Preserve inline styling inside tidy link labels

Author: st0012

lib/rdoc/markup/to_html.rb

@@ -158,19 +158,16 @@ def handle_regexp_RDOCLINK(target)
   def handle_regexp_TIDYLINK(target)
     text = target.text
 
-    return text unless
-      text =~ /^\{(.*)\}\[(.*?)\]$/ or text =~ /^(\S+)\[(.*?)\]$/
-
-    label = $1
-    url   = CGI.escapeHTML($2)
+    if tidy_link_capturing?
+      return finish_tidy_link(text)
+    end
 
-    if /^rdoc-image:/ =~ label
-      label = handle_RDOCLINK(label)
-    else
-      label = CGI.escapeHTML(label)
+    if text.start_with?('{') && !text.include?('}')
+      start_tidy_link text
+      return ''
     end
 
-    gen_url url, label
+    convert_complete_tidy_link(text)
   end
 
   # :section: Visitor
@@ -458,4 +455,156 @@ def to_html(item)
     super convert_flow @am.flow item
   end
 
+  private
+
+  def convert_flow(flow)
+    res = []
+
+    flow.each do |item|
+      case item
+      when String then
+        append_flow_fragment res, convert_string(item)
+      when RDoc::Markup::AttrChanger then
+        off_tags res, item
+        on_tags  res, item
+      when RDoc::Markup::RegexpHandling then
+        append_flow_fragment res, convert_regexp_handling(item)
+      else
+        raise "Unknown flow element: #{item.inspect}"
+      end
+    end
+
+    res.join
+  end
+
+  def append_flow_fragment(res, fragment)
+    return if fragment.nil? || fragment.empty?
+
+    emit_tidy_link_fragment(res, fragment)
+  end
+
+  def append_to_tidy_label(fragment)
+    @tidy_link_buffer << fragment
+  end
+
+  ##
+  # Matches an entire tidy link with a braced label "{label}[url]".
+  #
+  # Capture 1: label contents.
+  # Capture 2: URL text.
+  # Capture 3: trailing content.
+  TIDY_LINK_WITH_BRACES = /\A\{(.*?)\}\[(.*?)\](.*)\z/
+
+  ##
+  # Matches the tail of a braced tidy link when the opening brace was
+  # consumed earlier while accumulating the label text.
+  #
+  # Capture 1: remaining label content.
+  # Capture 2: URL text.
+  # Capture 3: trailing content.
+  TIDY_LINK_WITH_BRACES_TAIL = /\A(.*?)\}\[(.*?)\](.*)\z/
+
+  ##
+  # Matches a tidy link with a single-word label "label[url]".
+  #
+  # Capture 1: the single-word label (no whitespace).
+  # Capture 2: URL text between the brackets.
+  TIDY_LINK_SINGLE_WORD = /\A(\S+)\[(.*?)\](.*)\z/
+
+  def convert_complete_tidy_link(text)
+    return text unless
+      text =~ TIDY_LINK_WITH_BRACES or text =~ TIDY_LINK_SINGLE_WORD
+
+    label = $1
+    url   = CGI.escapeHTML($2)
+
+    label_html = if /^rdoc-image:/ =~ label
+                   handle_RDOCLINK(label)
+                 else
+                   render_tidy_link_label(label)
+                 end
+
+    gen_url url, label_html
+  end
+
+  def emit_tidy_link_fragment(res, fragment)
+    if tidy_link_capturing?
+      append_to_tidy_label fragment
+    else
+      res << fragment
+    end
+  end
+
+  def finish_tidy_link(text)
+    label_tail, url, trailing = extract_tidy_link_parts(text)
+
+    append_to_tidy_label CGI.escapeHTML(label_tail) unless label_tail.empty?
+
+    return '' unless url
+
+    label_html = @tidy_link_buffer
+
+    @tidy_link_buffer = nil
+
+    link = gen_url(url, label_html)
+
+    return link if trailing.empty?
+
+    link + CGI.escapeHTML(trailing)
+  end
+
+  def extract_tidy_link_parts(text)
+    if text =~ TIDY_LINK_WITH_BRACES
+      [$1, CGI.escapeHTML($2), $3]
+    elsif text =~ TIDY_LINK_WITH_BRACES_TAIL
+      [$1, CGI.escapeHTML($2), $3]
+    elsif text =~ TIDY_LINK_SINGLE_WORD
+      [$1, CGI.escapeHTML($2), $3]
+    else
+      [text, nil, '']
+    end
+  end
+
+  def on_tags(res, item)
+    each_attr_tag(item.turn_on) do |tag|
+      emit_tidy_link_fragment(res, annotate(tag.on))
+      @in_tt += 1 if tt? tag
+    end
+  end
+
+  def off_tags(res, item)
+    each_attr_tag(item.turn_off, true) do |tag|
+      emit_tidy_link_fragment(res, annotate(tag.off))
+      @in_tt -= 1 if tt? tag
+    end
+  end
+
+  def start_tidy_link(text)
+    @tidy_link_buffer = String.new
+    append_to_tidy_label CGI.escapeHTML(text.delete_prefix('{'))
+  end
+
+  def tidy_link_capturing?
+    !!@tidy_link_buffer
+  end
+
+  def render_tidy_link_label(label)
+    RDoc::Markup::LinkLabelToHtml.render(label, @options, @from_path)
+  end
+end
+
+##
+# Formatter dedicated to rendering tidy link labels without mutating the
+# calling formatter's state.
+
+class RDoc::Markup::LinkLabelToHtml < RDoc::Markup::ToHtml
+  def self.render(label, options, from_path)
+    new(options, from_path).to_html(label)
+  end
+
+  def initialize(options, from_path = nil)
+    super(options)
+
+    self.from_path = from_path if from_path
+  end
 end

lib/rdoc/markup/to_html_crossref.rb

@@ -184,38 +184,36 @@ def link(name, text, code = true, rdoc_ref: false)
     end
   end
 
