ceph/src/tools/ceph-diff-sorted.cc

   1 // -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
   2 // vim: ts=8 sw=2 smarttab
   3
   4 /*
   5  * diffsorted -- a utility to compute a line-by-line diff on two
   6  * sorted input files
   7  *
   8  * Copyright © 2019 Red Hat
   9  *
  10  * Author: J. Eric Ivancich
  11  *
  12  * This is free software; you can redistribute it and/or
  13  * modify it under the terms of the GNU Lesser General Public
  14  * License version 2.1, as published by the Free Software
  15  * Foundation.
  16  */
  17
  18
  19 /*
  20  * SUMMARY
  21  *
  22  * The `diffsorted` utility does a line-by-line diff on two sorted text
  23  * files and indicating lines that are in one file but not the other
  24  * using diff-style notation (although line numbers are not indicated).
  25  *
  26  * USAGE
  27  *
  28  *     rgw-diff-sorted file1.txt file2.txt
  29  *
  30  * NOTES
  31  *
  32  * Each files should have its lines in sorted order and should have no
  33  * empty lines.
  34  *
  35  * A potential input file can be sorted using the `sort` utility provided
  36  * that LANG=C to insure byte lexical order. For example:
  37  *
  38  *     LANG=C sort unsorted.txt >sorted.txt
  39  *
  40  * or:
  41  *
  42  *     export LANG=C
  43  *     sort unsorted.txt >sorted.txt
  44  *
  45  * EXIT STATUS
  46  *
  47  *     0 : files same
  48  *     1 : files different
  49  *     2 : usage problem (e.g., wrong number of command-line arguments)
  50  *     3 : problem opening input file
  51  *     4 : bad file content (e.g., unsorted order or empty lines)
  52  */
  53
  54
  55 #include <iostream>
  56 #include <fstream>
  57
  58
  59 struct FileOfLines {
  60   const char* filename;
  61   std::ifstream input;
  62   std::string this_line, prev_line;
  63   bool next_eof;
  64   bool is_eof;
  65
  66   FileOfLines(const char* _filename) :
  67     filename(_filename),
  68     input(filename),
  69     next_eof(false),
  70     is_eof(false)
  71   { }
  72
  73   void dump(const std::string& prefix) {
  74     do {
  75       std::cout << prefix << this_line << std::endl;
  76       advance();
  77     } while (!eof());
  78   }
  79
  80   bool eof() const {
  81     return is_eof;
  82   }
  83
  84   bool good() const {
  85     return input.good();
  86   }
  87
  88   void advance() {
  89     if (next_eof) {
  90       is_eof = true;
  91       return;
  92     }
  93
  94     prev_line = this_line;
  95     std::getline(input, this_line);
  96     if (this_line.empty()) {
  97       if (!input.eof()) {
  98         std::cerr << "Error: " << filename << " has an empty line." <<
  99           std::endl;
 100         exit(4);
 101       }
 102       is_eof = true;
 103       return;
 104     } else if (input.eof()) {
 105       next_eof = true;
 106     }
 107
 108     if (this_line < prev_line) {
 109       std::cerr << "Error: " << filename << " is not in sorted order; \"" <<
 110         this_line << "\" follows \"" << prev_line << "\"." << std::endl;
 111       exit(4);
 112     }
 113   }
 114
 115   const std::string line() const {
 116     return this_line;
 117   }
 118 };
 119
 120 int main(int argc, const char* argv[]) {
 121   if (argc != 3) {
 122     std::cerr << "Usage: " << argv[0] << " <file1> <file2>" << std::endl;
 123     exit(2);
 124   }
 125
 126   FileOfLines input1(argv[1]);
 127   if (!input1.good()) {
 128     std::cerr << "Error opening " << argv[1] <<
 129       "." << std::endl;
 130     exit(3);
 131   }
 132
 133   FileOfLines input2(argv[2]);
 134   if (!input2.good()) {
 135     std::cerr << "Error opening " << argv[2] <<
 136       "." << std::endl;
 137     exit(3);
 138   }
 139
 140   bool files_same = true;
 141
 142   input1.advance();
 143   input2.advance();
 144
 145   while (!input1.eof() && !input2.eof()) {
 146     if (input1.line() == input2.line()) {
 147       input1.advance();
 148       input2.advance();
 149     } else if (input1.line() < input2.line()) {
 150       files_same = false;
 151       std::cout << "< " << input1.line() << std::endl;
 152       input1.advance();
 153     } else {
 154       files_same = false;
 155       std::cout << "> " << input2.line() << std::endl;
 156       input2.advance();
 157     }
 158   }
 159
 160   if (!input1.eof()) {
 161     files_same = false;
 162     input1.dump("< ");
 163   } else if (!input2.eof()) {
 164     files_same = false;
 165     input2.dump("> ");
 166   }
 167
 168   if (files_same) {
 169     exit(0);
 170   } else {
 171     exit(1);
 172   }
 173 }