1/* Handling of recursive HTTP retrieving. 2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 3 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc. 4 5This file is part of GNU Wget. 6 7GNU Wget is free software; you can redistribute it and/or modify 8it under the terms of the GNU General Public License as published by 9the Free Software Foundation; either version 3 of the License, or 10 (at your option) any later version. 11 12GNU Wget is distributed in the hope that it will be useful, 13but WITHOUT ANY WARRANTY; without even the implied warranty of 14MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 15GNU General Public License for more details. 16 17You should have received a copy of the GNU General Public License 18along with Wget. If not, see <http://www.gnu.org/licenses/>. 19 20Additional permission under GNU GPL version 3 section 7 21 22If you modify this program, or any covered work, by linking or 23combining it with the OpenSSL project's OpenSSL library (or a 24modified version of that library), containing parts covered by the 25terms of the OpenSSL or SSLeay licenses, the Free Software Foundation 26grants you additional permission to convey the resulting work. 27Corresponding Source for a non-source form of such a combination 28shall include the source code for the parts of OpenSSL used as well 29as that of the covered work. */ 30 31#include "wget.h" 32 33#include <stdio.h> 34#include <stdlib.h> 35#include <string.h> 36#ifdef HAVE_UNISTD_H 37# include <unistd.h> 38#endif /* HAVE_UNISTD_H */ 39#include <errno.h> 40#include <assert.h> 41 42#include "url.h" 43#include "recur.h" 44#include "utils.h" 45#include "retr.h" 46#include "ftp.h" 47#include "host.h" 48#include "hash.h" 49#include "res.h" 50#include "convert.h" 51#include "html-url.h" 52#include "css-url.h" 53#include "spider.h" 54 55/* Functions for maintaining the URL queue. */ 56 57struct queue_element { 58 const char *url; /* the URL to download */ 59 const char *referer; /* the referring document */ 60 int depth; /* the depth */ 61 bool html_allowed; /* whether the document is allowed to 62 be treated as HTML. */ 63 struct iri *iri; /* sXXXav */ 64 bool css_allowed; /* whether the document is allowed to 65 be treated as CSS. */ 66 struct queue_element *next; /* next element in queue */ 67}; 68 69struct url_queue { 70 struct queue_element *head; 71 struct queue_element *tail; 72 int count, maxcount; 73}; 74 75/* Create a URL queue. */ 76 77static struct url_queue * 78url_queue_new (void) 79{ 80 struct url_queue *queue = xnew0 (struct url_queue); 81 return queue; 82} 83 84/* Delete a URL queue. */ 85 86static void 87url_queue_delete (struct url_queue *queue) 88{ 89 xfree (queue); 90} 91 92/* Enqueue a URL in the queue. The queue is FIFO: the items will be 93 retrieved ("dequeued") from the queue in the order they were placed 94 into it. */ 95 96static void 97url_enqueue (struct url_queue *queue, struct iri *i, 98 const char *url, const char *referer, int depth, 99 bool html_allowed, bool css_allowed) 100{ 101 struct queue_element *qel = xnew (struct queue_element); 102 qel->iri = i; 103 qel->url = url; 104 qel->referer = referer; 105 qel->depth = depth; 106 qel->html_allowed = html_allowed; 107 qel->css_allowed = css_allowed; 108 qel->next = NULL; 109 110 ++queue->count; 111 if (queue->count > queue->maxcount) 112 queue->maxcount = queue->count; 113 114 DEBUGP (("Enqueuing %s at depth %d\n", 115 quotearg_n_style (0, escape_quoting_style, url), depth)); 116 DEBUGP (("Queue count %d, maxcount %d.