isodoc 3.4.7 → 3.4.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6bf206774db883d4bf837cfbdccbf9d620d76f2252313d5ba0c81f5979ff40bb
-  data.tar.gz: d0c3dd48ec168524fef9799ce3455520d1d578b4e667cbb06a05f70e181c6d67
+  metadata.gz: e0934e0c0ea24dd7c91fe682c72855d5cd95c0dbf13bdedf43af86412c304885
+  data.tar.gz: 16ee08583e192dafd2d30ec886ef9eefd1703cefbeeefec2a22f042485721d2f
 SHA512:
-  metadata.gz: 9f6b1ab7e0be186198065832292a83fe32f33ce6ea4c6fe2faae9307bb5e86c03fe3b1e7e67412dd832af7316528e7d798680715a621583112f1486ec32484da
-  data.tar.gz: 56b8fb76537ea2b2ddec457558b905249bd8a699fb5a774eeb77c75e5366d00ff53d5977a7e5131b567785244c667083f46c3d7dd45e9fe5e70384bbf7f41305
+  metadata.gz: 2cfd83b6f9acc5496e2155cb52b4a4fc9e55d11ef8ba7b54405ed741de74e88b21cfd67b90def1e4515f397291c4eb6b3d7c133a8ff9b647c97b9246be041dd1
+  data.tar.gz: b873292c72723ae4359bf3aee7dafaf039f1d6bdc60df110c179b8150274e41c737122a6980a02819b7ca3b332a2e2d9d997e875822dc0ba4b90dd61f1d109e8

data/lib/isodoc/base_style/reset.scss

@@ -114,7 +114,7 @@ blockquote, q {
   }
 }
 
-  .h2Annex {
+.h2Annex {
   font-family: $headerfont;
 }
 

data/lib/isodoc/css.rb

@@ -75,31 +75,29 @@ module IsoDoc
     end
 
     def convert_scss(filename, stylesheet, stripwordcss)
-      load_scss_paths(filename)
+      load_paths = scss_load_paths(filename)
       Dir.mktmpdir do |dir|
         variables_file_path = File.join(dir, "variables.scss")
         File.write(variables_file_path, scss_fontheader(stripwordcss))
-        SassC.load_paths << dir
         modified_stylesheet = %( @use "variables" as *;\n#{stylesheet})
-        compile_scss(modified_stylesheet)
+        compile_scss(modified_stylesheet, load_paths + [dir])
       end
     end
 
-    def compile_scss(modified_stylesheet)
+    def compile_scss(modified_stylesheet, load_paths = [])
       SassC::Engine
         .new(modified_stylesheet, quiet_deps: true, syntax: :scss,
+                                  load_paths: load_paths,
                                   importer: SasscImporter)
         .render.gsub(/__WORD__/, "")
     end
 
-    def load_scss_paths(filename)
+    def scss_load_paths(filename)
       require "sassc-embedded"
       require "isodoc/sassc_importer"
       [File.join(Gem.loaded_specs["isodoc"].full_gem_path,
                  "lib", "isodoc"),
-       File.dirname(filename)].each do |name|
-        SassC.load_paths << name
-      end
+       File.dirname(filename)]
     end
 
     # stripwordcss if HTML stylesheet, !stripwordcss if DOC stylesheet

data/lib/isodoc/css_border_parser.rb

@@ -90,11 +90,17 @@ module IsoDoc
       # Extract rule sets from CSS string
       def extract_rule_sets(css_string)
         rule_sets = {}
-        css_string = css_string.gsub(/\/\*.*?\*\//m, "") # Remove comments
-        css_string.scan(/([^{]+)\{([^}]+)\}/m) do |selector, declarations|
+        # Safe C-style comment removal without polynomial backtracking:
+        # /\*[^*]*(?:\*+[^*/][^*]*)*\*+\/ matches /* ... */ in linear time
+        # by keeping the "star runs not followed by /" separate from non-star
+        # content, so no position can be matched by multiple paths.
+        css_string = css_string.gsub(%r{/\*[^*]*(?:\*+[^*/][^*]*)*\*+/}, "")
+        # [^{}]+ for selector and [^{}]* for declarations: both exclude braces,
+        # making the grouping unambiguous without the m flag.
+        css_string.scan(/([^{}]+)\{([^{}]*)\}/) do |selector, declarations|
           selector = selector.strip
           declarations_hash = {} # Extract declarations
-          declarations.scan(/([^:;]+):([^;]+);?/) do |property, value|
+          declarations.scan(/([^:;]+):([^;]*);?/) do |property, value|
             declarations_hash[property.strip] = value.strip
           end
           rule_sets[selector] = declarations_hash
@@ -149,8 +155,10 @@ module IsoDoc
           "color" => DEFAULT_VALUES["color"],
         }
         value.split(/\s+/).each do |part|
-          if width?(part) then components["width"] = part
-          elsif style?(part) then components["style"] = part
+          if width?(part)
+            components["width"] = part
+          elsif style?(part)
+            components["style"] = part
           elsif color?(part)
             components["color"] = convert_color_to_rgb(part)
           end

