| 1 |
/* ============================================================================ |
| 2 |
* Douglas Thrift's Search Engine License |
| 3 |
* |
| 4 |
* Copyright (C) 2002, 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 |
HttpHandler::HttpHandler() |
| 54 |
{ |
| 55 |
buffer = new char[BUFSIZ + 1]; |
| 56 |
|
| 57 |
#ifdef _WIN32 |
| 58 |
if (WSAStartup(MAKEWORD(2, 0), &data) != 0) |
| 59 |
{ |
| 60 |
error(program + ": WSAStartup"); |
| 61 |
exit(1); |
| 62 |
} |
| 63 |
#endif // _WIN32 |
| 64 |
|
| 65 |
begin = 0; |
| 66 |
} |
| 67 |
|
| 68 |
HttpHandler::~HttpHandler() |
| 69 |
{ |
| 70 |
delete [] buffer; |
| 71 |
|
| 72 |
#ifdef _WIN32 |
| 73 |
WSACleanup(); |
| 74 |
#endif // _WIN32 |
| 75 |
} |
| 76 |
|
| 77 |
bool HttpHandler::handle(URL &url, bool head) |
| 78 |
{ |
| 79 |
bool answer = false; |
| 80 |
|
| 81 |
if ((http = socket(PF_INET, SOCK_STREAM, 0)) == INVALID_SOCKET) |
| 82 |
{ |
| 83 |
error(program + ": Socket"); |
| 84 |
exit(1); |
| 85 |
} |
| 86 |
|
| 87 |
sockaddr_in address; |
| 88 |
hostent* host; |
| 89 |
|
| 90 |
address.sin_family = AF_INET; |
| 91 |
|
| 92 |
if ((host = gethostbyname(url.getAddress().c_str())) == NULL) |
| 93 |
{ |
| 94 |
error(program + ": Host: " + url.getAddress(), true); |
| 95 |
return answer; |
| 96 |
} |
| 97 |
|
| 98 |
address.sin_addr = *((in_addr*)*host->h_addr_list); |
| 99 |
address.sin_port = htons(url.getPort()); |
| 100 |
|
| 101 |
if (connect(http, (sockaddr*)&address, sizeof(sockaddr_in)) == |
| 102 |
SOCKET_ERROR) |
| 103 |
{ |
| 104 |
error(program + ": Connect"); |
| 105 |
return answer; |
| 106 |
} |
| 107 |
|
| 108 |
if (head) |
| 109 |
{ |
| 110 |
putline("HEAD " + url.getPath() + " HTTP/1.1"); |
| 111 |
} |
| 112 |
else |
| 113 |
{ |
| 114 |
putline("GET " + url.getPath() + " HTTP/1.1"); |
| 115 |
} |
| 116 |
|
| 117 |
putline("Accept: text/html; text/plain"); |
| 118 |
putline("User-Agent: " + agent(true) + ' ' + platform()); |
| 119 |
|
| 120 |
if (url.getPort() == 80) |
| 121 |
{ |
| 122 |
putline("Host: " + url.getAddress()); |
| 123 |
} |
| 124 |
else |
| 125 |
{ |
| 126 |
char* port = new char[1024]; |
| 127 |
sprintf(port, "%u", url.getPort()); |
| 128 |
|
| 129 |
putline("Host: " + url.getAddress() + ':' + port); |
| 130 |
|
| 131 |
delete [] port; |
| 132 |
} |
| 133 |
|
| 134 |
putline(); |
| 135 |
|
| 136 |
string line = getline(); |
| 137 |
|
| 138 |
if (line.find("HTTP/") != 0) |
| 139 |
{ |
| 140 |
return answer; |
| 141 |
} |
| 142 |
|
| 143 |
unsigned dot = line.find('.'); |
| 144 |
unsigned space = line.find(' '); |
| 145 |
|
| 146 |
unsigned major = strtoul(line.substr(5, dot - 5).c_str(), 0, 0); |
| 147 |
unsigned minor = strtoul(line.substr(dot + 1, space - dot - 1).c_str(), 0, |
| 148 |
0); |
| 149 |
|
| 150 |
if (major > 1 || minor < 1) |
