dev: introduce some helpers for introspecting on libxml2 patches

sparklemotion · Jan 3, 2022 · a6cb299 · a6cb299
1 parent b643ef6
commit a6cb299
Show file tree

Hide file tree

Showing 7 changed files with 35 additions and 19 deletions.
diff --git a/lib/nokogiri/version/info.rb b/lib/nokogiri/version/info.rb
@@ -187,20 +187,32 @@ def to_markdown
     end
   end
 
-  def self.uses_libxml?(requirement = nil) # :nodoc:
+  # :nodoc:
+  def self.uses_libxml?(requirement = nil)
     return false unless VersionInfo.instance.libxml2?
     return true unless requirement
     Gem::Requirement.new(requirement).satisfied_by?(VersionInfo.instance.loaded_libxml_version)
   end
 
-  def self.uses_gumbo? # :nodoc:
+  # :nodoc:
+  def self.uses_gumbo?
     uses_libxml? # TODO: replace with Gumbo functionality
   end
 
-  def self.jruby? # :nodoc:
+  # :nodoc:
+  def self.jruby?
     VersionInfo.instance.jruby?
   end
 
+  # :nodoc:
+  def self.libxml2_patches
+    if VersionInfo.instance.libxml2_using_packaged?
+      Nokogiri::VERSION_INFO["libxml"]["patches"]
+    else
+      []
+    end
+  end
+
   require_relative "../jruby/dependencies" if Nokogiri.jruby?
   require_relative "../extension"
 

diff --git a/test/helper.rb b/test/helper.rb
@@ -157,6 +157,16 @@ def skip_unless_libxml2(msg = "this test should only run with libxml2")
       skip(msg) unless Nokogiri.uses_libxml?
     end
 
+    def skip_unless_libxml2_patch(patch_name)
+      patch_dir = File.join(__dir__, "..", "patches", "libxml2")
+      if File.directory?(patch_dir) && !File.exist?(File.join(patch_dir, patch_name))
+        raise("checking for nonexistent patch file #{patch_name.inspect}")
+      end
+      unless Nokogiri.libxml2_patches.include?(patch_name)
+        skip("this test needs libxml2 patched with #{patch_name}")
+      end
+    end
+
     def skip_unless_jruby(msg = "this test should only run with jruby")
       skip(msg) unless Nokogiri.jruby?
     end

diff --git a/test/html4/test_attributes_properly_escaped.rb b/test/html4/test_attributes_properly_escaped.rb
@@ -6,9 +6,7 @@ module Nokogiri
   module HTML
     class TestAttributesProperlyEscaped < Nokogiri::TestCase
       def test_attribute_macros_are_escaped
-        if Nokogiri.uses_libxml? && !Nokogiri::VERSION_INFO["libxml"]["patches"]&.include?("0001-Remove-script-macro-support.patch")
-          skip("libxml2 has not been patched to be safe against attribute macros")
-        end
+        skip_unless_libxml2_patch("0001-Remove-script-macro-support.patch") if Nokogiri.uses_libxml?
 
         html = "<p><i for=\"&{<test>}\"></i></p>"
         document = Nokogiri::HTML::Document.new
@@ -18,9 +16,7 @@ def test_attribute_macros_are_escaped
       end
 
       def test_libxml_escapes_server_side_includes
-        if Nokogiri.uses_libxml? && !Nokogiri::VERSION_INFO["libxml"]["patches"]&.include?("0002-Update-entities-to-remove-handling-of-ssi.patch")
-          skip("libxml2 has not been patched to be safe against SSI")
-        end
+        skip_unless_libxml2_patch("0002-Update-entities-to-remove-handling-of-ssi.patch") if Nokogiri.uses_libxml?
 
         original_html = %(<p><a href='<!--"><test>-->'></a></p>)
         document = Nokogiri::HTML::Document.new
@@ -30,9 +26,7 @@ def test_libxml_escapes_server_side_includes
       end
 
       def test_libxml_escapes_server_side_includes_without_nested_quotes
-        if Nokogiri.uses_libxml? && !Nokogiri::VERSION_INFO["libxml"]["patches"]&.include?("0002-Update-entities-to-remove-handling-of-ssi.patch")
-          skip("libxml2 has not been patched to be safe against SSI")
-        end
+        skip_unless_libxml2_patch("0002-Update-entities-to-remove-handling-of-ssi.patch") if Nokogiri.uses_libxml?
 
         original_html = %(<p><i for="<!--<test>-->"></i></p>)
         document = Nokogiri::HTML::Document.new

diff --git a/test/html4/test_comments.rb b/test/html4/test_comments.rb
@@ -23,7 +23,7 @@ class TestComment < Nokogiri::TestCase
           let(:html) { "<html><body><div id=under-test><!--></div><div id=also-here></div></body></html>" }
 
           if Nokogiri.uses_libxml?
-            if Nokogiri::VersionInfo.instance.libxml2_using_packaged? && Nokogiri::VERSION_INFO["libxml"]["patches"]&.include?("0008-htmlParseComment-handle-abruptly-closed-comments.patch")
+            if Nokogiri.libxml2_patches.include?("0008-htmlParseComment-handle-abruptly-closed-comments.patch")
               it "behaves as if the comment is closed correctly" do # COMPLIANT
                 assert_equal 1, subject.children.length
                 assert subject.children.first.comment?
@@ -54,7 +54,7 @@ class TestComment < Nokogiri::TestCase
           let(:html) { "<html><body><div id=under-test><!---></div><div id=also-here></div></body></html>" }
 
           if Nokogiri.uses_libxml?
-            if Nokogiri::VersionInfo.instance.libxml2_using_packaged? && Nokogiri::VERSION_INFO["libxml"]["patches"]&.include?("0008-htmlParseComment-handle-abruptly-closed-comments.patch")
+            if Nokogiri.libxml2_patches.include?("0008-htmlParseComment-handle-abruptly-closed-comments.patch")
               it "behaves as if the comment is closed correctly" do # COMPLIANT
                 assert_equal 1, subject.children.length
                 assert subject.children.first.comment?

diff --git a/test/test_version.rb b/test/test_version.rb
@@ -33,6 +33,7 @@ def test_version_info_for_xerces_and_nekohtml
     skip_unless_jruby("xerces/nekohtml is only used for JRuby")
     assert_equal(Nokogiri::XERCES_VERSION, version_info["other_libraries"]["xerces"])
     assert_equal(Nokogiri::NEKO_VERSION, version_info["other_libraries"]["nekohtml"])
+    assert_empty(Nokogiri.libxml2_patches)
   end
 
   def test_version_info_for_libxml
@@ -41,12 +42,14 @@ def test_version_info_for_libxml
     if Nokogiri::VersionInfo.instance.libxml2_using_packaged?
       assert_equal("packaged", version_info["libxml"]["source"])
       assert(version_info["libxml"]["patches"])
+      assert_equal(Nokogiri.libxml2_patches, version_info["libxml"]["patches"])
       assert_equal(Nokogiri::VersionInfo.instance.libxml2_precompiled?, version_info["libxml"]["precompiled"])
     end
     if Nokogiri::VersionInfo.instance.libxml2_using_system?
       assert_equal("system", version_info["libxml"]["source"])
       refute(version_info["libxml"].key?("precompiled"))
       refute(version_info["libxml"].key?("patches"))
+      assert_empty(Nokogiri.libxml2_patches)
     end
 
     assert_equal(Nokogiri::LIBXML_COMPILED_VERSION, version_info["libxml"]["compiled"])

diff --git a/test/xml/sax/test_parser.rb b/test/xml/sax/test_parser.rb
@@ -396,8 +396,7 @@ def call_parse_io_with_encoding(encoding)
     end
 
     it :test_large_cdata_is_handled do
-      # see #2132 and https://gitlab.gnome.org/GNOME/libxml2/-/issues/200
-      skip("Upstream libxml2 <= 2.9.10 needs to be patched") if Nokogiri::VersionInfo.instance.libxml2_using_system?
+      skip("see #2132 and https://gitlab.gnome.org/GNOME/libxml2/-/issues/200") if Nokogiri.uses_libxml?("<=2.9.10")
 
       template = <<~EOF
         <?xml version="1.0" encoding="UTF-8"?>

diff --git a/test/xml/test_xpath.rb b/test/xml/test_xpath.rb
@@ -477,9 +477,7 @@ def test_xpath_syntax_error_should_not_display_line_and_column_if_both_are_zero
       end
 
       def test_huge_xpath_query
-        if Nokogiri.uses_libxml?("~>2.9.11") && !Nokogiri::VERSION_INFO["libxml"]["patches"]&.include?("0007-Fix-XPath-recursion-limit.patch")
-          skip("libxml2 under test is broken with respect to xpath query recusion depth")
-        end
+        skip_unless_libxml2_patch("0007-Fix-XPath-recursion-limit.patch") if Nokogiri.uses_libxml?("~>2.9.11")
 
         # real world example
         # from https://github.com/sparklemotion/nokogiri/issues/2257