From b4d986a3c9d805aa20a9667d805de9cccdd84905 Mon Sep 17 00:00:00 2001
From: Leonid Shevtsov <leonid@shevtsov.me>
Date: Fri, 23 Aug 2024 16:37:30 +0300
Subject: [PATCH] Improve parse_declarations! performance

---
 CHANGELOG.md               |  1 +
 lib/css_parser/rule_set.rb | 37 +++++++++++++++++++++++++++----------
 test/test_rule_set.rb      | 13 +++++++++++++
 3 files changed, 41 insertions(+), 10 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 557e7eb..4f49b2d 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,7 @@
 * RuleSet initialize now takes keyword argument, positional arguments are still supported but deprecated
 * Removed OffsetAwareRuleSet, it's a RuleSet with optional attributes filename and offset
 * Improved performance of block parsing by using StringScanner
+* Improve `RuleSet#parse_declarations!` performance by using substring search istead of regexps
 
 ### Version v1.18.0
 
diff --git a/lib/css_parser/rule_set.rb b/lib/css_parser/rule_set.rb
index 7f87816..82c7a99 100644
--- a/lib/css_parser/rule_set.rb
+++ b/lib/css_parser/rule_set.rb
@@ -26,6 +26,11 @@ class RuleSet
 
     WHITESPACE_REPLACEMENT = '___SPACE___'
 
+    # Tokens for parse_declarations!
+    COLON = ':'.freeze
+    SEMICOLON = ';'.freeze
+    LPAREN = '('.freeze
+    RPAREN = ')'.freeze
     class Declarations
       class Value
         attr_reader :value
@@ -647,20 +652,32 @@ def parse_declarations!(block) # :nodoc:
       return unless block
 
       continuation = nil
-      block.split(/[;$]+/m).each do |decs|
-        decs = (continuation ? continuation + decs : decs)
-        if decs =~ /\([^)]*\Z/ # if it has an unmatched parenthesis
-          continuation = "#{decs};"
-        elsif (matches = decs.match(/\s*(.[^:]*)\s*:\s*(?m:(.+))(?:;?\s*\Z)/i))
-          # skip end_of_declaration
-          property = matches[1]
-          value = matches[2]
-          add_declaration!(property, value)
-          continuation = nil
+      block.split(SEMICOLON) do |decs|
+        decs = (continuation ? "#{continuation};#{decs}" : decs)
+        if unmatched_open_parenthesis?(decs)
+          # Semicolon happened within parenthesis, so it is a part of the value
+          # the rest of the value is in the next segment
+          continuation = decs
+          next
         end
+
+        next unless (colon = decs.index(COLON))
+
+        property = decs[0, colon]
+        value = decs[(colon + 1)..]
+        property.strip!
+        value.strip!
+        next if property.empty? || value.empty?
+
+        add_declaration!(property, value)
+        continuation = nil
       end
     end
 
+    def unmatched_open_parenthesis?(declarations)
+      (lparen_index = declarations.index(LPAREN)) && !declarations.index(RPAREN, lparen_index)
+    end
+
     #--
     # TODO: way too simplistic
     #++
diff --git a/test/test_rule_set.rb b/test/test_rule_set.rb
index e4bfd5b..d1bf9aa 100644
--- a/test/test_rule_set.rb
+++ b/test/test_rule_set.rb
@@ -78,6 +78,19 @@ def test_each_declaration_containing_semicolons
     assert_equal('no-repeat;', rs['background-repeat'])
   end
 
+  def test_each_declaration_with_newlines
+    expected = Set[
+      {property: 'background-image', value: 'url(foo;bar)', is_important: false},
+      {property: 'font-weight', value: 'bold', is_important: true},
+    ]
+    rs = RuleSet.new(block: "background-image\n:\nurl(foo;bar);\n\n\n\n\n;;font-weight\n\n\n:bold\n\n\n!important")
+    actual = Set.new
+    rs.each_declaration do |prop, val, imp|
+      actual << {property: prop, value: val, is_important: imp}
+    end
+    assert_equal(expected, actual)
+  end
+
   def test_selector_sanitization
     selectors = "h1, h2,\nh3 "
     rs = RuleSet.new(selectors: selectors, block: "color: #fff;")