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, 2015-2025 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
  23 #include <diagnose.h>
  24 #include <dirname.h>
  25 #include <error.h>
  26 #include <exclude.h>
  27 #include <filenamecat.h>
  28 #include <mcel.h>
  29 #include <quote.h>
  30 #include <setjmp.h>
  31 #include <xalloc.h>
  32
  33 #ifndef HAVE_STRUCT_DIRENT_D_TYPE
  34 # define HAVE_STRUCT_DIRENT_D_TYPE false
  35 #endif
  36
  37 /* A sorted vector of file names obtained by reading a directory.
  38    If HAVE_STRUCT_DIRENT_D_TYPE, each name is preceded by a byte
  39    giving the file type as an enum detype value.  */
  40
  41 struct dirdata
  42 {
  43   idx_t nnames; /* Number of names.  */
  44   char const **names;   /* Sorted names of files in dir, followed by 0.  */
  45   char *data;   /* Allocated storage for file names.  */
  46 };
  47
  48 /* Whether file names in directories should be compared with
  49    locale-specific sorting.  */
  50 static bool locale_specific_sorting;
  51
  52 /* Where to go if locale-specific sorting fails.  */
  53 static jmp_buf failed_locale_specific_sorting;
  54
  55 static int compare_names (char const *, char const *);
  56 static bool dir_loop (struct comparison const *, int);
  57
  58
  59 /* Given the parent directory PARENTDIRFD (negative for current dir),
  60    read the directory named by DIR and store into DIRDATA a sorted
  61    vector of filenames for its contents.
  62    Use DIR's basename if PARENTDIRFD is nonnegative, for efficiency.
  63    If DIR->desc == NONEXISTENT, this directory is known to be
  64    nonexistent so set DIRDATA to an empty vector;
  65    otherwise, update DIR->desc and DIR->dirstream as needed.
  66    If STARTFILE, ignore directory entries less than STARTFILE, and if
  67    STARTFILE_ONLY, also ignore directory entries greater than STARTFILE.
  68    Return true if successful, false (setting errno) otherwise.  */
  69
  70 static bool
  71 dir_read (int parentdirfd, struct file_data *dir, struct dirdata *dirdata,
  72           char const *startfile, bool startfile_only)
  73 {
  74   /* Number of files in directory.  */
  75   idx_t nnames = 0;
  76
  77   /* Allocated and used storage for file name data.  */
  78   char *data;
  79
  80   dirdata->names = nullptr;
  81   dirdata->data = nullptr;
  82
  83   if (dir->desc != NONEXISTENT)
  84     {
  85       /* Open the directory and check for errors.  */
  86       int dirfd = dir->desc;
  87       if (dirfd < 0)
  88         {
  89           dirfd = openat (parentdirfd,
  90                           (parentdirfd < 0 ? dir->name
  91                            : last_component (dir->name)),
  92                           (O_RDONLY | O_CLOEXEC | O_DIRECTORY
  93                            | (no_dereference_symlinks ? O_NOFOLLOW : 0)));
  94           if (dirfd < 0)
  95             return false;
  96           dir->desc = dirfd;
  97         }
  98       DIR *reading = fdopendir (dirfd);
  99       if (!reading)
 100         return false;
 101       dir->dirstream = reading;
 102
 103       /* Initialize the table of filenames.  */
 104
 105       idx_t data_alloc = 512;
 106       idx_t data_used = 0;
 107       dirdata->data = data = ximalloc (data_alloc);
 108
 109       /* Read the directory entries, and insert the subfiles
 110          into the 'data' table.  */
 111
 112       while (true)
 113         {
 114           errno = 0;
 115           struct dirent *next = readdir (reading);
 116           if (!next)
 117             break;
 118
 119           char *d_name = next->d_name;
 120
 121           /* Ignore "." and "..".  */
 122           if (d_name[0] == '.'
