resolves #4306 disallow the use of the left square bracket in an attribute list on formatted text (PR #4307)

4 files changed, 28 insertions, 20 deletions

diff --git a/CHANGELOG.adoc b/CHANGELOG.adoc
index 707a6fb1..0fec8b7f 100644
--- a/CHANGELOG.adoc
+++ b/CHANGELOG.adoc
@@ -38,6 +38,7 @@ Compliance::
 
   * Don't promote level-0 special section at start of document to document title (#4151)
   * Disallow the use of dot (`.`) in the name of a named element attribute (#4147)
+  * Disallow the use of the left square bracket in an attribute list on formatted text (#4306)
 
 Improvements::
 
diff --git a/lib/asciidoctor.rb b/lib/asciidoctor.rb
index 581c8647..45c28f6a 100644
--- a/lib/asciidoctor.rb
+++ b/lib/asciidoctor.rb
@@ -436,6 +436,8 @@ module Asciidoctor
     CC_WORD = CG_WORD = '\p{Word}'
   end
 
+  QuoteAttributeListRxt = %(\\[([^\\[\\]]+)\\])
+
   QUOTE_SUBS = {}.tap do |accum|
     # unconstrained quotes:: can appear anywhere
     # constrained quotes:: must be bordered by non-word characters
@@ -443,46 +445,46 @@ module Asciidoctor
     # the order in which they are replaced is important
     accum[false] = normal = [
       # **strong**
-      [:strong, :unconstrained, /\\?(?:\[([^\]]+)\])?\*\*(#{CC_ALL}+?)\*\*/m],
+      [:strong, :unconstrained, /\\?(?:#{QuoteAttributeListRxt})?\*\*(#{CC_ALL}+?)\*\*/m],
       # *strong*
-      [:strong, :constrained, /(^|[^#{CC_WORD};:}])(?:\[([^\]]+)\])?\*(\S|\S#{CC_ALL}*?\S)\*(?!#{CG_WORD})/m],
+      [:strong, :constrained, /(^|[^#{CC_WORD};:}])(?:#{QuoteAttributeListRxt})?\*(\S|\S#{CC_ALL}*?\S)\*(?!#{CG_WORD})/m],
       # "`double-quoted`"
-      [:double, :constrained, /(^|[^#{CC_WORD};:}])(?:\[([^\]]+)\])?"`(\S|\S#{CC_ALL}*?\S)`"(?!#{CG_WORD})/m],
+      [:double, :constrained, /(^|[^#{CC_WORD};:}])(?:#{QuoteAttributeListRxt})?"`(\S|\S#{CC_ALL}*?\S)`"(?!#{CG_WORD})/m],
       # '`single-quoted`'
-      [:single, :constrained, /(^|[^#{CC_WORD};:`}])(?:\[([^\]]+)\])?'`(\S|\S#{CC_ALL}*?\S)`'(?!#{CG_WORD})/m],
+      [:single, :constrained, /(^|[^#{CC_WORD};:`}])(?:#{QuoteAttributeListRxt})?'`(\S|\S#{CC_ALL}*?\S)`'(?!#{CG_WORD})/m],
       # ``monospaced``
-      [:monospaced, :unconstrained, /\\?(?:\[([^\]]+)\])?``(#{CC_ALL}+?)``/m],
+      [:monospaced, :unconstrained, /\\?(?:#{QuoteAttributeListRxt})?``(#{CC_ALL}+?)``/m],
       # `monospaced`
-      [:monospaced, :constrained, /(^|[^#{CC_WORD};:"'`}])(?:\[([^\]]+)\])?`(\S|\S#{CC_ALL}*?\S)`(?![#{CC_WORD}"'`])/m],
+      [:monospaced, :constrained, /(^|[^#{CC_WORD};:"'`}])(?:#{QuoteAttributeListRxt})?`(\S|\S#{CC_ALL}*?\S)`(?![#{CC_WORD}"'`])/m],
       # __emphasis__
-      [:emphasis, :unconstrained, /\\?(?:\[([^\]]+)\])?__(#{CC_ALL}+?)__/m],
+      [:emphasis, :unconstrained, /\\?(?:#{QuoteAttributeListRxt})?__(#{CC_ALL}+?)__/m],
       # _emphasis_
-      [:emphasis, :constrained, /(^|[^#{CC_WORD};:}])(?:\[([^\]]+)\])?_(\S|\S#{CC_ALL}*?\S)_(?!#{CG_WORD})/m],
+      [:emphasis, :constrained, /(^|[^#{CC_WORD};:}])(?:#{QuoteAttributeListRxt})?_(\S|\S#{CC_ALL}*?\S)_(?!#{CG_WORD})/m],
       # ##mark## (referred to in AsciiDoc.py as unquoted)
-      [:mark, :unconstrained, /\\?(?:\[([^\]]+)\])?##(#{CC_ALL}+?)##/m],
+      [:mark, :unconstrained, /\\?(?:#{QuoteAttributeListRxt})?##(#{CC_ALL}+?)##/m],
       # #mark# (referred to in AsciiDoc.py as unquoted)
-      [:mark, :constrained, /(^|[^#{CC_WORD}&;:}])(?:\[([^\]]+)\])?#(\S|\S#{CC_ALL}*?\S)#(?!#{CG_WORD})/m],
+      [:mark, :constrained, /(^|[^#{CC_WORD}&;:}])(?:#{QuoteAttributeListRxt})?#(\S|\S#{CC_ALL}*?\S)#(?!#{CG_WORD})/m],
       # ^superscript^
-      [:superscript, :unconstrained, /\\?(?:\[([^\]]+)\])?\^(\S+?)\^/],
+      [:superscript, :unconstrained, /\\?(?:#{QuoteAttributeListRxt})?\^(\S+?)\^/],
       # ~subscript~
-      [:subscript, :unconstrained, /\\?(?:\[([^\]]+)\])?~(\S+?)~/],
+      [:subscript, :unconstrained, /\\?(?:#{QuoteAttributeListRxt})?~(\S+?)~/],
     ]
 