data/lib/isodoc/function/blocks.rb

@@ -150,7 +150,7 @@ module IsoDoc
         classtype = nil
         classtype = "MsoCommentText" if in_comment
         node["type"] == "floating-title" and
-          classtype = "h#{node['depth']}"
+          classtype = "h#{node['depth'] || '1'}"
         classtype ||= node["class"]
         classtype
       end
@@ -180,13 +180,19 @@ module IsoDoc
       end
 
       def quote_parse(node, out)
-        attrs = para_attrs(node)
-        attrs[:class] = "Quote"
+        attrs = quote_attrs(node)
         out.div(**attr_code(attrs)) do |p|
           node.children.each { |n| parse(n, p) unless n.name == "source" }
         end
       end
 
+      def quote_attrs(node)
+        ret = para_attrs(node).merge(class: "Quote")
+        node["type"] == "newcontent" and
+          ret[:class] += " AmendNewcontent"
+        ret
+      end
+
       def passthrough_parse(node, out)
         node["formats"] &&
           !(node["formats"].split.include? @format.to_s) and return

data/lib/isodoc/function/cleanup.rb

@@ -8,8 +8,8 @@ module IsoDoc
       def passthrough_cleanup(docxml)
         docxml.split(%r{(<passthrough>|</passthrough>)}).each_slice(4)
           .map do |a|
-            a.size > 2 and a[2] = HTMLEntities.new.decode(a[2])
-            [a[0], a[2]]
+          a.size > 2 and a[2] = HTMLEntities.new.decode(a[2])
+          [a[0], a[2]]
         end.join
       end
 
@@ -84,8 +84,11 @@ module IsoDoc
       end
 
       def remove_bottom_border(cell)
+        # [^;]* (not +): the preceding property name is the unambiguous
+        # delimiter, so zero-or-more is equivalent and avoids polynomial
+        # backtracking on the value portion.
         cell["style"] =
-          cell["style"].gsub(/border-bottom:[^;]+;/, "border-bottom:0pt;")
+          cell["style"].gsub(/border-bottom:[^;]*;/, "border-bottom:0pt;")
       end
 
       def table_get_or_make_tfoot(table)
@@ -112,7 +115,7 @@ module IsoDoc
       end
 
       TABLENOTE_CSS = "div[@class = 'Note' or @class = 'BlockSource' " \
-        "or @class = 'TableFootnote' or @class = 'figdl' or @class = 'key']"
+                      "or @class = 'TableFootnote' or @class = 'figdl' or @class = 'key']"
         .freeze
 
       def table_note_cleanup(docxml)
@@ -121,7 +124,7 @@ module IsoDoc
           insert_here = new_fullcolspan_row(t, tfoot)
           t.xpath("dl | p[@class = 'ListTitle'] | #{TABLENOTE_CSS}")
             .each do |d|
-              d.parent = insert_here
+            d.parent = insert_here
           end
         end
       end

data/lib/isodoc/function/lists.rb

@@ -14,6 +14,7 @@ module IsoDoc
 
       def ul_parse(node, out)
         out.div(**attr_code(class: "ul_wrap")) do |div|
+          n = node.at(ns("./fmt-ul")) and return fmt_ul_parse(node, n, out)
           list_title_parse(node, div)
           div.ul(**attr_code(ul_attrs(node))) do |ul|
             node.children.each { |n| n.name == "fmt-name" or parse(n, ul) }
@@ -21,6 +22,13 @@ module IsoDoc
         end
       end
 
