Annotation of parser3/src/main/pa_common.C, revision 1.199
1.15 paf 1: /** @file
1.16 paf 2: Parser: commonly functions.
3:
1.174 paf 4: Copyright(c) 2001-2004 ArtLebedev Group (http://www.artlebedev.com)
1.101 paf 5: Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
1.111 paf 6: */
1.16 paf 7:
1.199 ! paf 8: static const char * const IDENT_COMMON_C="$Date: 2004/12/10 07:41:06 $";
1.1 paf 9:
10: #include "pa_common.h"
1.4 paf 11: #include "pa_exception.h"
1.154 paf 12: #include "pa_hash.h"
1.14 paf 13: #include "pa_globals.h"
1.154 paf 14: #include "pa_request_charsets.h"
15: #include "pa_charsets.h"
16:
17: #define PA_HTTP
18:
19: #ifdef PA_HTTP
1.126 paf 20: #include "pa_vstring.h"
1.154 paf 21: #include "pa_vint.h"
1.155 paf 22: #include "pa_vhash.h"
23: #include "pa_vtable.h"
1.195 paf 24: #include "pa_socks.h"
1.154 paf 25:
26: #ifdef CYGWIN
27: #define _GNU_H_WINDOWS32_SOCKETS
28: // for PASCAL
29: #include <windows.h>
30: // SOCKET
31: typedef u_int SOCKET;
32: int PASCAL closesocket(SOCKET);
33: #else
34: # if defined(WIN32)
35: # include <windows.h>
36: # else
37: # define closesocket close
38: # endif
39: #endif
1.1 paf 40:
1.126 paf 41: #else
1.154 paf 42:
43: # if defined(WIN32)
44: # include <windows.h>
45: # endif
46:
1.98 paf 47: #endif
48:
1.93 paf 49: // some maybe-undefined constants
50:
1.82 paf 51: #ifndef _O_TEXT
52: # define _O_TEXT 0
53: #endif
54: #ifndef _O_BINARY
55: # define _O_BINARY 0
1.47 paf 56: #endif
1.80 paf 57:
1.138 paf 58: #ifdef HAVE_FTRUNCATE
59: # define PA_O_TRUNC 0
60: #else
61: # ifdef _O_TRUNC
62: # define PA_O_TRUNC _O_TRUNC
63: # else
64: # error you must have either ftruncate function or _O_TRUNC bit declared
65: # endif
1.154 paf 66: #endif
1.176 paf 67:
68: # ifndef INADDR_NONE
69: # define INADDR_NONE ((ulong) -1)
70: # endif
1.154 paf 71:
72: // defines for globals
73:
74: #define FILE_STATUS_NAME "status"
75:
76: // globals
77:
78: const String file_status_name(FILE_STATUS_NAME);
79:
1.178 paf 80: // defines
1.154 paf 81:
82: #define HTTP_METHOD_NAME "method"
1.180 paf 83: #define HTTP_FORM_NAME "form"
84: #define HTTP_BODY_NAME "body"
1.154 paf 85: #define HTTP_TIMEOUT_NAME "timeout"
86: #define HTTP_HEADERS_NAME "headers"
87: #define HTTP_ANY_STATUS_NAME "any-status"
88: #define HTTP_CHARSET_NAME "charset"
1.155 paf 89: #define HTTP_TABLES_NAME "tables"
1.178 paf 90: #define HTTP_USER "user"
91: #define HTTP_PASSWORD "password"
1.154 paf 92:
93: // defines
94:
1.127 paf 95: #define DEFAULT_USER_AGENT "parser3"
96:
1.154 paf 97: // functions
1.127 paf 98:
1.154 paf 99: void fix_line_breaks(char *str, size_t& length) {
1.87 paf 100: //_asm int 3;
1.154 paf 101: const char* const eob=str+length;
102: char* dest=str;
1.72 parser 103: // fix DOS: \r\n -> \n
104: // fix Macintosh: \r -> \n
1.154 paf 105: char* bol=str;
1.137 paf 106: while(char* eol=(char*)memchr(bol, '\r', eob -bol)) {
1.72 parser 107: size_t len=eol-bol;
108: if(dest!=bol)
1.126 paf 109: memcpy(dest, bol, len);
1.72 parser 110: dest+=len;
1.126 paf 111: *dest++='\n';
1.72 parser 112:
1.126 paf 113: if(&eol[1]<eob && eol[1]=='\n') { // \r, \n = DOS
1.72 parser 114: bol=eol+2;
1.154 paf 115: length--;
1.126 paf 116: } else // \r, not \n = Macintosh
1.72 parser 117: bol=eol+1;
118: }
1.154 paf 119: // last piece without \r
1.72 parser 120: if(dest!=bol)
1.126 paf 121: memcpy(dest, bol, eob-bol);
1.154 paf 122: str[length]=0; // terminating
1.72 parser 123: }
1.18 paf 124:
1.154 paf 125: char* file_read_text(Request_charsets& charsets,
126: const String& file_spec,
127: bool fail_on_read_problem,
128: HashStringValue* params/*, HashStringValue* * out_fields*/) {
129: File_read_result file=
130: file_read(charsets, file_spec, true, params, fail_on_read_problem);
131: return file.success?file.str:0;
1.126 paf 132: }
133:
1.178 paf 134: //http request stuff
1.154 paf 135: #ifdef PA_HTTP
136:
137: #undef CRLF
138: #define CRLF "\r\n"
1.126 paf 139:
