canon 0.3.44 → 0.3.46

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e7b141fefd0f254d40b0ba7621c7400d882f5a0bcee1fc2f96a51424ce4c7c36
-  data.tar.gz: '018e400f0949c85ca2db99aa362664e6434834b135e32efc891b9f4b1c794f46'
+  metadata.gz: 2dc57b150c1a0a2b9179f4b88a78c22aeb1529aa74d30857d8ca71a6159d6639
+  data.tar.gz: 65db8a7cadb8718b11a475858d3bfd5d4adcf0a4c738a7c940a6dfa60a6bfd3e
 SHA512:
-  metadata.gz: 0d25eaf6dfa12649678fb9e770762fbf1ad51923d7da03b168aa36acbb11fe3242edb958069b9738bf097aa0c90188498b3408f1fd900589bc091394c578d394
-  data.tar.gz: e519ee287d5016becf20d31ad057d825d9c6118273305ce713f13c8c0429172283b576fe6a0c477a91cfbda92328b8a4d4f3491b1e6c34e69425d24ff7f12eb1
+  metadata.gz: d411dc92e7978a904f303a95397f301286e9a12a3fc3732bb91919c2564375221b12cb13db64198f41942edeb5bbe2dfc8512b173cbfbdab44d27409d5a009a9
+  data.tar.gz: eaa339eae29ca147aa6ee9a5cecf9a328e69fc9b1f64b3acb742d11f4d451bef89efc11cde012e06e15f21b10ec09f95fbf36e0456c718ba925c015e47dba0ce

data/lib/canon/diff/diff_line_builder.rb

@@ -21,19 +21,22 @@ module Canon
       # @param text1 [String] The first document (preprocessed)
       # @param text2 [String] The second document (preprocessed)
       # @return [Array<DiffLine>] The assembled diff lines
-      def self.build(diff_nodes, text1, text2)
+      def self.build(diff_nodes, text1, text2, lines1: nil, lines2: nil)
         return [] if diff_nodes.nil? || diff_nodes.empty?
         return [] if text1.nil? || text2.nil?
 
-        new(diff_nodes, text1, text2).build
+        new(diff_nodes, text1, text2, lines1, lines2).build
       end
 
-      def initialize(diff_nodes, text1, text2)
+      def initialize(diff_nodes, text1, text2, lines1 = nil, lines2 = nil)
         @diff_nodes = diff_nodes
         @text1 = text1
         @text2 = text2
-        @lines1 = text1.split("\n")
-        @lines2 = text2.split("\n")
+        # The caller may hand in the already-split lines (the
+        # formatter splits these documents too); re-splitting both
+        # documents per pipeline stage was ~2x the line strings.
+        @lines1 = lines1 || text1.split("\n")
+        @lines2 = lines2 || text2.split("\n")
         # Build reverse indices for efficient content lookup in gap handling.
         # Maps content string to array of line indices where that content appears.
         @line_to_indices1 = build_line_index(@lines1)

data/lib/canon/diff/diff_node_enricher.rb

@@ -22,21 +22,22 @@ module Canon
       # @param text1 [String] The first document (preprocessed)
       # @param text2 [String] The second document (preprocessed)
       # @return [Array<DiffNode>] The same DiffNodes, enriched in place
-      def self.build(diff_nodes, text1, text2)
+      def self.build(diff_nodes, text1, text2, lines1: nil, lines2: nil)
         return diff_nodes if diff_nodes.nil? || diff_nodes.empty?
         return diff_nodes if text1.nil? || text2.nil?
 
-        new(diff_nodes, text1, text2).enrich
+        new(diff_nodes, text1, text2, lines1, lines2).enrich
       end
 
-      def initialize(diff_nodes, text1, text2)
+      def initialize(diff_nodes, text1, text2, lines1 = nil, lines2 = nil)
         @diff_nodes = diff_nodes
         @text1 = text1
         @text2 = text2
         @line_map1 = SourceLocator.build_line_map(text1)
         @line_map2 = SourceLocator.build_line_map(text2)
-        @lines1 = text1.split("\n")
-        @lines2 = text2.split("\n")
+        # Shared split — see DiffLineBuilder's note.
+        @lines1 = lines1 || text1.split("\n")
+        @lines2 = lines2 || text2.split("\n")
         # Track occurrences for text_content dimension to find correct element instance
         @text_occurrence1 = Hash.new(0)
         @text_occurrence2 = Hash.new(0)
@@ -927,7 +928,7 @@ module Canon
       # @return [Integer] the last line number
       def find_end_line(start_line, line_map, content)
         newline_count = content.count("\n")
-        [start_line + newline_count, line_map.length - 1].min
+        [start_line + newline_count, line_map.line_count - 1].min
       end
 
