llvm/unittests/Support/DJBTest.cpp

   1 //===---------- llvm/unittest/Support/DJBTest.cpp -------------------------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8
   9 #include "llvm/Support/DJB.h"
  10 #include "llvm/ADT/Twine.h"
  11 #include "gtest/gtest.h"
  12
  13 using namespace llvm;
  14
  15 TEST(DJBTest, caseFolding) {
  16   struct TestCase {
  17     StringLiteral One;
  18     StringLiteral Two;
  19   };
  20
  21   static constexpr TestCase Tests[] = {
  22       {{"ASDF"}, {"asdf"}},
  23       {{"qWeR"}, {"QwEr"}},
  24       {{"qqqqqqqqqqqqqqqqqqqq"}, {"QQQQQQQQQQQQQQQQQQQQ"}},
  25
  26       {{"I"}, {"i"}},
  27       // Latin Small Letter Dotless I
  28       {{/*U+130*/ "\xc4\xb0"}, {"i"}},
  29       // Latin Capital Letter I With Dot Above
  30       {{/*U+131*/ "\xc4\xb1"}, {"i"}},
  31
  32       // Latin Capital Letter A With Grave
  33       {{/*U+c0*/ "\xc3\x80"}, {/*U+e0*/ "\xc3\xa0"}},
  34       // Latin Capital Letter A With Macron
  35       {{/*U+100*/ "\xc4\x80"}, {/*U+101*/ "\xc4\x81"}},
  36       // Latin Capital Letter L With Acute
  37       {{/*U+139*/ "\xc4\xb9"}, {/*U+13a*/ "\xc4\xba"}},
  38       // Cyrillic Capital Letter Ie
  39       {{/*U+415*/ "\xd0\x95"}, {/*U+435*/ "\xd0\xb5"}},
  40       // Latin Capital Letter A With Circumflex And Grave
  41       {{/*U+1ea6*/ "\xe1\xba\xa6"}, {/*U+1ea7*/ "\xe1\xba\xa7"}},
  42       // Kelvin Sign
  43       {{/*U+212a*/ "\xe2\x84\xaa"}, {"k"}},
  44       // Glagolitic Capital Letter Chrivi
  45       {{/*U+2c1d*/ "\xe2\xb0\x9d"}, {/*U+2c4d*/ "\xe2\xb1\x8d"}},
  46       // Fullwidth Latin Capital Letter M
  47       {{/*U+ff2d*/ "\xef\xbc\xad"}, {/*U+ff4d*/ "\xef\xbd\x8d"}},
  48       // Old Hungarian Capital Letter Ej
  49       {{/*U+10c92*/ "\xf0\x90\xb2\x92"}, {/*U+10cd2*/ "\xf0\x90\xb3\x92"}},
  50   };
  51
  52   for (const TestCase &T : Tests) {
  53     SCOPED_TRACE("Comparing '" + T.One + "' and '" + T.Two + "'");
  54     EXPECT_EQ(caseFoldingDjbHash(T.One), caseFoldingDjbHash(T.Two));
  55   }
  56 }
  57
  58 TEST(DJBTest, knownValuesLowerCase) {
  59   struct TestCase {
  60     StringLiteral Text;
  61     uint32_t Hash;
  62   };
  63   static constexpr TestCase Tests[] = {
  64       {{""}, 5381u},
  65       {{"f"}, 177675u},
  66       {{"fo"}, 5863386u},
  67       {{"foo"}, 193491849u},
  68       {{"foob"}, 2090263819u},
  69       {{"fooba"}, 259229388u},
  70       {{"foobar"}, 4259602622u},
  71       {{"pneumonoultramicroscopicsilicovolcanoconiosis"}, 3999417781u},
  72   };
  73
  74   for (const TestCase &T : Tests) {
  75     SCOPED_TRACE("Text: '" + T.Text + "'");
  76     EXPECT_EQ(T.Hash, djbHash(T.Text));
  77     EXPECT_EQ(T.Hash, caseFoldingDjbHash(T.Text));
  78     EXPECT_EQ(T.Hash, caseFoldingDjbHash(T.Text.upper()));
  79   }
  80 }
  81
  82 TEST(DJBTest, knownValuesUnicode) {
  83   EXPECT_EQ(5866553u, djbHash(/*U+130*/ "\xc4\xb0"));
  84   EXPECT_EQ(177678u, caseFoldingDjbHash(/*U+130*/ "\xc4\xb0"));
  85   EXPECT_EQ(
  86       1302161417u,
  87       djbHash("\xc4\xb0\xc4\xb1\xc3\x80\xc3\xa0\xc4\x80\xc4\x81\xc4\xb9\xc4\xba"
  88               "\xd0\x95\xd0\xb5\xe1\xba\xa6\xe1\xba\xa7\xe2\x84\xaa\x6b\xe2\xb0"
  89               "\x9d\xe2\xb1\x8d\xef\xbc\xad\xef\xbd\x8d\xf0\x90\xb2\x92\xf0\x90"
  90               "\xb3\x92"));
  91   EXPECT_EQ(
  92       1145571043u,
  93       caseFoldingDjbHash(
  94           "\xc4\xb0\xc4\xb1\xc3\x80\xc3\xa0\xc4\x80\xc4\x81\xc4\xb9\xc4\xba"
  95           "\xd0\x95\xd0\xb5\xe1\xba\xa6\xe1\xba\xa7\xe2\x84\xaa\x6b\xe2\xb0"
  96           "\x9d\xe2\xb1\x8d\xef\xbc\xad\xef\xbd\x8d\xf0\x90\xb2\x92\xf0\x90"
  97           "\xb3\x92"));
  98 }