| 1 |
/* |
| 2 |
** Copyright (C) 1995, 1996, 1997, 1998 Hewlett-Packard Company |
| 3 |
** Originally by Kevin Hughes, kev@kevcom.com, 3/11/94 |
| 4 |
** |
| 5 |
** This program and library is free software; you can redistribute it and/or |
| 6 |
** modify it under the terms of the GNU (Library) General Public License |
| 7 |
** as published by the Free Software Foundation; either version 2 |
| 8 |
** of the License, or 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 (Library) General Public License for more details. |
| 14 |
** |
| 15 |
** You should have received a copy of the GNU (Library) General Public License |
| 16 |
** long with this program; if not, write to the Free Software |
| 17 |
** Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. |
| 18 |
**-------------------------------------------------------------------- |
| 19 |
** All the code in this file added by Ron Klachko ron@ckm.ucsf.edu 9/98 |
| 20 |
** |
| 21 |
** change sprintf to snprintf to avoid corruption, |
| 22 |
** test length of spiderdirectory before strcat to avoid corruption, |
| 23 |
** added safestrcpy() macro to avoid corruption from strcpy overflow, |
| 24 |
** define MAXPIDLEN instead of literal "32" - assumed return length from lgetpid() |
| 25 |
** SRE 11/17/99 |
| 26 |
** |
| 27 |
** added buffer size arg to grabStringValue - core dumping from overrun |
| 28 |
** SRE 2/22/00 |
| 29 |
** |
| 30 |
** 2000-11 jruiz,rasc some redesign |
| 31 |
*/ |
| 32 |
|
| 33 |
/* |
| 34 |
** http.c |
| 35 |
*/ |
| 36 |
|
| 37 |
#ifdef HAVE_CONFIG_H |
| 38 |
#include "acconfig.h" |
| 39 |
#endif |
| 40 |
|
| 41 |
#ifdef HAVE_UNISTD_H |
| 42 |
#include <unistd.h> |
| 43 |
#endif |
| 44 |
|
| 45 |
#ifdef HAVE_STDLIB_H |
| 46 |
#include <stdlib.h> |
| 47 |
#endif |
| 48 |
|
| 49 |
#ifdef HAVE_PROCESS_H |
| 50 |
#include <process.h> |
| 51 |
#endif |
| 52 |
|
| 53 |
#include <time.h> |
| 54 |
#include <stdarg.h> |
| 55 |
|
| 56 |
// for wait |
| 57 |
#ifndef _WIN32 |
| 58 |
#include <sys/types.h> |
| 59 |
#include <sys/wait.h> |
| 60 |
#endif |
| 61 |
|
| 62 |
#include "swish.h" |
| 63 |
#include "mem.h" |
| 64 |
#include "string.h" |
| 65 |
#include "index.h" |
| 66 |
#include "hash.h" |
| 67 |
#include "file.h" |
| 68 |
#include "check.h" |
| 69 |
#include "error.h" |
| 70 |
|
| 71 |
#include "http.h" |
| 72 |
#include "httpserver.h" |
| 73 |
|
| 74 |
#include "xml.h" |
| 75 |
#include "txt.h" |
| 76 |
#include "html.h" |
| 77 |
|
| 78 |
/* |
| 79 |
-- init structures for this module |
| 80 |
*/ |
| 81 |
|
| 82 |
void initModule_HTTP(SWISH * sw) |
| 83 |
{ |
| 84 |
struct MOD_HTTP *http; |
| 85 |
int i; |
| 86 |
|
| 87 |
http = (struct MOD_HTTP *) emalloc(sizeof(struct MOD_HTTP)); |
| 88 |
|
| 89 |
sw->HTTP = http; |
| 90 |
|
| 91 |
http->lenspiderdirectory = MAXSTRLEN; |
| 92 |
http->spiderdirectory = (char *) emalloc(http->lenspiderdirectory + 1); |