+      def fmt_ul_parse(elem, fmt_ul, out)
+        children_parse(fmt_ul, out)
+        output_elem = out.parent.elements.last
+        output_elem.parent["id"] = elem["id"]
+        s = keep_style(elem) and output_elem.parent["style"] = s
+      end
+
       OL_STYLE = {
         arabic: "1",
         roman: "i",
@@ -38,11 +46,13 @@ module IsoDoc
         { # type: node["type"] ? ol_style(node["type"].to_sym) : ol_depth(node),
           type: ol_style(node["type"]&.to_sym),
           start: node["start"],
-          id: node["id"], style: keep_style(node) }
+          id: node["id"], style: keep_style(node)
+        }
       end
 
       def ol_parse(node, out)
         out.div(**attr_code(class: "ol_wrap")) do |div|
+          n = node.at(ns("./fmt-ol")) and return fmt_ul_parse(node, n, out)
           list_title_parse(node, div)
           div.ol(**attr_code(ol_attrs(node))) do |ol|
             node.children.each { |n| n.name == "fmt-name" or parse(n, ol) }
@@ -53,10 +63,10 @@ module IsoDoc
       def li_checkbox(node)
         if node["uncheckedcheckbox"] == "true"
           '<span class="zzMoveToFollowing">' \
-                '<input type="checkbox" checked="checked"/></span>'
+            '<input type="checkbox" checked="checked"/></span>'
         elsif node["checkedcheckbox"] == "true"
           '<span class="zzMoveToFollowing">' \
-                '<input type="checkbox"/></span>'
+            '<input type="checkbox"/></span>'
         else ""
         end
       end

data/lib/isodoc/function/section_titles.rb

@@ -70,7 +70,8 @@ module IsoDoc
       def clause_name(_node, title, div, header_class)
         header_class = {} if header_class.nil?
         div.h1(**attr_code(header_class)) do |h1|
-          if title.is_a?(String) then h1 << title
+          if title.is_a?(String)
+            h1 << title
           elsif title
             children_parse(title, h1)
             clause_parse_subtitle(title, h1)
@@ -88,6 +89,7 @@ module IsoDoc
       end
 
       def variant_title(node, out)
+        # node["type"] == "toc" and return # used only for ToC!
         out.p(**attr_code(style: "display:none;",
                           class: "variant-title-#{node['type']}")) do |p|
           children_parse(node, p)

data/lib/isodoc/function/utils.rb

@@ -1,4 +1,5 @@
 require "metanorma-utils"
+require_relative "utils_img"
 
 module IsoDoc
   module Function
@@ -61,7 +62,7 @@ module IsoDoc
       end
 
       def to_xhtml_prep(xml)
-        xml.gsub!(/<\?xml[^<>]*>/, "")
+        xml.gsub!(/\A<\?xml[^<>]*>\n?/, "")
         xml.include?("<!DOCTYPE ") || (xml = DOCTYPE_HDR + xml)
         numeric_escapes(xml)
       end
@@ -117,9 +118,12 @@ module IsoDoc
       end
 
       def header_strip(hdr)
+        # \s[^<>]* rather than \s[^<>]+: \s already guarantees one whitespace
+        # char after <p, so [^<>]* (zero or more) suffices for any attributes,
+        # and removes the polynomial interaction between \s and [^<>]+.
         h1 = to_xhtml_fragment(hdr.to_s.gsub(%r{<br\s*/>}, " ")
-          .gsub(%r{</?p(\s[^<>]+)?>}, "")
-          .gsub(/<\/?h[123456][^<>]*>/, "").gsub(/<\/?b[^<>]*>/, "").dup)
+          .gsub(%r{</?p(\s[^<>]*)?>}, "")
+          .gsub(/<\/?h[123456][^<>]*>/, "").dup)
         h1.traverse do |x|
           if x.name == "span" && x["style"]&.include?("mso-tab-count")
             x.replace(" ")
@@ -162,41 +166,6 @@ module IsoDoc
           .gsub("&lt;", "&#x3c;").gsub("&gt;", "&#x3e;").gsub("&amp;", "&#x26;")
       end
 
