coradoc-adoc 2.0.26 → 2.0.28

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6d381a7e511028e9caa79826eccd74e14694968a18825cd179d28cbd785928c7
-  data.tar.gz: 0646f6bf33b86848c345fa772e23583def67169f01ae9b9a905ec0f8fa1922c7
+  metadata.gz: 4af9421f3be892d3f963f70789a03201faf12f95ba487207c3affb9f0dd94ba0
+  data.tar.gz: d1dfcfe239edf6ad09a67bf1dd9a39ac7992466e9aa979475d376b525e0472b5
 SHA512:
-  metadata.gz: 86b395057b440b17616f1743a4d4e72f8c9cffb21b344019b695233c81445450bfd2e17bfc77cb0e706d1284267c4e399680d1174818ebae709570c485c5d435
-  data.tar.gz: 25d8a142605cd3fe9df0dbd191c1004f4d9450ba79614a7b660cd629ef7c8673045143d4b5ed9fd7c0cdf25385d94ec615c2d2591479f41c228abe1fa16c77aa
+  metadata.gz: eb5d3978b3777a0b04fa51cf736fcee2cf4ce6de38bca4353eff5bd4b8cbf7dcce849f84fb60c05c2979f339fa6b182ff74662e491825a939ca29ea5424152c3
+  data.tar.gz: 0fc692a6bd815a8ee102f485c2724b1941fe90bffb09cfdcc066ab6ddf90479386f848fe4ad0ea63f66f8d75b12f40a53ed946a4e0343b3f9d201e860f3d042c

data/lib/coradoc/asciidoc/parser/block_header.rb

@@ -3,19 +3,29 @@
 module Coradoc
   module AsciiDoc
     module Parser
-      # Single Responsibility: parse the optional header that precedes a block.
+      # Single Responsibility: parse the optional header that precedes a
+      # structural element (block or section).
       #
-      # Encapsulates the canonical AsciiDoc block-header grammar in one place
-      # (DRY, MECE, single source of truth). Before this module existed, every
-      # block-like rule (block, table, section, paragraph, block_image) inlined
-      # its own header rule with subtly different slot orderings — and several
-      # of them captured the same Parslet key more than once in a single
-      # sequence, which triggered Parslet's "Duplicate subtrees while merging
-      # result" warning and silently discarded one of the captured values.
+      # Encapsulates the canonical AsciiDoc header grammar in one place
+      # (DRY, MECE, single source of truth). Two flavours live here because
+      # blocks and sections have different header grammars:
       #
-      # Canonical header shape, each component at most once:
+      #   block_header   = block_title? >> element_id? >> attribute_blocks?
+      #   section_header =                    element_id? >> attribute_blocks?
       #
-      #   block_title? >> element_id? >> attribute_blocks?
+      # Asciidoctor does NOT permit `.Title` to apply to a section — sections
+      # accept `[role]` attribute lists and `[[id]]` anchors but not block
+      # titles. The section's heading IS its title. Mixing the two shapes
+      # into one rule caused `section_block` to admit `.Foo` lines that
+      # collided with `section_title`'s `:title` capture, triggering
+      # Parslet's "Duplicate subtrees while merging result … (keys:
+      # [:title])" warning and silently dropping the block title.
+      #
+      # Before this module existed, every block-like rule inlined its own
+      # header rule with subtly different slot orderings — and several
+      # captured the same Parslet key more than once in a single sequence,
+      # which triggered Parslet's "Duplicate subtrees" warning and silently
+      # discarded one of the captured values.
       #
       # `attribute_blocks` accepts one or more consecutive `[...]` attribute
       # lists, captured as a Parslet sequence under the :attribute_list key.
@@ -31,8 +41,11 @@ module Coradoc
       # every attribute and lets the transformer merge them into a single
       # Coradoc::AsciiDoc::Model::AttributeList downstream.
       module BlockHeader
-        # Canonical block header rule. Single canonical order; each of title,
-        # id, and attribute_blocks is optional and matched at most once.
+        # Canonical block header rule. Includes the block title — every
+        # real block (delimited block, table, paragraph, image) accepts
+        # an optional `.Title` line before its body. Single canonical
+        # order; each of title, id, and attribute_blocks is optional and
+        # matched at most once.
         # @return [Parslet::Atoms::Base]
         def block_header
           block_title.maybe >>
@@ -40,6 +53,16 @@ module Coradoc
             attribute_blocks.maybe
         end
 
