xapian-applications/omega/diritor.cc

   1 /* diritor.cc: Iterator through entries in a directory.
   2  *
   3  * Copyright (C) 2007,2008,2010,2011,2012,2013,2014 Olly Betts
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License
  16  * along with this program; if not, write to the Free Software
  17  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301 USA
  18  */
  19
  20 #include <config.h>
  21
  22 #include "diritor.h"
  23
  24 #include "safeunistd.h"
  25 #include <sys/types.h>
  26
  27 #include <cerrno>
  28 #include <cstring>
  29
  30 using namespace std;
  31
  32 CommitAndExit::CommitAndExit(const char * msg_, const std::string & path,
  33                              int errno_)
  34 {
  35     msg = msg_;
  36     msg += " \"";
  37     msg += path;
  38     msg += "\" (";
  39     msg += strerror(errno_);
  40     msg += ")";
  41 }
  42
  43 CommitAndExit::CommitAndExit(const char * msg_, int errno_)
  44 {
  45     msg = msg_;
  46     msg += " (";
  47     msg += strerror(errno_);
  48     msg += ")";
  49 }
  50
  51 CommitAndExit::CommitAndExit(const char * msg_, const char * error)
  52 {
  53     msg = msg_;
  54     msg += " (";
  55     msg += error;
  56     msg += ")";
  57 }
  58
  59 #if defined O_NOATIME && O_NOATIME != 0
  60 uid_t DirectoryIterator::euid = geteuid();
  61 #endif
  62
  63 magic_t DirectoryIterator::magic_cookie = NULL;
  64
  65 void
  66 DirectoryIterator::call_stat()
  67 {
  68     build_path();
  69     int retval;
  70     if (fd >= 0) {
  71         retval = fstat(fd, &statbuf);
  72 #ifdef HAVE_LSTAT
  73     } else if (!follow_symlinks) {
  74         retval = lstat(path.c_str(), &statbuf);
  75 #endif
  76     } else {
  77         retval = stat(path.c_str(), &statbuf);
  78     }
  79     if (retval == -1) {
  80         if (errno == ENOENT || errno == ENOTDIR)
  81             throw FileNotFound();
  82         if (errno == EACCES)
  83             throw string(strerror(errno));
  84         // Commit changes to files processed so far.
  85         throw CommitAndExit("Can't stat", path, errno);
  86     }
  87 }
  88
  89 void
  90 DirectoryIterator::build_path()
  91 {
  92     if (path.length() == path_len) {
  93         path += '/';
  94         path += leafname();
  95     }
  96 }
  97
  98 void
  99 DirectoryIterator::start(const std::string & path_)
 100 {
 101     if (dir) closedir(dir);
 102     path = path_;
 103     path_len = path.length();
 104     dir = opendir(path.c_str());
 105     if (dir == NULL) {
 106         if (errno == ENOENT || errno == ENOTDIR)
 107             throw FileNotFound();
 108         if (errno == EACCES)
 109             throw string(strerror(errno));
 110         // Commit changes to files processed so far.
 111         throw CommitAndExit("Can't open directory", path, errno);
 112     }
 113 }
 114
 115 void
 116 DirectoryIterator::next_failed() const
 117 {
 118     // The Linux getdents() syscall (which readdir uses internally) is
 119     // documented as being able to return ENOENT and ENOTDIR.  Also,
 120     // EACCES has been observed here on CIFS mounts.
 121     if (errno == ENOENT || errno == ENOTDIR)
 122         throw FileNotFound();
 123     if (errno == EACCES)
 124         throw string(strerror(errno));
 125     throw CommitAndExit("Can't read next entry from directory", path, errno);
 126 }
 127
 128 string
 129 DirectoryIterator::get_magic_mimetype()
 130 {
 131     if (rare(magic_cookie == NULL)) {
 132 #ifdef MAGIC_MIME_TYPE
 133         magic_cookie = magic_open(MAGIC_SYMLINK|MAGIC_MIME_TYPE|MAGIC_ERROR);
 134 #else
 135         // MAGIC_MIME_TYPE was added in 4.22, released 2007-12-27.  If we don't
 136         // have it then use MAGIC_MIME instead and trim any encoding off below.
