| 1 |
/* |
| 2 |
$Id: file.c,v 1.43 2002/07/09 16:14:21 whmoseley Exp $ |
| 3 |
** |
| 4 |
** Copyright (C) 1995, 1996, 1997, 1998 Hewlett-Packard Company |
| 5 |
** Originally by Kevin Hughes, kev@kevcom.com, 3/11/94 |
| 6 |
** |
| 7 |
** This program and library is free software; you can redistribute it and/or |
| 8 |
** modify it under the terms of the GNU (Library) General Public License |
| 9 |
** as published by the Free Software Foundation; either version 2 |
| 10 |
** of the License, or any later version. |
| 11 |
** |
| 12 |
** This program is distributed in the hope that it will be useful, |
| 13 |
** but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 14 |
** MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 15 |
** GNU (Library) General Public License for more details. |
| 16 |
** |
| 17 |
** You should have received a copy of the GNU (Library) General Public License |
| 18 |
** along with this program; if not, write to the Free Software |
| 19 |
** Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. |
| 20 |
**------------------------------------------------------------- |
| 21 |
** Changed getdefaults to allow metaNames in the user |
| 22 |
** configuration file |
| 23 |
** G.Hill 4/16/97 ghill@library.berkeley.edu |
| 24 |
** |
| 25 |
** change sprintf to snprintf to avoid corruption, and use MAXSTRLEN from swish.h |
| 26 |
** added safestrcpy() macro to avoid corruption from strcpy overflow |
| 27 |
** SRE 11/17/99 |
| 28 |
** |
| 29 |
** added buffer size arg to grabStringValue - core dumping from overrun |
| 30 |
** fixed logical OR and other problems pointed out by "gcc -Wall" |
| 31 |
** SRE 2/22/00 |
| 32 |
** |
| 33 |
** counter modulo 128 had parens typo |
| 34 |
** SRE 2/23/00 |
| 35 |
** |
| 36 |
** read stopwords from file |
| 37 |
** Rainer Scherg (rasc) 2000-06-15 |
| 38 |
** |
| 39 |
** 2000-11-15 rasc |
| 40 |
** file_properties retrieves last mod date, filesize, and evals some swish |
| 41 |
** config flags for this file! |
| 42 |
** |
| 43 |
** 2001-02-12 rasc errormsg "print" changed... |
| 44 |
** 2001-03-16 rasc truncateDoc [read_stream] (if doc to large, truncate... ) |
| 45 |
** 2001-03-17 rasc fprop enhanced by "real_filename" |
| 46 |
** |
| 47 |
*/ |
| 48 |
|
| 49 |
#ifdef HAVE_CONFIG_H |
| 50 |
#include "acconfig.h" |
| 51 |
#endif |
| 52 |
|
| 53 |
#ifdef HAVE_STDLIB_H |
| 54 |
#include <stdlib.h> |
| 55 |
#endif |
| 56 |
#ifdef HAVE_UNISTD_H |
| 57 |
#include <unistd.h> |
| 58 |
#endif |
| 59 |
#include "swish.h" |
| 60 |
#include "mem.h" |
| 61 |
#include "string.h" |
| 62 |
#include "file.h" |
| 63 |
#include "error.h" |
| 64 |
#include "list.h" |
| 65 |
#include "hash.h" |
| 66 |
#include "check.h" |
| 67 |
#include "index.h" |
| 68 |
#include "filter.h" |
| 69 |
#include "metanames.h" |
| 70 |
|
| 71 |
|
| 72 |