+        # Section header rule. Asciidoctor does not permit `.Title` on
+        # sections — the section heading itself is the title. Sections
+        # still accept the same element_id and attribute_blocks slots
+        # that blocks do (`[[anchor]]`, `[appendix]`, `[role=x]`, etc.).
+        # @return [Parslet::Atoms::Base]
+        def section_header
+          element_id.maybe >>
+            attribute_blocks.maybe
+        end
+
         # One or more consecutive attribute_list + newline sequences, captured
         # as a Parslet sequence under :attribute_list. When multiple `[...]`
         # blocks precede a delimiter, all of them reach the transformer; when

data/lib/coradoc/asciidoc/parser/inline.rb

@@ -4,6 +4,27 @@ module Coradoc
   module AsciiDoc
     module Parser
       module Inline
+        # AsciiDoc typographic quote syntax: a 2-char pattern that
+        # Asciidoctor substitutes with the corresponding Unicode curly
+        # quote. Single source of truth for the pattern → Unicode char
+        # mapping; the transformer reads this table.
+        #
+        # The patterns MUST be recognised before +monospace_constrained+
+        # in the +inline+ alternation, otherwise the lone backtick in
+        # +`"`+ or +`"``+ fires monospace and the surrounding quote
+        # collapses to straight ASCII quotes wrapped around a spurious
+        # code span.
+        TYPOGRAPHIC_QUOTE_PATTERNS = {
+          '"`' => "“",  # U+201C left double
+          '`"' => "”",  # U+201D right double
+          "'`" => "‘",  # U+2018 left single
+          "`'" => "’"   # U+2019 right single
+        }.freeze
+
+        def typographic_quote
+          (str('"`') | str('`"') | str("'`") | str("`'")).as(:typographic_quote)
+        end
+
         def attribute_reference
           str('{').present? >> str('{') >>
             match('[a-zA-Z0-9_-]').repeat(1).as(:attribute_reference) >>
@@ -170,6 +191,7 @@ module Coradoc
 
         def inline_chars?
           match('[\[*#_{<^~`]').present? |
+            typographic_quote.present? |
             str('http').present? |
             str('https').present? |
             str('link:').present? |
@@ -198,7 +220,8 @@ module Coradoc
         end
 
         def inline
-          bold_unconstrained |
+          typographic_quote |
+            bold_unconstrained |
             bold_constrained |
             span_unconstrained |
             span_constrained |

data/lib/coradoc/asciidoc/parser/list.rb

@@ -20,13 +20,13 @@ module Coradoc
         def ordered_list(nesting_level = 1)
           attrs = (attribute_list >> newline).maybe
           r = olist_item(nesting_level)
-          attrs >> olist_item(nesting_level).present? >> r.repeat(1).as(:ordered)
+          attrs >> (empty_line.repeat(0) >> r).repeat(1).as(:ordered)
         end
 
         def unordered_list(nesting_level = 1)
           attrs = (attribute_list >> newline).maybe
           r = ulist_item(nesting_level)
-          attrs >> r.repeat(1).as(:unordered)
+          attrs >> (empty_line.repeat(0) >> r).repeat(1).as(:unordered)
         end
 
         def definition_list(_delimiter = nil)
@@ -53,7 +53,7 @@ module Coradoc
         def olist_item(nesting_level = 1)
           item = olist_marker(nesting_level).as(:marker) >>
                  match("\n").absent? >> space >>
-                 (text_line(true, unguarded: true) >>
+                 (text_line(false, unguarded: true) >>
                   list_item_continuation_lines).as(:lines)
 
