統合検索ページの module function 表示を ?. に統一しハイライトも修理する

lib/bitclust/search_index_generator.rb

@@ -92,11 +92,25 @@ def to_js(db, fdb = nil)
     # sorts after "3.4"), and entry order is the first appearance while
     # scanning versions in ascending order — independent of the caller's
     # argument order, so the generated file diffs stay stable.
+    #
+    # A module function's full_name is spelled ".#" by pre-4.0 databases and
+    # "?." by 4.0+ ones (bitclust#250), which would split it into two rows —
+    # confusing on a page that lists every version side by side. Fold the
+    # display to the current "?." spelling here, so both spellings collapse
+    # into one row with a unified versions list (name/type/path/match_name
+    # are spelling-independent already). Only entries carrying a match_name
+    # (dot-qualified methods) are folded: a prose heading whose *text*
+    # happens to contain ".#" must keep matching its page verbatim. The
+    # single-version in-page search keeps each version's own spelling —
+    # build_index is not affected. See the report in bitclust#279.
     def self.merge(version_indexes)
       merged = {} #: Hash[Array[String?], merged_entry]
       sorted = version_indexes.sort_by { |version, _| Gem::Version.new(version) }
       sorted.each do |version, index|
         index.each do |e|
+          if e[:match_name] && e[:full_name].include?('.#')
+            e = e.merge(full_name: e[:full_name].gsub('.#', '?.'))
+          end
           key = e.values_at(:name, :full_name, :type, :path)
           entry = (merged[key] ||= e.merge(versions: []))
           entry[:versions] << version

test/js/test_search.mjs

@@ -166,6 +166,30 @@ const index = [
          'search("each") still finds the ordinary method entry')
 }
 
