1 //===- TrigramIndexTest.cpp - Unit tests for TrigramIndex -----------------===//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 
10 #include "llvm/ADT/STLExtras.h"
11 #include "llvm/Support/TrigramIndex.h"
12 #include "gtest/gtest.h"
13 
14 #include <string>
15 #include <vector>
16 
17 using namespace llvm;
18 
19 namespace {
20 
21 class TrigramIndexTest : public ::testing::Test {
22 protected:
23   std::unique_ptr<TrigramIndex> makeTrigramIndex(
24       std::vector<std::string> Rules) {
25     std::unique_ptr<TrigramIndex> TI =
26         make_unique<TrigramIndex>();
27     for (auto &Rule : Rules)
28       TI->insert(Rule);
29     return TI;
30   }
31 };
32 
33 TEST_F(TrigramIndexTest, Empty) {
34   std::unique_ptr<TrigramIndex> TI =
35       makeTrigramIndex({});
36   EXPECT_FALSE(TI->isDefeated());
37   EXPECT_TRUE(TI->isDefinitelyOut("foo"));
38 }
39 
40 TEST_F(TrigramIndexTest, Basic) {
41   std::unique_ptr<TrigramIndex> TI =
42       makeTrigramIndex({"*hello*", "*wor.d*"});
43   EXPECT_FALSE(TI->isDefeated());
44   EXPECT_TRUE(TI->isDefinitelyOut("foo"));
45 }
46 
47 TEST_F(TrigramIndexTest, NoTrigramsInRules) {
48   std::unique_ptr<TrigramIndex> TI =
49       makeTrigramIndex({"b.r", "za*az"});
50   EXPECT_TRUE(TI->isDefeated());
51   EXPECT_FALSE(TI->isDefinitelyOut("foo"));
52   EXPECT_FALSE(TI->isDefinitelyOut("bar"));
53   EXPECT_FALSE(TI->isDefinitelyOut("zakaz"));
54 }
55 
56 TEST_F(TrigramIndexTest, NoTrigramsInARule) {
57   std::unique_ptr<TrigramIndex> TI =
58       makeTrigramIndex({"*hello*", "*wo.ld*"});
59   EXPECT_TRUE(TI->isDefeated());
60   EXPECT_FALSE(TI->isDefinitelyOut("foo"));
61 }
62 
63 TEST_F(TrigramIndexTest, RepetitiveRule) {
64   std::unique_ptr<TrigramIndex> TI =
65       makeTrigramIndex({"*bar*bar*bar*bar*bar", "bar*bar"});
66   EXPECT_FALSE(TI->isDefeated());
67   EXPECT_TRUE(TI->isDefinitelyOut("foo"));
68   EXPECT_TRUE(TI->isDefinitelyOut("bar"));
69   EXPECT_FALSE(TI->isDefinitelyOut("barbara"));
70   EXPECT_FALSE(TI->isDefinitelyOut("bar+bar"));
71 }
72 
73 TEST_F(TrigramIndexTest, PopularTrigram) {
74   std::unique_ptr<TrigramIndex> TI =
75       makeTrigramIndex({"*aaa*", "*aaaa*", "*aaaaa*", "*aaaaa*", "*aaaaaa*"});
76   EXPECT_TRUE(TI->isDefeated());
77 }
78 
79 TEST_F(TrigramIndexTest, PopularTrigram2) {
80   std::unique_ptr<TrigramIndex> TI =
81       makeTrigramIndex({"class1.h", "class2.h", "class3.h", "class4.h", "class.h"});
82   EXPECT_TRUE(TI->isDefeated());
83 }
84 
85 TEST_F(TrigramIndexTest, TooComplicatedRegex) {
86   std::unique_ptr<TrigramIndex> TI =
87       makeTrigramIndex({"[0-9]+"});
88   EXPECT_TRUE(TI->isDefeated());
89 }
90 
91 TEST_F(TrigramIndexTest, TooComplicatedRegex2) {
92   std::unique_ptr<TrigramIndex> TI =
93       makeTrigramIndex({"foo|bar"});
94   EXPECT_TRUE(TI->isDefeated());
95 }
96 
97 TEST_F(TrigramIndexTest, EscapedSymbols) {
98   std::unique_ptr<TrigramIndex> TI =
99       makeTrigramIndex({"*c\\+\\+*", "*hello\\\\world*", "a\\tb", "a\\0b"});
100   EXPECT_FALSE(TI->isDefeated());
101   EXPECT_FALSE(TI->isDefinitelyOut("c++"));
102   EXPECT_TRUE(TI->isDefinitelyOut("c\\+\\+"));
103   EXPECT_FALSE(TI->isDefinitelyOut("hello\\world"));
104   EXPECT_TRUE(TI->isDefinitelyOut("hello\\\\world"));
105   EXPECT_FALSE(TI->isDefinitelyOut("atb"));
106   EXPECT_TRUE(TI->isDefinitelyOut("a\\tb"));
107   EXPECT_TRUE(TI->isDefinitelyOut("a\tb"));
108   EXPECT_FALSE(TI->isDefinitelyOut("a0b"));
109 }
110 
111 TEST_F(TrigramIndexTest, Backreference1) {
112   std::unique_ptr<TrigramIndex> TI =
113       makeTrigramIndex({"*foo\\1*"});
114   EXPECT_TRUE(TI->isDefeated());
115 }
116 
117 TEST_F(TrigramIndexTest, Backreference2) {
118   std::unique_ptr<TrigramIndex> TI =
119       makeTrigramIndex({"*foo\\2*"});
120   EXPECT_TRUE(TI->isDefeated());
121 }
122 
123 TEST_F(TrigramIndexTest, Sequence) {
124   std::unique_ptr<TrigramIndex> TI =
125       makeTrigramIndex({"class1.h", "class2.h", "class3.h", "class4.h"});
126   EXPECT_FALSE(TI->isDefeated());
127   EXPECT_FALSE(TI->isDefinitelyOut("class1"));
128   EXPECT_TRUE(TI->isDefinitelyOut("class.h"));
129   EXPECT_TRUE(TI->isDefinitelyOut("class"));
130 }
131 
132 }  // namespace
133