           att = (list_continuation.present? >>
@@ -105,7 +105,7 @@ module Coradoc
           item = ulist_marker(nesting_level).as(:marker) >>
                  str(' [[[').absent? >>
                  match("\n").absent? >> space >>
-                 (text_line(true, unguarded: true) >>
+                 (text_line(false, unguarded: true) >>
                   list_item_continuation_lines).as(:lines)
 
           att = (list_continuation.present? >>
@@ -127,8 +127,15 @@ module Coradoc
         # list marker, block delimiter, attribute list, section, element
         # id, table boundary, list continuation, or list prefix).
         # `line_not_text?` (from Paragraph) is that exact lookahead.
+        #
+        # Each iteration matches exactly one source line via
+        # `text_line(false, ...)` (single-newline termination). The
+        # `.repeat(0)` walks multiple continuation lines one at a time.
+        # Using `text_line(true, ...)` here would let `line_ending.repeat(1)`
+        # greedy-match across blank lines, silently absorbing the
+        # follow-up paragraph into the last list item's content.
         def list_item_continuation_lines
-          (line_not_text? >> text_line(true, unguarded: true)).repeat(0)
+          (line_not_text? >> text_line(false, unguarded: true)).repeat(0)
         end
 
         def dlist_delimiter

data/lib/coradoc/asciidoc/parser/section.rb

@@ -27,7 +27,7 @@ module Coradoc
         def section_block(level = 2)
           return nil if level > 8
 
-          block_header >>
+          section_header >>
             section_title(level).as(:title) >>
             contents.as(:contents).maybe
         end

data/lib/coradoc/asciidoc/transform/element_transformers/document_transformer.rb

@@ -12,6 +12,7 @@ module Coradoc
               children = ToCoreModel.transform(doc.sections || doc.contents || [])
               children = CalloutMerger.call(children)
               children = prepend_frontmatter(children, doc.frontmatter)
+              children = insert_title_heading_after_frontmatter(children, title_text, doc.id)
               children = resolve_attribute_references(children, attributes)
 
               Coradoc::CoreModel::DocumentElement.new(
@@ -33,6 +34,41 @@ module Coradoc
               [block, *children]
             end
 
+            # Per the AsciiDoc spec, the document title (`= Title`) is the
+            # document's level-0 heading. Asciidoctor renders it as an
+            # <h1> at the top of the body by default. Emit a HeaderElement
+            # at level 0 as the first body child (after any FrontmatterBlock)
+            # so consumers that walk the body's children see the title —
+            # the standard ProseMirror/HTML rendering pattern — instead of
+            # having to read DocumentElement.title separately.
+            #
+            # The title attribute on DocumentElement is preserved for
+            # consumers that read it directly. The FromCoreModel Document
+            # transformer consumes the HeaderElement when constructing
+            # Model::Header so AsciiDoc round-trip does not double-render.
+            def insert_title_heading_after_frontmatter(children, title_text, doc_id)
+              return children if title_text.nil? || title_text.strip.empty?
+              return children if children.any? { |c| title_heading?(c) }
+
+              title_id = doc_id || Coradoc::CoreModel::IdGenerator
+                                    .generate_from_title(title_text)
+              title_heading = Coradoc::CoreModel::HeaderElement.new(
+                level: 0,
+                title: title_text,
+                content: title_text,
+                id: title_id
+              )
+              frontmatter_count = children.count do |child|
+                child.is_a?(Coradoc::CoreModel::FrontmatterBlock)
+              end
+              children.insert(frontmatter_count, title_heading)
+            end
+
+            def title_heading?(child)
+              child.is_a?(Coradoc::CoreModel::HeaderElement) &&
+                child.level.to_i.zero?
+            end
+
             # Resolve `{name}` attribute references in the body against the
             # document's own declared attributes. Single source of truth:
             # the resolver lives in core so other format gems can reuse it