+// --- wire up search_page.js (the cross-version /ja/search/ page) ---------
+// Like setupSearchBox, but for search_page.js: fresh (unwrapped) ranker
+// globals first, then the page script's own globals (a merged index whose
+// entries carry `versions`, the version list, and the path prefix), plus
+// the window shims search_page.js touches (location/history/focus).
+function setupSearchPage(index) {
+  loadRankerScripts()
+  const input = makeElement('input')
+  input.focus = function() {}
+  const result = makeElement('ul')
+  result.parentNode = makeElement('div')
+  globalThis.document = makeDocument({ 'search-field': input, 'search-results': result })
+  globalThis.search_data = { index }
+  globalThis.search_versions = ['3.4', '4.0']
+  globalThis.search_version_base = '../'
+  globalThis.window = {
+    location: { search: '', pathname: '/ja/search/' },
+    history: { replaceState() {} },
+  }
+  ;(0, eval)(std.loadFile(jsdir + 'search_page.js'))
+  document.dispatch('DOMContentLoaded')
+  return { input, result }
+}
+
 // End-to-end through search_init.js's DOM wiring: typing "defined?" and
 // dispatching keyup renders a result item carrying the new search-type-
 // heading badge, labeled via the formatType map added in this change.
@@ -237,9 +261,8 @@ const index = [
 // End-to-end through the DOM, same as the "defined?" test above: typing a
 // dot-qualified query renders a result whose *displayed* title is the
 // original "File.open" -- proving the match_name substitution never reaches
-// the page (highlightMatch degrades to no inline <em> marks for this query
-// shape, a known/documented limitation -- see the #279 report -- but the
-// text itself must still be exactly right and HTML-escaped normally).
+// the page -- highlighted whole by the highlightMatch wrap (see the
+// dedicated section below) and HTML-escaped normally.
 {
   const { input, result } = setupSearchBox(qualifiedIndex)
   input.value = 'File.open'
@@ -250,6 +273,88 @@ const index = [
          'the rendered result displays the literal "File.open" full_name')
   assert(!html.includes('File::open'), 'the rendered result never shows the "::" match_name form')
   assert(html.includes('method/-file/s/open.html'), 'the result links to the right path')
+  assert(html.includes('<em>File.open</em>'),
+         'the qualified query is highlighted as one contiguous <em> run (bitclust#279 comment)')
+}
+
+// --- bitclust#279 (comment): highlightMatch on qualified queries ---------
+// The vendored highlightMatch() compares q.normalized (with "." already
+// rewritten to "::", e.g. "file::open") against the literal-dot full_name
+// ("File.open"), so a dot-qualified query never earned an <em> highlight --
+// at best the fuzzy fallback gave up at the first ":". search_init.js wraps
+// it: whenever the vendored code found nothing to mark, retry with the
+// user's original query text, treating ".#" and "?." as the same
+// module-function spelling. Both marks are two characters long, so indexes
+// into the folded strings map 1:1 onto the displayed text.
+{
+  const q = parseQuery('File.open')
+  assert(highlightMatch('File.open', q) === '\u0001File.open\u0002',
+         'highlightMatch marks the whole contiguous match for "File.open"')
+}
+{
+  const q = parseQuery('file.o')
+  assert(highlightMatch('File.open', q) === '\u0001File.o\u0002pen',
+         'a partial qualified query highlights just its prefix, case-insensitively')
+}
+{
+  const q = parseQuery('Kernel?.open')
+  assert(highlightMatch('Kernel.#open', q) === '\u0001Kernel.#open\u0002',
+         'a "?."-spelled query highlights a ".#"-spelled entry (cross-notation)')
+  assert(highlightMatch('Kernel?.open', q) === '\u0001Kernel?.open\u0002',
+         'a "?."-spelled query highlights a "?."-spelled entry')
+}
+{
+  const q = parseQuery('Kernel.#open')
+  assert(highlightMatch('Kernel?.open', q) === '\u0001Kernel?.open\u0002',
+         'a ".#"-spelled query highlights a "?."-spelled entry (cross-notation)')
+}
+// Vendored behavior is preserved wherever it already worked: unqualified
+// queries keep the vendored contiguous highlight...
+{
+  const q = parseQuery('open')
+  assert(highlightMatch('File.open', q) === 'File.\u0001open\u0002',
+         'unqualified queries keep the vendored contiguous highlight')
+}
+// ...fuzzy scattered highlights pass through untouched...
+{
+  const q = parseQuery('fopen')
+  const r = highlightMatch('File.open', q)
+  assert(r.indexOf('\u0001') !== -1 && r.replace(/[\u0001\u0002]/g, '') === 'File.open',
+         'fuzzy highlights from the vendored code pass through unchanged')
+}
+// ...and a query that matches nothing still returns the text verbatim.
+{
+  const q = parseQuery('Dir.open')
+  assert(highlightMatch('File.open', q) === 'File.open',
+         'a non-matching qualified query leaves the text unhighlighted')
+}
+
+// --- bitclust#279 (comment): the cross-version page (search_page.js) -----
+// Same wraps as search_init.js, exercised end-to-end. The merged index only
+// carries the "?." spelling (SearchIndexGenerator.merge folds ".#" away);
+// an old-notation query must still find that entry, and the rendered row
+// shows the "?."-spelled title highlighted, linking to the newest version
+// with the full version list underneath.
+{
+  const mergedIndex = [
+    { name: 'open', full_name: 'Kernel?.open', match_name: 'Kernel::#open',
+      type: 'class_method', path: 'method/-kernel/m/open.html',
+      versions: ['3.4', '4.0'] },
+    { name: 'open', full_name: 'File.open', match_name: 'File::open',
+      type: 'class_method', path: 'method/-file/s/open.html',
+      versions: ['3.4', '4.0'] },
+  ]
+  const { input, result } = setupSearchPage(mergedIndex)
+  input.value = 'Kernel.#open'
+  input.dispatch('keyup', { key: 'n' })
+  assert(result.children.length === 1,
+         'a ".#"-spelled query finds the "?."-only merged entry on the cross-version page')
+  const html = result.children.length ? result.children[0].innerHTML : ''
+  assert(html.includes('<em>Kernel?.open</em>'),
+         '...rendered with the whole "?."-spelled title highlighted')
+  assert(html.includes('../4.0/method/-kernel/m/open.html'),
+         '...linking to the newest version')
+  assert(html.includes('>3.4<'), '...with the older version listed too')
 }
 
