nokogiri-html-ext 1.4.1 → 1.6.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 3c5f1ccd4ab43e822b1bc075f4b84c212b9ab539e07b206d2b65499a775ed539
-  data.tar.gz: 15ef0baaa308692662f39e238bedba2029e336b53d753190182b19b770e0ea6b
+  metadata.gz: 307b06335c0935cec4304b10c13ef113c514646d5a746f703ccb1567a39120ce
+  data.tar.gz: 26894c4ef4b3a018f2b776f4929035e590a8c9514af529fce581c23fb29e4cf0
 SHA512:
-  metadata.gz: 2ba7cba037735ea0323abae2828d0c8abd576886b97fef12b8319add735dedbe77ada5433f47e4f42a240855ff196dd178925a61bcbd99e7be834bf500423cd7
-  data.tar.gz: 53ddcdcc16b93fdf00947acaec6764d56f95665cac5bc96868a2a20a627b917907c1fa7182639d89d0359e5c0eed6ca4932cb112e2944de9403fca83d888a5e2
+  metadata.gz: 66057c43dd9cc165dcd12603f8546119f39292305b3d943f5930c138686acbec53d2dfd744565343fe1f2351a8aa9a46b84a6d7d1ba7cafe7f10dd17be19548e
+  data.tar.gz: 52e9543d90ed8dc1790458f660a02cdb7c4574cb8c3fc7a093b7780cbe4627a8badc6ef84d51aae7acdc7ad5aa2e88cdec38e5eb597acffd6f8e8652950b62f5

data/lib/nokogiri/html_ext/document.rb

@@ -3,6 +3,9 @@
 module Nokogiri
   module HTML4
     class Document < Nokogiri::XML::Document
+      # @see https://html.spec.whatwg.org/#image-candidate-string
+      IMAGE_CANDIDATE_STRING_PATTERN = /^(?<url>.+?)(?<descriptor>\s+.+)?$/.freeze
+
       # A map of HTML +srcset+ attributes and their associated element names.
       #
       # @see https://html.spec.whatwg.org/#srcset-attributes
@@ -56,37 +59,62 @@ module Nokogiri
         base["href"] = url
       end
 
-      # Convert a relative URL to an absolute URL based on the current document.
+      # Conditionally normalize a URL's empty path component.
       #
-      # @param url [String]
+      # @param [String]
       #
       # @return [String]
-      def resolve_relative_url(url)
-        strs = [document.url, base_href, url].filter_map { |str| str&.strip }
-
-        uri = Addressable::URI.join(*strs)
+      def normalize_empty_path(url)
+        uri = Addressable::URI.parse(url.strip)
 
-        # Normalize an empty path component.
-        #
-        # @see https://datatracker.ietf.org/doc/html/rfc3986#section-6.2.3
-        #   RFC 3986 Uniform Resource Identifier (URI): Generic Syntax § 6.2.3. Scheme-Based Normalization
-        uri.path = uri.normalized_path if uri.path.empty?
+        uri.path = uri.normalized_path if uri.path.empty? && uri.path != uri.normalized_path
 
         uri.to_s
       rescue Addressable::URI::InvalidURIError
         url
       end
 
-      # Convert the document's relative URLs to absolute URLs.
+      # Normalize the document's URLs whose path components are empty.
       #
       # @return [self]
-      def resolve_relative_urls!
+      def normalize_empty_paths!
         xpath(*ATTRIBUTES_XPATHS).each do |attr_node|
           attr_node.value =
             if SRCSET_ATTRIBUTES_MAP.key?(attr_node.name)
-              resolve_srcset_attributes(attr_node.value.split(/\s*,\s*/))
+              normalize_srcset_attribute(attr_node.value)
             else
-              resolve_relative_url(attr_node.value)
+              normalize_empty_path(attr_node.value)
+            end
+        end
+
+        self
+      end
+
+      # Resolve a URL based on the current document.
+      #
+      # @param url [String]
+      #
+      # @return [String]
+      def resolve_url(url)
+        uris = [document.url, base_href, url].filter_map { |uri| Addressable::URI.parse(uri&.strip) }
+
+        return uris.last.to_s if uris.last.absolute?
+
+        Addressable::URI.join(*uris).to_s
+      rescue Addressable::URI::InvalidURIError
+        url
+      end
+
+      # Resolve the document's URLs
+      #
+      # @return [self]
+      def resolve_urls!
+        xpath(*ATTRIBUTES_XPATHS).each do |attr_node|
+          attr_node.value =
+            if SRCSET_ATTRIBUTES_MAP.key?(attr_node.name)
+              resolve_srcset_attribute(attr_node.value)
+            else
+              resolve_url(attr_node.value)
             end
         end
 
