xref: /llvm-project/lldb/unittests/Core/DumpDataExtractorTest.cpp (revision 8fdfc1d64c51d949b84b2ee824f2ba9b9a3a6bd6)
1 //===-- DataDumpExtractorTest.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 "lldb/Core/DumpDataExtractor.h"
10 #include "lldb/Utility/DataBufferHeap.h"
11 #include "lldb/Utility/DataExtractor.h"
12 #include "lldb/Utility/Endian.h"
13 #include "lldb/Utility/StreamString.h"
14 #include "gtest/gtest.h"
15 #include <complex>
16 
17 using namespace lldb;
18 using namespace lldb_private;
19 
20 static void test_format_impl(const void *data, size_t data_size,
21                              size_t item_count, lldb::Format format,
22                              llvm::StringRef expected) {
23   StreamString result;
24   DataBufferHeap dumpbuffer(data, data_size);
25   DataExtractor extractor(dumpbuffer.GetBytes(), dumpbuffer.GetByteSize(),
26                           endian::InlHostByteOrder(),
27                           /*addr_size=*/4);
28   DumpDataExtractor(extractor, &result, 0, format, data_size, item_count, 1, 0,
29                     0, 0);
30   ASSERT_EQ(expected, result.GetString());
31 }
32 
33 template <typename T>
34 static void test_format(T data, lldb::Format format, llvm::StringRef expected) {
35   test_format_impl(&data, sizeof(T), 1, format, expected);
36 }
37 
38 static void test_format(llvm::StringRef str, lldb::Format format,
39                         llvm::StringRef expected) {
40   test_format_impl(str.bytes_begin(),
41                    // +1 to include the NULL char as the last byte
42                    str.size() + 1, 1, format, expected);
43 }
44 
45 template <typename T>
46 static void test_format(const std::vector<T> data, lldb::Format format,
47                         llvm::StringRef expected) {
48   test_format_impl(&data[0], data.size() * sizeof(T), data.size(), format,
49                    expected);
50 }
51 
52 TEST(DumpDataExtractorTest, Formats) {
53   test_format<uint8_t>(1, lldb::eFormatDefault, "0x00000000: 0x01");
54   test_format<uint8_t>(1, lldb::eFormatBoolean, "0x00000000: true");
55   test_format<uint8_t>(0xAA, lldb::eFormatBinary, "0x00000000: 0b10101010");
56   test_format<uint8_t>(1, lldb::eFormatBytes, "0x00000000: 01");
57   test_format<uint8_t>(1, lldb::eFormatBytesWithASCII, "0x00000000: 01  .");
58   test_format('?', lldb::eFormatChar, "0x00000000: '?'");
59   test_format('\x1A', lldb::eFormatCharPrintable, "0x00000000: .");
60   test_format('#', lldb::eFormatCharPrintable, "0x00000000: #");
61   test_format(std::complex<float>(1.2, 3.4), lldb::eFormatComplex,
62               "0x00000000: 1.2 + 3.4i");
63   test_format(std::complex<double>(4.5, 6.7), lldb::eFormatComplex,
64               "0x00000000: 4.5 + 6.7i");
65 
66   // long double is not tested here because for some platforms we treat it as 10
67   // bytes when the compiler allocates 16 bytes of space for it. (see
68   // DataExtractor::GetLongDouble) Meaning that when we extract the second one,
69   // it gets the wrong value (it's 6 bytes off). You could manually construct a
70   // set of bytes to match the 10 byte format but then if the test runs on a
71   // machine where we don't use 10 it'll break.
72 
73   test_format(llvm::StringRef("aardvark"), lldb::Format::eFormatCString,
74               "0x00000000: \"aardvark\"");
75   test_format<uint16_t>(99, lldb::Format::eFormatDecimal, "0x00000000: 99");
76   // Just prints as a signed integer.
77   test_format(-1, lldb::Format::eFormatEnum, "0x00000000: -1");
78   test_format(0xcafef00d, lldb::Format::eFormatHex, "0x00000000: 0xcafef00d");
79   test_format(0xcafef00d, lldb::Format::eFormatHexUppercase,
80               "0x00000000: 0xCAFEF00D");
81   test_format(0.456, lldb::Format::eFormatFloat, "0x00000000: 0.456");
82   test_format(9, lldb::Format::eFormatOctal, "0x00000000: 011");
83   // Chars packed into an integer.
84   test_format<uint32_t>(0x4C4C4442, lldb::Format::eFormatOSType,
85                         "0x00000000: 'LLDB'");
86   // Unicode8 doesn't have a specific formatter.
87   test_format<uint8_t>(0x34, lldb::Format::eFormatUnicode8, "0x00000000: 0x34");
88   test_format<uint16_t>(0x1122, lldb::Format::eFormatUnicode16,
89                         "0x00000000: U+1122");
90   test_format<uint32_t>(0x12345678, lldb::Format::eFormatUnicode32,
91                         "0x00000000: U+0x12345678");
92   test_format<unsigned int>(654321, lldb::Format::eFormatUnsigned,
93                             "0x00000000: 654321");
94   // This pointer is printed based on the size of uint64_t, so the test is the
95   // same for 32/64 bit host.