 123               && (d_name[1] == 0 || (d_name[1] == '.' && d_name[2] == 0)))
 124             continue;
 125
 126           if (startfile)
 127             {
 128               int cmp = compare_names (d_name, startfile);
 129               if (cmp < 0 || (startfile_only && !!cmp))
 130                 continue;
 131             }
 132
 133           if (excluded_file_name (excluded, d_name))
 134             continue;
 135
 136           idx_t d_size = HAVE_STRUCT_DIRENT_D_TYPE + _D_EXACT_NAMLEN (next) + 1;
 137           if (data_alloc - data_used < d_size)
 138             dirdata->data = data
 139               = xpalloc (data, &data_alloc,
 140                          d_size - (data_alloc - data_used), -1, 1);
 141 #if HAVE_STRUCT_DIRENT_D_TYPE
 142           char detype;
 143           switch (next->d_type)
 144             {
 145             case DT_BLK:  detype = DE_BLK;  break;
 146             case DT_CHR:  detype = DE_CHR;  break;
 147             case DT_DIR:  detype = DE_DIR;  break;
 148             case DT_FIFO: detype = DE_FIFO; break;
 149             case DT_LNK:  detype = DE_LNK;  break;
 150             case DT_REG:  detype = DE_REG;  break;
 151             case DT_SOCK: detype = DE_SOCK; break;
 152 # ifdef DT_WHT
 153             case DT_WHT:  detype = DE_WHT;  break;
 154 # endif
 155             case DT_UNKNOWN: detype = DE_UNKNOWN; break;
 156             default:         detype = DE_OTHER;   break;
 157             }
 158           data[data_used++] = detype;
 159           d_size--;
 160 #endif
 161           memcpy (data + data_used, d_name, d_size);
 162           data_used += d_size;
 163           nnames++;
 164         }
 165
 166       if (errno)
 167         return false;
 168     }
 169
 170   /* Create the 'names' table from the 'data' table.  */
 171   char const **names = xinmalloc (nnames + 1, sizeof *names);
 172   dirdata->names = names;
 173   dirdata->nnames = nnames;
 174   for (idx_t i = 0; i < nnames; i++)
 175     {
 176       data += HAVE_STRUCT_DIRENT_D_TYPE;
 177       names[i] = data;
 178       data += strlen (data) + 1;
 179     }
 180   names[nnames] = nullptr;
 181   return true;
 182 }
 183
 184 /* Compare strings in a locale-specific way, returning a value
 185    compatible with strcmp.  */
 186
 187 static int
 188 compare_collated (char const *name1, char const *name2)
 189 {
 190   errno = 0;
 191   int r = strcoll (name1, name2);
 192   if (errno)
 193     {
 194       error (0, errno, _("cannot compare file names %s and %s"),
 195              quote_n (0, name1), quote_n (1, name2));
 196       longjmp (failed_locale_specific_sorting, 1);
 197     }
 198   return r;
 199 }
 200
 201 /* Compare file names, returning a value compatible with strcmp.  */
 202
 203 static int
 204 compare_names (char const *name1, char const *name2)
 205 {
 206   if (ignore_file_name_case)
 207     return mbscasecmp (name1, name2);  /* Best we can do.  */
 208
 209   if (locale_specific_sorting)
 210     {
 211       int diff = compare_collated (name1, name2);
 212       if (diff)
 213         return diff;
 214     }
 215   return file_name_cmp (name1, name2);
 216 }
 217
 218 /* Compare names FILE1 and FILE2 when sorting a directory.
 219    Prefer filtered comparison, breaking ties with file_name_cmp.  */
 220
 221 static int
 222 compare_names_for_qsort (void const *file1, void const *file2)
 223 {
 224   char const *const *f1 = file1;
 225   char const *const *f2 = file2;
 226   return compare_names (*f1, *f2);
 227 }
 228 \f
 229 /* Compare the contents of two directories named in CMP.
 230    This is a top-level routine; it does everything necessary for diff
 231    on two directories.