/* Cough, hack, cough - convert slash to backslash for programs that are run via the shell */ |
| 73 |
#ifdef _WIN32 |
| 74 |
void make_windows_path( char *path ) |
| 75 |
{ |
| 76 |
char *c; |
| 77 |
|
| 78 |
for ( c = path; *c; c++ ) |
| 79 |
if ( '/' == *c ) |
| 80 |
*c = '\\'; |
| 81 |
} |
| 82 |
#endif |
| 83 |
|
| 84 |
/* Flip any backslashes to forward slashes, and remove trailing slash */ |
| 85 |
|
| 86 |
|
| 87 |
void normalize_path(char *path) |
| 88 |
{ |
| 89 |
int len = strlen( path ); |
| 90 |
char *c; |
| 91 |
|
| 92 |
/* For windows users */ |
| 93 |
for ( c = path; *c; c++ ) |
| 94 |
if ( '\\' == *c ) |
| 95 |
*c = '/'; |
| 96 |
|
| 97 |
while( len > 1 && path[len-1] == '/' ) |
| 98 |
{ |
| 99 |
path[len-1] = '\0'; |
| 100 |
len--; |
| 101 |
} |
| 102 |
} |
| 103 |
|
| 104 |
|
| 105 |
|
| 106 |
/* Is a file a directory? |
| 107 |
*/ |
| 108 |
|
| 109 |
int isdirectory(char *path) |
| 110 |
{ |
| 111 |
struct stat stbuf; |
| 112 |
|
| 113 |
if (stat(path, &stbuf)) |
| 114 |
return 0; |
| 115 |
return ((stbuf.st_mode & S_IFMT) == S_IFDIR) ? 1 : 0; |
| 116 |
} |
| 117 |
|
| 118 |
/* Is a file a regular file? |
| 119 |
*/ |
| 120 |
|
| 121 |
int isfile(char *path) |
| 122 |
{ |
| 123 |
struct stat stbuf; |
| 124 |
|
| 125 |
if (stat(path, &stbuf)) |
| 126 |
return 0; |
| 127 |
return ((stbuf.st_mode & S_IFMT) == S_IFREG) ? 1 : 0; |
| 128 |
} |
| 129 |
|
| 130 |
/* Is a file a link? |
| 131 |
*/ |
| 132 |
|
| 133 |
int islink(char *path) |
| 134 |
{ |
| 135 |
#ifndef NO_SYMBOLIC_FILE_LINKS |
| 136 |
struct stat stbuf; |
| 137 |
|
| 138 |
if (lstat(path, &stbuf)) |
| 139 |
return 0; |
| 140 |
return ((stbuf.st_mode & S_IFLNK) == S_IFLNK) ? 1 : 0; |
| 141 |
#else |
| 142 |
return 0; |
| 143 |
#endif |
| 144 |
} |
| 145 |
|
| 146 |
/* Get the size, in bytes, of a file. |
| 147 |
** Return -1 if there's a problem. |
| 148 |
*/ |
| 149 |
|
| 150 |
int getsize(char *path) |
| 151 |
{ |
| 152 |
struct stat stbuf; |
| 153 |
|
| 154 |
if (stat(path, &stbuf)) |
| 155 |
return -1; |
| 156 |
return stbuf.st_size; |
| 157 |
} |
| 158 |
|
| 159 |
|
| 160 |
|
| 161 |
FILE *openIndexFILEForWrite(char *filename) |
| 162 |
{ |
| 163 |
return fopen(filename, F_WRITE_BINARY); |
| 164 |
} |
| 165 |
|
| 166 |
FILE *openIndexFILEForRead(char *filename) |
| 167 |
{ |
| 168 |
return fopen(filename, F_READ_BINARY); |
| 169 |
} |
| 170 |
|
| 171 |
FILE *openIndexFILEForReadAndWrite(char *filename) |
| 172 |
{ |
| 173 |
return fopen(filename, F_READWRITE_BINARY); |
| 174 |
} |
| 175 |
|
| 176 |
void CreateEmptyFile(char *filename) |
| 177 |
{ |
| 178 |
FILE *fp; |
| 179 |
|
| 180 |
if (!(fp = openIndexFILEForWrite(filename))) |
| 181 |
{ |
| 182 |
progerrno("Couldn't write the file \"%s\": ", filename); |
| 183 |
} |
| 184 |
fclose(fp); |
| 185 |
} |
| 186 |
|
| 187 |
/* |
| 188 |
* Invoke the methods of the current Indexing Data Source |
| 189 |
*/ |
| 190 |
