8"?>'
-        expect(response.get_header('cache-control')).to eq 'must-revalidate, private, max-age=0'
+        expect(response.get_header('cache-control')).to eq 'must-revalidate, no-cache, no-store, private, max-age=0'
         expect(response.get_header('content-type')).to eq described_class::CONTENT_TYPE_RSS
       end
     end
@@ -93,7 +94,7 @@ def app = described_class
 
       it 'responds with Error', :aggregate_failures do
         expect(response.status).to eq 422
-        expect(response.body).to match(/UnknownStrategy/)
+        expect(response.body).to match(/Html2rss::Config::InvalidConfig/)
       end
     end
   end