 232
 233    If CMP->file[0].desc == NONEXISTENT, directory CMP->file[0] doesn't exist
 234    and pretend it is empty.  Otherwise, update CMP->file[0].desc and
 235    CMP->file[0].dirstream as needed.  Likewise for CMP->file[1].
 236
 237    Returns the maximum of all the values returned by compare_files,
 238    or EXIT_TROUBLE if trouble is encountered in opening files.  */
 239
 240 int
 241 diff_dirs (struct comparison *cmp)
 242 {
 243   if ((cmp->file[0].desc == NONEXISTENT || dir_loop (cmp, 0))
 244       && (cmp->file[1].desc == NONEXISTENT || dir_loop (cmp, 1)))
 245     {
 246       error (0, 0, _("%s: recursive directory loop"),
 247              squote (0, cmp->file[cmp->file[0].desc == NONEXISTENT].name));
 248       return EXIT_TROUBLE;
 249     }
 250
 251   /* Get contents of both dirs.  */
 252   struct dirdata dirdata[2];
 253   int val = EXIT_SUCCESS;
 254   for (int i = 0; i < 2; i++)
 255     if (! dir_read (cmp->parent->file[i].desc, &cmp->file[i], &dirdata[i],
 256                     cmp->parent == &noparent ? starting_file : nullptr, false))
 257       {
 258         perror_with_name (cmp->file[i].name);
 259         val = EXIT_TROUBLE;
 260       }
 261
 262   if (val == EXIT_SUCCESS)
 263     {
 264       /* Use locale-specific sorting if possible, else native byte order.  */
 265       locale_specific_sorting = true;
 266       if (! ignore_file_name_case)
 267         if (setjmp (failed_locale_specific_sorting))
 268           locale_specific_sorting = false;
 269
 270       /* Sort the directories.  */
 271       for (int i = 0; i < 2; i++)
 272         qsort (dirdata[i].names, dirdata[i].nnames, sizeof *dirdata[i].names,
 273                compare_names_for_qsort);
 274
 275       /* Loop while files remain in one or both dirs.  */
 276       char const **n0 = dirdata[0].names;
 277       char const **n1 = dirdata[1].names;
 278       while (*n0 || *n1)
 279         {
 280           /* Compare next name in dir 0 with next name in dir 1.
 281              At the end of a dir,
 282              pretend the "next name" in that dir is very large.  */
 283           int nameorder = !*n0 ? 1 : !*n1 ? -1 : compare_names (*n0, *n1);
 284
 285           /* Prefer a file_name_cmp match if available.  This algorithm is
 286              O(N**2), where N is the number of names in a directory
 287              that compare_names says are all equal, but in practice N
 288              is so small it's not worth tuning.  */
 289           if (nameorder == 0 && ignore_file_name_case)
 290             {
 291               int raw_order = file_name_cmp (*n0, *n1);
 292               if (raw_order != 0)
 293                 {
 294                   char const **lesser = raw_order < 0 ? n0 : n1;
 295                   char const *greater_name = *(raw_order < 0 ? n1 : n0);
 296
 297                   for (char const **p = lesser + 1;
 298                        *p && compare_names (*p, greater_name) == 0;
 299                        p++)
 300                     {
 301                       int c = file_name_cmp (*p, greater_name);
 302                       if (0 <= c)
 303                         {
 304                           if (c == 0)
 305                             {
 306                               memmove (lesser + 1, lesser,
 307                                        (char *) p - (char *) lesser);
 308                               *lesser = greater_name;
 309                             }
 310                           break;
 311                         }
 312                     }
 313                 }
 314             }
 315
 316           enum detype detypes[]
 317             = { HAVE_STRUCT_DIRENT_D_TYPE && *n0 ? (*n0)[-1] : DE_UNKNOWN,
 318                 HAVE_STRUCT_DIRENT_D_TYPE && *n1 ? (*n1)[-1] : DE_UNKNOWN };
 319           int v1 = compare_files (cmp, detypes,
 320                                   0 < nameorder ? nullptr : *n0++,
 321                                   nameorder < 0 ? nullptr : *n1++);
 322           if (val < v1)
 323             val = v1;
 324         }
 325     }
 326
 327   for (int i = 0; i < 2; i++)
 328     {
 329       free (dirdata[i].names);
 330       free (dirdata[i].data);
 331     }
 332
 333   return val;
 334 }
 335
 336 /* Return nonzero if CMP is looping recursively in argument I.  */
 337
 338 static bool ATTRIBUTE_PURE
 339 dir_loop (struct comparison const *cmp, int i)
 340 {
 341   for (struct comparison const *p = cmp; (p = p->parent) != &noparent; )
 342     if (same_file (&p->file[i].stat, &cmp->file[i].stat))
 343       return true;
 344   return false;
 345 }
 346
 347 /* Find a matching filename in a directory.  */
 348
 349 char *
 350 find_dir_file_pathname (struct file_data *dir, char const *file,
 351                         enum detype *detype)
 352 {
 353   char const *match = file;
 354
 355   struct dirdata dirdata;
 356   dirdata.names = nullptr;
 357   dirdata.data = nullptr;
 358
 359   if (ignore_file_name_case && dir_read (AT_FDCWD, dir, &dirdata, file, true))
 360     for (char const **p = dirdata.names; *p; p++)
 361       {
 362         if (file_name_cmp (*p, file) == 0)
 363           {
 364             match = *p;
 365             break;
 366           }
 367         if (match == file)
 368           match = *p;
 369       }
 370
 371   *detype = HAVE_STRUCT_DIRENT_D_TYPE && match != file ? match[-1] : DE_UNKNOWN;
 372   char *val = file_name_concat (dir->name, match, nullptr);
 373   free (dirdata.names);
 374   free (dirdata.data);
 375   return val;
 376 }