     accum[true] = compat = normal.drop 0
     # ``quoted''
-    compat[2] = [:double, :constrained, /(^|[^#{CC_WORD};:}])(?:\[([^\]]+)\])?``(\S|\S#{CC_ALL}*?\S)''(?!#{CG_WORD})/m]
+    compat[2] = [:double, :constrained, /(^|[^#{CC_WORD};:}])(?:#{QuoteAttributeListRxt})?``(\S|\S#{CC_ALL}*?\S)''(?!#{CG_WORD})/m]
     # `quoted'
-    compat[3] = [:single, :constrained, /(^|[^#{CC_WORD};:}])(?:\[([^\]]+)\])?`(\S|\S#{CC_ALL}*?\S)'(?!#{CG_WORD})/m]
+    compat[3] = [:single, :constrained, /(^|[^#{CC_WORD};:}])(?:#{QuoteAttributeListRxt})?`(\S|\S#{CC_ALL}*?\S)'(?!#{CG_WORD})/m]
     # ++monospaced++
-    compat[4] = [:monospaced, :unconstrained, /\\?(?:\[([^\]]+)\])?\+\+(#{CC_ALL}+?)\+\+/m]
+    compat[4] = [:monospaced, :unconstrained, /\\?(?:#{QuoteAttributeListRxt})?\+\+(#{CC_ALL}+?)\+\+/m]
     # +monospaced+
-    compat[5] = [:monospaced, :constrained, /(^|[^#{CC_WORD};:}])(?:\[([^\]]+)\])?\+(\S|\S#{CC_ALL}*?\S)\+(?!#{CG_WORD})/m]
+    compat[5] = [:monospaced, :constrained, /(^|[^#{CC_WORD};:}])(?:#{QuoteAttributeListRxt})?\+(\S|\S#{CC_ALL}*?\S)\+(?!#{CG_WORD})/m]
     # #unquoted#
     #compat[8] = [:unquoted, *compat[8][1..-1]]
     # ##unquoted##
     #compat[9] = [:unquoted, *compat[9][1..-1]]
     # 'emphasis'
-    compat.insert 3, [:emphasis, :constrained, /(^|[^#{CC_WORD};:}])(?:\[([^\]]+)\])?'(\S|\S#{CC_ALL}*?\S)'(?!#{CG_WORD})/m]
+    compat.insert 3, [:emphasis, :constrained, /(^|[^#{CC_WORD};:}])(?:#{QuoteAttributeListRxt})?'(\S|\S#{CC_ALL}*?\S)'(?!#{CG_WORD})/m]
   end
 
