| 1 |
/* ============================================================================ |
| 2 |
* Douglas Thrift's Search Engine License |
| 3 |
* |
| 4 |
* Copyright (C) 2002-2003, Douglas Thrift. All Rights Reserved. |
| 5 |
* Redistribution and use in source and binary forms, with or without |
| 6 |
* modification, are permitted provided that the following conditions are met: |
| 7 |
* |
| 8 |
* 1. Redistributions of source code must retain the above copyright notice, |
| 9 |
* this list of conditions and the following disclaimer. |
| 10 |
* |
| 11 |
* 2. Redistributions in binary form must reproduce the above copyright notice, |
| 12 |
* this list of conditions and the following disclaimer in the documentation |
| 13 |
* and/or other materials provided with the distribution. |
| 14 |
* |
| 15 |
* 3. The end-user documentation included with the redistribution, if any, must |
| 16 |
* include the following acknowledgment: |
| 17 |
* |
| 18 |
* "This product includes software developed by Douglas Thrift |
| 19 |
* (http://computers.douglasthrift.net/searchengine/)." |
| 20 |
* |
| 21 |
* Alternately, this acknowledgment may appear in the software itself, if |
| 22 |
* and wherever such third-party acknowledgments normally appear. |
| 23 |
* |
| 24 |
* 4. The names "Douglas Thrift" and "Douglas Thrift's Search Engine" must not |
| 25 |
* be used to endorse or promote products derived from this software without |
| 26 |
* specific prior written permission. For written permission, please visit |
| 27 |
* http://www.douglasthrift.net/contact.cgi for contact information. |
| 28 |
* |
| 29 |
* 5. Products derived from this software may not be called "Douglas Thrift's |
| 30 |
* Search Engine", nor may "Douglas Thrift's Search Engine" appear in their |
| 31 |
* name, without prior written permission. |
| 32 |
* |
| 33 |
* THIS SOFTWARE IS PROVIDED "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, |
| 34 |
* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND |
| 35 |
* FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE |
| 36 |
* COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, |
| 37 |
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT |
| 38 |
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, |
| 39 |
* OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF |
| 40 |
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING |
| 41 |
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, |
| 42 |
* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| 43 |
* ============================================================================ |
| 44 |
*/ |
| 45 |
// Douglas Thrift's Search Engine HTTP Handler |
| 46 |
// |
| 47 |
// Douglas Thrift |
| 48 |
// |
| 49 |
// HttpHandler.cpp |
| 50 |
|
| 51 |
#include "HttpHandler.h" |
| 52 |
|
| 53 |
#ifndef _WIN32 |
| 54 |
#include <unistd.h> |
| 55 |
#include <sys/types.h> |
| 56 |
#include <sys/socket.h> |
| 57 |