96   test_format<uint64_t>(0x4444555566667777, lldb::Format::eFormatPointer,
97                         "0x00000000: 0x4444555566667777");
98 
99   test_format(std::vector<char>{'A', '\x01', 'C'},
100               lldb::Format::eFormatVectorOfChar, "0x00000000: {A\\x01C}");
101   test_format(std::vector<int8_t>{0, -1, std::numeric_limits<int8_t>::max()},
102               lldb::Format::eFormatVectorOfSInt8, "0x00000000: {0 -1 127}");
103   test_format(std::vector<uint8_t>{12, 0xFF, 34},
104               lldb::Format::eFormatVectorOfUInt8,
105               "0x00000000: {0x0c 0xff 0x22}");
106   test_format(
107       std::vector<int16_t>{-1, 1234, std::numeric_limits<int16_t>::max()},
108       lldb::Format::eFormatVectorOfSInt16, "0x00000000: {-1 1234 32767}");
109   test_format(std::vector<uint16_t>{0xffff, 0xabcd, 0x1234},
110               lldb::Format::eFormatVectorOfUInt16,
111               "0x00000000: {0xffff 0xabcd 0x1234}");
112   test_format(std::vector<int32_t>{0, -1, std::numeric_limits<int32_t>::max()},
113               lldb::Format::eFormatVectorOfSInt32,
114               "0x00000000: {0 -1 2147483647}");
115   test_format(std::vector<uint32_t>{0, 0xffffffff, 0x1234abcd},
116               lldb::Format::eFormatVectorOfUInt32,
117               "0x00000000: {0x00000000 0xffffffff 0x1234abcd}");
118   test_format(std::vector<int64_t>{0, -1, std::numeric_limits<int64_t>::max()},
119               lldb::Format::eFormatVectorOfSInt64,
120               "0x00000000: {0 -1 9223372036854775807}");
121   test_format(std::vector<uint64_t>{0, 0xaaaabbbbccccdddd},
122               lldb::Format::eFormatVectorOfUInt64,
123               "0x00000000: {0x0000000000000000 0xaaaabbbbccccdddd}");
124 
125   // See half2float for format details.
126   test_format(std::vector<uint16_t>{0xabcd, 0x1234},
127               lldb::Format::eFormatVectorOfFloat16,
128               "0x00000000: {-0.0609436 0.000757217}");
129   test_format(std::vector<float>{std::numeric_limits<float>::min(),
130                                  std::numeric_limits<float>::max()},
131               lldb::Format::eFormatVectorOfFloat32,
132               "0x00000000: {1.17549e-38 3.40282e+38}");
133   test_format(std::vector<double>{std::numeric_limits<double>::min(),
134                                   std::numeric_limits<double>::max()},
135               lldb::Format::eFormatVectorOfFloat64,
136               "0x00000000: {2.2250738585072e-308 1.79769313486232e+308}");
137 
138   // Not sure we can rely on having uint128_t everywhere so emulate with
139   // uint64_t.
140   test_format(
141       std::vector<uint64_t>{0x1, 0x1111222233334444, 0xaaaabbbbccccdddd, 0x0},
142       lldb::Format::eFormatVectorOfUInt128,
143       "0x00000000: {0x11112222333344440000000000000001 "
144       "0x0000000000000000aaaabbbbccccdddd}");
145 
146   test_format(std::vector<int>{2, 4}, lldb::Format::eFormatComplexInteger,
147               "0x00000000: 2 + 4i");
148 
149   // Without an execution context this just prints the pointer on its own.
150   test_format<uint32_t>(0x11223344, lldb::Format::eFormatAddressInfo,
151                         "0x00000000: 0x11223344");
152 
153   // Input not written in hex form because that requires C++17.
154   test_format<float>(10, lldb::Format::eFormatHexFloat, "0x00000000: 0x1.4p3");
155 
156   // Can't disassemble without an execution context.
157   test_format<uint32_t>(0xcafef00d, lldb::Format::eFormatInstruction,
158                         "invalid target");
159 
160   // Has no special handling, intended for use elsewhere.
161   test_format<int>(99, lldb::Format::eFormatVoid, "0x00000000: 0x00000063");
162 }
163 
164 TEST(DumpDataExtractorTest, FormatCharArray) {
165   // Unlike the other formats, charArray isn't 1 array of N chars.
166   // It must be passed as N chars of 1 byte each.
167   // (eFormatVectorOfChar does this swap for you)
168   std::vector<char> data{'A', '\x01', '#'};
169   StreamString result;
170   DataBufferHeap dumpbuffer(&data[0], data.size());
171   DataExtractor extractor(dumpbuffer.GetBytes(), dumpbuffer.GetByteSize(),
172                           endian::InlHostByteOrder(), /*addr_size=*/4);
173 
174   DumpDataExtractor(extractor, &result, 0, lldb::Format::eFormatCharArray,
175                     /*item_byte_size=*/1,
176                     /*item_count=*/data.size(),
177                     /*num_per_line=*/data.size(), 0, 0, 0);
178   ASSERT_EQ("0x00000000: A\\x01#", result.GetString());
179 
180   result.Clear();
181   DumpDataExtractor(extractor, &result, 0, lldb::Format::eFormatCharArray, 1,
182                     data.size(), 1, 0, 0, 0);
183   // ASSERT macro thinks the split strings are multiple arguments so make a var.
184   const char *expected = "0x00000000: A\n"
185                          "0x00000001: \\x01\n"
186                          "0x00000002: #";
187   ASSERT_EQ(expected, result.GetString());
188 }
189