| 93 |
http->spiderdirectory[0] = '\0'; |
| 94 |
/* Initialize spider directory */ |
| 95 |
http->spiderdirectory = SafeStrCopy(http->spiderdirectory, SPIDERDIRECTORY, &http->lenspiderdirectory); |
| 96 |
|
| 97 |
for (i = 0; i < BIGHASHSIZE; i++) |
| 98 |
http->url_hash[i] = NULL; |
| 99 |
|
| 100 |
http->equivalentservers = NULL; |
| 101 |
|
| 102 |
/* http default system parameters */ |
| 103 |
http->maxdepth = 5; |
| 104 |
http->delay = 60; |
| 105 |
} |
| 106 |
|
| 107 |
void freeModule_HTTP(SWISH * sw) |
| 108 |
{ |
| 109 |
struct MOD_HTTP *http = sw->HTTP; |
| 110 |
|
| 111 |
if (http->spiderdirectory) |
| 112 |
efree(http->spiderdirectory); |
| 113 |
efree(http); |
| 114 |
sw->HTTP = NULL; |
| 115 |
} |
| 116 |
|
| 117 |
int configModule_HTTP(SWISH * sw, StringList * sl) |
| 118 |
{ |
| 119 |
struct MOD_HTTP *http = sw->HTTP; |
| 120 |
char *w0 = sl->word[0]; |
| 121 |
int retval = 1; |
| 122 |
|
| 123 |
int i; |
| 124 |
struct multiswline *list; |
| 125 |
struct swline *slist; |
| 126 |
|
| 127 |
if (strcasecmp(w0, "maxdepth") == 0) |
| 128 |
{ |
| 129 |
if (sl->n == 2) |
| 130 |
{ |
| 131 |
retval = 1; |
| 132 |
http->maxdepth = atoi(sl->word[1]); |
| 133 |
} |
| 134 |
else |
| 135 |
progerr("MaxDepth requires one value"); |
| 136 |
} |
| 137 |
else if (strcasecmp(w0, "delay") == 0) |
| 138 |
{ |
| 139 |
if (sl->n == 2) |
| 140 |
{ |
| 141 |
retval = 1; |
| 142 |
http->delay = atoi(sl->word[1]); |
| 143 |
} |
| 144 |
else |
| 145 |
progerr("Delay requires one value"); |
| 146 |
} |
| 147 |
else if (strcasecmp(w0, "spiderdirectory") == 0) |
| 148 |
{ |
| 149 |
if (sl->n == 2) |
| 150 |
{ |
| 151 |
retval = 1; |
| 152 |
http->spiderdirectory = erealloc( http->spiderdirectory, strlen(sl->word[1])+2); |
| 153 |
strcpy( http->spiderdirectory, sl->word[1] ); |
| 154 |
normalize_path( http->spiderdirectory ); |
| 155 |
|
| 156 |
|
| 157 |
if (!isdirectory(http->spiderdirectory)) |
| 158 |
{ |
| 159 |
progerr("SpiderDirectory. %s is not a directory", http->spiderdirectory); |
| 160 |
} |
| 161 |
|
| 162 |
if ( strlen( http->spiderdirectory ) != 1 || http->spiderdirectory[0] != '/' ) |
| 163 |
strcat(http->spiderdirectory, "/" ); /* In this case, we just add the delimiter */ |
| 164 |
|
| 165 |
} |
| 166 |
else |
| 167 |
progerr("SpiderDirectory requires one value"); |
| 168 |
} |
| 169 |
else if (strcasecmp(w0, "equivalentserver") == 0) |
| 170 |
{ |
| 171 |
if (sl->n > 1) |
| 172 |
{ |
| 173 |
retval = 1; |
| 174 |
/* Add a new list of equivalent servers */ |
| 175 |
list = (struct multiswline *) emalloc(sizeof(struct multiswline)); |
| 176 |
|
| 177 |
list->next = http->equivalentservers; |
| 178 |
list->list = 0; |
| 179 |
http->equivalentservers = list; |
| 180 |
|
| 181 |
for (i = 1; i < sl->n; i++) |
| 182 |
{ |
| 183 |
/* Add a new entry to this list */ |
| 184 |
