canon 0.3.48 → 0.3.50

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: fcfa53c5436e38941e6c789536bf9c268fb5b587930aeaca6020c22993def0a2
-  data.tar.gz: a814221d4dd934e35f8f6eb1169f8694fadf2e2fd60e49a9f8f61e6355ff9853
+  metadata.gz: 6bbfa0c65294f7a3ad7d77b678dab6fc1b9f5c0fc67bf55faba3371e6cfc6614
+  data.tar.gz: 28b3106753c227ad6973a2698ede0315d3e8812feed6af1cbccdf5cf08636184
 SHA512:
-  metadata.gz: 1ac2772108ab1d9cae92a479a98c94cbc02bff2a431a64e3dc9052ff4df5ec595a38c81849c3d9ddc1c1329b6a86bdfbb7d3aeae6b5bbacfa2d6d8d283acc0a6
-  data.tar.gz: 7794b9a9392eaa18bb43b85ec7a2e830d3c0f172bfd827e13b7e3f6290c82cb6b3491312090f8230671c1a7442e722a7c1b3d66a44398965b89ab4664ec11a4d
+  metadata.gz: 9d3c4a330ec9224eff5a8c84bd677ea1fc72500b06099c188b3c7b30e6b096d3897f8308a40fc003067887dd4f45986464c8786b8bee8096614035dc8189c8ea
+  data.tar.gz: 84549e0eb08a383bf14934b8b22f58e3ad4aeac8aaaeb56ce80c87d1127df9d3ebc7d2f17763cf6cfd1dda65702e90d2379e0ba5d64193f250bd1f409d292aa3

data/lib/canon/diff/diff_node_enricher.rb

@@ -41,6 +41,11 @@ module Canon
         # Track occurrences for text_content dimension to find correct element instance
         @text_occurrence1 = Hash.new(0)
         @text_occurrence2 = Hash.new(0)
+        # Opener-offset caches: one scan per unique element name per
+        # document replaces the per-call regex walk (which was also
+        # quadratic on repeated element names).
+        @opener_offsets1 = {}
+        @opener_offsets2 = {}
       end
 
       def enrich
@@ -1461,27 +1466,44 @@ range_start, range_end)
       # tag (no inside-the-element correction — see
       # locate_element_at_index).
       def count_elements_before_position_open(text, char_offset, opener)
-        count = 0
-        pos = 0
-        while (hit = text.index(opener, pos)) && hit < char_offset
-          count += 1
-          pos = hit + 1
-        end
-        count
+        count_openers_before(opener_offsets(text, opener), char_offset)
       end
 
       def count_elements_before_position(text, char_offset, element_name)
-        # Index loop instead of copying the prefix per call — the
-        # copy was O(offset) on every occurrence check.
-        opener = /<#{element_name}[>\s]/
-        count = 0
+        # Subtract 1 because the count includes the element we are inside
+        [count_openers_before(opener_offsets(text, opener_for(element_name)),
+                              char_offset) - 1, 0].max
+      end
+
+      # Count of opener offsets strictly before char_offset — binary
+      # search over the cached, ascending offset array.
+      def count_openers_before(offsets, char_offset)
+        idx = offsets.bsearch_index { |offset| offset >= char_offset }
+        idx || offsets.length
+      end
+
+      # All opener offsets for one opener regex in one document,
+      # cached per text (one linear scan per unique element name per
+      # document — callers previously re-scanned per occurrence
+      # check, which was also quadratic on repeated element names).
+      def opener_offsets(text, opener)
+        if text.equal?(@text1)
+          (@opener_offsets1[opener.source] ||= scan_opener_offsets(text, opener))
+        elsif text.equal?(@text2)
+          (@opener_offsets2[opener.source] ||= scan_opener_offsets(text, opener))
+        else
+          scan_opener_offsets(text, opener)
+        end
+      end
+
+      def scan_opener_offsets(text, opener)
+        offsets = []
         pos = 0
-        while (hit = text.index(opener, pos)) && hit < char_offset
-          count += 1
+        while (hit = text.index(opener, pos))
+          offsets << hit
           pos = hit + 1
         end
-        # Subtract 1 because the count includes the element we are inside
-        [count - 1, 0].max
+        offsets
       end
     end
   end

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

@@ -118,6 +118,13 @@ module Canon
           after = diff_node.serialized_after&.split("\n") || []
           fmt = diff_node.formatting?
           lines = []
+          header = parent_chain(diff_node.path)
+          if header
+            lines << Canon::Diff::DiffLine.new(
+              line_number: nil, new_position: nil,
+              content: header, type: :location_header
+            )
+          end
           require "diff/lcs" unless defined?(::Diff::LCS)
 
           ::Diff::LCS.sdiff(before, after).each do |change|
@@ -156,6 +163,27 @@ module Canon
           lines
         end
 
+        # Display chain of a DiffNode's ancestor elements, e.g.
+        # "bibitem > formattedref" — gives the blank-numbered
+        # unlocated lines a readable home (#86).
+        def parent_chain(path)
+          return nil if path.nil?
+
+          segments = path.split("/").reject(&:empty?)
+          return nil if segments.length < 2
+
+          chain = segments[0...-1].filter_map do |seg|
+            seg[/\A(?:\{[^}]+\})?([a-zA-Z0-9_:-]+)/, 1]
+          end
+          # Comparator paths repeat the element name before the node
+          # segment ("formattedref/formattedref/text()") — collapse
+          # consecutive repeats for display.
+          collapsed = chain.each_with_object([]) do |name, acc|
+            acc << name unless acc.last == name
+          end
+          collapsed.join(" > ")
+        end
+
         # The nearest preceding node (in @differences walk order)
         # that the enricher DID locate.
         def preceding_located(diff_node)
@@ -225,6 +253,9 @@ module Canon
               new_num = (diff_line.new_position || diff_line.line_number)&.+(1)
               output << format_unified_line(old_num, new_num, " ",
                                             diff_line.content)
+            when :location_header
+              output << format_location_header(diff_line.content)
+
             when :removed
               line_num = diff_line.line_number&.+(1)
               formatting = diff_line.formatting?
@@ -992,6 +1023,18 @@ module Canon
           end
         end
 
+        # Dimmed section header for unlocated-node contexts (#86):
+        # "⋯ in bibitem > formattedref"
+        def format_location_header(chain)
+          blank = " " * @line_num_width
+          text = "    ⋯ in #{chain}"
+          if @use_color
+            "#{blank}|#{blank}  | #{colorize(text, :light_black)}"
+          else
+            "#{blank}|#{blank}  | #{text}"
+          end
+        end
+
         def format_unified_line(old_num, new_num, marker, content, color = nil,
 informative: false, formatting: false)
           old_str = old_num ? "%#{@line_num_width}d" % old_num : " " * @line_num_width

data/lib/canon/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Canon
-  VERSION = "0.3.48"
+  VERSION = "0.3.50"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: canon
 version: !ruby/object:Gem::Version
-  version: 0.3.48
+  version: 0.3.50
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-09-16 00:00:00.000000000 Z
+date: 2026-09-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: diff-lcs