#include <netinet/in.h> |
| 58 |
#include <netdb.h> |
| 59 |
|
| 60 |
#define INVALID_SOCKET -1 |
| 61 |
#define SOCKET_ERROR -1 |
| 62 |
|
| 63 |
inline int closesocket(SOCKET s) { return close(s); } |
| 64 |
#endif |
| 65 |
|
| 66 |
HttpHandler::HttpHandler() |
| 67 |
{ |
| 68 |
buffer = new char[BUFSIZ + 1]; |
| 69 |
|
| 70 |
#ifdef _WIN32 |
| 71 |
if (WSAStartup(MAKEWORD(2, 0), &data) != 0) |
| 72 |
{ |
| 73 |
error(program + ": WSAStartup"); |
| 74 |
exit(1); |
| 75 |
} |
| 76 |
#endif // _WIN32 |
| 77 |
|
| 78 |
length = 0; |
| 79 |
chunked = false; |
| 80 |
} |
| 81 |
|
| 82 |
HttpHandler::~HttpHandler() |
| 83 |
{ |
| 84 |
delete [] buffer; |
| 85 |
|
| 86 |
#ifdef _WIN32 |
| 87 |
WSACleanup(); |
| 88 |
#endif // _WIN32 |
| 89 |
} |
| 90 |
|
| 91 |
bool HttpHandler::handle(URL &url, const string referer, bool head) |
| 92 |
{ |
| 93 |
bool answer = false; |
| 94 |
|
| 95 |
if ((http = socket(PF_INET, SOCK_STREAM, 0)) == INVALID_SOCKET) |
| 96 |
{ |
| 97 |
error(program + ": Socket"); |
| 98 |
exit(1); |
| 99 |
} |
| 100 |
|
| 101 |
sockaddr_in address; |
| 102 |
hostent* host; |
| 103 |
|
| 104 |
address.sin_family = AF_INET; |
| 105 |
|
| 106 |
if ((host = gethostbyname(url.getAddress().c_str())) == NULL) |
| 107 |
{ |
| 108 |
error(program + ": Host: " + url.getAddress(), true); |
| 109 |
return answer; |
| 110 |
} |
| 111 |
|
| 112 |
address.sin_addr = *((in_addr*)*host->h_addr_list); |
| 113 |
address.sin_port = htons(url.getPort()); |
| 114 |
|
| 115 |
if (connect(http, (sockaddr*)&address, sizeof(sockaddr_in)) == |
| 116 |
SOCKET_ERROR) |
| 117 |
{ |
| 118 |
error(program + ": Connect"); |
| 119 |
return answer; |
| 120 |
} |
| 121 |
|
| 122 |
if (head) |
| 123 |
{ |
| 124 |
putline("HEAD " + url.getPath() + " HTTP/1.1"); |
| 125 |
} |
| 126 |
else |
| 127 |
{ |
| 128 |
putline("GET " + url.getPath() + " HTTP/1.1"); |
| 129 |
} |
| 130 |
|
| 131 |
putline("Accept: text/html; text/plain"); |
| 132 |
putline("User-Agent: " + agent(true) + ' ' + platform()); |
| 133 |
|
| 134 |
if (url.getPort() == 80) |
| 135 |
{ |
| 136 |
putline("Host: " + url.getAddress()); |
| 137 |
} |
| 138 |
else |
| 139 |
{ |
| 140 |
char* port = new char[1024]; |
| 141 |
sprintf(port, "%u", url.getPort()); |
| 142 |
|
| 143 |
putline("Host: " + url.getAddress() + ':' + port); |
| 144 |
|
| 145 |
delete [] port; |
| 146 |
} |
| 147 |
|
| 148 |
if (referer != "") |
| 149 |
{ |
| 150 |
putline("Referer: " + referer); |
| 151 |
} |
| 152 |
|
| 153 |
putline("Connection: close"); |
| 154 |
putline(); |
| 155 |
|
| 156 |
code response; |
| 157 |
string line; |
| 158 |
|
| 159 |
do |
| 160 |
{ |
| 161 |
line = getline(); |
| 162 |
|
| 163 |
if (line.find("HTTP/") != 0) |
| 164 |
{ |
| 165 |
return answer; |
| 166 |
} |
| 167 |
|
| 168 |
unsigned dot = line.find('.'); |
| 169 |
unsigned space = line.find(' '); |
| 170 |
|
| 171 |
unsigned major = strtoul(line.substr(5, dot - 5).c_str(), 0, 10); |
| 172 |