-  def convert_flow(flow)
+  def convert_flow(flow, &block)
     res = []
 
     i = 0
     while i < flow.size
       item = flow[i]
       i += 1
       case item
-      when RDoc::Markup::AttrChanger then
+      when RDoc::Markup::AttrChanger
         # Make "+Class#method+" a cross reference
-        if tt_tag?(item.turn_on) and
-          String === (str = flow[i]) and
-          RDoc::Markup::AttrChanger === flow[i+1] and
-          tt_tag?(flow[i+1].turn_off, true) and
-          (@options.hyperlink_all ? ALL_CROSSREF_REGEXP : CROSSREF_REGEXP).match?(str) and
-          (text = cross_reference str) != str
-        then
-          text = yield text, res if defined?(yield)
-          res << text
-          i += 2
-          next
+        if tt_tag?(item.turn_on) && String === (str = flow[i]) && RDoc::Markup::AttrChanger === flow[i+1] &&
+          tt_tag?(flow[i+1].turn_off, true) && (@options.hyperlink_all ? ALL_CROSSREF_REGEXP : CROSSREF_REGEXP).match?(str)
+
+          unless tidy_link_capturing? || (text = cross_reference(str)) == str
+            text = block.call(text, res) if block
+            append_flow_fragment res, text
+            i += 2
+            next
+          end
         end
         off_tags res, item
         on_tags res, item
-      when String then
+      when String
         text = convert_string(item)
-        text = yield text, res if defined?(yield)
-        res << text
-      when RDoc::Markup::RegexpHandling then
+        text = block.call(text, res) if block
+        append_flow_fragment res, text
+      when RDoc::Markup::RegexpHandling
         text = convert_regexp_handling(item)
-        text = yield text, res if defined?(yield)
-        res << text
+        text = block.call(text, res) if block
+        append_flow_fragment res, text
       else
         raise "Unknown flow element: #{item.inspect}"
       end

test/rdoc/rdoc_markdown_test.rb

@@ -1263,6 +1263,26 @@ def test_gfm_table_with_backslashes_in_code_spans
     assert_equal expected, doc
   end
 
+  def test_markdown_link_with_styled_label
+    markdown = <<~MD
+      [Link to Foo](https://example.com)
+
+      [Link to `Foo`](https://example.com)
+
+      [Link to **Foo**](https://example.com)
+
+      [Link to `Foo` and `\Bar` and `Baz`](https://example.com)
+    MD
+
+    doc = parse markdown
+    html = @to_html.convert doc
+
+    assert_includes html, '<a href="https://example.com">Link to Foo</a>'
+    assert_includes html, '<a href="https://example.com">Link to <code>Foo</code></a>'
+    assert_includes html, '<a href="https://example.com">Link to <strong>Foo</strong></a>'
+    assert_includes html, '<a href="https://example.com">Link to <code>Foo</code> and <code>Bar</code> and <code>Baz</code></a>'
+  end
+
   def parse(text)
     @parser.parse text
   end

test/rdoc/rdoc_markup_to_html_crossref_test.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 require_relative 'xref_test_case'
+require 'rdoc/markdown'
 
 class RDocMarkupToHtmlCrossrefTest < XrefTestCase
 
@@ -270,6 +271,14 @@ def test_handle_regexp_TIDYLINK_label
                  link, 'C1#m@foo'
   end
 
+  def test_convert_TIDYLINK_markdown_with_crossrefs
+    markdown = RDoc::Markdown.parse('[Link to `C1` and `Foo` and `\\C1` and `Bar`](https://example.com)')
+
+    result = markdown.accept(@to)
+
+    assert_equal para('<a href="https://example.com">Link to <code>C1</code> and <code>Foo</code> and <code>\\C1</code> and <code>Bar</code></a>'), result
+  end
+
   def test_to_html_CROSSREF_email
     @options.hyperlink_all = false
 

test/rdoc/rdoc_markup_to_html_test.rb

@@ -734,6 +734,28 @@ def test_convert_TIDYLINK_multiple
     assert_equal expected, result
   end
 
+  def test_convert_TIDYLINK_with_code_label
+    result = @to.convert '{Link to +Foo+}[https://example.com]'
+
+    expected = "\n<p><a href=\"https://example.com\">Link to <code>Foo</code></a></p>\n"
+
+    assert_equal expected, result
+
+    result = @to.convert '{Link to +Foo+ and +Bar+ and +Baz+}[https://example.com]'
+
+    expected = "\n<p><a href=\"https://example.com\">Link to <code>Foo</code> and <code>Bar</code> and <code>Baz</code></a></p>\n"
+
+    assert_equal expected, result
+  end
+
+  def test_convert_TIDYLINK_with_bold_label
+    result = @to.convert '{Link to *Foo*}[https://example.com]'
+
+    expected = "\n<p><a href=\"https://example.com\">Link to <strong>Foo</strong></a></p>\n"
+
+    assert_equal expected, result
+  end
+
   def test_convert_TIDYLINK_image
     result =
       @to.convert '{rdoc-image:path/to/image.jpg}[http://example.com]'