| 151 |
{ |
| 152 |
cerr << program << ": Potentially Incompatible Server: HTTP/" << major |
| 153 |
<< "." << minor << "\n"; |
| 154 |
|
| 155 |
return answer; |
| 156 |
} |
| 157 |
|
| 158 |
code response = code(strtoul(line.substr(space + 1).c_str(), 0, 0)); |
| 159 |
|
| 160 |
do |
| 161 |
{ |
| 162 |
line = getline(); |
| 163 |
} |
| 164 |
while (line != ""); |
| 165 |
|
| 166 |
switch (response) |
| 167 |
{ |
| 168 |
case ok: |
| 169 |
answer = true; |
| 170 |
break; |
| 171 |
case choices: |
| 172 |
break; |
| 173 |
case moved: |
| 174 |
break; |
| 175 |
case found: |
| 176 |
break; |
| 177 |
case notfound: |
| 178 |
break; |
| 179 |
case internal: |
| 180 |
break; |
| 181 |
default: |
| 182 |
break; |
| 183 |
} |
| 184 |
|
| 185 |
return answer; |
| 186 |
} |
| 187 |
|
| 188 |
HttpHandler& HttpHandler::getline(string& line, char endline) |
| 189 |
{ |
| 190 |
int end = page.find(endline, begin); |
| 191 |
int newline = page.find('\n', begin); |
| 192 |
|
| 193 |
if (newline < end || end == string::npos) |
| 194 |
{ |
| 195 |
end = newline; |
| 196 |
} |
| 197 |
|
| 198 |
line = page.substr(begin, end - begin); |
| 199 |
|
| 200 |
if (end == string::npos) |
| 201 |
{ |
| 202 |
begin = end; |
| 203 |
} |
| 204 |
else |
| 205 |
{ |
| 206 |
begin = end + 1; |
| 207 |
} |
| 208 |
|
| 209 |
return *this; |
| 210 |
} |
| 211 |
|
| 212 |
bool HttpHandler::good() |
| 213 |
{ |
| 214 |
bool answer = true; |
| 215 |
|
| 216 |
if (begin >= page.length()) |
| 217 |
{ |
| 218 |
answer = false; |
| 219 |
} |
| 220 |
else if (begin == string::npos) |
| 221 |
{ |
| 222 |
answer = false; |
| 223 |
} |
| 224 |
|
| 225 |
return answer; |
| 226 |
} |
| 227 |
|
| 228 |
void HttpHandler::clear() |
| 229 |
{ |
| 230 |
type = ""; |
| 231 |
length = 0; |
| 232 |
location = ""; |
| 233 |
begin = 0; |
| 234 |
page = ""; |
| 235 |
} |
| 236 |
|
| 237 |
void HttpHandler::putline(const string line) |
| 238 |
{ |
| 239 |
sprintf(buffer, "%s\r\n", line.c_str()); |
| 240 |
if (send(http, buffer, strlen(buffer), 0) == SOCKET_ERROR) |
| 241 |
{ |
| 242 |
error(program + ": Send"); |
| 243 |
exit(1); |
| 244 |
} |
| 245 |
} |
| 246 |
|
| 247 |
string HttpHandler::getline() |
| 248 |
{ |
| 249 |
string line; |
| 250 |
char byte; |
| 251 |
|
| 252 |
do |
| 253 |
{ |
| 254 |
if (recv(http, &byte, 1, 0) == SOCKET_ERROR) |
| 255 |
{ |
| 256 |
error(program + ": Recv"); |
| 257 |
} |
| 258 |
|
| 259 |
if (byte != '\r' && byte != '\n') |
| 260 |
{ |
| 261 |
line += byte; |
| 262 |
} |
| 263 |
} |
| 264 |
while (byte != '\n'); |
| 265 |
|
| 266 |
return line; |
| 267 |
} |
| 268 |
|
| 269 |
void HttpHandler::error(const string prefix, bool host) |
| 270 |
{ |
| 271 |
#ifdef _WIN32 |
| 272 |
string error; |
| 273 |
|
| 274 |
switch (WSAGetLastError()) |
| 275 |
{ |
| 276 |