@@ -95,20 +123,40 @@ module Nokogiri
 
       private
 
-      # Resolve a set of +String+s that represent +srcset+ attribute image
-      # candidate strings.
+      # Normalize the URLs in a +srcset+ attribute's image candidate strings.
       #
-      # @param srcset_attributes [Array<String>]
+      # @param value [String]
       #
       # @return [String]
-      def resolve_srcset_attributes(srcset_attributes)
-        srcset_attributes.map! do |candidate_string|
-          # rubocop:disable Style/PerlBackrefs
-          candidate_string.sub(/^(.+?)(\s+.+)?$/) { "#{resolve_relative_url($1)}#{$2}" }
-          # rubocop:enable Style/PerlBackrefs
-        end
+      def normalize_srcset_attribute(value)
+        parse_image_candidate_strings(value)
+          .map! { |captures| "#{normalize_empty_path(captures["url"])}#{captures["descriptor"]}" }
+          .join(", ")
+      end
 
-        srcset_attributes.join(", ")
+      # Parse a +srcset+ attribute's value into a +Hash+ of image candidate
+      # strings.
+      #
+      # @see https://html.spec.whatwg.org/#image-candidate-string
+      #
+      # @param [String]
+      #
+      # @return [Array<Hash{String => String, nil}>]
+      def parse_image_candidate_strings(value)
+        value
+          .split(/\s*,\s*/)
+          .map! { |candidate_string| candidate_string.match(IMAGE_CANDIDATE_STRING_PATTERN).named_captures }
+      end
+
+      # Resolve the URLs in a +srcset+ attribute's image candidate strings.
+      #
+      # @param value [String]
+      #
+      # @return [String]
+      def resolve_srcset_attribute(value)
+        parse_image_candidate_strings(value)
+          .map! { |captures| "#{resolve_url(captures["url"])}#{captures["descriptor"]}" }
+          .join(", ")
       end
     end
   end

data/nokogiri-html-ext.gemspec

@@ -4,7 +4,7 @@ Gem::Specification.new do |spec|
   spec.required_ruby_version = ">= 2.7"
 
   spec.name          = "nokogiri-html-ext"
-  spec.version       = "1.4.1"
+  spec.version       = "1.6.0"
   spec.authors       = ["Jason Garber"]
   spec.email         = ["jason@sixtwothree.org"]
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: nokogiri-html-ext
 version: !ruby/object:Gem::Version
-  version: 1.4.1
+  version: 1.6.0
 platform: ruby
 authors:
 - Jason Garber
@@ -55,11 +55,11 @@ licenses:
 - MIT
 metadata:
   bug_tracker_uri: https://codeberg.org/jgarber/nokogiri-html-ext/issues
-  changelog_uri: https://codeberg.org/jgarber/nokogiri-html-ext/releases/tag/v1.4.1
-  documentation_uri: https://rubydoc.info/gems/nokogiri-html-ext/1.4.1
+  changelog_uri: https://codeberg.org/jgarber/nokogiri-html-ext/releases/tag/v1.6.0
+  documentation_uri: https://rubydoc.info/gems/nokogiri-html-ext/1.6.0
   homepage_uri: https://codeberg.org/jgarber/nokogiri-html-ext
   rubygems_mfa_required: 'true'
-  source_code_uri: https://codeberg.org/jgarber/nokogiri-html-ext/src/tag/v1.4.1
+  source_code_uri: https://codeberg.org/jgarber/nokogiri-html-ext/src/tag/v1.6.0
 rdoc_options: []
 require_paths:
 - lib