ViewVC Help

View File | Revision Log | Show Annotations | Download File | View Changeset | Root Listing

root/proj/trunk/Search/HttpHandler.cpp

Comparing trunk/Search/HttpHandler.cpp (file contents):
Revision 13 by douglas, 2002-12-06T19:56:56-08:00 vs.
Revision 179 by douglas, 2003-07-05T19:47:23-07:00

#	Line 1 | Line 1
1		/* ============================================================================
2		* Douglas Thrift's Search Engine License
3		*
4	<	* Copyright (C) 2002, Douglas Thrift. All Rights Reserved.
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		*
#	Line 50 | Line 50
50
51		#include "HttpHandler.h"
52
53	+	// Lovely C Sockets!
54	+	#ifndef _WIN32
55	+	// BSD Sockets
56	+	#include <unistd.h>
57	+	#include <sys/types.h>
58	+	#include <sys/socket.h>
59	+	#include <netinet/in.h>
60	+	#include <netdb.h>
61	+
62	+	#define INVALID_SOCKET -1
63	+	#define SOCKET_ERROR -1
64	+
65	+	inline int closesocket(SOCKET s) { return close(s); }
66	+	#endif
67	+
68		HttpHandler::HttpHandler()
69		{
70	+	buffer = new char[BUFSIZ + 1];
71	+
72		#ifdef _WIN32
73	<	WSADATA data;
57	<	if (WSAStartup(MAKEWORD(2, 0) != 0, &data))
73	>	if (WSAStartup(MAKEWORD(2, 0), &data) != 0)
74		{
75	<	error(program + ": WSAStartup()");
75	>	error(program + ": WSAStartup");
76		exit(1);
77		}
78		#endif // _WIN32
79
80	<	begin = 0;
80	>	length = 0;
81	>	chunked = false;
82		}
83
84		HttpHandler::~HttpHandler()
85		{
86	+	delete [] buffer;
87	+
88		#ifdef _WIN32
89		WSACleanup();
90		#endif // _WIN32
91		}
92
93	<	bool HttpHandler::connect(URL &url, bool head)
93	>	bool HttpHandler::handle(URL &url, const string referer, bool head)
94		{
95		bool answer = false;
96
97	<	//      if (status != JNI_ERR)
98	<	//      {
99	<	//              if (cls != 0)
100	<	//              {
101	<	//                      if (mid != 0)
102	<	//                      {
103	<	//                              jstring addressJ = env->NewStringUTF(url.getAddress().c_str());
104	<	//                              jint portJ = url.getPort();
105	<	//                              jstring pathJ = env->NewStringUTF(url.getPath().c_str());
106	<	//                              jstring programNameJ = env->NewStringUTF(programName.c_str());
107	<	//                              jstring programVersionJ =
108	<	//                                      env->NewStringUTF(programVersion.c_str());
109	<
110	<	//                              jstring pageJ = (jstring)env->CallStaticObjectMethod(cls, mid,
111	<	//                                      addressJ, portJ, pathJ, programNameJ, programVersionJ);
112	<
113	<	//                              const char* pageC = env->GetStringUTFChars(pageJ, 0);
114	<	//                              page = pageC;
115	<	//                              env->ReleaseStringUTFChars(pageJ, pageC);
116	<
117	<	//                              if (page != "unknown host\n" && page != "io exception\n" &&
118	<	//                                      page != "bad headers\n") answer = true;
119	<	//                      }
120	<	//              }
121	<	//      }
97	>	if ((http = socket(PF_INET, SOCK_STREAM, 0)) == INVALID_SOCKET)
98	>	{
99	>	error(program + ": Socket");
100	>	exit(1);
101	>	}
102	>
103	>	sockaddr_in address;
104	>	hostent* host;
105	>
106	>	address.sin_family = AF_INET;
107	>
108	>	if ((host = gethostbyname(url.getAddress().c_str())) == NULL)
109	>	{
110	>	error(program + ": Host: " + url.getAddress(), true);
111	>	return answer;
112	>	}
113	>
114	>	address.sin_addr = *((in_addr*)*host->h_addr_list);
115	>	address.sin_port = htons(url.getPort());
116	>
117	>	if (connect(http, (sockaddr*)&address, sizeof(sockaddr_in)) ==
118	>	SOCKET_ERROR)
119	>	{
120	>	error(program + ": Connect");