-      def save_dataimage(uri, _relative_dir = true)
-        %r{^data:(?<imgclass>image|application)/(?<imgtype>[^;]+);(?:charset=[^;]+;)?base64,(?<imgdata>.+)$} =~ uri
-        imgtype = "emf" if emf?("#{imgclass}/#{imgtype}")
-        imgtype = imgtype.sub(/\+[a-z0-9]+$/, "") # svg+xml
-        imgtype = "png" unless /^[a-z0-9]+$/.match? imgtype
-        imgtype == "postscript" and imgtype = "eps"
-        Tempfile.open(["image", ".#{imgtype}"],
-                      mode: File::BINARY | File::SHARE_DELETE) do |f|
-          f.binmode
-          f.write(Base64.strict_decode64(imgdata))
-          @tempfile_cache << f # persist to the end
-          f.path
-        end
-      end
-
-      def save_svg(img)
-        Tempfile.open(["image", ".svg"],
-                      mode: File::BINARY | File::SHARE_DELETE) do |f|
-          f.write(img.to_xml)
-          @tempfile_cache << f # persist to the end
-          f.path
-        end
-      end
-
-      def image_localfile(img)
-        img.name == "svg" && !img["src"] and
-          return save_svg(img)
-        case img["src"]
-        when /^data:/ then save_dataimage(img["src"], false)
-        when %r{^([A-Z]:)?/} then img["src"]
-        when nil then nil
-        else File.join(@localdir, img["src"])
-        end
-      end
-
       LABELLED_ANCESTOR_ELEMENTS =
         %w(example requirement recommendation permission
            note table figure sourcecode).freeze
@@ -206,15 +175,6 @@ module IsoDoc
           .intersection(LABELLED_ANCESTOR_ELEMENTS - exceptions).empty?
       end
 
-      def emf?(type)
-        %w(application/emf application/x-emf image/x-emf image/x-mgx-emf
-           application/x-msmetafile image/x-xbitmap image/emf).include? type
-      end
-
-      def eps?(type)
-        %w(application/postscript image/x-eps).include? type
-      end
-
       def cleanup_entities(text, is_xml: true)
         c = HTMLEntities.new
         if is_xml
@@ -234,10 +194,6 @@ module IsoDoc
         path[/\s/] ? "\"#{path}\"" : path
       end
 
-      def imgfile_suffix(uri, suffix)
-        "#{File.join(File.dirname(uri), File.basename(uri, '.*'))}.#{suffix}"
-      end
-
       # Unescape &#x26; to & in href attributes only
       # This ensures URLs work correctly while preserving &#x26; in text content
       # This operates on the final string output after all Nokogiri processing
@@ -254,6 +210,34 @@ module IsoDoc
           end
         end
       end
+
+      # parse CSV-encoded key=value attribute
+      COMMA_PLACEHOLDER = "##COMMA##".freeze
+
+      # Temporarily replace commas inside quotes with a placeholder
+      def comma_placeholder(options)
+        processed = ""
+        in_quotes = false
+        options.each_char do |c|
+          c == "'" and in_quotes = !in_quotes
+          processed << if c == "," && in_quotes
+                         COMMA_PLACEHOLDER
+                       else c
+                       end
+        end
+        processed
+      end
+
+      def csv_attribute_extract(options)
+        options.gsub!(/([a-z_]+)='/, %('\\1=))
+        processed = comma_placeholder(options)
+        CSV.parse_line(processed, quote_char: "'")
+          .each_with_object({}) do |x, acc|
+          x.gsub!(COMMA_PLACEHOLDER, ",")
+          m = /^(.+?)=(.+)?$/.match(x) or next
+          acc[m[1].to_sym] = m[2].sub(/^(["'])(.+)\1$/, "\\2")
+        end
+      end
     end
   end
 end