 // --- bitclust#279: defensive fallback ------------------------------------
@@ -266,8 +371,10 @@ const index = [
 {
   parseQuery = undefined
   computeScore = undefined
-  assert(typeof parseQuery !== 'function' && typeof computeScore !== 'function',
-         '(fallback setup) parseQuery/computeScore are hidden, simulating a future closure-based ranker')
+  highlightMatch = undefined
+  assert(typeof parseQuery !== 'function' && typeof computeScore !== 'function' &&
+         typeof highlightMatch !== 'function',
+         '(fallback setup) parseQuery/computeScore/highlightMatch are hidden, simulating a future closure-based ranker')
 
   // A closure-hiding stand-in ranker: same public shape as SearchRanker, but
   // its matching logic never touches globalThis.

test/test_search_index_generator.rb

@@ -416,6 +416,44 @@ def test_merge_orders_entries_by_first_appearance_in_version_order
     assert_equal %w[B A], merged.map { |e| e[:full_name] }
   end
 
+  def test_merge_folds_module_function_display_to_question_dot
+    # bitclust#279 コメント対応: 統合ページ(/ja/search/)で同じ module
+    # function が "Kernel.#open"(4.0 より前の版)と "Kernel?.open"
+    # (4.0 以降)の2行に分かれて出るのはわかりにくい。表示は現行表記の
+    # "?." に畳んで1エントリに合流させる(name/type/path/match_name は
+    # 表記に依らず同一なので、full_name さえ畳めばキーが一致する)。
+    # 単一版ページの表示はその版の表記のまま(build_index 側は触らない)。
+    pre40  = { name: 'open', full_name: 'Kernel.#open', type: 'class_method',
+               path: 'method/-kernel/m/open.html', match_name: 'Kernel::#open' }
+    post40 = { name: 'open', full_name: 'Kernel?.open', type: 'class_method',
+               path: 'method/-kernel/m/open.html', match_name: 'Kernel::#open' }
+    merged = BitClust::SearchIndexGenerator.merge(
+      [['3.4', [pre40]], ['4.0', [post40]]])
+    assert_equal 1, merged.size
+    assert_equal 'Kernel?.open', merged[0][:full_name]
+    assert_equal %w[3.4 4.0], merged[0][:versions]
+    assert_equal 'Kernel::#open', merged[0][:match_name]
+  end
+
+  def test_merge_folds_display_even_when_only_old_notation_exists
+    # 4.0 以降に存在しないまま消えたメソッドでも、統合ページの表示は
+    # "?." に揃える(znz さんの依頼は「?. だけ出てくるように」)
+    pre40 = { name: 'gone', full_name: 'Kernel.#gone', type: 'class_method',
+              path: 'method/-kernel/m/gone.html', match_name: 'Kernel::#gone' }
+    merged = BitClust::SearchIndexGenerator.merge([['3.4', [pre40]]])
+    assert_equal 'Kernel?.gone', merged[0][:full_name]
+  end
+
+  def test_merge_keeps_dot_hash_in_entries_without_match_name
+    # ".#" を字面に含んでいても match_name を持たないエントリ(散文
+    # 見出しなど)は表示を畳まない — 見出しはページ本文の字面と一致
+    # していることに意味がある
+    heading = { name: '.#', full_name: '.# (プログラム・文・式)', type: 'heading',
+                path: 'doc/symref.html#dot-hash' }
+    merged = BitClust::SearchIndexGenerator.merge([['3.4', [heading]]])
+    assert_equal '.# (プログラム・文・式)', merged[0][:full_name]
+  end
+
   def test_merged_js_format
     js = BitClust::SearchIndexGenerator.merged_js([['3.4', [entry]]])
     assert_match(/\Avar search_data = \{/, js)

test/test_searchpage_command.rb

@@ -86,6 +86,35 @@ def test_merged_index_versions
     assert_equal %w[3.4 4.1], klass['versions']
   end
 