unsigned minor = strtoul(line.substr(dot + 1, space - dot - 1).c_str(), |
| 173 |
0, 10); |
| 174 |
|
| 175 |
if (major > 1) |
| 176 |
{ |
| 177 |
cerr << program << ": Potentially Incompatible Server: HTTP/" << |
| 178 |
major << "." << minor << "\n"; |
| 179 |
|
| 180 |
return answer; |
| 181 |
} |
| 182 |
|
| 183 |
response = code(strtoul(line.substr(space + 1).c_str(), 0, 10)); |
| 184 |
|
| 185 |
if (response < ok) do line = getline(); while (line != ""); |
| 186 |
} |
| 187 |
while (response < ok); |
| 188 |
|
| 189 |
do |
| 190 |
{ |
| 191 |
line = getline(); |
| 192 |
|
| 193 |
if (line != "") |
| 194 |
{ |
| 195 |
unsigned colon = line.find(':'); |
| 196 |
|
| 197 |
string field = line.substr(0, colon); |
| 198 |
string value = line.substr(colon + 1); |
| 199 |
|
| 200 |
while (isspace(value[0])) value.erase(0, 1); |
| 201 |
|
| 202 |
if (field == "Content-Type") |
| 203 |
{ |
| 204 |
type = value; |
| 205 |
} |
| 206 |
else if (field == "Content-Length") |
| 207 |
{ |
| 208 |
length = strtoul(value.c_str(), 0, 10); |
| 209 |
} |
| 210 |
else if (field == "Location") |
| 211 |
{ |
| 212 |
location = value; |
| 213 |
} |
| 214 |
else if (field == "Transfer-Encoding") |
| 215 |
{ |
| 216 |
chunked = value == "chunked"; |
| 217 |
} |
| 218 |
} |
| 219 |
} |
| 220 |
while (line != ""); |
| 221 |
|
| 222 |
switch (response) |
| 223 |
{ |
| 224 |
case ok: |
| 225 |
if (debug) cerr << "response = " << response << "\n"; |
| 226 |
answer = true; |
| 227 |
break; |
| 228 |
case choices: |
| 229 |
case moved: |
| 230 |
case found: |
| 231 |
if (debug) cerr << "response = " << response << "\n" |
| 232 |
<< "location = " << location << "\n"; |
| 233 |
location = getLink(location, url); |
| 234 |
break; |
| 235 |
case notfound: |
| 236 |
case internal: |
| 237 |
if (debug) cerr << "response = " << response << "\n"; |
| 238 |
break; |
| 239 |
default: |
| 240 |
if (debug) cerr << "response = " << response << "\n"; |
| 241 |
if (response <= 299) |
| 242 |
{ |
| 243 |
answer = true; |
| 244 |
} |
| 245 |
else if (response <= 399) |
| 246 |
{ |
| 247 |
location = getLink(location, url); |
| 248 |
} |
| 249 |
break; |
| 250 |
} |
| 251 |
|
| 252 |
if (!head && answer) populate(); |
| 253 |
|
| 254 |
return answer; |
| 255 |
} |
| 256 |
|
| 257 |
HttpHandler& HttpHandler::getline(string& line, char endline) |
| 258 |
{ |
| 259 |
unsigned end = page.find(endline); |
| 260 |
unsigned newline = page.find('\n'); |
| 261 |
|
| 262 |
if (newline < end || end == string::npos) |
| 263 |
{ |
| 264 |
end = newline; |
| 265 |
} |
| 266 |
|
| 267 |
line = page.substr(0, end); |
| 268 |
page.erase(0, (end == string::npos ? end : end + 1)); |
| 269 |
|
| 270 |
return *this; |
| 271 |
} |
| 272 |
|
| 273 |
void HttpHandler::clear() |
| 274 |
{ |
| 275 |
closesocket(http); |
| 276 |
|
| 277 |
type = ""; |
| 278 |
length = 0; |
| 279 |
location = ""; |
| 280 |
page = ""; |
| 281 |
chunked = false; |
| 282 |
} |
| 283 |
