jekyll-l10n 1.3.14 → 1.3.16

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 26ebc278a54393fb2293a6f73ba9d68373a20705180d3e29127f701ba1f6f714
-  data.tar.gz: c25fd4fe9fa8ccbff69eb563480e54e57566464b7e7b283850ba3e4c55806e0a
+  metadata.gz: e9b11bbd19ea0a06c6282c11b22319d49ec9a25060d7324d667e14c78f397881
+  data.tar.gz: 0b7c89696c593e771c2694090d0fcebb4c9414a07fc2474f568c5d298953667b
 SHA512:
-  metadata.gz: 47e9cc0ee57f3bf1fdf5c78049232e88c05f7b5b57dc9ff221a098cb071515112cc192412419df12321e188dcb7b52aa4756f6b44511901472c2f24e0757ada8
-  data.tar.gz: 87c2d77e2397c5b3705dd59a351320457ecd77ba50578c0ba766dcc7101c7ee3ce5a985d4d41d4fb162e1b56f6b98d50cd4b11eee63f040e2601290b55ce6a1c
+  metadata.gz: fee23453215a6576799f774d9de4f94e870554b0c9fe9044ec7b0fd20477f20cb99d4d56d026005224774cdb4d332f0e493eb0bf8e8cf0750289a0028850fd8b
+  data.tar.gz: 0ac080acebbd08f462ee480f771804890111603e94f9c17aae24a12ec7e7d984659096c64e513e7c34c535fc5c2f0d127dad9a6038704ca9f4107b0c6d110115

data/lib/jekyll-l10n/po_file/writer.rb

@@ -189,7 +189,7 @@ module Jekyll
       end
 
       def self.escape_backslashes(value)
-        value.gsub('\\', '\\\\')
+        value.gsub('\\') { '\\\\' }
       end
 
       def self.escape_quotes_and_get_delimiter(escaped)

data/lib/jekyll-l10n/translation/block_text_extractor.rb

@@ -19,7 +19,6 @@ module Jekyll
     # * Remove nested block elements from text
     # * Remove empty icon tags (external link markers)
     # * Normalize and validate extracted text
-    # * Decode HTML entities
     #
     # @example
     #   text = BlockTextExtractor.extract(paragraph_node)
@@ -31,7 +30,8 @@ module Jekyll
       #
       # Returns nil if element is not extractable or if extracted text fails
       # validation. Clones the node, removes nested block elements and empty
-      # icon tags, normalizes whitespace, decodes HTML entities, and validates.
+      # icon tags, normalizes whitespace, and validates. HTML entities are
+      # preserved verbatim to match the keys produced by the extraction pipeline.
       #
       # @param node [Nokogiri::XML::Element] DOM element to extract from
       # @return [String, nil] Normalized text from element, or nil if not valid
@@ -43,7 +43,6 @@ module Jekyll
         HtmlTextUtils.remove_empty_icon_tags(clone)
 
         text = TextNormalizer.normalize(clone.inner_html).strip
-        text = HtmlTextUtils.decode_html_entities(text)
 
         TextValidator.valid?(text) ? text : nil
       end

data/lib/jekyll-l10n/translation/html_translator.rb

@@ -108,7 +108,7 @@ module Jekyll
         normalized_text = TextNormalizer.normalize(text)
         translated = TranslationResolver.resolve(node, normalized_text, translations)
 
-        return if apply_block_level_translation?(node, normalized_text, translated)
+        return if apply_block_level_translation?(node, normalized_text, translations)
 
         if @debug_logging
           log_translation_debug_info(text, normalized_text, translated,
@@ -147,8 +147,10 @@ module Jekyll
                              "[HtmlTranslator]   Translation: #{translation[0..truncate_length]}..."
         end
 
+        preserved_anchors = parent.css('.heading-anchor').map(&:dup)
         parent.children.each(&:remove)
         parent.inner_html = translation
+        preserved_anchors.each { |a| parent.add_child(a) } if parent.css('.heading-anchor').empty?
       end
 
       # Apply fallback when translation is missing or empty.
@@ -202,18 +204,45 @@ module Jekyll
         @debug_logging && text.include?('attribute')
       end
 
-      def apply_block_level_translation?(node, normalized_text, translated)
-        return false unless translated && node.parent && content_element?(node.parent)
+      def apply_block_level_translation?(node, normalized_text, translations)
+        ancestor = find_content_element_ancestor(node)
+        return false unless ancestor
 