data/lib/isodoc/function/utils_img.rb

@@ -0,0 +1,55 @@
+require "metanorma-utils"
+
+module IsoDoc
+  module Function
+    module Utils
+      def save_dataimage(uri, _relative_dir = true)
+        %r{^data:(?<imgclass>image|application)/(?<imgtype>[^;]+);(?:charset=[^;]+;)?base64,(?<imgdata>.+)$} =~ uri
+        imgtype = "emf" if emf?("#{imgclass}/#{imgtype}")
+        imgtype = imgtype.sub(/\+[a-z0-9]+$/, "") # svg+xml
+        imgtype = "png" unless /^[a-z0-9]+$/.match? imgtype
+        imgtype == "postscript" and imgtype = "eps"
+        Tempfile.open(["image", ".#{imgtype}"],
+                      mode: File::BINARY | File::SHARE_DELETE) do |f|
+          f.binmode
+          f.write(Base64.strict_decode64(imgdata))
+          @tempfile_cache << f # persist to the end
+          f.path
+        end
+      end
+
+      def save_svg(img)
+        Tempfile.open(["image", ".svg"],
+                      mode: File::BINARY | File::SHARE_DELETE) do |f|
+          f.write(img.to_xml)
+          @tempfile_cache << f # persist to the end
+          f.path
+        end
+      end
+
+      def image_localfile(img)
+        img.name == "svg" && !img["src"] and
+          return save_svg(img)
+        case img["src"]
+        when /^data:/ then save_dataimage(img["src"], false)
+        when %r{^([A-Z]:)?/} then img["src"]
+        when nil then nil
+        else File.join(@localdir, img["src"])
+        end
+      end
+
+      def emf?(type)
+        %w(application/emf application/x-emf image/x-emf image/x-mgx-emf
+           application/x-msmetafile image/x-xbitmap image/emf).include? type
+      end
+
+      def eps?(type)
+        %w(application/postscript image/x-eps).include? type
+      end
+
+      def imgfile_suffix(uri, suffix)
+        "#{File.join(File.dirname(uri), File.basename(uri, '.*'))}.#{suffix}"
+      end
+    end
+  end
+end

data/lib/isodoc/gem_tasks.rb

@@ -111,15 +111,18 @@ module IsoDoc
     end
 
     def compile_scss(filename)
-      load_scss_paths(filename)
+      load_paths = load_scss_paths(filename)
       Dir.mktmpdir do |dir|
         File.write(File.join(dir, "variables.scss"), fonts_placeholder)
-        SassC.load_paths << dir
         sheet_content = File.read(filename, encoding: "UTF-8")
-          .gsub(%r<([a-z])\.([0-9])(?=[^{}]*{)>m, "\\1.__WORD__\\2")
+          # Drop the m flag and use [^{}\n]* so the lookahead is bounded to
+          # one line, preventing O(n^2) scanning across multi-line inputs.
+          .gsub(/([a-z])\.([0-9])(?=[^{}\n]*{)/, "\\1.__WORD__\\2")
         SassC::Engine.new(%<@use "variables" as *;\n#{sheet_content}>,
-                          syntax: :scss, quiet_deps: true, importer: SasscImporter)
-          .render.gsub(/__WORD__/, "")
+                          syntax: :scss, quiet_deps: true,
+                          load_paths: load_paths + [dir],
+                          importer: SasscImporter)
+          .render.gsub("__WORD__", "")
       end
     end
 
@@ -131,10 +134,7 @@ module IsoDoc
                                 "lib", "isodoc")
                     else File.join("lib", "isodoc")
                     end
-      [isodoc_path,
-       File.dirname(filename)].each do |name|
-         SassC.load_paths << name
-       end
+      [isodoc_path, File.dirname(filename)]
     end
 
     def compile_scss_task(current_task)

data/lib/isodoc/html_function/postprocess.rb

@@ -24,8 +24,8 @@ module IsoDoc
       end
 
       def html5(doc)
-        doc.sub(%r{<!DOCTYPE html [^<>]+>}, "<!DOCTYPE html>")
-          .sub(%r{<\?xml[^<>]+>}, "")
+        doc.sub(%r{\A<!DOCTYPE html [^<>]+>}, "<!DOCTYPE html>")
+          .sub(%r{\A<\?xml[^<>]+>}, "")
       end
 
       def html_cleanup(html)