\n", queue->count, queue->maxcount)); 117 118 if (i) 119 DEBUGP (("[IRI Enqueuing %s with %s\n", quote_n (0, url), 120 i->uri_encoding ? quote_n (1, i->uri_encoding) : "None")); 121 122 if (queue->tail) 123 queue->tail->next = qel; 124 queue->tail = qel; 125 126 if (!queue->head) 127 queue->head = queue->tail; 128} 129 130/* Take a URL out of the queue. Return true if this operation 131 succeeded, or false if the queue is empty. */ 132 133static bool 134url_dequeue (struct url_queue *queue, struct iri **i, 135 const char **url, const char **referer, int *depth, 136 bool *html_allowed, bool *css_allowed) 137{ 138 struct queue_element *qel = queue->head; 139 140 if (!qel) 141 return false; 142 143 queue->head = queue->head->next; 144 if (!queue->head) 145 queue->tail = NULL; 146 147 *i = qel->iri; 148 *url = qel->url; 149 *referer = qel->referer; 150 *depth = qel->depth; 151 *html_allowed = qel->html_allowed; 152 *css_allowed = qel->css_allowed; 153 154 --queue->count; 155 156 DEBUGP (("Dequeuing %s at depth %d\n", 157 quotearg_n_style (0, escape_quoting_style, qel->url), qel->depth)); 158 DEBUGP (("Queue count %d, maxcount %d.\n", queue->count, queue->maxcount)); 159 160 xfree (qel); 161 return true; 162} 163 164static bool download_child_p (const struct urlpos *, struct url *, int, 165 struct url *, struct hash_table *, struct iri *); 166static bool descend_redirect_p (const char *, struct url *, int, 167 struct url *, struct hash_table *, struct iri *); 168 169 170/* Retrieve a part of the web beginning with START_URL. This used to 171 be called "recursive retrieval", because the old function was 172 recursive and implemented depth-first search. retrieve_tree on the 173 other hand implements breadth-search traversal of the tree, which 174 results in much nicer ordering of downloads. 175 176 The algorithm this function uses is simple: 177 178 1. put START_URL in the queue. 179 2. while there are URLs in the queue: 180 181 3. get next URL from the queue. 182 4. download it. 183 5. if the URL is HTML and its depth does not exceed maximum depth, 184 get the list of URLs embedded therein. 185 6. for each of those URLs do the following: 186 187 7. if the URL is not one of those downloaded before, and if it 188 satisfies the criteria specified by the various command-line 189 options, add it to the queue. */ 190 191uerr_t 192retrieve_tree (struct url *start_url_parsed, struct iri *pi) 193{ 194 uerr_t status = RETROK; 195 196 /* The queue of URLs we need to load. */ 197 struct url_queue *queue; 198 199 /* The URLs we do not wish to enqueue, because they are already in 200 the queue, but haven't been downloaded yet. */ 201 struct hash_table *blacklist; 202 203 int up_error_code; 204 struct iri *i = iri_new (); 205 206#define COPYSTR(x) (x) ? xstrdup(x) : NULL; 207 /* Duplicate pi struct if not NULL */ 208 if (pi) 209 { 210 i->uri_encoding = COPYSTR (pi->uri_encoding); 211 i->content_encoding = COPYSTR (pi->content_encoding); 212 i->utf8_encode = pi->utf8_encode; 213 } 214 else 215 set_uri_encoding (i, opt.locale, true); 216#undef COPYSTR 217 218 queue = url_queue_new (); 219 blacklist = make_string_hash_table (0); 220 221 /* Enqueue the starting URL. Use start_url_parsed->url rather than 222 just URL so we enqueue the canonical form of the URL. */ 223 url_enqueue (queue, i, xstrdup (start_url_parsed->url), NULL, 0, true, 224 false); 225 string_set_add (blacklist, start_url_parsed->url); 226 227 while (1) 228 { 229 bool descend = false; 230 char *url, *referer, *file = NULL; 231 