slist = (struct swline *) emalloc(sizeof(struct swline)); |
| 185 |
|
| 186 |
slist->line = estrdup( sl->word[i] ); |
| 187 |
slist->next = list->list; |
| 188 |
list->list = slist; |
| 189 |
} |
| 190 |
|
| 191 |
} |
| 192 |
else |
| 193 |
progerr("EquivalentServers requires at least one value"); |
| 194 |
} |
| 195 |
else |
| 196 |
{ |
| 197 |
retval = 0; |
| 198 |
} |
| 199 |
|
| 200 |
return retval; |
| 201 |
} |
| 202 |
typedef struct urldepth |
| 203 |
{ |
| 204 |
char *url; |
| 205 |
int depth; |
| 206 |
struct urldepth *next; |
| 207 |
} |
| 208 |
urldepth; |
| 209 |
|
| 210 |
|
| 211 |
int http_already_indexed(SWISH * sw, char *url); |
| 212 |
urldepth *add_url(SWISH * sw, urldepth * list, char *url, int depth, char *baseurl); |
| 213 |
|
| 214 |
|
| 215 |
urldepth *add_url(SWISH * sw, urldepth * list, char *url, int depth, char *baseurl) |
| 216 |
{ |
| 217 |
urldepth *item; |
| 218 |
struct MOD_HTTP *http = sw->HTTP; |
| 219 |
|
| 220 |
|
| 221 |
if (!equivalentserver(sw, url, baseurl)) |
| 222 |
{ |
| 223 |
if (sw->verbose >= 3) |
| 224 |
printf("Skipping %s: %s\n", url, "Wrong method or server."); |
| 225 |
|
| 226 |
|
| 227 |
} |
| 228 |
else if (http->maxdepth && (depth >= http->maxdepth)) |
| 229 |
{ |
| 230 |
if (sw->verbose >= 3) |
| 231 |
printf("Skipping %s: %s\n", url, "Too deep."); |
| 232 |
} |
| 233 |
else if (sw->nocontentslist && isoksuffix(url, sw->nocontentslist)) |
| 234 |
{ |
| 235 |
if (sw->verbose >= 3) |
| 236 |
printf("Skipping %s: %s\n", url, "Wrong suffix."); |
| 237 |
|
| 238 |
} |
| 239 |
else if (urldisallowed(sw, url)) |
| 240 |
{ |
| 241 |
if (sw->verbose >= 3) |
| 242 |
printf("Skipping %s: %s\n", url, "URL disallowed by robots.txt."); |
| 243 |
} |
| 244 |
else if (!http_already_indexed(sw, url)) |
| 245 |
{ |
| 246 |
item = (urldepth *) emalloc(sizeof(urldepth)); |
| 247 |
item->url = estrdup(url); |
| 248 |
item->depth = depth; |
| 249 |
#if 0 |
| 250 |
/* Depth first searching |
| 251 |
* */ |
| 252 |
item->next = list; |
| 253 |
list = item; |
| 254 |
#else |
| 255 |
/* Breadth first searching |
| 256 |
* */ |
| 257 |
item->next = 0; |
| 258 |
if (!list) |
| 259 |
{ |
| 260 |
list = item; |
| 261 |
} |
| 262 |
else |
| 263 |
{ |
| 264 |
urldepth *walk; |
| 265 |
|
| 266 |
for (walk = list; walk->next; walk = walk->next) |
| 267 |
{ |
| 268 |
} |
| 269 |
walk->next = item; |
| 270 |
} |
| 271 |
#endif |
| 272 |
} |
| 273 |
|
| 274 |
return list; |
| 275 |
} |
| 276 |
|
| 277 |
|
| 278 |
/* Have we already indexed a file or directory? |
| 279 |
** This function is used to avoid multiple index entries |
| 280 |
** or endless looping due to symbolic links. |
| 281 |
*/ |
| 282 |
|
| 283 |
int http_already_indexed(SWISH * sw, char *url) |
| 284 |
{ |
| 285 |
struct url_info *p; |
| 286 |
|
| 287 |
int len; |
| 288 |
unsigned hashval; |
| 289 |
struct MOD_HTTP *http = sw->HTTP; |