       # Find the occurrence of a value at a specific element index.
@@ -1016,7 +1017,7 @@ module Canon
         # so the prefix scan counts exactly the preceding siblings —
         # no "inside the element" correction (count_elements_before_
         # position subtracts one for text-node offsets).
-        opener = /<#{element_name}[>\s]/
+        opener = opener_for(element_name)
         occurrences = SourceLocator.locate_all(value, text, line_map)
         occurrences.each do |occ|
           count = count_elements_before_position_open(text,
@@ -1093,7 +1094,7 @@ module Canon
                                                       line_map)
         return nil unless line_idx
 
-        col = search_start - line_map[line_idx][:start_offset]
+        col = search_start - line_map.start_at(line_idx)
         { char_offset: search_start, line_number: line_idx, col: col }
       end
 
@@ -1265,7 +1266,7 @@ range_start, range_end)
                                                               line_map)
                 return nil unless line_idx
 
-                col = value_pos - line_map[line_idx][:start_offset]
+                col = value_pos - line_map.start_at(line_idx)
                 return { char_offset: value_pos, line_number: line_idx,
                          col: col }
               end
@@ -1280,7 +1281,7 @@ range_start, range_end)
                                                         line_map)
           return nil unless line_idx
 
-          col = value_pos - line_map[line_idx][:start_offset]
+          col = value_pos - line_map.start_at(line_idx)
           return { char_offset: value_pos, line_number: line_idx, col: col }
         end
 
@@ -1332,7 +1333,7 @@ range_start, range_end)
                                                             line_map)
               return nil unless line_idx
 
-              col = value_pos - line_map[line_idx][:start_offset]
+              col = value_pos - line_map.start_at(line_idx)
               return { char_offset: value_pos, line_number: line_idx, col: col }
             end
           end
@@ -1386,7 +1387,7 @@ range_start, range_end)
                                                             line_map)
               return nil unless line_idx
 
-              col = anchor_pos - line_map[line_idx][:start_offset]
+              col = anchor_pos - line_map.start_at(line_idx)
               return { char_offset: anchor_pos, line_number: line_idx,
                        col: col }
             else
@@ -1395,7 +1396,7 @@ range_start, range_end)
                                                             line_map)
               return nil unless line_idx
 
-              col = tag_end_pos - line_map[line_idx][:start_offset]
+              col = tag_end_pos - line_map.start_at(line_idx)
               return { char_offset: tag_end_pos, line_number: line_idx,
                        col: col }
             end
@@ -1447,6 +1448,15 @@ range_start, range_end)
       # @param char_offset [Integer] character offset to check before
       # @param element_name [String] name of element to count
       # @return [Integer] element index (0-based) of the element containing the position
+      # Opener regexes memoized per element name — compiling
+      # /<name[>\s]/ per call was an allocation per occurrence check.
+      OPENER_CACHE = {}.compare_by_identity
+      private_constant :OPENER_CACHE
+
+      def opener_for(element_name)
+        OPENER_CACHE[element_name] ||= /<#{element_name}[>\s]/
+      end
+
       # Occurrence count with offsets AT the element's own opening
       # tag (no inside-the-element correction — see
       # locate_element_at_index).