int depth; 232 bool html_allowed, css_allowed; 233 bool is_css = false; 234 bool dash_p_leaf_HTML = false; 235 236 if (opt.quota && total_downloaded_bytes > opt.quota) 237 break; 238 if (status == FWRITEERR) 239 break; 240 241 /* Get the next URL from the queue... */ 242 243 if (!url_dequeue (queue, (struct iri **) &i, 244 (const char **)&url, (const char **)&referer, 245 &depth, &html_allowed, &css_allowed)) 246 break; 247 248 /* ...and download it. Note that this download is in most cases 249 unconditional, as download_child_p already makes sure a file 250 doesn't get enqueued twice -- and yet this check is here, and 251 not in download_child_p. This is so that if you run `wget -r 252 URL1 URL2', and a random URL is encountered once under URL1 253 and again under URL2, but at a different (possibly smaller) 254 depth, we want the URL's children to be taken into account 255 the second time. */ 256 if (dl_url_file_map && hash_table_contains (dl_url_file_map, url)) 257 { 258 file = xstrdup (hash_table_get (dl_url_file_map, url)); 259 260 DEBUGP (("Already downloaded \"%s\", reusing it from \"%s\".\n", 261 url, file)); 262 263 /* this sucks, needs to be combined! */ 264 if (html_allowed 265 && downloaded_html_set 266 && string_set_contains (downloaded_html_set, file)) 267 { 268 descend = true; 269 is_css = false; 270 } 271 if (css_allowed 272 && downloaded_css_set 273 && string_set_contains (downloaded_css_set, file)) 274 { 275 descend = true; 276 is_css = true; 277 } 278 } 279 else 280 { 281 int dt = 0, url_err; 282 char *redirected = NULL; 283 struct url *url_parsed = url_parse (url, &url_err, i, true); 284 285 status = retrieve_url (url_parsed, url, &file, &redirected, referer, 286 &dt, false, i, true); 287 288 if (html_allowed && file && status == RETROK 289 && (dt & RETROKF) && (dt & TEXTHTML)) 290 { 291 descend = true; 292 is_css = false; 293 } 294 295 /* a little different, css_allowed can override content type 296 lots of web servers serve css with an incorrect content type 297 */ 298 if (file && status == RETROK 299 && (dt & RETROKF) && 300 ((dt & TEXTCSS) || css_allowed)) 301 { 302 descend = true; 303 is_css = true; 304 } 305 306 if (redirected) 307 { 308 /* We have been redirected, possibly to another host, or 309 different path, or wherever. Check whether we really 310 want to follow it. */ 311 if (descend) 312 { 313 if (!descend_redirect_p (redirected, url_parsed, depth, 314 start_url_parsed, blacklist, i)) 315 descend = false; 316 else 317 /* Make sure that the old pre-redirect form gets 318 blacklisted. */ 319 string_set_add (blacklist, url); 320 } 321 322 xfree (url); 323 url = redirected; 324 } 325 else 326 { 327 xfree (url); 328 url = xstrdup (url_parsed->url); 329 } 330 url_free(url_parsed); 331 } 332 333 if (opt.spider) 334 { 335 visited_url (url, referer); 336 } 337 338 if (descend 339 && depth >= opt.reclevel && opt.reclevel != INFINITE_RECURSION) 340 { 341 if (opt.page_requisites 342 && (depth == opt.reclevel || depth == opt.reclevel + 1)) 343 { 344 /* When -p is specified, we are allowed to exceed the 345 maximum depth, but only for the "inline" links, 346 i.e. those that are needed to display the page. 347 Originally this could exceed the depth at most by 348 one, but we allow one more level so that the leaf 349 pages that contain frames can be loaded 350 correctly. */ 351 dash_p_leaf_HTML = true; 352 } 353 else 354 { 355 /* Either -p wasn't specified or it was and we've 356 already spent the two extra (pseudo-)levels that it 357 affords us, so we need to bail out. */ 358 DEBUGP (("Not descending further; at depth %d, max. %d.