| 290 |
|
| 291 |
/* Hash with via the uri alone. Depending on the equivalent |
| 292 |
** servers, we may or may not make the decision of the entire |
| 293 |
** url or just the uri. |
| 294 |
*/ |
| 295 |
hashval = bighash(url_uri(url, &len)); /* Search hash for this file. */ |
| 296 |
for (p = http->url_hash[hashval]; p != NULL; p = p->next) |
| 297 |
if ((strcmp(url, p->url) == 0) || (equivalentserver(sw, url, p->url) && (strcmp(url_uri(url, &len), url_uri(p->url, &len)) == 0))) |
| 298 |
{ /* We found it. */ |
| 299 |
if (sw->verbose >= 3) |
| 300 |
printf("Skipping %s: %s\n", url, "Already indexed."); |
| 301 |
return 1; |
| 302 |
} |
| 303 |
|
| 304 |
/* Not found, make new entry. */ |
| 305 |
p = (struct url_info *) emalloc(sizeof(struct url_info)); |
| 306 |
|
| 307 |
p->url = estrdup(url); |
| 308 |
p->next = http->url_hash[hashval]; |
| 309 |
http->url_hash[hashval] = p; |
| 310 |
|
| 311 |
return 0; |
| 312 |
} |
| 313 |
|
| 314 |
|
| 315 |
char *url_method(char *url, int *plen) |
| 316 |
{ |
| 317 |
char *end; |
| 318 |
|
| 319 |
if ((end = strstr(url, "://")) == NULL) |
| 320 |
{ |
| 321 |
return NULL; |
| 322 |
} |
| 323 |
*plen = end - url; |
| 324 |
return url; |
| 325 |
} |
| 326 |
|
| 327 |
|
| 328 |
char *url_serverport(char *url, int *plen) |
| 329 |
{ |
| 330 |
int methodlen; |
| 331 |
char *serverstart; |
| 332 |
char *serverend; |
| 333 |
|
| 334 |
if (url_method(url, &methodlen) == NULL) |
| 335 |
{ |
| 336 |
return NULL; |
| 337 |
} |
| 338 |
|
| 339 |
/* +3 for |
| 340 |
* */ |
| 341 |
serverstart = url + methodlen + 3; |
| 342 |
if ((serverend = strchr(serverstart, '/')) == NULL) |
| 343 |
{ |
| 344 |
*plen = strlen(serverstart); |
| 345 |
} |
| 346 |
else |
| 347 |
{ |
| 348 |
*plen = serverend - serverstart; |
| 349 |
} |
| 350 |
|
| 351 |
return serverstart; |
| 352 |
} |
| 353 |
|
| 354 |
|
| 355 |
char *url_uri(char *url, int *plen) |
| 356 |
{ |
| 357 |
if ((url = url_serverport(url, plen)) == 0) |
| 358 |
{ |
| 359 |
return 0; |
| 360 |
} |
| 361 |
url += *plen; |
| 362 |
*plen = strlen(url); |
| 363 |
return url; |
| 364 |
} |
| 365 |
/************************************************************ |
| 366 |
* |
| 367 |
* Fork and exec a program, and wait for child to exit. |
| 368 |
* Returns |
| 369 |
* |
| 370 |
*************************************************************/ |
| 371 |
#ifndef _WIN32 |
| 372 |
static void run_program(char* prog, char** args) |
| 373 |
{ |
| 374 |
pid_t pid = fork(); |
| 375 |
int status; |
| 376 |
|
| 377 |
/* In parent, wait for child */ |
| 378 |
if ( pid ) |
| 379 |
{ |
| 380 |
wait( &status ); |
| 381 |
if ( WIFEXITED( status ) ) // exited normally if non-zero |
| 382 |
return; |
| 383 |
|
| 384 |
progerr("%s exited with non-zero status (%d)", prog, WEXITSTATUS(status) ); |
| 385 |
} |
| 386 |
|
| 387 |
