From 9e3f5b980655817993682e409cbda72956d865cb Mon Sep 17 00:00:00 2001
From: Joel Drapper <joel@drapper.me>
Date: Tue, 16 Apr 2024 22:31:35 +0100
Subject: [PATCH] Fix XSS

---
 lib/phlex/sgml.rb                   |  2 +-
 test/phlex/view/naughty_business.rb | 72 +++++++++++++++++++++++++++++
 2 files changed, 73 insertions(+), 1 deletion(-)

diff --git a/lib/phlex/sgml.rb b/lib/phlex/sgml.rb
index 161be940..994adc38 100644
--- a/lib/phlex/sgml.rb
+++ b/lib/phlex/sgml.rb
@@ -382,7 +382,7 @@ def __attributes__(attributes, buffer = +"")
 				end
 
 				lower_name = name.downcase
-				next if lower_name == "href" && v.start_with?(/\s*javascript:/i)
+				next if lower_name == "href" && v.to_s.downcase.tr("\t \n", "").start_with?("javascript:")
 
 				# Detect unsafe attribute names. Attribute names are considered unsafe if they match an event attribute or include unsafe characters.
 				if HTML::EVENT_ATTRIBUTES.include?(lower_name) || name.match?(/[<>&"']/)
diff --git a/test/phlex/view/naughty_business.rb b/test/phlex/view/naughty_business.rb
index 6599191f..ad5d692e 100644
--- a/test/phlex/view/naughty_business.rb
+++ b/test/phlex/view/naughty_business.rb
@@ -85,6 +85,78 @@ def view_template
 		end
 	end
 
+	with "naughty javascript link protocol with a hidden tab character" do
+		view do
+			def view_template
+				a(href: "\tjavascript:alert(1)") { "XSS" }
+				a(href: "j\tavascript:alert(1)") { "XSS" }
+				a(href: "ja\tvascript:alert(1)") { "XSS" }
+				a(href: "jav\tascript:alert(1)") { "XSS" }
+				a(href: "java\tscript:alert(1)") { "XSS" }
+				a(href: "javas\tcript:alert(1)") { "XSS" }
+				a(href: "javasc\tript:alert(1)") { "XSS" }
+				a(href: "javascr\tipt:alert(1)") { "XSS" }
+				a(href: "javascri\tpt:alert(1)") { "XSS" }
+				a(href: "javascrip\tt:alert(1)") { "XSS" }
+				a(href: "javascript\t:alert(1)") { "XSS" }
+				a(href: "javascript:\talert(1)") { "XSS" }
+			end
+		end
+
+		it "strips the javascript protocol" do
+			expect(output.scan("<a>").size).to be == 12
+			expect(output.scan("href").size).to be == 0
+		end
+	end
+
+	with "naughty javascript link protocol with a hidden newline character" do
+		view do
+			def view_template
+				a(href: "\njavascript:alert(1)") { "XSS" }
+				a(href: "j\navascript:alert(1)") { "XSS" }
+				a(href: "ja\nvascript:alert(1)") { "XSS" }
+				a(href: "jav\nascript:alert(1)") { "XSS" }
+				a(href: "java\nscript:alert(1)") { "XSS" }
+				a(href: "javas\ncript:alert(1)") { "XSS" }
+				a(href: "javasc\nript:alert(1)") { "XSS" }
+				a(href: "javascr\nipt:alert(1)") { "XSS" }
+				a(href: "javascri\npt:alert(1)") { "XSS" }
+				a(href: "javascrip\nt:alert(1)") { "XSS" }
+				a(href: "javascript\n:alert(1)") { "XSS" }
+				a(href: "javascript:\nalert(1)") { "XSS" }
+			end
+		end
+
+		it "strips the javascript protocol" do
+			expect(output.scan("<a>").size).to be == 12
+			expect(output.scan("href").size).to be == 0
+		end
+	end
+
+	with "naughty javascript link protocol with a hidden whitespace character" do
+		view do
+			def view_template
+				a(href: " javascript:alert(1)") { "XSS" }
+				a(href: "j avascript:alert(1)") { "XSS" }
+				a(href: "ja vascript:alert(1)") { "XSS" }
+				a(href: "jav ascript:alert(1)") { "XSS" }
+				a(href: "java script:alert(1)") { "XSS" }
+				a(href: "javas cript:alert(1)") { "XSS" }
+				a(href: "javasc ript:alert(1)") { "XSS" }
+				a(href: "javascr ipt:alert(1)") { "XSS" }
+				a(href: "javascri pt:alert(1)") { "XSS" }
+				a(href: "javascrip t:alert(1)") { "XSS" }
+				a(href: "javascript :alert(1)") { "XSS" }
+				a(href: "javascript: alert(1)") { "XSS" }
+			end
+		end
+
+		it "strips the javascript protocol" do
+			expect(output.scan("<a>").size).to be == 12
+			expect(output.scan("href").size).to be == 0
+		end
+	end
+
 	Phlex::HTML::EVENT_ATTRIBUTES.each do |event_attribute|
 		with "with naughty #{event_attribute} attribute" do
 			naughty_attributes = { event_attribute => "alert(1);" }