moxml 0.1.10 → 0.1.12

data/spec/moxml/namespace_uri_validation_spec.rb

@@ -0,0 +1,140 @@
+# frozen_string_literal: true
+
+RSpec.describe "Namespace URI validation" do
+  let(:context) { Moxml.new }
+  let(:doc) { context.create_document }
+  let(:element) { doc.create_element("test") }
+
+  describe "RFC 3986 URI-reference validation" do
+    context "with valid absolute URIs" do
+      %w[
+        http://example.com
+        https://www.w3.org/2001/XMLSchema
+        urn:isbn:12345
+        ftp://ftp.is.co.za/rfc/rfc1808.txt
+        mailto:John.Doe@example.com
+        mailto:bar
+        tel:+1-816-555-1212
+        zarquon://example.org/namespace
+        http://example.org/namespace#apples
+        data:text/plain;base64,SGVsbG8=
+        tag:example.com,2000:test
+      ].each do |uri|
+        it "accepts #{uri.inspect}" do
+          expect { element.add_namespace("ns", uri) }.not_to raise_error
+        end
+      end
+    end
+
+    context "with valid relative URI-references" do
+      %w[
+        my-custom-ns
+        ../relative
+        namespaces/zaphod
+        path/to/resource
+        hello%20world
+      ].each do |uri|
+        it "accepts #{uri.inspect}" do
+          expect { element.add_namespace("ns", uri) }.not_to raise_error
+        end
+      end
+    end
+
+    context "with valid fragment-only references" do
+      %w[
+        #fragment
+        #beeblebrox
+      ].each do |uri|
+        it "accepts #{uri.inspect}" do
+          expect { element.add_namespace("ns", uri) }.not_to raise_error
+        end
+      end
+    end
+
+    context "with invalid URIs" do
+      [
+        "invalid uri",
+        "has space",
+        "two  spaces",
+      ].each do |uri|
+        it "rejects #{uri.inspect}" do
+          expect do
+            element.add_namespace("ns", uri)
+          end.to raise_error(Moxml::NamespaceError, /Invalid URI/)
+        end
+      end
+    end
+
+    context "with control characters" do
+      [
+        "invalid\x00uri",
+        "bad\x01char",
+        "control\x1Fchar",
+      ].each do |uri|
+        it "rejects URI containing control characters" do
+          expect do
+            element.add_namespace("ns", uri)
+          end.to raise_error(Moxml::NamespaceError)
+        end
+      end
+    end
+  end
+
+  describe "empty URI constraint" do
+    it "accepts empty URI for default namespace undeclaration" do
+      expect { element.add_namespace(nil, "") }.not_to raise_error
+    end
+
+    it "rejects empty URI for prefixed namespace declarations" do
+      expect do
+        element.add_namespace("xs", "")
+      end.to raise_error(Moxml::NamespaceError, /empty URI/)
+    end
+  end
+
+  describe "lenient namespace_uri_mode" do
+    let(:context) do
+      Moxml.new do |config|
+        config.namespace_uri_mode = :lenient
+      end
+    end
+
+    context "with non-URI strings" do
+      [
+        "invalid uri",
+        "has space",
+        "two  spaces",
+        "my custom namespace",
+        "not-a-uri but still valid in lenient mode",
+      ].each do |uri|
+        it "accepts #{uri.inspect}" do
+          expect { element.add_namespace("ns", uri) }.not_to raise_error
+        end
+      end
+    end
+
+    context "with control characters" do
+      [
+        "invalid\x00uri",
+        "bad\x01char",
+        "control\x1Fchar",
+      ].each do |uri|
+        it "still rejects #{uri.inspect}" do
+          expect do
+            element.add_namespace("ns", uri)
+          end.to raise_error(Moxml::NamespaceError)
+        end
+      end
+    end
+
+    it "still rejects empty URI for prefixed namespace declarations" do
+      expect do
+        element.add_namespace("xs", "")
+      end.to raise_error(Moxml::NamespaceError, /empty URI/)
+    end
+
+    it "still accepts valid URIs" do
+      expect { element.add_namespace("ns", "http://example.com") }.not_to raise_error
+    end
+  end
+end

data/spec/moxml/xpath/axes_spec.rb

@@ -36,13 +36,12 @@ RSpec.describe "XPath Axes" do
         compiler.respond_to?(:"on_axis_#{axis}", true)
       end
 
-      puts "\nImplemented Axes: #{implemented.size}/13"
-      implemented.each { |a| puts "  ✓ #{a}" }
+      implemented.each { |a| }
 
       missing = axes - implemented
       if missing.any?
-        puts "\nMissing Axes: #{missing.size}/13"
-        missing.each { |a| puts "  ✗ #{a}" }
+
+        missing.each { |a| }
       end
 
       # Expect at least 6 axes (3 original + 3 new critical axes)

data/spec/performance/xpath_benchmark_spec.rb

@@ -6,7 +6,7 @@ require "yaml"
 require "fileutils"
 require "time"
 
-RSpec.describe "XPath Performance Benchmark" do
+RSpec.describe "XPath Performance Benchmark", :performance do
   if ENV["SKIP_BENCHMARKS"]
     it "skips benchmarks when SKIP_BENCHMARKS is set" do
       skip "Benchmarks skipped. To run benchmarks, use: bundle exec rspec " \
@@ -59,14 +59,7 @@ RSpec.describe "XPath Performance Benchmark" do
 
     describe "XPath query performance" do
       it "benchmarks XPath operations across all adapters" do
-        puts "\n#{'=' * 80}"
-        puts "XPath Performance Benchmark - All Adapters"
-        puts "=" * 80
-
-        xpath_patterns.each do |pattern_name, xpath|
-          puts "\nPattern: #{pattern_name}"
-          puts "-" * 80
-
+        xpath_patterns.each_value do |xpath|
           Benchmark.ips do |x|
             x.config(time: 5, warmup: 2)
 
@@ -83,20 +76,10 @@ RSpec.describe "XPath Performance Benchmark" do
 
             x.compare!
           end
-
-          puts "\n"
         end
-
-        puts "=" * 80
-        puts "Benchmark complete"
-        puts "=" * 80
       end
 
       it "generates detailed performance comparison table" do
-        puts "\n#{'=' * 80}"
-        puts "Detailed XPath Performance Comparison"
-        puts "=" * 80
-
         results_table = {}
 
         xpath_patterns.each do |pattern_name, xpath|
@@ -121,28 +104,17 @@ RSpec.describe "XPath Performance Benchmark" do
           end
         end
 
-        puts "\nResults (operations per second):"
-        puts "-" * 80
-
         adapters.each do |adapter|
-          puts "\n#{adapter.to_s.capitalize}:"
           xpath_patterns.each_key do |pattern_name|
             ops = results_table[pattern_name][adapter]
             if ops
-              puts "  #{pattern_name}: #{ops.round(2)} ops/sec"
-            else
-              puts "  #{pattern_name}: Not supported"
+
             end
           end
         end
 
-        puts "\n#{'=' * 80}"
-        puts "Relative Performance (fastest = 1.0x baseline):"
-        puts "-" * 80
-
         relative_results = {}
         xpath_patterns.each_key do |pattern_name|
-          puts "\n#{pattern_name}:"
           valid_results = results_table[pattern_name].compact
           next if valid_results.empty?
 
@@ -152,11 +124,10 @@ RSpec.describe "XPath Performance Benchmark" do
             if ops
               relative = ops / fastest
               relative_results[pattern_name][adapter] = relative
-              puts "  #{adapter}: #{relative.round(3)}x " \
-                   "(#{ops.round(2)} ops/sec)"
+
             else
               relative_results[pattern_name][adapter] = nil
-              puts "  #{adapter}: Not supported"
+
             end
           end
         end
@@ -187,12 +158,6 @@ RSpec.describe "XPath Performance Benchmark" do
         }
 
         File.write(output_file, YAML.dump(yaml_data))
-
-        puts "\n#{'=' * 80}"
-        puts "Results saved to: #{output_file}"
-        puts "=" * 80
-        puts "Test complete - see output above for results"
-        puts "=" * 80
       end
     end
 
@@ -221,18 +186,9 @@ RSpec.describe "XPath Performance Benchmark" do
       end
 
       it "benchmarks namespace-aware XPath" do
-        puts "\n#{'=' * 80}"
-        puts "Namespace-Aware XPath Performance"
-        puts "=" * 80
-        puts "\nNote: REXML and Ox do not support namespace-aware XPath"
-        puts "-" * 80
-
         namespace_capable_adapters = %i[nokogiri libxml oga]
 
-        namespace_patterns.each do |pattern_name, (xpath, namespaces)|
-          puts "\nPattern: #{pattern_name}"
-          puts "-" * 80
-
+        namespace_patterns.each_value do |(xpath, namespaces)|
           Benchmark.ips do |x|
             x.config(time: 5, warmup: 2)
 
@@ -248,11 +204,7 @@ RSpec.describe "XPath Performance Benchmark" do
 
             x.compare!
           end
-
-          puts "\n"
         end
-
-        puts "=" * 80
       end
     end
   end

data/spec/support/w3c_namespace_helpers.rb

@@ -0,0 +1,41 @@
+# frozen_string_literal: true
+
+require "rexml/document"
+
+W3C_NS_FIXTURES_DIR = File.expand_path("../fixtures/w3c/namespaces/1.0", __dir__)
+
+# Parse the test catalog to get test metadata
+def load_w3c_namespace_tests
+  catalog = File.read(File.join(W3C_NS_FIXTURES_DIR, "rmt-ns10.xml"))
+  doc = REXML::Document.new(catalog)
+  tests = []
+  doc.elements.each("TESTCASES/TEST") do |test_el|
+    tests << {
+      id: test_el.attributes["ID"],
+      uri: test_el.attributes["URI"],
+      type: test_el.attributes["TYPE"],
+      sections: test_el.attributes["SECTIONS"],
+      description: test_el.text.strip,
+    }
+  end
+  tests
+end
+
+W3C_NAMESPACE_TESTS = load_w3c_namespace_tests
+
+# Known adapter-level limitations for specific tests.
+# These are parser bugs/limitations, not moxml issues.
+ADAPTER_SKIP_TESTS = {
+  # Test 006: ISO-8859-1 encoded IRI — adapters receive binary-read content
+  # and may fail on encoding before namespace processing begins.
+  "rmt-ns10-006" => :all,
+  # Test 047: DOCTYPE with colon in element name — Oga parser limitation.
+  "ht-ns10-047" => [:oga],
+}.freeze
+
+def skip_for_adapter?(test_id, adapter)
+  skip_config = ADAPTER_SKIP_TESTS[test_id]
+  return false unless skip_config
+
+  skip_config == :all || skip_config.include?(adapter)
+end

data/spec/unit/rexml_isolated_test.rb

@@ -0,0 +1,271 @@
+# frozen_string_literal: true
+
+require "rspec"
+require "moxml"
+
+RSpec.describe "REXML Adapter Isolated Test" do
+  let(:rexml_context) { Moxml.new(:rexml) }
+
+  describe "text extraction behavior" do
+    it "extracts simple text correctly" do
+      xml = <<~XML
+        <root>Hello World</root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      expect(text).to eq("Hello World")
+    end
+
+    it "demonstrates BMJBMJ concatenation issue - should NOT add spaces" do
+      xml = <<~XML
+        <root>
+          <journal>BMJ</journal>
+          <journal>BMJ</journal>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # This should FAIL to demonstrate the round-trip issue
+      # Other adapters produce: "BMJBMJ"
+      # REXML currently produces: "BMJ BMJ" (with space)
+      # For round-trip compatibility, REXML should produce "BMJBMJ"
+      expect(text).to eq("BMJBMJ")
+    end
+
+    it "demonstrates mixed case transition issue - should NOT add spaces" do
+      xml = <<~XML
+        <root>
+          <issn>0959-8138</issn>
+          <publisher>BMJ</publisher>
+          <author>j</author>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # This should FAIL to demonstrate the round-trip issue
+      # Other adapters produce: "0959-8138BMJj"
+      # REXML currently produces: "0959-8138 BMJ j" (with spaces)
+      # For round-trip compatibility, REXML should produce "0959-8138BMJj"
+      expect(text).to eq("0959-8138BMJj")
+    end
+
+    it "demonstrates digit transition issue - should NOT add spaces" do
+      xml = <<~XML
+        <root>
+          <volume>324</volume>
+          <issue>i7342</issue>
+          <page>pg880</page>
+          <id>11950738</id>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # This should FAIL to demonstrate the round-trip issue
+      # Other adapters produce: "324i7342pg88011950738"
+      # REXML currently produces: "324 i7342 pg880 11950738" (with spaces)
+      # For round-trip compatibility, REXML should produce "324i7342pg88011950738"
+      expect(text).to eq("324i7342pg88011950738")
+    end
+
+    it "demonstrates word boundary issue - should NOT add spaces" do
+      xml = <<~XML
+        <root>
+          <article-type>version-of-record</article-type>
+          <title>Primary</title>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # This should FAIL to demonstrate the round-trip issue
+      # Other adapters produce: "version-of-recordPrimary"
+      # REXML currently produces: "version-of-record Primary" (with space)
+      # For round-trip compatibility, REXML should produce "version-of-recordPrimary"
+      expect(text).to eq("version-of-recordPrimary")
+    end
+
+    it "demonstrates complex mixed content issue - should NOT add spaces" do
+      xml = <<~XML
+        <root>
+          <section>Primary</section>
+          <year>190</year>
+          <page>102</page>
+          <id>18219355357</id>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # This should FAIL to demonstrate the round-trip issue
+      # Other adapters produce: "Primary19010218219355357"
+      # REXML currently produces: "Primary 190 102 18219355357" (with spaces)
+      # For round-trip compatibility, REXML should produce "Primary19010218219355357"
+      expect(text).to eq("Primary19010218219355357")
+    end
+
+    it "demonstrates all patterns together - should NOT add spaces" do
+      xml = <<~XML
+        <root>
+          <journal>BMJ</journal>
+          <journal>BMJ</journal>
+          <issn>0959-8138</issn>
+          <publisher>BMJ</publisher>
+          <author>j</author>
+          <volume>324</volume>
+          <issue>i7342</issue>
+          <page>pg880</page>
+          <id>11950738</id>
+          <article-type>version-of-record</article-type>
+          <section>Primary</section>
+          <year>190</year>
+          <page>102</page>
+          <id>18219355357</id>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # This should FAIL to demonstrate the round-trip issue
+      # Other adapters produce: "BMJBMJ0959-8138BMJj324i7342pg88011950738version-of-recordPrimary19010218219355357"
+      # REXML currently produces: "BMJ BMJ 0959-8138 BMJ j 324 i7342 pg880 11950738 version-of-record Primary 190 102 18219355357" (with spaces)
+      # For round-trip compatibility, REXML should produce concatenated version
+      expect(text).to eq("BMJBMJ0959-8138BMJj324i7342pg88011950738version-of-recordPrimary19010218219355357")
+    end
+
+    it "demonstrates specific round-trip failure patterns - BMJ.v" do
+      xml = <<~XML
+        <root>
+          <journal>BMJ</journal>
+          <volume>v</volume>
+          <number>324</number>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # Based on actual adapter behavior: both nokogiri and rexml produce "BMJv324"
+      # The test expectation was wrong - should expect "BMJv324" not "BMJ.v324"
+      expect(text).to eq("BMJv324")
+    end
+
+    it "demonstrates specific round-trip failure patterns - i7342.pg" do
+      xml = <<~XML
+        <root>
+          <issue>i7342</issue>
+          <page>pg880</page>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # Based on actual adapter behavior: both nokogiri and rexml produce "i7342pg880"
+      # The test expectation was wrong - should expect "i7342pg880" not "i7342.pg880"
+      expect(text).to eq("i7342pg880")
+    end
+
+    it "demonstrates specific round-trip failure patterns - 190102" do
+      xml = <<~XML
+        <root>
+          <year>190</year>
+          <page>102</page>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # Based on round-trip failure: expected "190102" but got "190 102"
+      expect(text).to eq("190102")
+    end
+
+    it "demonstrates round-trip failure pattern - bmj BMJ" do
+      xml = <<~XML
+        <root>
+          <journal>bmj</journal>
+          <journal>BMJ</journal>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # Based on round-trip failure: expected "bmjBMJ" but got "bmj BMJ"
+      expect(text).to eq("bmjBMJ")
+    end
+
+    it "demonstrates round-trip failure pattern - 8138 BMJ" do
+      xml = <<~XML
+        <root>
+          <issn>8138</issn>
+          <publisher>BMJ</publisher>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # Based on round-trip failure: expected "8138BMJ" but got "8138 BMJ"
+      expect(text).to eq("8138BMJ")
+    end
+
+    it "demonstrates round-trip failure pattern - BMJ.v 324" do
+      xml = <<~XML
+        <root>
+          <journal>BMJ</journal>
+          <volume>v</volume>
+          <number>324</number>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # Based on actual adapter behavior: both nokogiri and rexml produce "BMJv324"
+      # The test expectation was wrong - should expect "BMJv324" not "BMJ.v324"
+      expect(text).to eq("BMJv324")
+    end
+
+    it "demonstrates round-trip failure pattern - 7342 880" do
+      xml = <<~XML
+        <root>
+          <issue>7342</issue>
+          <page>880</page>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # Based on round-trip failure: expected "7342880" but got "7342 880"
+      expect(text).to eq("7342880")
+    end
+
+    it "demonstrates round-trip failure pattern - version-of-record Primary" do
+      xml = <<~XML
+        <root>
+          <article-type>version-of-record</article-type>
+          <section>Primary</section>
+        </root>
+      XML
+
+      doc = rexml_context.parse(xml.dup)
+      text = doc.root.text
+
+      # Based on round-trip failure: expected "version-of-recordPrimary" but got "version-of-record Primary"
+      expect(text).to eq("version-of-recordPrimary")
+    end
+  end
+end