execvp (prog, args); |
| 388 |
progerrno("Failed to fork '%s'. Error: ", prog ); |
| 389 |
} |
| 390 |
#endif |
| 391 |
|
| 392 |
/************************************************************ |
| 393 |
* |
| 394 |
* Fetch a URL |
| 395 |
* Side effect that it appends to "response_file" |
| 396 |
* -- lazy programmer hoping that -S http will go away... |
| 397 |
* |
| 398 |
* Under Windows system() is used to call "perl" |
| 399 |
* Otherwise, exec is called on the swishspider program |
| 400 |
* |
| 401 |
*************************************************************/ |
| 402 |
|
| 403 |
int get(SWISH * sw, char *contenttype_or_redirect, time_t *last_modified, time_t * plastretrieval, char *file_prefix, char *url) |
| 404 |
{ |
| 405 |
int code = 500; |
| 406 |
FILE *fp; |
| 407 |
struct MOD_HTTP *http = sw->HTTP; |
| 408 |
|
| 409 |
/* Build path to swishspider program */ |
| 410 |
char *spider_prog = emalloc( strlen(http->spiderdirectory) + strlen("swishspider+fill") ); |
| 411 |
sprintf(spider_prog, "%sswishspider", http->spiderdirectory ); // note that spiderdir MUST be set. |
| 412 |
|
| 413 |
|
| 414 |
/* Sleep a little so we don't overwhelm the server */ |
| 415 |
if ((time(0) - *plastretrieval) < http->delay) |
| 416 |
{ |
| 417 |
int num_sec = http->delay - (time(0) - *plastretrieval); |
| 418 |
sleep(num_sec); |
| 419 |
} |
| 420 |
|
| 421 |
*plastretrieval = time(0); |
| 422 |
|
| 423 |
|
| 424 |
#ifdef _WIN32 |
| 425 |
/* Should be in autoconf or obsoleted by extprog. - DLN 2001-11-05 */ |
| 426 |
{ |
| 427 |
int retval; |
| 428 |
char commandline[] = "perl %s %s \"%s\""; |
| 429 |
char *command = emalloc( strlen(commandline) + strlen(spider_prog) + strlen(file_prefix) + strlen(url) + 1 ); |
| 430 |
|
| 431 |
sprintf(command, commandline, spider_prog, file_prefix, url); |
| 432 |
|
| 433 |
retval = system( command ); |
| 434 |
efree( command ); |
| 435 |
efree( spider_prog ); |
| 436 |
|
| 437 |
printf("Returned %d\n", retval ); |
| 438 |
|
| 439 |
if ( retval ) |
| 440 |
return 500; |
| 441 |
} |
| 442 |
#else |
| 443 |
{ |
| 444 |
char *args[4]; |
| 445 |
|
| 446 |
args[0] = spider_prog; |
| 447 |
args[1] = file_prefix; |
| 448 |
args[2] = url; |
| 449 |
args[3] = NULL; |
| 450 |
run_program( spider_prog, args ); |
| 451 |
efree( spider_prog ); |
| 452 |
} |
| 453 |
#endif |
| 454 |
|
| 455 |
|
| 456 |
/* NAUGHTY SIDE EFFECT */ |
| 457 |
strcat( file_prefix, ".response" ); |
| 458 |
|
| 459 |
if ( !(fp = fopen(file_prefix, F_READ_TEXT)) ) |
| 460 |
{ |
| 461 |
progerrno("Failed to open file '%s': ", file_prefix ); |
| 462 |
} |
| 463 |
else |
| 464 |
{ |
| 465 |
char buffer[500]; |
| 466 |
|
| 467 |
fgets(buffer, 400, fp); |
| 468 |
code = atoi(buffer); |
| 469 |
if ((code == 200) || ((code / 100) == 3)) |
| 470 |
{ |
| 471 |
/* read content-type redirect */ |
| 472 |