-        return false if TranslationResolver.contains_protected_elements?(node.parent)
+        return false if TranslationResolver.contains_protected_elements?(ancestor)
 
-        block_text = BlockTextExtractor.extract(node.parent)
+        block_text = BlockTextExtractor.extract(ancestor)
         return false unless block_text && block_text != normalized_text
 
-        apply_block_level_translation(node.parent, translated)
+        translation = select_block_translation(node, ancestor, normalized_text, block_text,
+                                               translations)
+        return false unless translation
+
+        apply_block_level_translation(ancestor, translation)
         true
       end
 
+      # When text is directly inside a content element, prefer its direct translation
+      # (the msgstr may itself contain HTML, e.g. a heading with an anchor).
+      # When text is nested inside inline element(s), only apply if the full block
+      # text is an explicit translation key.
+      def select_block_translation(node, ancestor, normalized_text, block_text, translations)
+        if node.parent == ancestor
+          translations[normalized_text] || translations[block_text]
+        else
+          translations[block_text]
+        end
+      end
+
+      def find_content_element_ancestor(node)
+        current = node&.parent
+        while current&.element?
+          return current if content_element?(current)
+
+          current = current.parent
+        end
+        nil
+      end
+
       def log_translation_debug_info(text, normalized_text, translated, translations)
         translation_data = DebugLogger::TranslationData.new(text: text,
                                                             normalized_text: normalized_text,
@@ -223,7 +252,8 @@ module Jekyll
       end
 
       private :log_text_node_debug, :should_skip_translation?, :should_log_text_debug?,
-              :apply_block_level_translation?, :log_translation_debug_info
+              :apply_block_level_translation?, :select_block_translation,
+              :find_content_element_ancestor, :log_translation_debug_info
     end
   end
 end

data/lib/jekyll-l10n/utils/html_text_utils.rb

@@ -89,8 +89,10 @@ module Jekyll
       # Extract text with inline tags preserved.
       #
       # Extracts text from an element, removes block elements and empty icons,
-      # normalizes whitespace, and decodes HTML entities. Returns plain text
-      # suitable for translation.
+      # and normalizes whitespace. HTML entities (e.g. <, >) are preserved
+      # verbatim so that entity-encoded content inside inline elements (such as
+      # <code>&lt;p&gt;</code>) is written to PO msgids as-is and does not
+      # become a live HTML tag when the msgstr is later injected via inner_html.
       #
       # @param node [Nokogiri::XML::Node] Element to extract from
       # @return [String] Extracted and normalized text
@@ -100,7 +102,6 @@ module Jekyll
         remove_empty_icon_tags(clone)
 
         text = TextNormalizer.normalize(clone.inner_html)
-        text = decode_html_entities(text)
         text&.then { |t| TextNormalizer.normalize(t).strip }
       end
 

data/lib/jekyll-l10n/utils/translation_resolver.rb

@@ -59,13 +59,12 @@ module Jekyll
       #   # Returns nil (protected element present, prevents block translation)
       #   TranslationResolver.try_block_level_translation(node, "text", translations)
       def self.try_block_level_translation(node, text, translations)
-        return nil unless node.parent && content_element?(node.parent)
+        ancestor = find_content_element_ancestor(node)
+        return nil unless ancestor
 
-        # Don't attempt block-level translation if parent contains protected elements
-        # (script, style, pre). These cannot be safely applied at block level.
-        return nil if contains_protected_elements?(node.parent)
+        return nil if contains_protected_elements?(ancestor)
 
-        block_text = BlockTextExtractor.extract(node.parent)
+        block_text = BlockTextExtractor.extract(ancestor)
         return nil unless block_text && block_text != text
 
         translations[block_text]
@@ -78,6 +77,17 @@ module Jekyll
         HtmlElements::CONTENT_ELEMENTS.include?(node.name)
       end
 
+      def self.find_content_element_ancestor(node)
+        current = node&.parent
+        while current&.element?
+          return current if content_element?(current)
+
+          current = current.parent
+        end
+        nil
+      end
+      private_class_method :find_content_element_ancestor
+
       # Check if an element contains protected child elements that block translations.
       #
       # Protected elements (script, style, pre) cannot have their surrounding text

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: jekyll-l10n
 version: !ruby/object:Gem::Version
-  version: 1.3.14
+  version: 1.3.16
 platform: ruby
 authors:
 - ReleaseBot