html-proofer 4.4.0 → 4.4.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 98bde6dd5e32f42e5983fd1aaa7bf02d3359b40f2b4a8b5360a5826abf2cc674
-  data.tar.gz: 9b7d57fd18e625ab7cc91c46faceac510b05ba5290f665fb5dd333b92446b00b
+  metadata.gz: 4043366fa1c0d7aaa58fc0b577d25d2d0fba7e72afca2b43ae688e90c55dddef
+  data.tar.gz: a93621d2169757d3665e5950e8c99d23a1b64cc9014fbc806a986fcdcb906e26
 SHA512:
-  metadata.gz: a666be806bbb70028488b3ef89f1325d7b5faaec9f48e105220db2d2f7dd022c12dabac828949ffd2100bdc2b4e3bd4c1f9ce9df48bab17b9106d9453eb49f01
-  data.tar.gz: 8a58bb5b120ab3cf3eb0e692f3f52f4385bd5923d3a7c91e95dddd17d8b0e25d20dd207872099665ac37737afb6b4ed86d04d39bc7793c6af2abcd35390d2051
+  metadata.gz: 1636df520fa7d4035ca482eeafd0b5e059b0fc77a0219f340e7c47b8f2105c8dd107bf03e0d78ca582f5d7abef9d5853edae498a2133b75eab6f92b7e77b636f
+  data.tar.gz: 5310df2250a5db3c8864b12ae01e49a0a2d043fc15f5bece45615cbad1d2b9100c899fa35902db5536b2be85908f9208ebfe65e31318cf6a6db47e945a745f17

data/lib/html_proofer/url_validator/internal.rb

@@ -22,22 +22,39 @@ module HTMLProofer
       end
 
       def run_internal_link_checker(links)
+        # collect urls and metadata for hashes to be checked in the same target file
+        file_paths_hashes_to_check = {}
         to_add = []
-        links.each_pair do |link, matched_files|
+        links.each_with_index do |(link, matched_files), i|
+          matched_count_to_log = pluralize(matched_files.count, "reference", "references")
+          @logger.log(:debug, "(#{i + 1} / #{links.count}) Internal link #{link}: Checking #{matched_count_to_log}")
           matched_files.each do |metadata|
             url = HTMLProofer::Attribute::Url.new(@runner, link, base_url: metadata[:base_url])
 
             @runner.current_source = metadata[:source]
             @runner.current_filename = metadata[:filename]
 
-            unless file_exists?(url)
+            target_file_path = url.absolute_path
+            unless file_exists?(target_file_path)
               @failed_checks << Failure.new(@runner.current_filename, "Links > Internal",
                 "internally linking to #{url}, which does not exist", line: metadata[:line], status: nil, content: nil)
               to_add << [url, metadata, false]
               next
             end
 
-            unless hash_exists?(url)
+            hash_exists = hash_exists_for_url?(url)
+            if hash_exists.nil?
+              # the hash needs to be checked in the target file, we collect the url and metadata
+              unless file_paths_hashes_to_check.key?(target_file_path)
+                file_paths_hashes_to_check[target_file_path] = {}
+              end
+              unless file_paths_hashes_to_check[target_file_path].key?(url.hash)
+                file_paths_hashes_to_check[target_file_path][url.hash] = []
+              end
+              file_paths_hashes_to_check[target_file_path][url.hash] << [url, metadata]
+              next
+            end
+            unless hash_exists
               @failed_checks << Failure.new(@runner.current_filename, "Links > Internal",
                 "internally linking to #{url}; the file exists, but the hash '#{url.hash}' does not", line: metadata[:line], status: nil, content: nil)
               to_add << [url, metadata, false]
@@ -48,6 +65,24 @@ module HTMLProofer
           end
         end
 