fgets(contenttype_or_redirect, MAXSTRLEN, fp); /* more yuck */ |
| 473 |
*(contenttype_or_redirect + strlen(contenttype_or_redirect) - 1) = '\0'; |
| 474 |
} |
| 475 |
if (code == 200) |
| 476 |
{ |
| 477 |
/* read last-mod time */ |
| 478 |
fgets(buffer, 400, fp); /* more yuck */ |
| 479 |
*last_modified = (time_t)strtol(buffer, NULL, 10); // go away http.c -- no error checking |
| 480 |
} |
| 481 |
|
| 482 |
|
| 483 |
fclose(fp); |
| 484 |
} |
| 485 |
|
| 486 |
return code; |
| 487 |
} |
| 488 |
|
| 489 |
int cmdf(int (*cmd) (const char *), char *fmt, char *string, pid_t pid) |
| 490 |
{ |
| 491 |
int rc; |
| 492 |
char *buffer; |
| 493 |
|
| 494 |
buffer = emalloc(strlen(fmt) + strlen(string) + sizeof(pid_t) * 8 + 1); |
| 495 |
|
| 496 |
sprintf(buffer, fmt, string, pid); |
| 497 |
|
| 498 |
rc = cmd(buffer); |
| 499 |
efree(buffer); |
| 500 |
return rc; |
| 501 |
} |
| 502 |
|
| 503 |
char *readline(FILE * fp) |
| 504 |
{ |
| 505 |
static char *buffer = 0; |
| 506 |
static int buffersize = 512; |
| 507 |
|
| 508 |
if (buffer == 0) |
| 509 |
{ |
| 510 |
buffer = (char *) emalloc(buffersize); |
| 511 |
} |
| 512 |
/* |
| 513 |
*Try to read in the line |
| 514 |
*/ |
| 515 |
|
| 516 |
if (fgets(buffer, buffersize, fp) == NULL) |
| 517 |
{ |
| 518 |
return NULL; |
| 519 |
} |
| 520 |
|
| 521 |
/* |
| 522 |
* Make sure we read the entire line. If not, double the buffer |
| 523 |
* size and try to read the rest |
| 524 |
*/ |
| 525 |
while (buffer[strlen(buffer) - 1] != '\n') |
| 526 |
{ |
| 527 |
buffer = (char *) erealloc(buffer, buffersize * 2); |
| 528 |
|
| 529 |
/* |
| 530 |
* The easiest way to verify that this line is okay is to consider |
| 531 |
* the situation where the buffer is 2 bytes longs. Since fgets() |
| 532 |
* always guarantees to put the trailing NULL, it will have essentially |
| 533 |
* used only 1 bytes. We double it to four, so we now have the left |
| 534 |
* over byte (that currently contains NULL) in addition to the doubling |
| 535 |
* which gets us to read buffersize + 1. |
| 536 |
*/ |
| 537 |
if (fgets(buffer + buffersize - 1, buffersize + 1, fp) == 0) |
| 538 |
{ |
| 539 |
break; |
| 540 |
} |
| 541 |
buffersize *= 2; |
| 542 |
} |
| 543 |
|
| 544 |
return buffer; |
| 545 |
} |
| 546 |
|
| 547 |
|
| 548 |
/* A local version of getpid() so that we don't have to suffer |
| 549 |
** a system call each time we need it. |
| 550 |
*/ |
| 551 |
pid_t lgetpid() |
| 552 |
{ |
| 553 |
static pid_t pid = -1; |
| 554 |
|
| 555 |
if (pid == -1) |
| 556 |
{ |
| 557 |
pid = getpid(); |
| 558 |
} |
| 559 |
return pid; |
| 560 |
} |
| 561 |
|
| 562 |
#if 0 |
| 563 |
|
| 564 |
/* Testing the robot rules parsing code... |
| 565 |
**/ |
| 566 |
void http_indexpath(char *url) |
| 567 |
{ |
| 568 |
httpserverinfo *server = getserverinfo(url); |
| 569 |
robotrules *robotrule; |
| 570 |
|
| 571 |
