contrib/diffutils/src/dir.c

   1 /* Read, sort and compare two directories.  Used for GNU DIFF.
   2
   3    Copyright (C) 1988-1989, 1992-1995, 1998, 2001-2002, 2004, 2006-2007,
   4    2009-2013, 2015-2018 Free Software Foundation, Inc.
   5
   6    This file is part of GNU DIFF.
   7
   8    This program is free software: you can redistribute it and/or modify
   9    it under the terms of the GNU General Public License as published by
  10    the Free Software Foundation, either version 3 of the License, or
  11    (at your option) any later version.
  12
  13    This program is distributed in the hope that it will be useful,
  14    but WITHOUT ANY WARRANTY; without even the implied warranty of
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16    GNU General Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License
  19    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  20
  21 #include "diff.h"
  22 #include <error.h>
  23 #include <exclude.h>
  24 #include <filenamecat.h>
  25 #include <setjmp.h>
  26 #include <xalloc.h>
  27
  28 /* Read the directory named by DIR and store into DIRDATA a sorted vector
  29    of filenames for its contents.  DIR->desc == -1 means this directory is
  30    known to be nonexistent, so set DIRDATA to an empty vector.
  31    Return -1 (setting errno) if error, 0 otherwise.  */
  32
  33 struct dirdata
  34 {
  35   size_t nnames;        /* Number of names.  */
  36   char const **names;   /* Sorted names of files in dir, followed by 0.  */
  37   char *data;   /* Allocated storage for file names.  */
  38 };
  39
  40 /* Whether file names in directories should be compared with
  41    locale-specific sorting.  */
  42 static bool locale_specific_sorting;
  43
  44 /* Where to go if locale-specific sorting fails.  */
  45 static jmp_buf failed_locale_specific_sorting;
  46
  47 static bool dir_loop (struct comparison const *, int);
  48
  49
  50 /* Read a directory and get its vector of names.  */
  51
  52 static bool
  53 dir_read (struct file_data const *dir, struct dirdata *dirdata)
  54 {
  55   register struct dirent *next;
  56   register size_t i;
  57
  58   /* Address of block containing the files that are described.  */
  59   char const **names;
  60
  61   /* Number of files in directory.  */
  62   size_t nnames;
  63
  64   /* Allocated and used storage for file name data.  */
  65   char *data;
  66   size_t data_alloc, data_used;
  67
  68   dirdata->names = 0;
  69   dirdata->data = 0;
  70   nnames = 0;
  71   data = 0;
  72
  73   if (dir->desc != -1)
  74     {
  75       /* Open the directory and check for errors.  */
  76       register DIR *reading = opendir (dir->name);
  77       if (!reading)
  78         return false;
  79
  80       /* Initialize the table of filenames.  */
  81
  82       data_alloc = 512;
  83       data_used = 0;
  84       dirdata->data = data = xmalloc (data_alloc);
  85
  86       /* Read the directory entries, and insert the subfiles
  87          into the 'data' table.  */
  88
  89       while ((errno = 0, (next = readdir (reading)) != 0))
  90         {
  91           char *d_name = next->d_name;
  92           size_t d_size = _D_EXACT_NAMLEN (next) + 1;
  93
  94           /* Ignore "." and "..".  */
  95           if (d_name[0] == '.'