|
| 284 |
void HttpHandler::populate() |
| 285 |
{ |
| 286 |
if (!chunked) |
| 287 |
{ |
| 288 |
unsigned left = length; |
| 289 |
|
| 290 |
while (left > 0) |
| 291 |
{ |
| 292 |
memset(buffer, 0, BUFSIZ + 1); |
| 293 |
|
| 294 |
unsigned bytes = left > BUFSIZ ? BUFSIZ : left; |
| 295 |
unsigned received; |
| 296 |
|
| 297 |
while (true) |
| 298 |
{ |
| 299 |
if ((received = recv(http, buffer, bytes, 0)) == SOCKET_ERROR) |
| 300 |
{ |
| 301 |
error(program + ": Recv"); |
| 302 |
exit(1); |
| 303 |
} |
| 304 |
else if (received != bytes) |
| 305 |
{ |
| 306 |
left -= received; |
| 307 |
page += buffer; |
| 308 |
|
| 309 |
memset(buffer, 0, BUFSIZ + 1); |
| 310 |
|
| 311 |
bytes -= received; |
| 312 |
} |
| 313 |
else |
| 314 |
{ |
| 315 |
break; |
| 316 |
} |
| 317 |
} |
| 318 |
|
| 319 |
page += buffer; |
| 320 |
left -= bytes; |
| 321 |
} |
| 322 |
} |
| 323 |
else |
| 324 |
{ |
| 325 |
unsigned chunk; |
| 326 |
|
| 327 |
do |
| 328 |
{ |
| 329 |
chunk = strtoul(getline().c_str(), 0, 16); |
| 330 |
|
| 331 |
unsigned left = chunk; |
| 332 |
|
| 333 |
while (left > 0) |
| 334 |
{ |
| 335 |
memset(buffer, 0, BUFSIZ + 1); |
| 336 |
|
| 337 |
unsigned bytes = left > BUFSIZ ? BUFSIZ : left; |
| 338 |
unsigned received; |
| 339 |
|
| 340 |
while (true) |
| 341 |
{ |
| 342 |
if ((received = recv(http, buffer, bytes, 0)) == |
| 343 |
SOCKET_ERROR) |
| 344 |
{ |
| 345 |
error(program + ": Recv"); |
| 346 |
exit(1); |
| 347 |
} |
| 348 |
else if (received != bytes) |
| 349 |
{ |
| 350 |
left -= received; |
| 351 |
page += buffer; |
| 352 |
|
| 353 |
memset(buffer, 0, BUFSIZ + 1); |
| 354 |
|
| 355 |
bytes -= received; |
| 356 |
} |
| 357 |
else |
| 358 |
{ |
| 359 |
break; |
| 360 |
} |
| 361 |
} |
| 362 |
|
| 363 |
page += buffer; |
| 364 |
left -= bytes; |
| 365 |
} |
| 366 |
|
| 367 |
getline(); |
| 368 |
length += chunk; |
| 369 |
} |
| 370 |
while (chunk > 0); |
| 371 |
} |
| 372 |
|
| 373 |
for (unsigned index = 0; index < page.length(); index++) |
| 374 |
{ |
| 375 |
if (page[index] == '\r' && (index + 1 < page.length()) ? page[index + |
| 376 |
1] == '\n' : false) |
| 377 |
{ |
| 378 |
page.erase(index, 1); |
| 379 |
} |
| 380 |
else if (page[index] == '\r') |
| 381 |
{ |
| 382 |
page[index] = '\n'; |
| 383 |
} |
| 384 |
} |
| 385 |
} |
| 386 |
|
| 387 |
void HttpHandler::putline(const string line) |
| 388 |
{ |
| 389 |
sprintf(buffer, "%s\r\n", line.c_str()); |
| 390 |
if (send(http, buffer, strlen(buffer), 0) == SOCKET_ERROR) |
| 391 |
{ |
| 392 |
error(program + ": Send"); |
| 393 |
exit(1); |
| 394 |
} |
| 395 |
} |
| 396 |
|
| 397 |
string HttpHandler::getline() |
| 398 |
{ |
| 399 |
string line; |
| 400 |
char byte; |
| 401 |
|
| 402 |
do |
| 403 |
{ |
| 404 |
if (recv(http, &byte, 1, 0) == SOCKET_ERROR) |
| 405 |
{ |
| 406 |
error(program + ": Recv"); |
| 407 |
} |
| 408 |
|
| 409 |
if (byte != '\r' && byte != '\n') |
| 410 |
{ |
| 411 |
line += byte; |
| 412 |
} |
| 413 |