case WSAEACCES: |
| 277 |
error = "Permission denied."; |
| 278 |
break; |
| 279 |
case WSAEADDRINUSE: |
| 280 |
error = "Address already in use."; |
| 281 |
break; |
| 282 |
case WSAEADDRNOTAVAIL: |
| 283 |
error = "Cannot assign requested address."; |
| 284 |
break; |
| 285 |
case WSAEAFNOSUPPORT: |
| 286 |
error = "Address family not supported by protocol family."; |
| 287 |
break; |
| 288 |
case WSAEALREADY: |
| 289 |
error = "Operation already in progress."; |
| 290 |
break; |
| 291 |
case WSAECONNABORTED: |
| 292 |
error = "Software caused connection abort."; |
| 293 |
break; |
| 294 |
case WSAECONNREFUSED: |
| 295 |
error = "Connection refused."; |
| 296 |
break; |
| 297 |
case WSAECONNRESET: |
| 298 |
error = "Connection reset by peer."; |
| 299 |
break; |
| 300 |
case WSAEDESTADDRREQ: |
| 301 |
error = "Destination address required."; |
| 302 |
break; |
| 303 |
case WSAEFAULT: |
| 304 |
error = "Bad address."; |
| 305 |
break; |
| 306 |
case WSAEHOSTDOWN: |
| 307 |
error = "Host is down."; |
| 308 |
break; |
| 309 |
case WSAEHOSTUNREACH: |
| 310 |
error = "No route to host."; |
| 311 |
break; |
| 312 |
case WSAEINPROGRESS: |
| 313 |
error = "Operation now in progress."; |
| 314 |
break; |
| 315 |
case WSAEINTR: |
| 316 |
error = "Interrupted function call."; |
| 317 |
break; |
| 318 |
case WSAEINVAL: |
| 319 |
error = "Invalid argument."; |
| 320 |
break; |
| 321 |
case WSAEISCONN: |
| 322 |
error = "Socket is already connected."; |
| 323 |
break; |
| 324 |
case WSAEMFILE: |
| 325 |
error = "Too many open files."; |
| 326 |
break; |
| 327 |
case WSAEMSGSIZE: |
| 328 |
error = "Message too long."; |
| 329 |
break; |
| 330 |
case WSAENETDOWN: |
| 331 |
error = "Network is down."; |
| 332 |
break; |
| 333 |
case WSAENETRESET: |
| 334 |
error = "Network dropped connection on reset."; |
| 335 |
break; |
| 336 |
case WSAENETUNREACH: |
| 337 |
error = "Network is unreachable."; |
| 338 |
break; |
| 339 |
case WSAENOBUFS: |
| 340 |
error = "No buffer space available."; |
| 341 |
break; |
| 342 |
case WSAENOPROTOOPT: |
| 343 |
error = "Bad protocol option."; |
| 344 |
break; |
| 345 |
case WSAENOTCONN: |
| 346 |
error = "Socket is not connected."; |
| 347 |
break; |
| 348 |
case WSAENOTSOCK: |
| 349 |
error = "Socket operation on non-socket."; |
| 350 |
break; |
| 351 |
case WSAEOPNOTSUPP: |
| 352 |
error = "Operation not supported."; |
| 353 |
break; |
| 354 |
case WSAEPFNOSUPPORT: |
| 355 |
error = "Protocol family not supported."; |
| 356 |
break; |
| 357 |
case WSAEPROCLIM: |
| 358 |
error = "Too many processes."; |
| 359 |
break; |
| 360 |
case WSAEPROTONOSUPPORT: |
| 361 |
error = "Protocol not supported."; |
| 362 |
break; |
| 363 |
case WSAEPROTOTYPE: |
| 364 |
error = "Protocol wrong type for socket."; |
| 365 |
break; |
| 366 |