  96               && (d_name[1] == 0 || (d_name[1] == '.' && d_name[2] == 0)))
  97             continue;
  98
  99           if (excluded_file_name (excluded, d_name))
 100             continue;
 101
 102           while (data_alloc < data_used + d_size)
 103             {
 104               if (PTRDIFF_MAX / 2 <= data_alloc)
 105                 xalloc_die ();
 106               dirdata->data = data = xrealloc (data, data_alloc *= 2);
 107             }
 108
 109           memcpy (data + data_used, d_name, d_size);
 110           data_used += d_size;
 111           nnames++;
 112         }
 113       if (errno)
 114         {
 115           int e = errno;
 116           closedir (reading);
 117           errno = e;
 118           return false;
 119         }
 120 #if CLOSEDIR_VOID
 121       closedir (reading);
 122 #else
 123       if (closedir (reading) != 0)
 124         return false;
 125 #endif
 126     }
 127
 128   /* Create the 'names' table from the 'data' table.  */
 129   if (PTRDIFF_MAX / sizeof *names - 1 <= nnames)
 130     xalloc_die ();
 131   dirdata->names = names = xmalloc ((nnames + 1) * sizeof *names);
 132   dirdata->nnames = nnames;
 133   for (i = 0;  i < nnames;  i++)
 134     {
 135       names[i] = data;
 136       data += strlen (data) + 1;
 137     }
 138   names[nnames] = 0;
 139   return true;
 140 }
 141
 142 /* Compare strings in a locale-specific way, returning a value
 143    compatible with strcmp.  */
 144
 145 static int
 146 compare_collated (char const *name1, char const *name2)
 147 {
 148   int r;
 149   errno = 0;
 150   if (ignore_file_name_case)
 151     r = strcasecoll (name1, name2);
 152   else
 153     r = strcoll (name1, name2);
 154   if (errno)
 155     {
 156       error (0, errno, _("cannot compare file names '%s' and '%s'"),
 157              name1, name2);
 158       longjmp (failed_locale_specific_sorting, 1);
 159     }
 160   return r;
 161 }
 162
 163 /* Compare file names, returning a value compatible with strcmp.  */
 164
 165 static int
 166 compare_names (char const *name1, char const *name2)
 167 {
 168   if (locale_specific_sorting)
 169     {
 170       int diff = compare_collated (name1, name2);
 171       if (diff || ignore_file_name_case)
 172         return diff;
 173     }
 174   return file_name_cmp (name1, name2);
 175 }
 176
 177 /* Compare names FILE1 and FILE2 when sorting a directory.
 178    Prefer filtered comparison, breaking ties with file_name_cmp.  */
 179
 180 static int
 181 compare_names_for_qsort (void const *file1, void const *file2)
 182 {
 183   char const *const *f1 = file1;
 184   char const *const *f2 = file2;
 185   char const *name1 = *f1;
 186   char const *name2 = *f2;
 187   if (locale_specific_sorting)
 188     {
 189       int diff = compare_collated (name1, name2);
 190       if (diff)
 191         return diff;
 192     }
 193   return file_name_cmp (name1, name2);
 194 }
 195 \f
 196 /* Compare the contents of two directories named in CMP.
 197    This is a top-level routine; it does everything necessary for diff
 198    on two directories.
 199
 200    CMP->file[0].desc == -1 says directory CMP->file[0] doesn't exist,
 201    but pretend it is empty.  Likewise for CMP->file[1].
 202
 203    HANDLE_FILE is a caller-provided subroutine called to handle each file.
 204    It gets three operands: CMP, name of file in dir 0, name of file in dir 1.
 205    These names are relative to the original working directory.
 206
 207    For a file that appears in only one of the dirs, one of the name-args
 208    to HANDLE_FILE is zero.
 209
 210    Returns the maximum of all the values returned by HANDLE_FILE,
 211    or EXIT_TROUBLE if trouble is encountered in opening files.  */
 212
 213 int
 214 diff_dirs (struct comparison const *cmp,
 215            int (*handle_file) (struct comparison const *,
 216                                char const *, char const *))
 217 {
 218   struct dirdata dirdata[2];
 219   int volatile val = EXIT_SUCCESS;
 220   int i;
 221
 222   if ((cmp->file[0].desc == -1 || dir_loop (cmp, 0))
 223       && (cmp->file[1].desc == -1 || dir_loop (cmp, 1)))
 224     {
 225       error (0, 0, _("%s: recursive directory loop"),
 226              cmp->file[cmp->file[0].desc == -1].name);
 227       return EXIT_TROUBLE;
 228     }
 229
 230   /* Get contents of both dirs.  */
 231   for (i = 0; i < 2; i++)
 232     if (! dir_read (&cmp->file[i], &dirdata[i]))
 233       {
 234         perror_with_name (cmp->file[i].name);
 235         val = EXIT_TROUBLE;
 236       }
 237
 238   if (val == EXIT_SUCCESS)
 239     {
 240       char const **volatile names[2];
 241       names[0] = dirdata[0].names;
 242       names[1] = dirdata[1].names;
 243
 244       /* Use locale-specific sorting if possible, else native byte order.  */
 245       locale_specific_sorting = true;
 246       if (setjmp (failed_locale_specific_sorting))
 247         locale_specific_sorting = false;
 248
 249       /* Sort the directories.  */
 250       for (i = 0; i < 2; i++)
 251         qsort (names[i], dirdata[i].nnames, sizeof *dirdata[i].names,
 252                compare_names_for_qsort);
 253
 254       /* If '-S name' was given, and this is the topmost level of comparison,
 255          ignore all file names less than the specified starting name.  */
 256
 257       if (starting_file && ! cmp->parent)
 258         {
 259           while (*names[0] && compare_names (*names[0], starting_file) < 0)
 260             names[0]++;
 261           while (*names[1] && compare_names (*names[1], starting_file) < 0)
 262             names[1]++;
 263         }
 264
 265       /* Loop while files remain in one or both dirs.  */
 266       while (*names[0] || *names[1])
 267         {
 268           /* Compare next name in dir 0 with next name in dir 1.