\n", 359 depth, opt.reclevel)); 360 descend = false; 361 } 362 } 363 364 /* If the downloaded document was HTML or CSS, parse it and enqueue the 365 links it contains. */ 366 367 if (descend) 368 { 369 bool meta_disallow_follow = false; 370 struct urlpos *children 371 = is_css ? get_urls_css_file (file, url) : 372 get_urls_html (file, url, &meta_disallow_follow, i); 373 374 if (opt.use_robots && meta_disallow_follow) 375 { 376 free_urlpos (children); 377 children = NULL; 378 } 379 380 if (children) 381 { 382 struct urlpos *child = children; 383 struct url *url_parsed = url_parse (url, NULL, i, true); 384 struct iri *ci; 385 char *referer_url = url; 386 bool strip_auth = (url_parsed != NULL 387 && url_parsed->user != NULL); 388 assert (url_parsed != NULL); 389 390 /* Strip auth info if present */ 391 if (strip_auth) 392 referer_url = url_string (url_parsed, URL_AUTH_HIDE); 393 394 for (; child; child = child->next) 395 { 396 if (child->ignore_when_downloading) 397 continue; 398 if (dash_p_leaf_HTML && !child->link_inline_p) 399 continue; 400 if (download_child_p (child, url_parsed, depth, start_url_parsed, 401 blacklist, i)) 402 { 403 ci = iri_new (); 404 set_uri_encoding (ci, i->content_encoding, false); 405 url_enqueue (queue, ci, xstrdup (child->url->url), 406 xstrdup (referer_url), depth + 1, 407 child->link_expect_html, 408 child->link_expect_css); 409 /* We blacklist the URL we have enqueued, because we 410 don't want to enqueue (and hence download) the 411 same URL twice. */ 412 string_set_add (blacklist, child->url->url); 413 } 414 } 415 416 if (strip_auth) 417 xfree (referer_url); 418 url_free (url_parsed); 419 free_urlpos (children); 420 } 421 } 422 423 if (file 424 && (opt.delete_after 425 || opt.spider /* opt.recursive is implicitely true */ 426 || !acceptable (file))) 427 { 428 /* Either --delete-after was specified, or we loaded this 429 (otherwise unneeded because of --spider or rejected by -R) 430 HTML file just to harvest its hyperlinks -- in either case, 431 delete the local file. */ 432 DEBUGP (("Removing file due to %s in recursive_retrieve():\n", 433 opt.delete_after ? "--delete-after" : 434 (opt.spider ? "--spider" : 435 "recursive rejection criteria"))); 436 logprintf (LOG_VERBOSE, 437 (opt.delete_after || opt.spider 438 ? _("Removing %s.\n") 439 : _("Removing %s since it should be rejected.\n")), 440 file); 441 if (unlink (file)) 442 logprintf (LOG_NOTQUIET, "unlink: %s\n", strerror (errno)); 443 logputs (LOG_VERBOSE, "\n"); 444 register_delete_file (file); 445 } 446 447 xfree (url); 448 xfree_null (referer); 449 xfree_null (file); 450 iri_free (i); 451 } 452 453 /* If anything is left of the queue due to a premature exit, free it 454 now. */ 455 { 456 char *d1, *d2; 457 int d3; 458 bool d4, d5; 459 struct iri *d6; 460 while (url_dequeue (queue, (struct iri **)&d6, 461 (const char **)&d1, (const char **)&d2, &d3, &d4, &d5)) 462 { 463 iri_free (d6); 464 xfree (d1); 465 xfree_null (d2); 466 } 467 } 468 url_queue_delete (queue); 469 470 string_set_free (blacklist); 471 472 if (opt.quota && total_downloaded_bytes > opt.quota) 473 return QUOTEXC; 474 else if (status == FWRITEERR) 475 return FWRITEERR; 476 else 477 return RETROK; 478} 479 480/* Based on the context provided by retrieve_tree, decide whether a 481 URL is to be descended to. This is only ever called from 482 retrieve_tree, but is in a separate function for clarity. 