+  def test_merged_index_shows_module_functions_as_question_dot_only
+    # bitclust#279 コメント対応: 版によって表示が ".#"(4.0 より前)と
+    # "?."(4.0 以降)に割れる module function は、統合ページでは "?."
+    # 表記の1エントリに合流させる(両表記が併記されるとわかりにくい)
+    db33 = build_db('3.3', <<~'RD')
+      description
+
+      = module Kernel
+      == Module Functions
+      --- at_exit{ ... } -> Proc
+      aaa
+    RD
+    db40 = build_db('4.0', <<~'RD')
+      description
+
+      = module Kernel
+      == Module Functions
+      --- at_exit{ ... } -> Proc
+      aaa
+    RD
+    run_command(["--outputdir=#{@out}", db40, db33])
+    js = File.read("#{@out}/js/search_data.js")
+    data = JSON.parse(js.sub(/\Avar search_data = /, '').sub(/;\z/, ''))
+    entries = data['index'].select { |e| e['name'] == 'at_exit' }
+    assert_equal ['Kernel?.at_exit'], entries.map { |e| e['full_name'] }
+    assert_equal %w[3.3 4.0], entries[0]['versions']
+    assert_not_match(/Kernel\.#/, js)
+  end
+
   def test_index_html_embeds_versions_and_ui
     run_command(["--outputdir=#{@out}", @db41, @db34])
     html = File.read("#{@out}/index.html")

theme/default/js/search_init.js

@@ -67,6 +67,35 @@
     };
   }
 
+  // bitclust#279 (continued): highlightMatch() compares q.normalized (with
+  // "." already rewritten to "::" and "?." folded to ".#" by the wrap
+  // above) against the displayed full_name, which keeps its literal
+  // "."/".#"/"?." — so the dot-qualified queries fixed above matched their
+  // entry but never earned an <em> highlight (at best the vendored fuzzy
+  // fallback gave up at the first ":"). Whenever the vendored code found
+  // nothing to mark, retry a contiguous match with the user's *original*
+  // query text (q.original, restored by the parseQuery wrap), treating
+  // ".#" and "?." as the same module-function spelling. Both marks are two
+  // characters long, so indexes into the folded strings map 1:1 onto the
+  // displayed text and the markers can be spliced straight into it.
+  if (typeof highlightMatch === 'function') {
+    var alikiHighlightMatch = highlightMatch;
+    highlightMatch = function(text, q) {
+      var marked = alikiHighlightMatch(text, q);
+      if (marked !== text) return marked;         // the vendored code managed
+      if (!text || !q || !q.original) return marked;
+      var canonText = text.toLowerCase().replace(/\.#/g, '?.');
+      var canonQuery = q.original.toLowerCase().replace(/\.#/g, '?.');
+      if (!canonQuery) return marked;
+      var start = canonText.indexOf(canonQuery);
+      if (start === -1) return marked;
+      var end = start + canonQuery.length;
+      return text.substring(0, start) +
+        '\u0001' + text.substring(start, end) + '\u0002' +
+        text.substring(end);
+    };
+  }
+
   function createSearchInstance(input, result) {
     if (!input || !result) return null;
 

theme/default/js/search_page.js

@@ -57,6 +57,33 @@
     };
   }
 
+  // Same highlightMatch wrap as search_init.js: when the vendored code
+  // found nothing to mark (dot-qualified queries never contiguous-match a
+  // literal-dot full_name once "." has been rewritten to "::"), retry with
+  // the user's original query text, treating ".#" and "?." as the same
+  // module-function spelling. On this page the merged index only carries
+  // the "?." spelling (SearchIndexGenerator.merge folds ".#" away), so
+  // this is what lets a query typed in the pre-4.0 ".#" notation still
+  // light up the "?."-spelled row it matched. Both marks are two
+  // characters, so indexes into the folded strings map 1:1 onto the text.
+  if (typeof highlightMatch === 'function') {
+    var alikiHighlightMatch = highlightMatch;
+    highlightMatch = function(text, q) {
+      var marked = alikiHighlightMatch(text, q);
+      if (marked !== text) return marked;         // the vendored code managed
+      if (!text || !q || !q.original) return marked;
+      var canonText = text.toLowerCase().replace(/\.#/g, '?.');
+      var canonQuery = q.original.toLowerCase().replace(/\.#/g, '?.');
+      if (!canonQuery) return marked;
+      var start = canonText.indexOf(canonQuery);
+      if (start === -1) return marked;
+      var end = start + canonQuery.length;
+      return text.substring(0, start) +
+        '\u0001' + text.substring(start, end) + '\u0002' +
+        text.substring(end);
+    };
+  }
+
   function versionHref(version, path) {
     return search_version_base + version + '/' + path;
   }