data/lib/isodoc/html_function/postprocess_cover.rb

@@ -6,14 +6,17 @@ module IsoDoc
   module HtmlFunction
     module Html
       def script_cdata(result)
-        result.gsub(%r{<script([^<>]*)>\s*<!\[CDATA\[}m, "<script\\1>")
-          .gsub(%r{\]\]>\s*</script>}, "</script>")
-          .gsub(%r{<!\[CDATA\[\s*<script([^<>]*)>}m, "<script\\1>")
-          .gsub(%r{</script>\s*\]\]>}, "</script>")
-          .gsub(%r{<style([^<>]*)>\s*<!\[CDATA\[}m, "<style\\1>")
-          .gsub(%r{\]\]>\s*</style>}, "</style>")
-          .gsub(%r{<!\[CDATA\[\s*<style([^<>]*)>}m, "<style\\1>")
-          .gsub(%r{</style>\s*\]\]>}, "</style>")
+        # Use explicit [ \t\n\r]* rather than \s* with the m flag:
+        # [^<>]* and [ \t\n\r]* are disjoint character classes (whitespace
+        # chars are not < or >), preventing polynomial backtracking.
+        result.gsub(%r{<script([^<>]*)>[ \t\n\r]*<!\[CDATA\[}, "<script\\1>")
+          .gsub(%r{\]\]>[ \t\n\r]*</script>}, "</script>")
+          .gsub(%r{<!\[CDATA\[[ \t\n\r]*<script([^<>]*)>}, "<script\\1>")
+          .gsub(%r{</script>[ \t\n\r]*\]\]>}, "</script>")
+          .gsub(%r{<style([^<>]*)>[ \t\n\r]*<!\[CDATA\[}, "<style\\1>")
+          .gsub(%r{\]\]>[ \t\n\r]*</style>}, "</style>")
+          .gsub(%r{<!\[CDATA\[[ \t\n\r]*<style([^<>]*)>}, "<style\\1>")
+          .gsub(%r{</style>[ \t\n\r]*\]\]>}, "</style>")
       end
 
       def htmlstylesheet(file)
@@ -31,7 +34,7 @@ module IsoDoc
         head << Nokogiri::HTML.fragment("<style>#{htmlstylesheet(@htmlstylesheet)}</style>")
         s = htmlstylesheet(@htmlstylesheet_override) and head << Nokogiri::HTML.fragment("<style>#{s}</style>")
         s = @meta.get[:code_css] and
-          head << Nokogiri::HTML.fragment("<style>#{s.gsub(/sourcecode/,
+          head << Nokogiri::HTML.fragment("<style>#{s.gsub('sourcecode',
                                                            'pre.sourcecode')}</style>")
         @bare and
           head << "<style>body {margin-left: 2em; margin-right: 2em;}</style>"

data/lib/isodoc/presentation_function/annex.rb

@@ -0,0 +1,97 @@
+module IsoDoc
+  class PresentationXMLConvert < ::IsoDoc::Convert
+    def annex(docxml)
+      docxml.xpath(ns("//annex")).each do |f|
+        @xrefs.klass.single_term_clause?(f) and single_term_clause_retitle(f)
+        annex1(f)
+        @xrefs.klass.single_term_clause?(f) and single_term_clause_unnest(f)
+      end
+      @xrefs.parse_inclusions(clauses: true).parse(docxml)
+    end
+
+    def annex1(elem)
+      lbl = @xrefs.anchor(elem["id"], :label, false)
+      orig_title = annex1_title_fmt(elem)
+      %w(title variant-title).each do |k|
+        k == "variant-title" && (t = elem.at(ns("./variant-title"))) &&
+          t["type"] != "_toc_temp" and next
+        # we only prefix variant-title we have just copied from title
+        annex1_title_prefix(elem, lbl, k)
+      end
+      annex1_title_postprocess(elem, orig_title)
+    end
+
+    def annex1_title_fmt(elem)
+      t = elem.at(ns("./title")) or return nil
+      unless elem.at(ns("./variant-title[@type='toc']"))
+        t.next = to_xml(t)
+        v = t.next
+        v.name = "variant-title"
+        v["type"] = "_toc_temp"
+      end
+      orig_title = to_xml(t.children)
+      t.children = annex1_title_fmt_inline(t)
+      orig_title
+    end
+
+    def annex1_title_fmt_inline(title)
+      "<strong>#{to_xml(title.children)}</strong>"
+    end
+
+    def annex1_title_prefix(elem, lbl, tag)
+      delim, lbl = annex1_title_prefix_prep(elem, lbl, tag)
+      if unnumbered_clause?(elem)
+        prefix_name(elem, {}, nil, tag)
+      else
+        prefix_name(elem, { caption: delim }, lbl, tag,
+                    fmt_xref_label: tag == "title")
+      end
+    end
+
+    def annex1_title_prefix_prep(elem, lbl, tag)
+      delim = annex_delim_override(elem)
+      if tag == "variant-title" # strip obligation, boldface
+        lbl &&= lbl.gsub("<strong>", "").gsub("</strong>", "")
+          # Use (?:<br/>)* (non-capturing) and [^<]* instead of .+? so the
+          # obligation-span content is matched without polynomial backtracking.
+          # fmt-obligation spans contain plain text, so [^<]* is equivalent.
+          .sub(%r{(?:<br/>)*<span class=.fmt-obligation.>[^<]*</span>}, "")
+        delim = "<tab/>"
+      end
+      [delim, lbl]
+    end
+
+    # only accept variant-title we have just copied from title and prefixed
+    # into fmt-variant-title: that replaces the variant-title we have inserted
+    def annex1_title_postprocess(elem, orig_title)
+      v, t, v1 = annex1_variant_title_postprocess_prep(elem)
+      t&.children = orig_title
+      v1&.name == "fmt-variant-title" or return
+      v&.remove
+      v1.name = "variant-title"
+      v1["type"] = "toc"
+      s = v1.at(ns(".//semx[@element = 'variant-title']")) or return
+      s["element"] = "title"
+      t and s["source"] = t["id"]
+    end
+
+    def annex1_variant_title_postprocess_prep(elem)
+      v = elem.at(ns("./variant-title[@type='_toc_temp']"))
+      t = elem.at(ns("./title"))
+      # fmt-variant-title generated from variant-title
+      v1 = v&.next_element ||
+        #  fmt-variant-title generated only from lbl
+        elem.at(ns("./fmt-variant-title"))
+      [v, t, v1]
+    end
+
+    def annex_delim_override(elem)
+      m = elem.document.root.at(ns("//presentation-metadata/annex-delim"))
+      m ? to_xml(m.children) : annex_delim(elem)
+    end
+
+    def annex_delim(_elem)
+      "<br/><br/>"
+    end
+  end
+end

