jekyll-auto-thumbnails 1.0.0 → 2.0.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 69d56f34b0465c7143f8785a0426c2bac56a268b633b64d901d1365c6baab476
-  data.tar.gz: ec8b83991caa1e78aadf337e4ef9ff09c0f480767e6052abe74706e020197a38
+  metadata.gz: d593b4ae0f39bd0020f6656a021327666afa8ee4776a2147d36ce3f8283a1840
+  data.tar.gz: af035e37d27382e9a8d1d15b20ea14e3a0994f2fe1d1ff312c09b4dac924829f
 SHA512:
-  metadata.gz: c280d35426d31fa2b9d062b085f482622a8162127d4ddad31b48d0173698067405dd00f6ab40c422a02b84444278a8b888622a6cf58b2fee6c48573f6e421d4d
-  data.tar.gz: b0f71661e914d71dff921f6b802c425a80e3ef7289263aff18686cb61b3299059e8a2733045ee26f9ba2d42aab2e0e26ce0ecb61f5ed3021ee62e61a3432d7b0
+  metadata.gz: 89945f98a6e83a64ec2d4f20b96e827d4e408ffdda0f3340da9e99086ef156ef146a6617db1f3cb1ba7de8f34d4790f47783924fc3e02afb5e1f9a9c90506d33
+  data.tar.gz: 0b0a29fc0985c1a1930f2b1a644ed574fb4abae81fe52f0722b0a329c2731020e7cc915b7fdf8155b63e1767f2227ac409a5e2697152648d61cdca6816f59582

data/CHANGELOG.md

@@ -1,5 +1,29 @@
 # Changelog
 
+## [2.0.1](https://github.com/Texarkanine/jekyll-auto-thumbnails/compare/v2.0.0...v2.0.1) (2026-05-14)
+
+
+### Bug Fixes
+
+* **deps:** bump addressable from 2.8.8 to 2.9.0 ([#31](https://github.com/Texarkanine/jekyll-auto-thumbnails/issues/31)) ([e4c0312](https://github.com/Texarkanine/jekyll-auto-thumbnails/commit/e4c03124bbe1bb18183b0ffec1039cdc4f4cefb1))
+* **deps:** bump nokogiri from 1.19.2 to 1.19.3 in the production-deps group ([#34](https://github.com/Texarkanine/jekyll-auto-thumbnails/issues/34)) ([1248eb3](https://github.com/Texarkanine/jekyll-auto-thumbnails/commit/1248eb3b81b3269210fa7d1e2fc98928803030ce))
+* **deps:** bump nokogiri in the production-deps group ([1248eb3](https://github.com/Texarkanine/jekyll-auto-thumbnails/commit/1248eb3b81b3269210fa7d1e2fc98928803030ce))
+
+## [2.0.0](https://github.com/Texarkanine/jekyll-auto-thumbnails/compare/v1.0.0...v2.0.0) (2026-04-24)
+
+
+### ⚠ BREAKING CHANGES
+
+* Incorrectly round-tripped all content through HTML4 parser ([#30](https://github.com/Texarkanine/jekyll-auto-thumbnails/issues/30))
+
+### Bug Fixes
+
+* **deps:** bump nokogiri from 1.19.0 to 1.19.1 in the production-deps group ([#20](https://github.com/Texarkanine/jekyll-auto-thumbnails/issues/20)) ([0ed6c0c](https://github.com/Texarkanine/jekyll-auto-thumbnails/commit/0ed6c0cf65aa17ed0182ff7eb1f85d8b053188e1))
+* **deps:** bump nokogiri from 1.19.1 to 1.19.2 in the production-deps group across 1 directory ([#24](https://github.com/Texarkanine/jekyll-auto-thumbnails/issues/24)) ([023882c](https://github.com/Texarkanine/jekyll-auto-thumbnails/commit/023882c815b26f1d1143635c56f2adc78a124b85))
+* **deps:** bump nokogiri in the production-deps group ([023882c](https://github.com/Texarkanine/jekyll-auto-thumbnails/commit/023882c815b26f1d1143635c56f2adc78a124b85))
+* **deps:** bump nokogiri in the production-deps group ([0ed6c0c](https://github.com/Texarkanine/jekyll-auto-thumbnails/commit/0ed6c0cf65aa17ed0182ff7eb1f85d8b053188e1))
+* Incorrectly round-tripped all content through HTML4 parser ([#30](https://github.com/Texarkanine/jekyll-auto-thumbnails/issues/30)) ([75640c8](https://github.com/Texarkanine/jekyll-auto-thumbnails/commit/75640c8c0de097e1ebb9e615e1acf8a00c0ef7e6))
+
 ## [1.0.0](https://github.com/Texarkanine/jekyll-auto-thumbnails/compare/v0.3.3...v1.0.0) (2026-01-18)
 
 

data/README.md

@@ -43,8 +43,25 @@ auto_thumbnails:
   