void indexpath(SWISH * sw, char *path) |
| 191 |
{ |
| 192 |
/* invoke routine to index a "path" */ |
| 193 |
(*IndexingDataSource->indexpath_fn) (sw, path); |
| 194 |
} |
| 195 |
|
| 196 |
|
| 197 |
/* |
| 198 |
-- read file into a buffer |
| 199 |
-- truncate file if necessary (truncateDocSize) |
| 200 |
-- return: buffer |
| 201 |
-- 2001-03-16 rasc truncateDoc |
| 202 |
*/ |
| 203 |
|
| 204 |
/* maybe some day this could be chunked reading? */ |
| 205 |
|
| 206 |
char *read_stream(SWISH *sw, char *name, FILE * fp, long filelen, long max_size) |
| 207 |
{ |
| 208 |
long c, |
| 209 |
offset; |
| 210 |
long bufferlen; |
| 211 |
unsigned char *buffer, *tmp = NULL; |
| 212 |
size_t bytes_read; |
| 213 |
|
| 214 |
|
| 215 |
if (filelen) |
| 216 |
{ |
| 217 |
|
| 218 |
/* truncate doc? */ |
| 219 |
if (max_size && (filelen > max_size)) |
| 220 |
{ |
| 221 |
filelen = max_size; |
| 222 |
} |
| 223 |
|
| 224 |
buffer = (unsigned char *)Mem_ZoneAlloc(sw->Index->perDocTmpZone, filelen + 1); |
| 225 |
*buffer = '\0'; |
| 226 |
bytes_read = fread(buffer, 1, filelen, fp); |
| 227 |
|
| 228 |
|
| 229 |
buffer[filelen] = '\0'; |
| 230 |
|
| 231 |
/* JFP - substitute null chars, VFC record may have null char in reclen word, try to discard them */ |
| 232 |
if ( strlen( (char *)buffer ) < bytes_read ) |
| 233 |
{ |
| 234 |
int i; |
| 235 |
progwarn("Substituted possible embedded null character(s) in file '%s'\n", name); |
| 236 |
for (i = 0; i < bytes_read; ++i) |
| 237 |
if (buffer[i] == '\0') buffer[i] = '\n'; |
| 238 |
} |
| 239 |
|
| 240 |
} |
| 241 |
else |
| 242 |
{ /* if we are reading from a popen call, filelen is 0 */ |
| 243 |
|
| 244 |
buffer = (unsigned char *)Mem_ZoneAlloc(sw->Index->perDocTmpZone,(bufferlen = RD_BUFFER_SIZE) + 1); |
| 245 |
*buffer = '\0'; |
| 246 |
for (offset = 0; (c = fread(buffer + offset, 1, RD_BUFFER_SIZE, fp)) == RD_BUFFER_SIZE; offset += RD_BUFFER_SIZE) |
| 247 |
{ |
| 248 |
/* truncate? break if to much read */ |
| 249 |
if (max_size && (bufferlen > max_size)) |
| 250 |
{ |
| 251 |
break; |
| 252 |
} |
| 253 |
tmp = (unsigned char *)Mem_ZoneAlloc(sw->Index->perDocTmpZone, bufferlen + RD_BUFFER_SIZE + 1); |
| 254 |
memcpy(tmp,buffer,bufferlen+1); |
| 255 |
buffer = tmp; |
| 256 |
bufferlen += RD_BUFFER_SIZE; |
| 257 |
} |
| 258 |
filelen = offset + c; |
| 259 |
|
| 260 |
if (max_size && (filelen > max_size)) |
| 261 |
{ |
| 262 |
filelen = max_size; |
| 263 |
} |
| 264 |
buffer[filelen] = '\0'; |
| 265 |
} |
| 266 |
return (char *) buffer; |
| 267 |
} |
| 268 |
|
| 269 |
/* Sept 25, 2001 - moseley |
| 270 |
* Flush the file -- for use with -S prog, when either Truncate is in use, or |
| 271 |
* the parser aborted for some reason (e.g. !isoktitle). |
| 272 |
*/ |
| 273 |
|
| 274 |
void flush_stream( FileProp *fprop ) |
| 275 |
{ |
| 276 |
char tmpbuf[4096]; |
| 277 |
int read; |
| 278 |
|