data/lib/coradoc/asciidoc/transform/from_core_model.rb

@@ -27,23 +27,24 @@ module Coradoc
           def transform_structural_element(element)
             case element
             when CoreModel::DocumentElement
-              header = if element.title
-                         Coradoc::AsciiDoc::Model::Header.new(
-                           title: Coradoc::AsciiDoc::Model::Title.new(
-                             content: element.title,
-                             level_int: 0
-                           )
-                         )
-                       else
-                         Coradoc::AsciiDoc::Model::Header.new(title: '')
-                       end
-
               without_frontmatter, frontmatter = extract_frontmatter(Array(element.children))
+              without_title_heading, title_text = extract_title_heading(without_frontmatter, element.title)
+
+              header = if title_text
+                          Coradoc::AsciiDoc::Model::Header.new(
+                            title: Coradoc::AsciiDoc::Model::Title.new(
+                              content: title_text,
+                              level_int: 0
+                            )
+                          )
+                        else
+                          Coradoc::AsciiDoc::Model::Header.new(title: '')
+                        end
 
               Coradoc::AsciiDoc::Model::Document.new(
                 id: element.id,
                 header: header,
-                sections: flatten_children(without_frontmatter),
+                sections: flatten_children(without_title_heading),
                 frontmatter: frontmatter
               )
             when CoreModel::SectionElement
@@ -456,6 +457,26 @@ module Coradoc
             [children.drop(1), yaml.nil? || yaml.empty? ? nil : yaml]
           end
 
+          # Pull the level-0 HeaderElement (the document title) out of the
+          # body so the AsciiDoc serializer can render it as `= Title`
+          # via Model::Header instead of double-rendering it as a body
+          # section. Returns [remaining_children, title_text]. Falls
+          # back to +fallback_title+ when no HeaderElement is present
+          # (preserves backward compatibility with callers that build
+          # DocumentElement programmatically and put the title only on
+          # DocumentElement#title rather than in the children list).
+          def extract_title_heading(children, fallback_title)
+            title_idx = children.index do |c|
+              c.is_a?(CoreModel::HeaderElement) && c.level.to_i.zero?
+            end
+            return [children, fallback_title] unless title_idx
+
+            heading = children[title_idx]
+            title_text = heading.title || fallback_title
+            remaining = children[0...title_idx] + children[(title_idx + 1)..]
+            [remaining, title_text]
+          end
+
           def resolve_semantic_type(block)
             semantic = block.resolve_semantic_type
             return semantic if semantic

data/lib/coradoc/asciidoc/transformer/inline_rules.rb

@@ -67,6 +67,19 @@ module Coradoc
               )
             end
 
+            # Typographic quotes — Asciidoctor's curly-quote substitution.
+            # Each 2-char pattern collapses to a single Unicode character
+            # at parse time so downstream consumers (HTML, Markdown) see
+            # the literal typographic char without any post-processing.
+            # The mapping is the single source of truth declared on the
+            # parser side (Parser::Inline::TYPOGRAPHIC_QUOTE_PATTERNS).
+            rule(typographic_quote: simple(:pattern)) do
+              char = Coradoc::AsciiDoc::Parser::Inline::TYPOGRAPHIC_QUOTE_PATTERNS.fetch(
+                pattern.to_s, pattern.to_s
+              )
+              Model::TextElement.new(content: char)
+            end
+
             # Hard line break (` +\n` or `\\n`). Emitted as a dedicated
             # AsciiDoc model (Inline::HardLineBreak) distinct from
             # Model::LineBreak, which only represents paragraph-separator

data/lib/coradoc/asciidoc/version.rb

@@ -2,6 +2,6 @@
 
 module Coradoc
   module AsciiDoc
-    VERSION = '2.0.26'
+    VERSION = '2.0.28'
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: coradoc-adoc
 version: !ruby/object:Gem::Version
-  version: 2.0.26
+  version: 2.0.28
 platform: ruby
 authors:
 - Ribose Inc.