case WSAESHUTDOWN: |
| 367 |
error = "Cannot send after socket shutdown."; |
| 368 |
break; |
| 369 |
case WSAESOCKTNOSUPPORT: |
| 370 |
error = "Socket type not supported."; |
| 371 |
break; |
| 372 |
case WSAETIMEDOUT: |
| 373 |
error = "Connection timed out."; |
| 374 |
break; |
| 375 |
case WSATYPE_NOT_FOUND: |
| 376 |
error = "Class type not found."; |
| 377 |
break; |
| 378 |
case WSAEWOULDBLOCK: |
| 379 |
error = "Resource temporarily unavailable."; |
| 380 |
break; |
| 381 |
case WSAHOST_NOT_FOUND: |
| 382 |
error = "Host not found."; |
| 383 |
break; |
| 384 |
case WSA_INVALID_HANDLE: |
| 385 |
error = "Specified event object handle is invalid."; |
| 386 |
break; |
| 387 |
case WSA_INVALID_PARAMETER: |
| 388 |
error = "One or more parameters are invalid."; |
| 389 |
break; |
| 390 |
// case WSAINVALIDPROCTABLE: |
| 391 |
// error = "Invalid procedure table from service provider."; |
| 392 |
// break; |
| 393 |
// case WSAINVALIDPROVIDER: |
| 394 |
// error = "Invalid service provider version number."; |
| 395 |
// break; |
| 396 |
case WSA_IO_INCOMPLETE: |
| 397 |
error = "Overlapped I/O event object not in signaled state."; |
| 398 |
break; |
| 399 |
case WSA_IO_PENDING: |
| 400 |
error = "Overlapped operations will complete later."; |
| 401 |
break; |
| 402 |
case WSA_NOT_ENOUGH_MEMORY: |
| 403 |
error = "Insufficient memory available."; |
| 404 |
break; |
| 405 |
case WSANOTINITIALISED: |
| 406 |
error = "Successful WSAStartup not yet performed."; |
| 407 |
break; |
| 408 |
case WSANO_DATA: |
| 409 |
error = "Valid name, no data record of requested type."; |
| 410 |
break; |
| 411 |
case WSANO_RECOVERY: |
| 412 |
error = "This is a non-recoverable error."; |
| 413 |
break; |
| 414 |
// case WSAPROVIDERFAILEDINIT: |
| 415 |
// error = "Unable to initialize a service provider."; |
| 416 |
// break; |
| 417 |
case WSASYSCALLFAILURE: |
| 418 |
error = "System call failure."; |
| 419 |
break; |
| 420 |
case WSASYSNOTREADY: |
| 421 |
error = "Network subsystem is unavailable."; |
| 422 |
break; |
| 423 |
case WSATRY_AGAIN: |
| 424 |
error = "Non-authoritative host not found."; |
| 425 |
break; |
| 426 |
case WSAVERNOTSUPPORTED: |
| 427 |
error = "WINSOCK.DLL version out of range."; |
| 428 |
break; |
| 429 |
case WSAEDISCON: |
| 430 |
error = "Graceful shutdown in progress."; |
| 431 |
break; |
| 432 |
case WSA_OPERATION_ABORTED: |
| 433 |
error = "Overlapped operation aborted."; |
| 434 |
break; |
| 435 |
default: |
| 436 |
error = "Unknown error."; |
| 437 |
break; |
| 438 |
} |
| 439 |
|
| 440 |
cerr << prefix << ": " << error << "\n"; |
| 441 |
#else |
| 442 |
if (host) |
| 443 |
{ |
| 444 |
herror(prefix.c_str()); |
| 445 |
} |
| 446 |
else |
| 447 |
{ |
| 448 |
perror(prefix.c_str()); |
| 449 |
} |
| 450 |
#endif // _WIN32 |
| 451 |
} |