data/lib/canon/diff/source_locator.rb

@@ -15,23 +15,51 @@ module Canon
     #   SourceLocator.locate("line2", "line1\nline2\nline3", line_map)
     #   # => { char_offset: 6, line_number: 1, col: 0 }
     class SourceLocator
+      # Line offset map: two flat Integer arrays (starts, ends) — a
+      # hash-per-line was one allocation per document line on every
+      # enrichment, the presentation stage's quiet constant.
+      class LineMap
+        attr_reader :starts, :ends
+
+        def initialize(starts, ends)
+          @starts = starts
+          @ends = ends
+        end
+
+        def empty?
+          @starts.empty?
+        end
+
+        def line_count
+          @starts.length
+        end
+
+        def start_at(index)
+          @starts[index]
+        end
+
+        def end_at(index)
+          @ends[index]
+        end
+      end
+
       # Build a line offset map from source text.
-      # Each entry records the start and end character offset of a line.
       #
       # @param text [String] the full source text
-      # @return [Array<Hash>] array of { start_offset:, end_offset: } hashes,
-      #   one per line (0-indexed)
+      # @return [LineMap] flat offset arrays, one entry per line
+      #   (0-indexed)
       def self.build_line_map(text)
-        return [] if text.nil? || text.empty?
+        return LineMap.new([], []) if text.nil? || text.empty?
 
-        map = []
+        starts = []
         offset = 0
         text.each_line do |line|
-          line_end = offset + line.length
-          map << { start_offset: offset, end_offset: line_end }
-          offset = line_end
+          starts << offset
+          offset += line.length
         end
-        map
+        ends = starts[1..] || []
+        ends << text.length
+        LineMap.new(starts, ends)
       end
 
       # Locate a substring within source text and return its position.
@@ -55,7 +83,7 @@ module Canon
         line_idx = find_line_for_offset(char_offset, line_map)
         return nil if line_idx.nil?
 
-        col = char_offset - line_map[line_idx][:start_offset]
+        col = char_offset - line_map.start_at(line_idx)
 
         { char_offset: char_offset, line_number: line_idx, col: col }
       end
@@ -77,7 +105,7 @@ module Canon
           line_idx = find_line_for_offset(pos, line_map)
           break if line_idx.nil?
 
-          col = pos - line_map[line_idx][:start_offset]
+          col = pos - line_map.start_at(line_idx)
           results << { char_offset: pos, line_number: line_idx, col: col }
           offset = pos + 1
         end
@@ -92,8 +120,8 @@ module Canon
         # @param line_map [Array<Hash>] the line offset map
         # @return [Integer, nil] the 0-based line index, or nil
         def find_line_for_offset(char_offset, line_map)
-          line_map.bsearch_index do |entry|
-            entry[:end_offset] > char_offset
+          line_map.ends.bsearch_index do |end_offset|
+            end_offset > char_offset
           end
         end
       end

data/lib/canon/diff_formatter/by_line/xml_formatter.rb

@@ -44,12 +44,22 @@ module Canon
           # Compute line number width BEFORE formatting
           compute_line_num_width(doc1, doc2)
 
+          # The pipeline splits both documents once here; every stage
+          # downstream reuses the arrays (three redundant full-doc
+          # splits per render before).
+          @pipeline_lines1 = doc1.split("\n")
+          @pipeline_lines2 = doc2.split("\n")
+
           # Phase 1: Enrich DiffNodes with character positions
-          Canon::Diff::DiffNodeEnricher.build(@differences, doc1, doc2)
+          Canon::Diff::DiffNodeEnricher.build(@differences, doc1, doc2,
+                                              lines1: @pipeline_lines1,
+                                              lines2: @pipeline_lines2)
 
           # Phase 2: Assemble DiffLines from enriched DiffNodes
           diff_lines = Canon::Diff::DiffLineBuilder.build(@differences, doc1,
-                                                          doc2)
+                                                          doc2,
+                                                          lines1: @pipeline_lines1,
+                                                          lines2: @pipeline_lines2)
 
           # Layers 3-5: Build report through pipeline
           report = Canon::Diff::DiffReportBuilder.build(
@@ -178,8 +188,8 @@ module Canon
         def format_report(report, doc1, doc2)
           return "" if report.contexts.empty?
 
-          lines1 = doc1.split("\n")
-          lines2 = doc2.split("\n")
+          lines1 = @pipeline_lines1 || doc1.split("\n")
+          lines2 = @pipeline_lines2 || doc2.split("\n")
 
           output = []
 

data/lib/canon/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Canon
-  VERSION = "0.3.44"
+  VERSION = "0.3.46"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: canon
 version: !ruby/object:Gem::Version
-  version: 0.3.44
+  version: 0.3.46
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-09-15 00:00:00.000000000 Z
+date: 2026-09-16 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: diff-lcs