} |
| 414 |
while (byte != '\n'); |
| 415 |
|
| 416 |
return line; |
| 417 |
} |
| 418 |
|
| 419 |
void HttpHandler::error(const string& prefix, bool host) |
| 420 |
{ |
| 421 |
#ifdef _WIN32 |
| 422 |
string error; |
| 423 |
|
| 424 |
switch (WSAGetLastError()) |
| 425 |
{ |
| 426 |
case WSAEACCES: |
| 427 |
error = "Permission denied."; |
| 428 |
break; |
| 429 |
case WSAEADDRINUSE: |
| 430 |
error = "Address already in use."; |
| 431 |
break; |
| 432 |
case WSAEADDRNOTAVAIL: |
| 433 |
error = "Cannot assign requested address."; |
| 434 |
break; |
| 435 |
case WSAEAFNOSUPPORT: |
| 436 |
error = "Address family not supported by protocol family."; |
| 437 |
break; |
| 438 |
case WSAEALREADY: |
| 439 |
error = "Operation already in progress."; |
| 440 |
break; |
| 441 |
case WSAECONNABORTED: |
| 442 |
error = "Software caused connection abort."; |
| 443 |
break; |
| 444 |
case WSAECONNREFUSED: |
| 445 |
error = "Connection refused."; |
| 446 |
break; |
| 447 |
case WSAECONNRESET: |
| 448 |
error = "Connection reset by peer."; |
| 449 |
break; |
| 450 |
case WSAEDESTADDRREQ: |
| 451 |
error = "Destination address required."; |
| 452 |
break; |
| 453 |
case WSAEFAULT: |
| 454 |
error = "Bad address."; |
| 455 |
break; |
| 456 |
case WSAEHOSTDOWN: |
| 457 |
error = "Host is down."; |
| 458 |
break; |
| 459 |
case WSAEHOSTUNREACH: |
| 460 |
error = "No route to host."; |
| 461 |
break; |
| 462 |
case WSAEINPROGRESS: |
| 463 |
error = "Operation now in progress."; |
| 464 |
break; |
| 465 |
case WSAEINTR: |
| 466 |
error = "Interrupted function call."; |
| 467 |
break; |
| 468 |
case WSAEINVAL: |
| 469 |
error = "Invalid argument."; |
| 470 |
break; |
| 471 |
case WSAEISCONN: |
| 472 |
error = "Socket is already connected."; |
| 473 |
break; |
| 474 |
case WSAEMFILE: |
| 475 |
error = "Too many open files."; |
| 476 |
break; |
| 477 |
case WSAEMSGSIZE: |
| 478 |
error = "Message too long."; |
| 479 |
break; |
| 480 |
case WSAENETDOWN: |
| 481 |
error = "Network is down."; |
| 482 |
break; |
| 483 |
case WSAENETRESET: |
| 484 |
error = "Network dropped connection on reset."; |
| 485 |
break; |
| 486 |
case WSAENETUNREACH: |
| 487 |
error = "Network is unreachable."; |
| 488 |
break; |
| 489 |
case WSAENOBUFS: |
| 490 |
error = "No buffer space available."; |
| 491 |
break; |
| 492 |
case WSAENOPROTOOPT: |
| 493 |
error = "Bad protocol option."; |
| 494 |
break; |
| 495 |
case WSAENOTCONN: |
| 496 |
error = "Socket is not connected."; |
| 497 |
break; |
| 498 |
case WSAENOTSOCK: |
| 499 |
error = "Socket operation on non-socket."; |
| 500 |
break; |
| 501 |
case WSAEOPNOTSUPP: |
| 502 |
error = "Operation not supported."; |
| 503 |
break; |
| 504 |
case WSAEPFNOSUPPORT: |
| 505 |
error = "Protocol family not supported."; |
| 506 |
break; |
| 507 |
case WSAEPROCLIM: |
| 508 |
error = "Too many processes."; |
| 509 |
break; |
| 510 |
case WSAEPROTONOSUPPORT: |
| 511 |
error = "Protocol not supported."; |