121	>	return answer;
122	>	}
123	>
124	>	if (head)
125	>	{
126	>	putline("HEAD " + url.getPath() + " HTTP/1.1");
127	>	}
128	>	else
129	>	{
130	>	putline("GET " + url.getPath() + " HTTP/1.1");
131	>	}
132	>
133	>	putline("Accept: text/html; text/plain");
134	>	putline("User-Agent: " + agent(true) + ' ' + platform());
135	>
136	>	if (url.getPort() == 80)
137	>	{
138	>	putline("Host: " + url.getAddress());
139	>	}
140	>	else
141	>	{
142	>	char* port = new char[1024];
143	>	sprintf(port, "%u", url.getPort());
144	>
145	>	putline("Host: " + url.getAddress() + ':' + port);
146	>
147	>	delete [] port;
148	>	}
149	>
150	>	if (referer != "")
151	>	{
152	>	putline("Referer: " + referer);
153	>	}
154	>
155	>	putline("Connection: close");
156	>	putline();
157	>
158	>	code response;
159	>	string line;
160	>
161	>	do
162	>	{
163	>	line = getline();
164	>
165	>	if (line.find("HTTP/") != 0)
166	>	{
167	>	return answer;
168	>	}
169	>
170	>	unsigned dot = line.find('.');
171	>	unsigned space = line.find(' ');
172	>
173	>	unsigned major = strtoul(line.substr(5, dot - 5).c_str(), 0, 10);
174	>	unsigned minor = strtoul(line.substr(dot + 1, space - dot - 1).c_str(),
175	>	0, 10);
176	>
177	>	if (major > 1)
178	>	{
179	>	cerr << program << ": Potentially Incompatible Server: HTTP/" <<
180	>	major << "." << minor << "\n";
181	>
182	>	return answer;
183	>	}
184	>
185	>	response = code(strtoul(line.substr(space + 1).c_str(), 0, 10));
186	>
187	>	if (response < ok) do line = getline(); while (line != "");
188	>	}
189	>	while (response < ok);
190	>
191	>	do
192	>	{
193	>	line = getline();
194	>
195	>	if (line != "")
196	>	{
197	>	unsigned colon = line.find(':');
198	>
199	>	string field = line.substr(0, colon);
200	>	string value = line.substr(colon + 1);
201	>
202	>	while (isspace(value[0])) value.erase(0, 1);
203	>
204	>	if (field == "Content-Type")
205	>	{
206	>	type = value;
207	>	}
208	>	else if (field == "Content-Length")
209	>	{
210	>	length = strtoul(value.c_str(), 0, 10);
211	>	}
212	>	else if (field == "Location")
213	>	{
214	>	location = value;
215	>	}
216	>	else if (field == "Transfer-Encoding")
217	>	{
218	>	chunked = value == "chunked";
219	>	}
220	>	}
221	>	}
222	>	while (line != "");
223	>
224	>	switch (response)
225	>	{
226	>	case ok:
227	>	if (debug) cerr << "response = " << response << "\n";
228	>	answer = true;
229	>	break;
230	>	case choices:
231	>	case moved:
232	>	case found:
233	>	if (debug) cerr << "response = " << response << "\n"
234	>	<< "location = " << location << "\n";
235	>	location = getLink(location, url);
236	>	break;
237	>	case notfound:
238	>	case internal:
239	>	if (debug) cerr << "response = " << response << "\n";
240	>	break;
241	>	default:
242	>	if (debug) cerr << "response = " << response << "\n";
243	>	if (response <= 299)
244	>	{
245	>	answer = true;
246	>	}
247	>	else if (response <= 399)
248	>	{
249	>	location = getLink(location, url);
250	>	}
251	>	break;
252	>	}
253	>
254	>	if (!head && answer) populate();
255
256		return answer;
257		}
258
259		HttpHandler& HttpHandler::getline(string& line, char endline)
260		{
261	<	int end = page.find(endline, begin);
262	<	int newline = page.find('\n', begin);
261	>	unsigned end = page.find(endline);
262	>	unsigned newline = page.find('\n');
263
264		if (newline < end || end == string::npos)
265		{
266		end = newline;
267		}
268
269	<	line = page.substr(begin, end - begin);
269	>	line = page.substr(0, end);
270	>	page.erase(0, (end == string::npos ? end : end + 1));
271	>
272	>	return *this;
273	>	}
274
275	<	if (end == string::npos)
275	>	void HttpHandler::clear()
276	>	{
277	>	closesocket(http);
278	>
279	>	type = "";
280	>	length = 0;
281	>	location = "";
282	>	page = "";
