From a4d43ae72aacdacd98b830a81938a4dfbf21cf3d Mon Sep 17 00:00:00 2001
From: evazion <noizave@gmail.com>
Date: Wed, 6 Apr 2022 17:15:02 -0500
Subject: [PATCH] post queries: track whether metatag values are quoted.

This is necessary for the `commentary:` metatag, which has different
behavior depending on whether the metatag value is quoted. For example,
`commentary:translated` finds translated commentaries, while
`commentary:"translated"` finds commentaries containing the literal word
"translated".
---
 app/logical/post_query/ast.rb       | 15 ++++++++++-----
 app/logical/post_query/parser.rb    | 10 +++++-----
 test/unit/post_query_parser_test.rb |  4 ++--
 3 files changed, 17 insertions(+), 12 deletions(-)

diff --git a/app/logical/post_query/ast.rb b/app/logical/post_query/ast.rb
index 284b23997..4e5eefda5 100644
--- a/app/logical/post_query/ast.rb
+++ b/app/logical/post_query/ast.rb
@@ -60,8 +60,8 @@ class PostQuery
           AST.new(:tag, [name])
         end
 
-        def metatag(name, value)
-          AST.new(:metatag, [name, value])
+        def metatag(name, value, quoted = false)
+          AST.new(:metatag, [name, value, quoted])
         end
       end
 
@@ -211,7 +211,7 @@ class PostQuery
           "none"
         in [:tag, name]
           name
-        in [:metatag, name, value]
+        in [:metatag, name, value, quoted]
           "#{name}:#{quoted_value}"
         in [:wildcard, name]
           "(wildcard #{name})"
@@ -231,7 +231,7 @@ class PostQuery
           name
         in [:tag, name]
           name
-        in [:metatag, name, value]
+        in [:metatag, name, value, quoted]
           "#{name}:#{quoted_value}"
         in :not, child
           child.term? ? "-#{child.to_infix}" : "-(#{child.to_infix})"
@@ -354,11 +354,16 @@ class PostQuery
         args.second if metatag?
       end
 
+      # @return [String, nil] True if the metatag's value was enclosed in quotes.
+      def quoted?
+        args.third if metatag?
+      end
+
       # @return [String, nil] The value of the metatag as a quoted string, if a metatag node.
       def quoted_value
         return nil unless metatag?
 
-        if value.include?(" ") || value.starts_with?('"') || value.starts_with?("'") || value.empty?
+        if quoted?
           %Q{"#{value.gsub(/"/, '\\"')}"}
         else
           value
diff --git a/app/logical/post_query/parser.rb b/app/logical/post_query/parser.rb
index 160362bec..5f8f631ca 100644
--- a/app/logical/post_query/parser.rb
+++ b/app/logical/post_query/parser.rb
@@ -164,8 +164,8 @@ class PostQuery
         if accept(METATAG_NAME_REGEX)
           name = @scanner.matched.delete_suffix(":").downcase
           name = name.singularize + "_count" if name.in?(PostQueryBuilder::COUNT_METATAG_SYNONYMS)
-          value = quoted_string
-          node(:metatag, name, value)
+          quoted, value = quoted_string
+          node(:metatag, name, value, quoted)
         end
       end
 
@@ -173,13 +173,13 @@ class PostQuery
         if accept('"')
           a = accept(/([^"\\]|\\")*/).gsub(/\\"/, '"') # handle backslash escaped quotes
           expect('"')
-          a
+          [true, a]
         elsif accept("'")
           a = accept(/([^'\\]|\\')*/).gsub(/\\'/, "'") # handle backslash escaped quotes
           expect("'")
-          a
+          [true, a]
         else
-          string(/[^ ]+/)
+          [false, string(/[^ ]+/)]
         end
       end
 
diff --git a/test/unit/post_query_parser_test.rb b/test/unit/post_query_parser_test.rb
index 7de999e2e..328df2d4c 100644
--- a/test/unit/post_query_parser_test.rb
+++ b/test/unit/post_query_parser_test.rb
@@ -69,14 +69,14 @@ class PostQueryParserTest < ActiveSupport::TestCase
       assert_parse_equals('source:""', 'source:""')
       assert_parse_equals('source:"\""', 'source:"\""')
       assert_parse_equals(%q{source:"don't say \"lazy\" okay"}, %q{source:"don't say \"lazy\" okay"})
-      assert_parse_equals(%q{(and source:foo)bar a)}, %q{(a (source:"foo)bar"))})
+      assert_parse_equals(%q{(and source:"foo)bar" a)}, %q{(a (source:"foo)bar"))})
 
       assert_parse_equals('source:"foo bar"', "source:'foo bar'")
       assert_parse_equals("source:foobar'(", "source:foobar'(")
       assert_parse_equals('source:""', "source:''")
       assert_parse_equals('source:"\'"', "source:'\\''")
       assert_parse_equals(%q{source:"don't say \"lazy\" okay"}, %q{source:'don\'t say "lazy" okay'})
-      assert_parse_equals(%q{(and source:foo)bar a)}, %q{(a (source:'foo)bar'))})
+      assert_parse_equals(%q{(and source:"foo)bar" a)}, %q{(a (source:'foo)bar'))})
     end
 
     should "parse metatag synonyms correctly" do