| 279 |
while ( fprop->bytes_read < fprop->fsize ) |
| 280 |
{ |
| 281 |
if ( ( fprop->fsize - fprop->bytes_read ) > 4096 ) |
| 282 |
{ |
| 283 |
if ( !(read = fread(tmpbuf, 1, 4096, fprop->fp))) |
| 284 |
break; |
| 285 |
|
| 286 |
fprop->bytes_read += read; |
| 287 |
} |
| 288 |
else |
| 289 |
{ |
| 290 |
read = fread(tmpbuf, 1, fprop->fsize - fprop->bytes_read, fprop->fp); |
| 291 |
break; |
| 292 |
} |
| 293 |
} |
| 294 |
} |
| 295 |
|
| 296 |
|
| 297 |
/* Mar 27, 2001 - moseley |
| 298 |
* Separate out the creation of the file properties |
| 299 |
* |
| 300 |
*/ |
| 301 |
|
| 302 |
FileProp *init_file_properties(SWISH * sw) |
| 303 |
{ |
| 304 |
FileProp *fprop; |
| 305 |
|
| 306 |
fprop = (FileProp *) emalloc(sizeof(FileProp)); |
| 307 |
/* emalloc checks fail and aborts... */ |
| 308 |
|
| 309 |
memset( fprop, 0, sizeof(FileProp) ); |
| 310 |
|
| 311 |
return fprop; |
| 312 |
} |
| 313 |
|
| 314 |
|
| 315 |
/* Mar 27, 2001 - moseley |
| 316 |
* Separate out the adjusting of file properties by config settings |
| 317 |
* 2001-04-09 rasc changed filters |
| 318 |
*/ |
| 319 |
|
| 320 |
void init_file_prop_settings(SWISH * sw, FileProp * fprop) |
| 321 |
{ |
| 322 |
|
| 323 |
/* Basename of document path => document filename */ |
| 324 |
fprop->real_filename = str_basename(fprop->real_path); |
| 325 |
|
| 326 |
|
| 327 |
/* -- get Doc Type as is in IndexContents or Defaultcontents |
| 328 |
-- doctypes by jruiz |
| 329 |
*/ |
| 330 |
|
| 331 |
/* Might already be set by a header in extpro.c */ |
| 332 |
if ( !fprop->doctype ) |
| 333 |
{ |
| 334 |
/* Get the type by file extension -- or return NODOCTYPE */ |
| 335 |
fprop->doctype = getdoctype(fprop->real_path, sw->indexcontents); |
| 336 |
|
| 337 |
/* If was not set by getdoctype() then assign it the default parser */ |
| 338 |
/* This could still be NODOCTYPE, or it might be something set by DefaultContents */ |
| 339 |
|
| 340 |
if (fprop->doctype == NODOCTYPE) |
| 341 |
fprop->doctype = sw->DefaultDocType; |
| 342 |
} |
| 343 |
|
| 344 |
|
| 345 |
/* -- index just the filename (or doc title tags)? |
| 346 |
-- this param was "wrongly" named indextitleonly */ |
| 347 |
|
| 348 |
fprop->index_no_content = (sw->nocontentslist != NULL) && isoksuffix(fprop->real_path, sw->nocontentslist); |
| 349 |
|
| 350 |
/* -- Any filter for this file type? |
| 351 |
-- NULL = No Filter, (char *) path to filter prog. |
| 352 |
*/ |
| 353 |
|
| 354 |
fprop->hasfilter = hasfilter(sw, fprop->real_path); |
| 355 |
|
| 356 |
fprop->stordesc = hasdescription(fprop->doctype, sw->storedescription); |
| 357 |
|
| 358 |
} |
| 359 |
|
| 360 |
|
| 361 |
|
| 362 |
/* |
| 363 |
-- file_properties |
| 364 |
-- Get/eval information about a file and return it. |
| 365 |
-- Some flags are calculated from swish configs for this "real_path" |
| 366 |
-- Structure has to be freed using free_file_properties |