283	>	chunked = false;
284	>	}
285	>
286	>	void HttpHandler::populate()
287	>	{
288	>	if (!chunked)
289		{
290	<	begin = end;
290	>	unsigned left = length;
291	>
292	>	while (left > 0)
293	>	{
294	>	memset(buffer, 0, BUFSIZ + 1);
295	>
296	>	unsigned bytes = left > BUFSIZ ? BUFSIZ : left;
297	>	unsigned received;
298	>
299	>	while (true)
300	>	{
301	>	if ((received = recv(http, buffer, bytes, 0)) == SOCKET_ERROR)
302	>	{
303	>	error(program + ": Recv");
304	>	exit(1);
305	>	}
306	>	else if (received != bytes)
307	>	{
308	>	left -= received;
309	>	page += buffer;
310	>
311	>	memset(buffer, 0, BUFSIZ + 1);
312	>
313	>	bytes -= received;
314	>	}
315	>	else
316	>	{
317	>	break;
318	>	}
319	>	}
320	>
321	>	page += buffer;
322	>	left -= bytes;
323	>	}
324		}
325		else
326		{
327	<	begin = end + 1;
327	>	unsigned chunk;
328	>
329	>	do
330	>	{
331	>	chunk = strtoul(getline().c_str(), 0, 16);
332	>
333	>	unsigned left = chunk;
334	>
335	>	while (left > 0)
336	>	{
337	>	memset(buffer, 0, BUFSIZ + 1);
338	>
339	>	unsigned bytes = left > BUFSIZ ? BUFSIZ : left;
340	>	unsigned received;
341	>
342	>	while (true)
343	>	{
344	>	if ((received = recv(http, buffer, bytes, 0)) ==
345	>	SOCKET_ERROR)
346	>	{
347	>	error(program + ": Recv");
348	>	exit(1);
349	>	}
350	>	else if (received != bytes)
351	>	{
352	>	left -= received;
353	>	page += buffer;
354	>
355	>	memset(buffer, 0, BUFSIZ + 1);
356	>
357	>	bytes -= received;
358	>	}
359	>	else
360	>	{
361	>	break;
362	>	}
363	>	}
364	>
365	>	page += buffer;
366	>	left -= bytes;
367	>	}
368	>
369	>	getline();
370	>	length += chunk;
371	>	}
372	>	while (chunk > 0);
373		}
374
375	<	return *this;
375	>	for (unsigned index = 0; index < page.length(); index++)
376	>	{
377	>	if (page[index] == '\r' && (index + 1 < page.length()) ? page[index +
378	>	1] == '\n' : false)
379	>	{
380	>	page.erase(index, 1);
381	>	}
382	>	else if (page[index] == '\r')
383	>	{
384	>	page[index] = '\n';
385	>	}
386	>	}
387		}
388
389	<	bool HttpHandler::good()
389	>	void HttpHandler::putline(const string line)
390		{
391	<	bool answer = true;
392	<
135	<	if (begin >= page.length())
391	>	sprintf(buffer, "%s\r\n", line.c_str());
392	>	if (send(http, buffer, strlen(buffer), 0) == SOCKET_ERROR)
393		{
394	<	answer = false;
395	<	}
139	<	else if (begin == string::npos)
140	<	{
141	<	answer = false;
394	>	error(program + ": Send");
395	>	exit(1);
396		}
143	–
144	–	return answer;
397		}
398
399	<	void HttpHandler::clear()
399	>	string HttpHandler::getline()
400		{
401	<	begin = 0;
402	<	page = "";
401	>	string line;
402	>	char byte;
403	>
404	>	do
405	>	{
406	>	if (recv(http, &byte, 1, 0) == SOCKET_ERROR)
407	>	{
408	>	error(program + ": Recv");
409	>	}
410	>
411	>	if (byte != '\r' && byte != '\n')
412	>	{
413	>	line += byte;
414	>	}
415	>	}
416	>	while (byte != '\n');
417	>
418	>	return line;
419		}
420
421	<	void HttpHandler::error(const string prefix, bool host)
421	>	void HttpHandler::error(const string& prefix, bool host)
422		{
423		#ifdef _WIN32
424		string error;
#	Line 325 | Line 593 | void HttpHandler::error(const string pre
593		#else
594		if (host)
595		{
596	<	herror(prefix.c_str());
596	>	string error;
597	>
598	>	switch (h_errno)
599	>	{
600	>	case HOST_NOT_FOUND:
601	>	error = "Unknown host";
602	>	break;
603	>	case TRY_AGAIN:
604	>	error = "Host name lookup failure";
605	>	break;
606	>	case NO_RECOVERY:
607	>	error = "Unknown server error";
608	>	break;
609	>	case NO_DATA:
610	>	error = "No address associated with name";
611	>	break;
612	>	default:
613	>	error = "Unknown error";
614	>	break;
615	>	}
616	>
617	>	cerr << prefix << ": " << error << "\n";
618		}
619		else
620		{

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines