Fixes bchavez#210. Randomizer.Utf16String generates technically valid…

… UTF16 strings. (bchavez#214) * Fixes bchavez#210. Randomizer.Utf16String generates technically valid UTF16 strings.
wregulski · Mar 25, 2019 · ebcd260 · ebcd260
1 parent 19aba84
commit ebcd260
Show file tree

Hide file tree

Showing 4 changed files with 500 additions and 2 deletions.
diff --git a/HISTORY.md b/HISTORY.md
@@ -1,6 +1,7 @@
 ## v26.0.3 - /master current
 Release Date: UNRELEASED
-* Added `placeholder.com` image service. 
+* Issue 210: Added `Randomizer.Utf16String` that generates technically valid Unicode with paired high/low surrogates.
+* Added `placeholder.com` image service.
 
 ## v26.0.2
 Release Date: 2019-03-22

diff --git a/Source/Bogus.Tests/RandomizerTest.cs b/Source/Bogus.Tests/RandomizerTest.cs
@@ -5,6 +5,7 @@
 using FluentAssertions;
 using Xunit;
 using Xunit.Abstractions;
+using Z.ExtensionMethods;
 
 namespace Bogus.Tests
 {
@@ -404,5 +405,66 @@ public void can_get_a_weighted_true_value()
 
          truths.Should().BeLessThan(0.25f); // roughly
       }
+
+
+      public static IEnumerable<object[]> ExactLenUtf16(int maxTest)
+      {
+         return Enumerable.Range(1, maxTest)
+            .Select(i => new object[] {i, i});
+      }
+
+      public static IEnumerable<object[]> VarLenUtf16(int maxTest)
+      {
+         return Enumerable.Range(1, maxTest)
+            .Select(i => new object[] { i, i+20 });
+      }
+
+      [Theory]
+      [MemberData(nameof(ExactLenUtf16), parameters: 100)]
+      [MemberData(nameof(VarLenUtf16), parameters: 100)]
+      public void can_generate_valid_utf16_string_with_surrogates(int min, int max)
+      {
+         var x = r.Utf16String(min, max);
+
+         x.Length.Should().BeGreaterOrEqualTo(min)
+            .And
+            .BeLessOrEqualTo(max);
+
+         for( int i = 0; i < x.Length; i++ )
+         {
+            var current = x[i];
+            if( char.IsSurrogate(current) )
+            {
+               char.IsLetterOrDigit(x, i).Should().BeTrue();
+               i++;
+            }
+            else
+            {
+               char.IsLetterOrDigit(current).Should().BeTrue();
+            }
+         }
+      }
+
+      [Theory]
+      [MemberData(nameof(ExactLenUtf16), parameters: 100)]
+      [MemberData(nameof(VarLenUtf16), parameters: 100)]
+      public void can_generate_valid_utf16_without_surrogates(int min, int max)
+      {
+         var x = r.Utf16String(min, max, excludeSurrogates: true);
+
+         x.Length.Should().BeGreaterOrEqualTo(min)
+            .And
+            .BeLessOrEqualTo(max);
+
+         x.Any(char.IsSurrogate).Should().BeFalse();
+         x.All(char.IsLetterOrDigit).Should().BeTrue();
+      }
+
+      [Fact]
+      public void static_utf16_tests()
+      {
+         r.Utf16String().Should().Be("𦊕𡰴ဩਢۥ侀ゞᜡﷴ𠸮𝒾ⶱﹱຂⱪ𝒟𝚪𝒩𝝵ೡཧΐ뢫ⱜੳ𣽇𐨖ਆ𐀼ฏളﾺ𐏈𦵻𐠸𝕋ꡨ𝔇ずరᢔﬣ𐨟𝔚ઐష");
+         r.Utf16String().Should().Be("னమወ𡂑ப𐠁ஞ𦂰ઐ𢂍ᬒ𒀓𨴽𝜅𧊆𦑆ආ𝜂ଭ𐀪ಋΊ౦౨ㆳꠁⶺ𝛈𡓎𡯸ᜑ𐁁𝜹લ𩠺ଐ𦕲ﬔჁⶂム𝐾𣭄ງヾ༤𝒪ᙵͼ၅𦛃𩕾ﷸ𝜦ⱶ");
+      }
    }
 }