| 512 |
break; |
| 513 |
case WSAEPROTOTYPE: |
| 514 |
error = "Protocol wrong type for socket."; |
| 515 |
break; |
| 516 |
case WSAESHUTDOWN: |
| 517 |
error = "Cannot send after socket shutdown."; |
| 518 |
break; |
| 519 |
case WSAESOCKTNOSUPPORT: |
| 520 |
error = "Socket type not supported."; |
| 521 |
break; |
| 522 |
case WSAETIMEDOUT: |
| 523 |
error = "Connection timed out."; |
| 524 |
break; |
| 525 |
case WSATYPE_NOT_FOUND: |
| 526 |
error = "Class type not found."; |
| 527 |
break; |
| 528 |
case WSAEWOULDBLOCK: |
| 529 |
error = "Resource temporarily unavailable."; |
| 530 |
break; |
| 531 |
case WSAHOST_NOT_FOUND: |
| 532 |
error = "Host not found."; |
| 533 |
break; |
| 534 |
case WSA_INVALID_HANDLE: |
| 535 |
error = "Specified event object handle is invalid."; |
| 536 |
break; |
| 537 |
case WSA_INVALID_PARAMETER: |
| 538 |
error = "One or more parameters are invalid."; |
| 539 |
break; |
| 540 |
// case WSAINVALIDPROCTABLE: |
| 541 |
// error = "Invalid procedure table from service provider."; |
| 542 |
// break; |
| 543 |
// case WSAINVALIDPROVIDER: |
| 544 |
// error = "Invalid service provider version number."; |
| 545 |
// break; |
| 546 |
case WSA_IO_INCOMPLETE: |
| 547 |
error = "Overlapped I/O event object not in signaled state."; |
| 548 |
break; |
| 549 |
case WSA_IO_PENDING: |
| 550 |
error = "Overlapped operations will complete later."; |
| 551 |
break; |
| 552 |
case WSA_NOT_ENOUGH_MEMORY: |
| 553 |
error = "Insufficient memory available."; |
| 554 |
break; |
| 555 |
case WSANOTINITIALISED: |
| 556 |
error = "Successful WSAStartup not yet performed."; |
| 557 |
break; |
| 558 |
case WSANO_DATA: |
| 559 |
error = "Valid name, no data record of requested type."; |
| 560 |
break; |
| 561 |
case WSANO_RECOVERY: |
| 562 |
error = "This is a non-recoverable error."; |
| 563 |
break; |
| 564 |
// case WSAPROVIDERFAILEDINIT: |
| 565 |
// error = "Unable to initialize a service provider."; |
| 566 |
// break; |
| 567 |
case WSASYSCALLFAILURE: |
| 568 |
error = "System call failure."; |
| 569 |
break; |
| 570 |
case WSASYSNOTREADY: |
| 571 |
error = "Network subsystem is unavailable."; |
| 572 |
break; |
| 573 |
case WSATRY_AGAIN: |
| 574 |
error = "Non-authoritative host not found."; |
| 575 |
break; |
| 576 |
case WSAVERNOTSUPPORTED: |
| 577 |
error = "WINSOCK.DLL version out of range."; |
| 578 |
break; |
| 579 |
case WSAEDISCON: |
| 580 |
error = "Graceful shutdown in progress."; |
| 581 |
break; |
| 582 |
case WSA_OPERATION_ABORTED: |
| 583 |
error = "Overlapped operation aborted."; |
| 584 |
break; |
| 585 |
default: |
| 586 |
error = "Unknown error."; |
| 587 |
break; |
| 588 |
} |
| 589 |
|
| 590 |
cerr << prefix << ": " << error << "\n"; |
| 591 |
#else |
| 592 |
if (host) |
| 593 |
{ |
| 594 |
herror(prefix.c_str()); |
| 595 |
} |
| 596 |
else |
| 597 |
{ |
| 598 |
perror(prefix.c_str()); |
| 599 |
} |
| 600 |
#endif // _WIN32 |
| 601 |
} |