483 484 The most expensive checks (such as those for robots) are memoized 485 by storing these URLs to BLACKLIST. This may or may not help. It 486 will help if those URLs are encountered many times. */ 487 488static bool 489download_child_p (const struct urlpos *upos, struct url *parent, int depth, 490 struct url *start_url_parsed, struct hash_table *blacklist, 491 struct iri *iri) 492{ 493 struct url *u = upos->url; 494 const char *url = u->url; 495 bool u_scheme_like_http; 496 497 DEBUGP (("Deciding whether to enqueue \"%s\".\n", url)); 498 499 if (string_set_contains (blacklist, url)) 500 { 501 if (opt.spider) 502 { 503 char *referrer = url_string (parent, URL_AUTH_HIDE_PASSWD); 504 DEBUGP (("download_child_p: parent->url is: %s\n", quote (parent->url))); 505 visited_url (url, referrer); 506 xfree (referrer); 507 } 508 DEBUGP (("Already on the black list.\n")); 509 goto out; 510 } 511 512 /* Several things to check for: 513 1. if scheme is not http, and we don't load it 514 2. check for relative links (if relative_only is set) 515 3. check for domain 516 4. check for no-parent 517 5. check for excludes && includes 518 6. check for suffix 519 7. check for same host (if spanhost is unset), with possible 520 gethostbyname baggage 521 8. check for robots.txt 522 523 Addendum: If the URL is FTP, and it is to be loaded, only the 524 domain and suffix settings are "stronger". 525 526 Note that .html files will get loaded regardless of suffix rules 527 (but that is remedied later with unlink) unless the depth equals 528 the maximum depth. 529 530 More time- and memory- consuming tests should be put later on 531 the list. */ 532 533 /* Determine whether URL under consideration has a HTTP-like scheme. */ 534 u_scheme_like_http = schemes_are_similar_p (u->scheme, SCHEME_HTTP); 535 536 /* 1. Schemes other than HTTP are normally not recursed into. */ 537 if (!u_scheme_like_http && !(u->scheme == SCHEME_FTP && opt.follow_ftp)) 538 { 539 DEBUGP (("Not following non-HTTP schemes.\n")); 540 goto out; 541 } 542 543 /* 2. If it is an absolute link and they are not followed, throw it 544 out. */ 545 if (u_scheme_like_http) 546 if (opt.relative_only && !upos->link_relative_p) 547 { 548 DEBUGP (("It doesn't really look like a relative link.\n")); 549 goto out; 550 } 551 552 /* 3. If its domain is not to be accepted/looked-up, chuck it 553 out. */ 554 if (!accept_domain (u)) 555 { 556 DEBUGP (("The domain was not accepted.\n")); 557 goto out; 558 } 559 560 /* 4. Check for parent directory. 561 562 If we descended to a different host or changed the scheme, ignore 563 opt.no_parent. Also ignore it for documents needed to display 564 the parent page when in -p mode. */ 565 if (opt.no_parent 566 && schemes_are_similar_p (u->scheme, start_url_parsed->scheme) 567 && 0 == strcasecmp (u->host, start_url_parsed->host) 568 && u->port == start_url_parsed->port 569 && !(opt.page_requisites && upos->link_inline_p)) 570 { 571 if (!subdir_p (start_url_parsed->dir, u->dir)) 572 { 573 DEBUGP (("Going to \"%s\" would escape \"%s\" with no_parent on.\n", 574 u->dir, start_url_parsed->dir)); 575 goto out; 576 } 577 } 578 579 /* 5. If the file does not match the acceptance list, or is on the 580 rejection list, chuck it out. The same goes for the directory 581 exclusion and inclusion lists. */ 582 if (opt.includes || opt.excludes) 583 { 584 if (!accdir (u->dir)) 585 { 586 DEBUGP (("%s (%s) is excluded/not-included.