1 //===- split-file.cpp - Input splitting 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 // Split input into multipe parts separated by regex '^(.|//)--- ' and extract
10 // the specified part.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #include "llvm/ADT/DenseMap.h"
15 #include "llvm/ADT/StringExtras.h"
16 #include "llvm/ADT/StringRef.h"
17 #include "llvm/Support/CommandLine.h"
18 #include "llvm/Support/FileOutputBuffer.h"
19 #include "llvm/Support/FileSystem.h"
20 #include "llvm/Support/LineIterator.h"
21 #include "llvm/Support/MemoryBuffer.h"
22 #include "llvm/Support/Path.h"
23 #include "llvm/Support/ToolOutputFile.h"
24 #include "llvm/Support/WithColor.h"
25 #include <string>
26 #include <system_error>
27 
28 using namespace llvm;
29 
30 static cl::OptionCategory cat("split-file Options");
31 
32 static cl::opt<std::string> input(cl::Positional, cl::desc("filename"),
33                                   cl::cat(cat));
34 
35 static cl::opt<std::string> output(cl::Positional, cl::desc("directory"),
36                                    cl::value_desc("directory"), cl::cat(cat));
37 
38 static cl::opt<bool> noLeadingLines("no-leading-lines",
39                                     cl::desc("Don't preserve line numbers"),
40                                     cl::cat(cat));
41 
42 static StringRef toolName;
43 static int errorCount;
44 
45 LLVM_ATTRIBUTE_NORETURN static void fatal(StringRef filename,
46                                           const Twine &message) {
47   if (filename.empty())
48     WithColor::error(errs(), toolName) << message << '\n';
49   else
50     WithColor::error(errs(), toolName) << filename << ": " << message << '\n';
51   exit(1);
52 }
53 
54 static void error(StringRef filename, int64_t line, const Twine &message) {
55   ++errorCount;
56   errs() << filename << ':' << line << ": ";
57   WithColor::error(errs()) << message << '\n';
58 }
59 
60 namespace {
61 struct Part {
62   const char *begin = nullptr;
63   const char *end = nullptr;
64   int64_t leadingLines = 0;
65 };
66 } // namespace
67 
68 static int handle(MemoryBuffer &inputBuf, StringRef input) {
69   DenseMap<StringRef, Part> partToBegin;
70   StringRef lastPart, separator;
71   for (line_iterator i(inputBuf, /*SkipBlanks=*/false, '\0'); !i.is_at_eof();) {
72     const int64_t lineNo = i.line_number();
73     const StringRef line = *i++;
74     const size_t markerLen = line.startswith("//") ? 6 : 5;
75     if (!(line.size() >= markerLen &&
76           line.substr(markerLen - 4).startswith("--- ")))
77       continue;
78     separator = line.substr(0, markerLen);
79     const StringRef partName = line.substr(markerLen);
80     if (partName.empty()) {
81       error(input, lineNo, "empty part name");
82       continue;
83     }
84     if (isSpace(partName.front()) || isSpace(partName.back())) {
85       error(input, lineNo, "part name cannot have leading or trailing space");
86       continue;
87     }
88 
89     auto res = partToBegin.try_emplace(partName);
90     if (!res.second) {
91       error(input, lineNo,
92             "'" + separator + partName + "' occurs more than once");
93       continue;
94     }
95     if (!lastPart.empty())
96       partToBegin[lastPart].end = line.data();
97     Part &cur = res.first->second;
98     if (!i.is_at_eof())
99       cur.begin = i->data();
100     // If --no-leading-lines is not specified, numEmptyLines is 0. Append
101     // newlines so that the extracted part preserves line numbers.
102     cur.leadingLines = noLeadingLines ? 0 : i.line_number() - 1;
103 
104     lastPart = partName;
105   }
106   if (lastPart.empty())
107     fatal(input, "no part separator was found");
108   if (errorCount)
109     return 1;
110   partToBegin[lastPart].end = inputBuf.getBufferEnd();
111 
112   std::vector<std::unique_ptr<ToolOutputFile>> outputFiles;
113   SmallString<256> partPath;
114   for (auto &keyValue : partToBegin) {
115     partPath.clear();
116     sys::path::append(partPath, output, keyValue.first);
117     std::error_code ec =
118         sys::fs::create_directories(sys::path::parent_path(partPath));
119     if (ec)
120       fatal(input, ec.message());
121     auto f = std::make_unique<ToolOutputFile>(partPath.str(), ec,
122                                               llvm::sys::fs::OF_None);
123     if (!f)
124       fatal(input, ec.message());
125 
126     Part &part = keyValue.second;
127     for (int64_t i = 0; i != part.leadingLines; ++i)
128       (*f).os().write('\n');
129     if (part.begin)
130       (*f).os().write(part.begin, part.end - part.begin);
131     outputFiles.push_back(std::move(f));
132   }
133 
134   for (std::unique_ptr<ToolOutputFile> &outputFile : outputFiles)
135     outputFile->keep();
136   return 0;
137 }
138 
139 int main(int argc, const char **argv) {
140   toolName = sys::path::stem(argv[0]);
141   cl::HideUnrelatedOptions({&cat});
142   cl::ParseCommandLineOptions(
143       argc, argv,
144       "Split input into multiple parts separated by regex '^(.|//)--- ' and "
145       "extract the part specified by '^(.|//)--- <part>'\n",
146       nullptr,
147       /*EnvVar=*/nullptr,
148       /*LongOptionsUseDoubleDash=*/true);
149 
150   if (input.empty())
151     fatal("", "input filename is not specified");
152   if (output.empty())
153     fatal("", "output directory is not specified");
154   ErrorOr<std::unique_ptr<MemoryBuffer>> bufferOrErr =
155       MemoryBuffer::getFileOrSTDIN(input);
156   if (std::error_code ec = bufferOrErr.getError())
157     fatal(input, ec.message());
158 
159   // Delete output if it is a file or an empty directory, so that we can create
160   // a directory.
161   sys::fs::file_status status;
162   if (std::error_code ec = sys::fs::status(output, status))
163     if (ec.value() != static_cast<int>(std::errc::no_such_file_or_directory))
164       fatal(output, ec.message());
165   if (status.type() != sys::fs::file_type::file_not_found &&
166       status.type() != sys::fs::file_type::directory_file &&
167       status.type() != sys::fs::file_type::regular_file)
168     fatal(output, "output cannot be a special file");
169   if (std::error_code ec = sys::fs::remove(output, /*IgnoreNonExisting=*/true))
170     if (ec.value() != static_cast<int>(std::errc::directory_not_empty) &&
171         ec.value() != static_cast<int>(std::errc::file_exists))
172       fatal(output, ec.message());
173   return handle(**bufferOrErr, input);
174 }
175