 269              At the end of a dir,
 270              pretend the "next name" in that dir is very large.  */
 271           int nameorder = (!*names[0] ? 1 : !*names[1] ? -1
 272                            : compare_names (*names[0], *names[1]));
 273
 274           /* Prefer a file_name_cmp match if available.  This algorithm is
 275              O(N**2), where N is the number of names in a directory
 276              that compare_names says are all equal, but in practice N
 277              is so small it's not worth tuning.  */
 278           if (nameorder == 0 && ignore_file_name_case)
 279             {
 280               int raw_order = file_name_cmp (*names[0], *names[1]);
 281               if (raw_order != 0)
 282                 {
 283                   int greater_side = raw_order < 0;
 284                   int lesser_side = 1 - greater_side;
 285                   char const **lesser = names[lesser_side];
 286                   char const *greater_name = *names[greater_side];
 287                   char const **p;
 288
 289                   for (p = lesser + 1;
 290                        *p && compare_names (*p, greater_name) == 0;
 291                        p++)
 292                     {
 293                       int c = file_name_cmp (*p, greater_name);
 294                       if (0 <= c)
 295                         {
 296                           if (c == 0)
 297                             {
 298                               memmove (lesser + 1, lesser,
 299                                        (char *) p - (char *) lesser);
 300                               *lesser = greater_name;
 301                             }
 302                           break;
 303                         }
 304                     }
 305                 }
 306             }
 307
 308           int v1 = (*handle_file) (cmp,
 309                                    0 < nameorder ? 0 : *names[0]++,
 310                                    nameorder < 0 ? 0 : *names[1]++);
 311           if (val < v1)
 312             val = v1;
 313         }
 314     }
 315
 316   for (i = 0; i < 2; i++)
 317     {
 318       free (dirdata[i].names);
 319       free (dirdata[i].data);
 320     }
 321
 322   return val;
 323 }
 324
 325 /* Return nonzero if CMP is looping recursively in argument I.  */
 326
 327 static bool _GL_ATTRIBUTE_PURE
 328 dir_loop (struct comparison const *cmp, int i)
 329 {
 330   struct comparison const *p = cmp;
 331   while ((p = p->parent))
 332     if (0 < same_file (&p->file[i].stat, &cmp->file[i].stat))
 333       return true;
 334   return false;
 335 }
 336
 337 /* Find a matching filename in a directory.  */
 338
 339 char *
 340 find_dir_file_pathname (char const *dir, char const *file)
 341 {
 342   /* The 'IF_LINT (volatile)' works around what appears to be a bug in
 343      gcc 4.8.0 20120825; see
 344      <http://lists.gnu.org/archive/html/bug-diffutils/2012-08/msg00007.html>.
 345      */
 346   char const * IF_LINT (volatile) match = file;
 347
 348   char *val;
 349   struct dirdata dirdata;
 350   dirdata.names = NULL;
 351   dirdata.data = NULL;
 352
 353   if (ignore_file_name_case)
 354     {
 355       struct file_data filedata;
 356       filedata.name = dir;
 357       filedata.desc = 0;
 358
 359       if (dir_read (&filedata, &dirdata))
 360         {
 361           locale_specific_sorting = true;
 362           if (setjmp (failed_locale_specific_sorting))
 363             match = file; /* longjmp may mess up MATCH.  */
 364           else
 365             {
 366               for (char const **p = dirdata.names; *p; p++)
 367                 if (compare_names (*p, file) == 0)
 368                   {
 369                     if (file_name_cmp (*p, file) == 0)
 370                       {
 371                         match = *p;
 372                         break;
 373                       }
 374                     if (match == file)
 375                       match = *p;
 376                   }
 377             }
 378         }
 379     }
 380
 381   val = file_name_concat (dir, match, NULL);
 382   free (dirdata.names);
 383   free (dirdata.data);
 384   return val;
 385 }