1 //===-- llvm-strings.cpp - Printable String dumping utility ---------------===//
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 // This program is a utility that works like binutils "strings", that is, it
10 // prints out printable strings in a binary, objdump, or archive file.
11 //
12 //===----------------------------------------------------------------------===//
13
14 #include "Opts.inc"
15 #include "llvm/ADT/StringExtras.h"
16 #include "llvm/Object/Binary.h"
17 #include "llvm/Option/Arg.h"
18 #include "llvm/Option/ArgList.h"
19 #include "llvm/Option/Option.h"
20 #include "llvm/Support/CommandLine.h"
21 #include "llvm/Support/Error.h"
22 #include "llvm/Support/Format.h"
23 #include "llvm/Support/InitLLVM.h"
24 #include "llvm/Support/MemoryBuffer.h"
25 #include "llvm/Support/Program.h"
26 #include "llvm/Support/WithColor.h"
27 #include <cctype>
28 #include <string>
29
30 using namespace llvm;
31 using namespace llvm::object;
32
33 namespace {
34 enum ID {
35 OPT_INVALID = 0, // This is not an option ID.
36 #define OPTION(...) LLVM_MAKE_OPT_ID(__VA_ARGS__),
37 #include "Opts.inc"
38 #undef OPTION
39 };
40
41 #define OPTTABLE_STR_TABLE_CODE
42 #include "Opts.inc"
43 #undef OPTTABLE_STR_TABLE_CODE
44
45 #define OPTTABLE_PREFIXES_TABLE_CODE
46 #include "Opts.inc"
47 #undef OPTTABLE_PREFIXES_TABLE_CODE
48
49 using namespace llvm::opt;
50 static constexpr opt::OptTable::Info InfoTable[] = {
51 #define OPTION(...) LLVM_CONSTRUCT_OPT_INFO(__VA_ARGS__),
52 #include "Opts.inc"
53 #undef OPTION
54 };
55
56 class StringsOptTable : public opt::GenericOptTable {
57 public:
StringsOptTable()58 StringsOptTable()
59 : GenericOptTable(OptionStrTable, OptionPrefixesTable, InfoTable) {
60 setGroupedShortOptions(true);
61 setDashDashParsing(true);
62 }
63 };
64 } // namespace
65
66 static StringRef ToolName;
67
68 static cl::list<std::string> InputFileNames(cl::Positional,
69 cl::desc("<input object files>"));
70
71 static int MinLength = 4;
72 static bool PrintFileName;
73
74 enum radix { none, octal, hexadecimal, decimal };
75 static radix Radix;
76
reportCmdLineError(const Twine & Message)77 [[noreturn]] static void reportCmdLineError(const Twine &Message) {
78 WithColor::error(errs(), ToolName) << Message << "\n";
79 exit(1);
80 }
81
82 template <typename T>
parseIntArg(const opt::InputArgList & Args,int ID,T & Value)83 static void parseIntArg(const opt::InputArgList &Args, int ID, T &Value) {
84 if (const opt::Arg *A = Args.getLastArg(ID)) {
85 StringRef V(A->getValue());
86 if (!llvm::to_integer(V, Value, 0) || Value <= 0)
87 reportCmdLineError("expected a positive integer, but got '" + V + "'");
88 }
89 }
90
strings(raw_ostream & OS,StringRef FileName,StringRef Contents)91 static void strings(raw_ostream &OS, StringRef FileName, StringRef Contents) {
92 auto print = [&OS, FileName](unsigned Offset, StringRef L) {
93 if (L.size() < static_cast<size_t>(MinLength))
94 return;
95 if (PrintFileName)
96 OS << FileName << ": ";
97 switch (Radix) {
98 case none:
99 break;
100 case octal:
101 OS << format("%7o ", Offset);
102 break;
103 case hexadecimal:
104 OS << format("%7x ", Offset);
105 break;
106 case decimal:
107 OS << format("%7u ", Offset);
108 break;
109 }
110 OS << L << '\n';
111 };
112
113 const char *B = Contents.begin();
114 const char *P = nullptr, *E = nullptr, *S = nullptr;
115 for (P = Contents.begin(), E = Contents.end(); P < E; ++P) {
116 if (isPrint(*P) || *P == '\t') {
117 if (S == nullptr)
118 S = P;
119 } else if (S) {
120 print(S - B, StringRef(S, P - S));
121 S = nullptr;
122 }
123 }
124 if (S)
125 print(S - B, StringRef(S, E - S));
126 }
127
main(int argc,char ** argv)128 int main(int argc, char **argv) {
129 InitLLVM X(argc, argv);
130 BumpPtrAllocator A;
131 StringSaver Saver(A);
132 StringsOptTable Tbl;
133 ToolName = argv[0];
134 opt::InputArgList Args =
135 Tbl.parseArgs(argc, argv, OPT_UNKNOWN, Saver,
136 [&](StringRef Msg) { reportCmdLineError(Msg); });
137 if (Args.hasArg(OPT_help)) {
138 Tbl.printHelp(
139 outs(),
140 (Twine(ToolName) + " [options] <input object files>").str().c_str(),
141 "llvm string dumper");
142 // TODO Replace this with OptTable API once it adds extrahelp support.
143 outs() << "\nPass @FILE as argument to read options from FILE.\n";
144 return 0;
145 }
146 if (Args.hasArg(OPT_version)) {
147 outs() << ToolName << '\n';
148 cl::PrintVersionMessage();
149 return 0;
150 }
151
152 parseIntArg(Args, OPT_bytes_EQ, MinLength);
153 PrintFileName = Args.hasArg(OPT_print_file_name);
154 StringRef R = Args.getLastArgValue(OPT_radix_EQ);
155 if (R.empty())
156 Radix = none;
157 else if (R == "o")
158 Radix = octal;
159 else if (R == "d")
160 Radix = decimal;
161 else if (R == "x")
162 Radix = hexadecimal;
163 else
164 reportCmdLineError("--radix value should be one of: '' (no offset), 'o' "
165 "(octal), 'd' (decimal), 'x' (hexadecimal)");
166
167 if (MinLength == 0) {
168 errs() << "invalid minimum string length 0\n";
169 return EXIT_FAILURE;
170 }
171
172 std::vector<std::string> InputFileNames = Args.getAllArgValues(OPT_INPUT);
173 if (InputFileNames.empty())
174 InputFileNames.push_back("-");
175
176 for (const auto &File : InputFileNames) {
177 ErrorOr<std::unique_ptr<MemoryBuffer>> Buffer =
178 MemoryBuffer::getFileOrSTDIN(File, /*IsText=*/true);
179 if (std::error_code EC = Buffer.getError())
180 errs() << File << ": " << EC.message() << '\n';
181 else
182 strings(llvm::outs(), File == "-" ? "{standard input}" : File,
183 Buffer.get()->getMemBufferRef().getBuffer());
184 }
185
186 return EXIT_SUCCESS;
187 }
188