printf("User-agent: %s\n", server->useragent ? server->useragent : "(none)"); |
| 572 |
for (robotrule = server->robotrules; robotrule; robotrule = robotrule->next) |
| 573 |
{ |
| 574 |
printf("Disallow: %s\n", robotrule->disallow); |
| 575 |
} |
| 576 |
} |
| 577 |
|
| 578 |
#else |
| 579 |
|
| 580 |
/********************************************************/ |
| 581 |
/* "Public" functions */ |
| 582 |
/********************************************************/ |
| 583 |
|
| 584 |
/* The main entry point for the module. For fs.c, decides whether this |
| 585 |
** is a file or directory and routes to the correct routine. |
| 586 |
*/ |
| 587 |
void http_indexpath(SWISH * sw, char *url) |
| 588 |
{ |
| 589 |
urldepth *urllist = 0; |
| 590 |
urldepth *item; |
| 591 |
static int lentitle = 0; |
| 592 |
static char *title = NULL; |
| 593 |
char *tmptitle; |
| 594 |
static int lencontenttype = 0; |
| 595 |
static char *contenttype = NULL; |
| 596 |
int code; |
| 597 |
time_t last_modified = 0; |
| 598 |
|
| 599 |
httpserverinfo *server; |
| 600 |
char *link; |
| 601 |
char *p; |
| 602 |
FileProp *fprop; |
| 603 |
FILE *fp; |
| 604 |
struct MOD_Index *idx = sw->Index; |
| 605 |
|
| 606 |
char *file_prefix; // prefix for use with files written by swishspider -- should just be on the stack! |
| 607 |
char *file_suffix; // where to copy the suffix |
| 608 |
|
| 609 |
|
| 610 |
/* Initialize buffers */ |
| 611 |
|
| 612 |
|
| 613 |
file_prefix = emalloc( strlen(idx->tmpdir) + MAXPIDLEN + strlen("/swishspider@.contents+fill") ); |
| 614 |
sprintf(file_prefix, "%s/swishspider@%ld", idx->tmpdir, (long) lgetpid()); |
| 615 |
file_suffix = file_prefix + strlen( file_prefix ); |
| 616 |
|
| 617 |
|
| 618 |
if (!lentitle) |
| 619 |
title = emalloc((lentitle = MAXSTRLEN) + 1); |
| 620 |
|
| 621 |
if (!lencontenttype) |
| 622 |
contenttype = emalloc((lencontenttype = MAXSTRLEN) + 1); |
| 623 |
|
| 624 |
|
| 625 |
|
| 626 |
/* prime the pump with the first url */ |
| 627 |
urllist = add_url(sw, urllist, url, 0, url); |
| 628 |
|
| 629 |
|
| 630 |
|
| 631 |
/* retrieve each url and add urls to a certain depth */ |
| 632 |
|
| 633 |
while (urllist) |
| 634 |
{ |
| 635 |
item = urllist; |
| 636 |
urllist = urllist->next; |
| 637 |
|
| 638 |
if (sw->verbose >= 2) |
| 639 |
{ |
| 640 |
printf("retrieving %s (%d)...\n", item->url, item->depth); |
| 641 |
fflush(stdout); |
| 642 |
} |
| 643 |
|
| 644 |
/* We don't check if this url is legal here, because we do that before adding to the list. */ |
| 645 |
server = getserverinfo(sw, item->url); |
| 646 |
|
| 647 |
strcpy( file_suffix, "" ); // reset to just the prefix |
| 648 |
|
| 649 |
if ((code = get(sw, contenttype, &last_modified, &server->lastretrieval, file_prefix, item->url)) == 200) |
| 650 |
{ |
| 651 |
/* Set the file_prefix to be the path to "contents" */ |
| 652 |
strcpy( file_suffix, ".contents" ); |