\n", url, u->dir)); 587 goto out; 588 } 589 } 590 591 /* 6. Check for acceptance/rejection rules. We ignore these rules 592 for directories (no file name to match) and for non-leaf HTMLs, 593 which can lead to other files that do need to be downloaded. (-p 594 automatically implies non-leaf because with -p we can, if 595 necesary, overstep the maximum depth to get the page requisites.) */ 596 if (u->file[0] != '\0' 597 && !(has_html_suffix_p (u->file) 598 /* The exception only applies to non-leaf HTMLs (but -p 599 always implies non-leaf because we can overstep the 600 maximum depth to get the requisites): */ 601 && (/* non-leaf */ 602 opt.reclevel == INFINITE_RECURSION 603 /* also non-leaf */ 604 || depth < opt.reclevel - 1 605 /* -p, which implies non-leaf (see above) */ 606 || opt.page_requisites))) 607 { 608 if (!acceptable (u->file)) 609 { 610 DEBUGP (("%s (%s) does not match acc/rej rules.\n", 611 url, u->file)); 612 goto out; 613 } 614 } 615 616 /* 7. */ 617 if (schemes_are_similar_p (u->scheme, parent->scheme)) 618 if (!opt.spanhost && 0 != strcasecmp (parent->host, u->host)) 619 { 620 DEBUGP (("This is not the same hostname as the parent's (%s and %s).\n", 621 u->host, parent->host)); 622 goto out; 623 } 624 625 /* 8. */ 626 if (opt.use_robots && u_scheme_like_http) 627 { 628 struct robot_specs *specs = res_get_specs (u->host, u->port); 629 if (!specs) 630 { 631 char *rfile; 632 if (res_retrieve_file (url, &rfile, iri)) 633 { 634 specs = res_parse_from_file (rfile); 635 636 /* Delete the robots.txt file if we chose to either delete the 637 files after downloading or we're just running a spider. */ 638 if (opt.delete_after || opt.spider) 639 { 640 logprintf (LOG_VERBOSE, "Removing %s.\n", rfile); 641 if (unlink (rfile)) 642 logprintf (LOG_NOTQUIET, "unlink: %s\n", 643 strerror (errno)); 644 } 645 646 xfree (rfile); 647 } 648 else 649 { 650 /* If we cannot get real specs, at least produce 651 dummy ones so that we can register them and stop 652 trying to retrieve them. */ 653 specs = res_parse ("", 0); 654 } 655 res_register_specs (u->host, u->port, specs); 656 } 657 658 /* Now that we have (or don't have) robots.txt specs, we can 659 check what they say. */ 660 if (!res_match_path (specs, u->path)) 661 { 662 DEBUGP (("Not following %s because robots.txt forbids it.\n", url)); 663 string_set_add (blacklist, url); 664 goto out; 665 } 666 } 667 668 /* The URL has passed all the tests. It can be placed in the 669 download queue. */ 670 DEBUGP (("Decided to load it.\n")); 671 672 return true; 673 674 out: 675 DEBUGP (("Decided NOT to load it.\n")); 676 677 return false; 678} 679 680/* This function determines whether we will consider downloading the 681 children of a URL whose download resulted in a redirection, 682 possibly to another host, etc. It is needed very rarely, and thus 683 it is merely a simple-minded wrapper around download_child_p. */ 684 685static bool 686descend_redirect_p (const char *redirected, struct url *orig_parsed, int depth, 687 struct url *start_url_parsed, struct hash_table *blacklist, 688 struct iri *iri) 689{ 690 struct url *new_parsed; 691 struct urlpos *upos; 692 bool success; 693 694 assert (orig_parsed != NULL); 695 696 new_parsed = url_parse (redirected, NULL, NULL, false); 697 assert (new_parsed != NULL); 698 699 upos = xnew0 (struct urlpos); 700 upos->url = new_parsed; 701 702 success = download_child_p (upos, orig_parsed, depth, 703 start_url_parsed, blacklist, iri); 704 705 url_free (new_parsed); 706 xfree (upos); 707 708 if (!success) 709 DEBUGP (("Redirection \"%s\" failed the test.\n", redirected)); 710 711 return success; 712} 713 714/* vim:set sts=2 sw=2 cino+={s: */ 715