   # NOTE order of replacements is significant
diff --git a/lib/asciidoctor/rx.rb b/lib/asciidoctor/rx.rb
index 6cc378ea..5e73b13f 100644
--- a/lib/asciidoctor/rx.rb
+++ b/lib/asciidoctor/rx.rb
@@ -573,8 +573,8 @@ module Asciidoctor
   #
   # NOTE we always capture the attributes so we know when to use compatible (i.e., legacy) behavior
   InlinePassRx = {
-    false => ['+', '`', /(^|[^#{CC_WORD};:])(?:\[([^\]]+)\])?(\\?(\+|`)(\S|\S#{CC_ALL}*?\S)\4)(?!#{CG_WORD})/m],
-    true => ['`', nil, /(^|[^`#{CC_WORD}])(?:\[([^\]]+)\])?(\\?(`)([^`\s]|[^`\s]#{CC_ALL}*?\S)\4)(?![`#{CC_WORD}])/m],
+    false => ['+', '`', /(^|[^#{CC_WORD};:])(?:#{QuoteAttributeListRxt})?(\\?(\+|`)(\S|\S#{CC_ALL}*?\S)\4)(?!#{CG_WORD})/m],
+    true => ['`', nil, /(^|[^`#{CC_WORD}])(?:#{QuoteAttributeListRxt})?(\\?(`)([^`\s]|[^`\s]#{CC_ALL}*?\S)\4)(?![`#{CC_WORD}])/m],
   }
 
   # Matches an inline plus passthrough spanning multiple lines, but only when it occurs directly
@@ -595,7 +595,7 @@ module Asciidoctor
   #   pass:quotes[text]
   #
   # NOTE we have to support an empty pass:[] for compatibility with AsciiDoc.py
-  InlinePassMacroRx = /(?:(?:(\\?)\[([^\]]+)\])?(\\{0,2})(\+\+\+?|\$\$)(#{CC_ALL}*?)\4|(\\?)pass:([a-z]+(?:,[a-z-]+)*)?\[(|#{CC_ALL}*?[^\\])\])/m
+  InlinePassMacroRx = /(?:(?:(\\?)#{QuoteAttributeListRxt})?(\\{0,2})(\+\+\+?|\$\$)(#{CC_ALL}*?)\4|(\\?)pass:([a-z]+(?:,[a-z-]+)*)?\[(|#{CC_ALL}*?[^\\])\])/m
 
   # Matches an xref (i.e., cross-reference) inline macro, which may span multiple lines.
   #
diff --git a/test/substitutions_test.rb b/test/substitutions_test.rb
index b4b4ab09..f75abfa4 100644
--- a/test/substitutions_test.rb
+++ b/test/substitutions_test.rb
@@ -203,6 +203,11 @@ context 'Substitutions' do
       assert_equal '<span class="statement">a few words</span>', para.sub_quotes(para.source)
     end
 
+    test 'does not recognize attribute list with left square bracket on formatted text' do
+      para = block_from_string 'key: [ *before [.redacted]#redacted# after* ]'
+      assert_equal 'key: [ <strong>before <span class="redacted">redacted</span> after</strong> ]', para.sub_quotes(para.source)
+    end
+
     test 'single-line constrained strong string' do
       para = block_from_string '*a few strong words*'
       assert_equal '<strong>a few strong words</strong>', para.sub_quotes(para.source)