| 653 |
|
| 654 |
|
| 655 |
/* Patch from Steve van der Burg */ |
| 656 |
/* change from strcmp to strncmp */ |
| 657 |
|
| 658 |
|
| 659 |
/* Fetch title from doc if it's HTML */ |
| 660 |
|
| 661 |
if (strncmp(contenttype, "text/html", 9) == 0) |
| 662 |
title = SafeStrCopy(title, (char *) (tmptitle = parseHTMLtitle(sw , file_prefix)), &lentitle); |
| 663 |
else |
| 664 |
if ((p = strrchr(item->url, '/'))) |
| 665 |
title = SafeStrCopy(title, p + 1, &lentitle); |
| 666 |
else |
| 667 |
title = SafeStrCopy(title, item->url, &lentitle); |
| 668 |
|
| 669 |
|
| 670 |
/* Now index the file */ |
| 671 |
|
| 672 |
/* What to do with non text files?? */ |
| 673 |
if ( strncmp(contenttype, "text/", 5) == 0 ) |
| 674 |
{ |
| 675 |
fprop = file_properties(item->url, file_prefix, sw); |
| 676 |
fprop->mtime = last_modified; |
| 677 |
|
| 678 |
/* only index contents of text docs */ |
| 679 |
// this would just index the path name |
| 680 |
//fprop->index_no_content = strncmp(contenttype, "text/", 5); |
| 681 |
|
| 682 |
do_index_file(sw, fprop); |
| 683 |
free_file_properties(fprop); |
| 684 |
} |
| 685 |
else if (sw->verbose >= 3) |
| 686 |
printf("Skipping %s: Wrong content type: %s.\n", url, contenttype); |
| 687 |
|
| 688 |
|
| 689 |
|
| 690 |
|
| 691 |
/* add new links as extracted by the spider */ |
| 692 |
|
| 693 |
if (strncmp(contenttype, "text/html", 9) == 0) |
| 694 |
{ |
| 695 |
strcpy( file_suffix, ".links" ); |
| 696 |
|
| 697 |
if ((fp = fopen(file_prefix, F_READ_TEXT)) != NULL) |
| 698 |
{ |
| 699 |
/* URLs can get quite large so don't depend on a fixed size buffer */ |
| 700 |
|
| 701 |
while ((link = readline(fp)) != NULL) |
| 702 |
{ |
| 703 |
*(link + strlen(link) - 1) = '\0'; |
| 704 |
urllist = add_url(sw, urllist, link, item->depth + 1, url); |
| 705 |
} |
| 706 |
fclose(fp); |
| 707 |
} |
| 708 |
} |
| 709 |
|
| 710 |
} |
| 711 |
else if ((code / 100) == 3) |
| 712 |
{ |
| 713 |
if ( *contenttype ) |
| 714 |
urllist = add_url(sw, urllist, contenttype, item->depth, url); |
| 715 |
else |
| 716 |
if (sw->verbose >= 3) |
| 717 |
printf("URL '%s' returned redirect code %d without a Location.\n", url, code); |
| 718 |
} |
| 719 |
|
| 720 |
|
| 721 |
|
| 722 |
/* Clean up the files left by swishspider */ |
| 723 |
cmdf(unlink, "%s/swishspider@%ld.response", idx->tmpdir, lgetpid()); |
| 724 |
cmdf(unlink, "%s/swishspider@%ld.contents", idx->tmpdir, lgetpid()); |
| 725 |
cmdf(unlink, "%s/swishspider@%ld.links", idx->tmpdir, lgetpid()); |
| 726 |
} |
| 727 |
efree(file_prefix); |
| 728 |
} |
| 729 |
|
| 730 |
#endif |
| 731 |
|
| 732 |
|
| 733 |
|
| 734 |
|
| 735 |
struct _indexing_data_source_def HTTPIndexingDataSource = { |
| 736 |
"HTTP-Crawler", |
| 737 |
"http", |
| 738 |
http_indexpath, |
| 739 |
configModule_HTTP |
| 740 |
}; |
Parent Directory
| 
Revision Log
| 
Revision Graph