+        # check hashes by target file
+        @logger.log(:info, "Checking internal link hashes in #{pluralize(file_paths_hashes_to_check.count, "file", "files")}")
+        file_paths_hashes_to_check.each_with_index do |(file_path, hashes_to_check), i|
+          hash_count_to_log = pluralize(hashes_to_check.count, "hash", "hashes")
+          @logger.log(:debug, "(#{i + 1} / #{file_paths_hashes_to_check.count}) Checking #{hash_count_to_log} in #{file_path}")
+          html = create_nokogiri(file_path)
+          hashes_to_check.each_pair do |href_hash, url_metadata|
+            exists = hash_exists_in_html?(href_hash, html)
+            url_metadata.each do |(url, metadata)|
+              unless exists
+                @failed_checks << Failure.new(metadata[:filename], "Links > Internal",
+                  "internally linking to #{url}; the file exists, but the hash '#{href_hash}' does not", line: metadata[:line], status: nil, content: nil)
+              end
+              to_add << [url, metadata, exists]
+            end
+          end
+        end
+
         # adding directly to the cache above results in an endless loop
         to_add.each do |(url, metadata, exists)|
           @cache.add_internal(url.to_s, metadata, exists)
@@ -56,15 +91,15 @@ module HTMLProofer
         @failed_checks
       end
 
-      private def file_exists?(url)
-        absolute_path = url.absolute_path
-        return @runner.checked_paths[url.absolute_path] if @runner.checked_paths.key?(absolute_path)
+      private def file_exists?(absolute_path)
+        return @runner.checked_paths[absolute_path] if @runner.checked_paths.key?(absolute_path)
 
-        @runner.checked_paths[url.absolute_path] = File.exist?(absolute_path)
+        @runner.checked_paths[absolute_path] = File.exist?(absolute_path)
       end
 
-      # verify the target hash
-      private def hash_exists?(url)
+      # verify the hash w/o just based on the URL, w/o looking at the target file
+      # => returns nil if the has could not be verified
+      private def hash_exists_for_url?(url)
         href_hash = url.hash
         return true if blank?(href_hash)
         return true unless @runner.options[:check_internal_hash]
@@ -76,10 +111,18 @@ module HTMLProofer
         decoded_href_hash = Addressable::URI.unescape(href_hash)
         fragment_ids = [href_hash, decoded_href_hash]
         # https://www.w3.org/TR/html5/single-page.html#scroll-to-fragid
-        fragment_ids.include?("top") || !find_fragments(fragment_ids, url).empty?
+        return true if fragment_ids.include?("top")
+
+        nil
+      end
+
+      private def hash_exists_in_html?(href_hash, html)
+        decoded_href_hash = Addressable::URI.unescape(href_hash)
+        fragment_ids = [href_hash, decoded_href_hash]
+        !find_fragments(fragment_ids, html).empty?
       end
 
-      private def find_fragments(fragment_ids, url)
+      private def find_fragments(fragment_ids, html)
         xpaths = fragment_ids.uniq.flat_map do |frag_id|
           escaped_frag_id = "'#{frag_id.split("'").join("', \"'\", '")}', ''"
           [
@@ -89,7 +132,6 @@ module HTMLProofer
         end
         xpaths << XpathFunctions.new
 
-        html = create_nokogiri(url.absolute_path)
         html.xpath(*xpaths)
       end
     end

data/lib/html_proofer/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module HTMLProofer
-  VERSION = "4.4.0"
+  VERSION = "4.4.1"
 end

data/lib/html_proofer.rb

@@ -4,7 +4,7 @@ require "zeitwerk"
 lib_dir = File.join(File.dirname(__dir__), "lib")
 gem_loader = Zeitwerk::Loader.for_gem
 gem_loader.inflector.inflect(
-  "html_proofer" => "HTMLProofer"
+  "html_proofer" => "HTMLProofer",
 )
 gem_loader.ignore(File.join(lib_dir, "html-proofer.rb"))
 gem_loader.setup

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: html-proofer
 version: !ruby/object:Gem::Version
-  version: 4.4.0
+  version: 4.4.1
 platform: ruby
 authors:
 - Garen Torikian
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-08-13 00:00:00.000000000 Z
+date: 2022-09-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: addressable