data/lib/isodoc/presentation_function/autonum.rb

@@ -1,10 +1,10 @@
 module IsoDoc
   class PresentationXMLConvert < ::IsoDoc::Convert
-    def prefix_name(node, delims, label, elem)
+    def prefix_name(node, delims, label, elem, fmt_xref_label: true)
       sem_xml_descendant?(node) and return
       label, delims = prefix_name_defaults(node, delims, label, elem)
       name, ins, ids, number = prefix_name_prep(node, elem)
-      ins.next = fmt_xref_label(label, number, ids)
+      fmt_xref_label and ins.next = fmt_xref_label(label, number, ids)
       # autonum can be empty, e.g single note in clause: "NOTE []"
       number and node["autonum"] = number.gsub(/<[^>]+>/, "")
       !node.at(ns("./fmt-#{elem}")) &&
@@ -40,6 +40,7 @@ module IsoDoc
       node.at(ns("./sentinel"))&.remove
       strip_duplicate_ids(node, node.at(ns("./#{elem}")),
                           node.at(ns("./fmt-#{elem}")))
+      node.at(ns("./fmt-#{elem}"))
     end
 
     def transfer_id(old, new)
@@ -67,7 +68,7 @@ module IsoDoc
     # remove ids duplicated between sem_title and pres_title
     # index terms are assumed transferred to pres_title from sem_title
     def strip_duplicate_ids(_node, sem_title, pres_title)
-      sem_title && pres_title or return
+      (sem_title && pres_title) or return
       ids = gather_all_ids(pres_title)
       sem_title.xpath(".//*[@id]").each do |x|
         ids.include?(x["id"]) or next
@@ -78,7 +79,7 @@ module IsoDoc
     end
 
     def semx(node, label, element = "autonum")
-      id = node["id"] || node[:id] || elem["original-id"]
+      id = node["id"] || node[:id] || node["original-id"]
       /<semx element='[^']+' source='#{id}'/.match?(label) and return label
       l = stripsemx(label)
       %(<semx element='#{element}' source='#{id}'>#{l}</semx>)
@@ -202,7 +203,7 @@ module IsoDoc
     def sem_xml_descendant_inline?(_node, ancestor_names)
       %w[xref eref origin link name title newcontent]
         .any? do |name|
-          ancestor_names.include?(name)
+        ancestor_names.include?(name)
       end and return true
     end
 
@@ -213,7 +214,7 @@ module IsoDoc
       block?(node) and return false
       %w[preferred admitted deprecated related definition source]
         .any? do |name|
-          ancestor_names.include?(name)
+        ancestor_names.include?(name)
       end
     end