 137         magic_cookie = magic_open(MAGIC_SYMLINK|MAGIC_MIME|MAGIC_ERROR);
 138 #endif
 139         if (magic_cookie == NULL) {
 140             // Commit changes to files processed so far.
 141             throw CommitAndExit("Failed to initialise the file magic library",
 142                                 errno);
 143         }
 144         if (magic_load(magic_cookie, NULL) == -1) {
 145             // Commit changes to files processed so far.
 146             const char * err = magic_error(magic_cookie);
 147             throw CommitAndExit("Failed to load the file magic database", err);
 148         }
 149     }
 150
 151     const char * res = NULL;
 152     // Prior to 5.15, magic_descriptor() closed the fd passed, so avoid it.
 153 #if defined MAGIC_VERSION && MAGIC_VERSION - 0 >= 515
 154     if (fd >= 0) {
 155         if (lseek(fd, 0, SEEK_SET) == 0)
 156             res = magic_descriptor(magic_cookie, fd);
 157     } else
 158 #endif
 159     {
 160         build_path();
 161         res = magic_file(magic_cookie, path.c_str());
 162     }
 163     if (!res) {
 164         const char * err = magic_error(magic_cookie);
 165         if (rare(err)) {
 166             int eno = magic_errno(magic_cookie);
 167             if (eno == ENOENT || eno == ENOTDIR)
 168                 throw FileNotFound();
 169             string m("Failed to use magic on file: ");
 170             m += err;
 171             throw m;
 172         }
 173         return string();
 174     }
 175
 176     // Sometimes libmagic returns this string instead of a mime-type for some
 177     // Microsoft documents, so pick a suitable MIME content-type based on the
 178     // extension.  Newer versions seem to return "application/CDFV2-corrupt"
 179     // instead for this case (on Debian, file 5.11 gives the former and file
 180     // 5.18 the latter).
 181 #define COMPOSITE_DOC "Composite Document File V2 Document"
 182     if (strncmp(res, COMPOSITE_DOC, sizeof(COMPOSITE_DOC) - 1) == 0 ||
 183         strcmp(res, "application/CDFV2-corrupt") == 0) {
 184         // Default to something self-explanatory.
 185         res = "application/x-compound-document-file";
 186         const char * leaf = leafname();
 187         const char * ext = strrchr(leaf, '.');
 188         if (ext && strlen(++ext) == 3) {
 189             char e[3];
 190             for (int i = 0; i != 3; ++i) {
 191                 if (ext[i] <= 'Z' && ext[i] >= 'A')
 192                     e[i] = ext[i] + ('a' - 'A');
 193                 else
 194                     e[i] = ext[i];
 195             }
 196             switch (e[0]) {
 197                 case 'd':
 198                     if (e[1] == 'o')
 199                         res = "application/msword";
 200                     break;
 201                 case 'm':
 202                     if (e[1] == 's' && e[2] == 'g')
 203                         res = "application/vnd.ms-outlook";
 204                     break;
 205                 case 'p':
 206                     if (e[1] == 'p' || e[1] == 'o')
 207                         res = "application/vnd.ms-powerpoint";
 208                     else if (e[1] == 'u' && e[2] == 'b')
 209                         res = "application/x-mspublisher";
 210                     break;
 211                 case 'x':
 212                     if (e[1] == 'l')
 213                         res = "application/vnd.ms-excel";
 214                     break;
 215                 case 'w':
 216                     if (e[1] == 'p' && e[2] != 'd')
 217                         res = "application/vnd.ms-works";
 218                     break;
 219             }
 220         }
 221     } else {
 222 #ifndef MAGIC_MIME_TYPE
 223         // Discard any encoding returned.
 224         char * spc = strchr(res, ' ');
 225         if (spc)
 226             *spc = '\0';
 227 #endif
 228     }
 229
 230     return res;
 231 }