| 367 |
-- 2000-11-15 rasc |
| 368 |
-- return: (FileProp *) |
| 369 |
-- A failed stat returns an empty (default) structure |
| 370 |
|
| 371 |
-- 2000-12 |
| 372 |
-- Added StoreDescription |
| 373 |
*/ |
| 374 |
|
| 375 |
FileProp *file_properties(char *real_path, char *work_file, SWISH * sw) |
| 376 |
{ |
| 377 |
FileProp *fprop; |
| 378 |
struct stat stbuf; |
| 379 |
|
| 380 |
/* create an initilized fprop structure */ |
| 381 |
|
| 382 |
fprop = init_file_properties(sw); |
| 383 |
|
| 384 |
|
| 385 |
/* Dup these, since the real_path may be reallocated by FileRules */ |
| 386 |
fprop->real_path = estrdup( real_path ); |
| 387 |
fprop->work_path = estrdup( work_file ? work_file : real_path ); |
| 388 |
fprop->orig_path = estrdup( real_path ); |
| 389 |
|
| 390 |
|
| 391 |
/* Stat the file */ |
| 392 |
/* This is really the wrong place for this, as it's really only useful for fs.c method */ |
| 393 |
/* for http.c it means the last mod date is the temp file date */ |
| 394 |
/* Probably this entire function isn't needed - moseley */ |
| 395 |
|
| 396 |
if (!stat(fprop->work_path, &stbuf)) |
| 397 |
{ |
| 398 |
fprop->fsize = (long) stbuf.st_size; |
| 399 |
fprop->mtime = stbuf.st_mtime; |
| 400 |
} |
| 401 |
|
| 402 |
|
| 403 |
/* Now set various fprop settings based mostly on file name */ |
| 404 |
|
| 405 |
init_file_prop_settings(sw, fprop); |
| 406 |
|
| 407 |
|
| 408 |
|
| 409 |
#ifdef DEBUG |
| 410 |
fprintf(stderr, "file_properties: path=%s, (workpath=%s), fsize=%ld, last_mod=%ld Doctype: %d Filter: %p\n", |
| 411 |
fprop->real_path, fprop->work_path, (long) fprop->fsize, (long) fprop->mtime, fprop->doctype, fprop->filterprog); |
| 412 |
#endif |
| 413 |
|
| 414 |
return fprop; |
| 415 |
} |
| 416 |
|
| 417 |
|
| 418 |
/* -- Free FileProp structure |
| 419 |
-- unless no alloc for strings simple free structure |
| 420 |
*/ |
| 421 |
|
| 422 |
void free_file_properties(FileProp * fprop) |
| 423 |
{ |
| 424 |
efree( fprop->real_path ); |
| 425 |
efree( fprop->work_path ); |
| 426 |
efree( fprop->orig_path ); |
| 427 |
efree(fprop); |
| 428 |
} |
| 429 |
|
| 430 |
|
| 431 |
static char *temp_file_template = "XXXXXX"; |
| 432 |
/*********************************************************************** |
| 433 |
* Create a temporary file |
| 434 |
* |
| 435 |
* Call With: |
| 436 |
* *SWISH = to get at the TmpDir config setting which I don't like |
| 437 |
* *prefix = chars to prepend to the file name |
| 438 |
* **file_name_buffer = where to store address of file name |
| 439 |
* unlink = if true, will unlink file |
| 440 |
* if not unlinked, then caller must free the name |
| 441 |
* Return: |
| 442 |
* *FILE |
| 443 |
* modified file_name_buffer |
| 444 |
* |
| 445 |
* Will create temp files in the directory specified by environment vars |
| 446 |