   # JPEG quality for generated thumbnails (0-100)
   quality: 85  # default: 85
+
+  # HTML parser used for scanning and rewriting.
+  # - html5 (default, CRuby only): standards-compliant, does not inject a
+  #   <meta http-equiv="Content-Type"> when serializing.
+  # - html4: legacy libxml2-based parser; re-inserts the encoding meta tag
+  #   on serialize. Required on JRuby.
+  parser: html5  # default: html5
 ```
 
+### HTML parser (`parser: html5 | html4`)
+
+As of `v2.0.0`, `jekyll-auto-thumbnails` parses and serializes HTML with Nokogiri's HTML5 parser by default.
+
+Set `parser: html4` to restore the previous behavior (the libxml2-based `Nokogiri::HTML` parser) if you depended on its serialization quirks.
+
+**JRuby note**: `Nokogiri::HTML5` is not available on JRuby. Under JRuby, you **must** explicitly set `parser: html4`; otherwise the plugin will raise a configuration error at build time.
+
+Pages that contain no replaceable `<article><img>` are no longer re-serialized at all - they are passed through byte-for-byte, regardless of which parser is selected.
+
 ## How It Works
 
 1. **HTML Scanning**: After all plugins run, scans `<article>` tags for images

data/lib/jekyll-auto-thumbnails/configuration.rb

@@ -6,7 +6,9 @@ module JekyllAutoThumbnails
   # Parses Jekyll site config for auto_thumbnails options and provides
   # accessor methods with appropriate defaults and validation.
   class Configuration
-    attr_reader :max_width, :max_height, :quality, :cache_dir
+    VALID_PARSERS = %i[html4 html5].freeze
+
+    attr_reader :max_width, :max_height, :quality, :cache_dir, :parser
 
     # Initialize configuration from Jekyll site
     #
@@ -18,6 +20,7 @@ module JekyllAutoThumbnails
       @max_width = parse_dimension(config_hash["max_width"])
       @max_height = parse_dimension(config_hash["max_height"])
       @quality = parse_quality(config_hash.fetch("quality", 85))
+      @parser = parse_parser(config_hash.fetch("parser", "html5"))
       @cache_dir = File.join(site.source, ".jekyll-cache", "jekyll-auto-thumbnails")
     end
 
@@ -47,5 +50,39 @@ module JekyllAutoThumbnails
       val = value.to_i
       val.between?(0, 100) ? val : 85
     end
+
+    # Parse and validate the HTML parser choice.
+    #
+    # Accepts "html4" or "html5" (case-insensitive). Raises on any other
+    # input so HTML correctness is never silently traded for availability.
+    # JRuby cannot use the HTML5 parser (Nokogiri::HTML5 is CRuby-only),
+    # so selecting it under JRuby also raises, directing the user to set
+    # `auto_thumbnails.parser: html4` explicitly.
+    #
+    # @param value [Object] parser value from config
+    # @return [Symbol] :html4 or :html5
+    # @raise [ArgumentError] for invalid values or JRuby + :html5
+    def parse_parser(value)
+      unless value.is_a?(String)
+        raise ArgumentError,
+              "auto_thumbnails: parser must be a string (\"html4\" or \"html5\"); got #{value.inspect}"
+      end
+
+      symbol = value.downcase.to_sym
+      unless VALID_PARSERS.include?(symbol)
+        raise ArgumentError,
+              "auto_thumbnails: parser must be one of #{VALID_PARSERS.join(", ")}; " \
+              "got #{value.inspect}"
+      end
+
+      if symbol == :html5 && RUBY_ENGINE == "jruby"
+        raise ArgumentError,
+              "auto_thumbnails: parser: html5 is not supported on JRuby " \
+              "(Nokogiri::HTML5 is CRuby-only). " \
+              "Set auto_thumbnails.parser: html4 in _config.yml to run on JRuby."
+      end
+
+      symbol
+    end
   end
 end

data/lib/jekyll-auto-thumbnails/hooks.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require_relative "html_parser"
+
 module JekyllAutoThumbnails
   # Jekyll hook integration
   module Hooks
@@ -72,7 +74,7 @@ module JekyllAutoThumbnails
         next unless doc.output
         next unless html_document?(doc)
 
-        doc.output = replace_urls(doc.output, url_map)
+        doc.output = replace_urls(doc.output, url_map, parser: config.parser)
       end
 
       Jekyll.logger.info "AutoThumbnails:", "Generated #{url_map.size} thumbnails"
@@ -105,28 +107,37 @@ module JekyllAutoThumbnails
       Jekyll.logger.info "AutoThumbnails:", "All thumbnails copied"
     end
 
-    # Replace image URLs in HTML
+    # Replace image URLs in HTML.
+    #
+    # Returns the input string unchanged (object identity) when no replacement
+    # was actually made. This is both a perf win for pages with no matching
+    # images and a correctness win: the HTML4 path otherwise round-trips every
+    # page through libxml2's serializer, which injects a spurious
+    # `<meta http-equiv="Content-Type">` on HTML5 sites.
     #
     # @param html [String] HTML content
     # @param url_map [Hash] original URL => thumbnail URL
-    # @return [String] modified HTML
-    def self.replace_urls(html, url_map)
-      # Return early if no replacements needed
+    # @param parser [Symbol] :html5 (default) or :html4
+    # @return [String] modified HTML (or the input itself, unchanged)
+    def self.replace_urls(html, url_map, parser: :html5)
       return html if url_map.empty?
+      return html unless html.match?(/<img/i)
 
-      doc = Nokogiri::HTML(html)
+      doc = HtmlParser.parse(html, parser)
 
+      modified = false
       doc.css("article img").each do |img|
         src = img["src"]
         next unless src
 
-        # Find thumbnail URL for this image
         thumb_url = url_map[src]
-        img["src"] = thumb_url if thumb_url
+        next unless thumb_url
+
+        img["src"] = thumb_url
+        modified = true
       end
 
-      # Serialize with encoding declaration to match Jekyll output
-      doc.to_html
+      modified ? doc.to_html : html
     end
 
     # Check if a document outputs HTML (not CSS, JS, etc.)

data/lib/jekyll-auto-thumbnails/html_parser.rb

@@ -0,0 +1,39 @@
+# frozen_string_literal: true
+
+require "nokogiri"
+# Nokogiri::HTML5 ships with Nokogiri on CRuby but is not available on JRuby.
+# Configuration rejects `parser: :html5` under JRuby, so this require only
+# needs to succeed on CRuby. On JRuby, the HTML4 branch is the only path
+# that ever runs.
+require "nokogiri/html5" unless RUBY_ENGINE == "jruby"
+
+module JekyllAutoThumbnails
+  # Dispatch between Nokogiri's HTML5 and libxml2-based HTML4 parsers.
+  #
+  # HTML5 is the default throughout the gem; HTML4 is kept as an opt-in for
+  # users who depended on libxml2's serialization quirks (notably the
+  # injected `<meta http-equiv="Content-Type">`).
+  module HtmlParser
+    module_function
+
+    # Parse an HTML string with the selected parser.
+    #
+    # @param html [String] HTML source
+    # @param parser [Symbol] :html5 or :html4
+    # @return [Nokogiri::HTML5::Document, Nokogiri::HTML4::Document]
+    # @raise [ArgumentError] for an unknown parser symbol
+    # @raise [NameError] if :html5 is requested under JRuby (should be
+    #   prevented by Configuration validation; this is a belt-and-suspenders
+    #   guard)
+    def parse(html, parser)
+      case parser
+      when :html5
+        Nokogiri::HTML5.parse(html)
+      when :html4
+        Nokogiri::HTML(html)
+      else
+        raise ArgumentError, "Unknown HTML parser: #{parser.inspect}"
+      end
+    end
+  end
+end

data/lib/jekyll-auto-thumbnails/imagemagick_wrapper.rb

@@ -83,7 +83,7 @@ module JekyllAutoThumbnails
     #
     # @param cmd [String] command name
     # @return [Boolean] true if command found
-    def self.command_exists?(cmd)
+    private_class_method def self.command_exists?(cmd)
       cmd_name = Gem.win_platform? ? "#{cmd}.exe" : cmd
       path_dirs = ENV["PATH"].to_s.split(File::PATH_SEPARATOR)
 

data/lib/jekyll-auto-thumbnails/scanner.rb

@@ -2,6 +2,7 @@
 
 require "nokogiri"
 require "open3"
+require_relative "html_parser"
 require_relative "imagemagick_wrapper"
 
 module JekyllAutoThumbnails
@@ -14,7 +15,7 @@ module JekyllAutoThumbnails
     # @param config [Configuration] configuration
     # @param site_source [String] Jekyll site source (optional, for unsized image checking)
     def self.scan_html(html, registry, config, site_source = nil)
-      doc = Nokogiri::HTML(html)
+      doc = HtmlParser.parse(html, config.parser)
 
       # Only process images in <article> tags
       doc.css("article img").each do |img|

data/lib/jekyll-auto-thumbnails/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module JekyllAutoThumbnails
-  VERSION = "1.0.0"
+  VERSION = "2.0.1"
 end

data/lib/jekyll-auto-thumbnails.rb

@@ -4,6 +4,7 @@ require "jekyll"
 require "nokogiri"
 
 require_relative "jekyll-auto-thumbnails/version"
+require_relative "jekyll-auto-thumbnails/html_parser"
 require_relative "jekyll-auto-thumbnails/configuration"
 require_relative "jekyll-auto-thumbnails/url_resolver"
 require_relative "jekyll-auto-thumbnails/digest_calculator"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: jekyll-auto-thumbnails
 version: !ruby/object:Gem::Version
-  version: 1.0.0
+  version: 2.0.1
 platform: ruby
 authors:
 - Texarkanine
@@ -158,6 +158,7 @@ files:
 - lib/jekyll-auto-thumbnails/digest_calculator.rb
 - lib/jekyll-auto-thumbnails/generator.rb
 - lib/jekyll-auto-thumbnails/hooks.rb
+- lib/jekyll-auto-thumbnails/html_parser.rb
 - lib/jekyll-auto-thumbnails/imagemagick_wrapper.rb
 - lib/jekyll-auto-thumbnails/registry.rb
 - lib/jekyll-auto-thumbnails/scanner.rb