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
13using namespace llvm;
14
15TEST(DJBTest, caseFolding) {
16 struct TestCase {
17 StringLiteral One;
18 StringLiteral Two;
19 };
20
21 static constexpr TestCase Tests[] = {
22 {.One: {"ASDF"}, .Two: {"asdf"}},
23 {.One: {"qWeR"}, .Two: {"QwEr"}},
24 {.One: {"qqqqqqqqqqqqqqqqqqqq"}, .Two: {"QQQQQQQQQQQQQQQQQQQQ"}},
25
26 {.One: {"I"}, .Two: {"i"}},
27 // Latin Small Letter Dotless I
28 {.One: {/*U+130*/ "\xc4\xb0"}, .Two: {"i"}},
29 // Latin Capital Letter I With Dot Above
30 {.One: {/*U+131*/ "\xc4\xb1"}, .Two: {"i"}},
31
32 // Latin Capital Letter A With Grave
33 {.One: {/*U+c0*/ "\xc3\x80"}, .Two: {/*U+e0*/ "\xc3\xa0"}},
34 // Latin Capital Letter A With Macron
35 {.One: {/*U+100*/ "\xc4\x80"}, .Two: {/*U+101*/ "\xc4\x81"}},
36 // Latin Capital Letter L With Acute
37 {.One: {/*U+139*/ "\xc4\xb9"}, .Two: {/*U+13a*/ "\xc4\xba"}},
38 // Cyrillic Capital Letter Ie
39 {.One: {/*U+415*/ "\xd0\x95"}, .Two: {/*U+435*/ "\xd0\xb5"}},
40 // Latin Capital Letter A With Circumflex And Grave
41 {.One: {/*U+1ea6*/ "\xe1\xba\xa6"}, .Two: {/*U+1ea7*/ "\xe1\xba\xa7"}},
42 // Kelvin Sign
43 {.One: {/*U+212a*/ "\xe2\x84\xaa"}, .Two: {"k"}},
44 // Glagolitic Capital Letter Chrivi
45 {.One: {/*U+2c1d*/ "\xe2\xb0\x9d"}, .Two: {/*U+2c4d*/ "\xe2\xb1\x8d"}},
46 // Fullwidth Latin Capital Letter M
47 {.One: {/*U+ff2d*/ "\xef\xbc\xad"}, .Two: {/*U+ff4d*/ "\xef\xbd\x8d"}},
48 // Old Hungarian Capital Letter Ej
49 {.One: {/*U+10c92*/ "\xf0\x90\xb2\x92"}, .Two: {/*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
58TEST(DJBTest, knownValuesLowerCase) {
59 struct TestCase {
60 StringLiteral Text;
61 uint32_t Hash;
62 };
63 static constexpr TestCase Tests[] = {
64 {.Text: {""}, .Hash: 5381u},
65 {.Text: {"f"}, .Hash: 177675u},
66 {.Text: {"fo"}, .Hash: 5863386u},
67 {.Text: {"foo"}, .Hash: 193491849u},
68 {.Text: {"foob"}, .Hash: 2090263819u},
69 {.Text: {"fooba"}, .Hash: 259229388u},
70 {.Text: {"foobar"}, .Hash: 4259602622u},
71 {.Text: {"pneumonoultramicroscopicsilicovolcanoconiosis"}, .Hash: 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
82TEST(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}
99

source code of llvm/unittests/Support/DJBTest.cpp