* TMPDIR and TMP, and by the config.h setting of TMPDIR in that order. |
| 447 |
* |
| 448 |
* Note: |
| 449 |
* It's expected that swish is not run suid, so |
| 450 |
* (getuid()==geteuid()) && (getgid()==getegid()) |
| 451 |
* if not checked. I'm not sure if that would choke on other platforms. |
| 452 |
* |
| 453 |
* |
| 454 |
* Source: |
| 455 |
* http://www.linuxdoc.org/HOWTO/Secure-Programs-HOWTO/avoid-race.html |
| 456 |
* |
| 457 |
* Questions: |
| 458 |
* Can non-unix OS unlink the file and continue to hold the fd? |
| 459 |
* |
| 460 |
***********************************************************************/ |
| 461 |
|
| 462 |
FILE *create_tempfile(SWISH *sw, const char *f_mode, char *prefix, char **file_name_buffer, int remove_file_name ) |
| 463 |
{ |
| 464 |
int temp_fd; |
| 465 |
mode_t old_mode; |
| 466 |
FILE *temp_file; |
| 467 |
char *file_name; |
| 468 |
int file_name_len; |
| 469 |
struct MOD_Index *idx = sw->Index; |
| 470 |
char *tmpdir = NULL; |
| 471 |
file_name_len = (prefix ? strlen(prefix) : 0) + strlen( temp_file_template ) + strlen( TEMP_FILE_PREFIX ); |
| 472 |
|
| 473 |
|
| 474 |
|
| 475 |
/* Perl is nice sometimes */ |
| 476 |
if ( !( tmpdir = getenv("TMPDIR")) ) |
| 477 |
if ( !(tmpdir = getenv("TMP")) ) |
| 478 |
if( !(tmpdir = getenv("TEMP")) ) |
| 479 |
tmpdir = idx->tmpdir; |
| 480 |
|
| 481 |
if ( tmpdir && !*tmpdir ) |
| 482 |
tmpdir = NULL; // just in case it's the empty string |
| 483 |
|
| 484 |
if ( tmpdir ) |
| 485 |
file_name_len += strlen( tmpdir ) + 1; // for path separator |
| 486 |
|
| 487 |
|
| 488 |
|
| 489 |
file_name = emalloc( file_name_len + 1 ); |
| 490 |
|
| 491 |
*file_name = '\0'; |
| 492 |
|
| 493 |
if ( tmpdir ) |
| 494 |
{ |
| 495 |
strcat( file_name, tmpdir ); |
| 496 |
normalize_path( file_name ); |
| 497 |
strcat( file_name, "/" ); |
| 498 |
} |
| 499 |
|
| 500 |
strcat( file_name, TEMP_FILE_PREFIX ); |
| 501 |
|
| 502 |
if ( prefix ) |
| 503 |
strcat( file_name, prefix ); |
| 504 |
|
| 505 |
strcat( file_name, temp_file_template ); |
| 506 |
|
| 507 |
old_mode = umask(077); /* Create file with restrictive permissions */ |
| 508 |
|
| 509 |
temp_fd = mkstemp( file_name ); |
| 510 |
|
| 511 |
(void) umask(old_mode); |
| 512 |
|
| 513 |
if (temp_fd == -1) |
| 514 |
progerrno("Couldn't open temporary file '%s': ", file_name ); |
| 515 |
|
| 516 |
if (!(temp_file = fdopen(temp_fd, f_mode))) |
| 517 |
progerrno("Couldn't create temporary file '%s' file descriptor: ", file_name); |
| 518 |
|
| 519 |
if ( remove_file_name ) |
| 520 |
{ |
| 521 |
if ( remove( file_name ) == -1 ) |
| 522 |
progerrno("Couldn't unlink temporary file '%s' :", file_name); |
| 523 |
|
| 524 |
efree( file_name ); |
| 525 |
} |
| 526 |
else |
| 527 |
*file_name_buffer = file_name; |
| 528 |
|
| 529 |
|
| 530 |
return temp_file; |
| 531 |
} |
| 532 |
|
| 533 |
|
Parent Directory
| 
Revision Log
| 
Revision Graph