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