140: static bool set_addr(struct sockaddr_in *addr, const char* host, const short port){
141: memset(addr, 0, sizeof(*addr));
142: addr->sin_family=AF_INET;
143: addr->sin_port=htons(port);
144: if(host) {
1.175 paf 145: ulong packed_ip=inet_addr(host);
1.184 paf 146: if(packed_ip!=INADDR_NONE)
1.185 paf 147: memcpy(&addr->sin_addr, &packed_ip, sizeof(packed_ip));
1.184 paf 148: else {
149: struct hostent *hostIP=gethostbyname(host);
150: if(hostIP)
151: memcpy(&addr->sin_addr, hostIP->h_addr, hostIP->h_length);
152: else
153: return false;
154: }
1.126 paf 155: } else
156: addr->sin_addr.s_addr=INADDR_ANY;
157: return true;
158: }
159:
1.171 paf 160: static int http_read_response(char*& response, size_t& response_size, int sock, bool fail_on_status_ne_200){
161: response=(char*)pa_malloc_atomic(1/*terminator*/); // setting memory block type
162: response[(response_size=0)]=0;
1.154 paf 163: int result=0;
1.171 paf 164: char* EOLat=0;
1.126 paf 165: while(true) {
1.171 paf 166: char buf[MAX_STRING*10];
167: ssize_t received_size=recv(sock, buf, sizeof(buf), 0);
1.196 paf 168: if(received_size<=0) {
169: if(int no=pa_socks_errno())
1.197 paf 170: throw Exception("http.timeout",
1.196 paf 171: 0,
172: "error receiving response: %s (%d)", pa_socks_strerr(no), no);
1.126 paf 173: break;
1.196 paf 174: }
1.171 paf 175: response=(char*)pa_realloc(response, response_size+received_size+1/*terminator*/);
176: memcpy(response+response_size, buf, received_size);
177: response_size+=received_size;
178: response[response_size]=0;
179:
1.192 paf 180: if(!result && (EOLat=strstr(response, "\n"))) { // checking status in first response
1.171 paf 181: const String status_line(pa_strdup(response, EOLat-response));
1.154 paf 182: ArrayString astatus;
183: size_t pos_after=0;
184: status_line.split(astatus, pos_after, " ");
1.182 paf 185: const String& status_code=*astatus.get(astatus.count()>1?1:0);
1.154 paf 186: result=status_code.as_int();
1.142 paf 187:
1.154 paf 188: if(fail_on_status_ne_200 && result!=200)
1.142 paf 189: throw Exception("http.status",
1.154 paf 190: &status_code,
1.142 paf 191: "invalid HTTP response status");
192: }
193: }
1.154 paf 194: if(result)
195: return result;
1.142 paf 196: else
197: throw Exception("http.response",
198: 0,
1.173 paf 199: "bad response from host - no status found (size=%u)", response_size);
1.126 paf 200: }
201:
202: /* ********************** request *************************** */
203:
204: #if defined(SIGALRM) && defined(HAVE_SIGSETJMP) && defined(HAVE_SIGLONGJMP)
1.145 paf 205: # define PA_USE_ALARM
1.126 paf 206: #endif
207:
1.145 paf 208: #ifdef PA_USE_ALARM
1.126 paf 209: static sigjmp_buf timeout_env;
1.199 ! paf 210: static void timeout_handler(int /*sig*/){
1.126 paf 211: siglongjmp(timeout_env, 1);
212: }
213: #endif
214:
1.171 paf 215: static int http_request(char*& response, size_t& response_size,
1.193 paf 216: const char* host, short port,
1.152 paf 217: const char* request,
1.196 paf 218: int timeout_secs,
1.152 paf 219: bool fail_on_status_ne_200) {
1.126 paf 220: if(!host)
221: throw Exception("http.host",
1.154 paf 222: 0,
1.126 paf 223: "zero hostname"); //never
224:
1.199 ! paf 225: volatile int sock=-1;
1.145 paf 226: #ifdef PA_USE_ALARM
1.146 paf 227: signal(SIGALRM, timeout_handler);
1.126 paf 228: #endif
1.145 paf 229: #ifdef PA_USE_ALARM
230: if(sigsetjmp(timeout_env, 1)) {
231: // stupid gcc [2.95.4] generated bad code
232: // which failed to handle sigsetjmp+throw: crashed inside of pre-throw code.
1.199 ! paf 233: // rewritten simplier [athough duplicating closesocket code]
1.145 paf 234: if(sock>=0)
235: closesocket(sock);
236: throw Exception("http.timeout",
1.199 ! paf 237: 0,
1.145 paf 238: "timeout occured while retrieving document");
1.146 paf 239: return 0; // never
1.145 paf 240: } else {
1.196 paf 241: alarm(timeout_secs);
1.145 paf 242: #endif
1.146 paf 243: try {
244: int result;
1.126 paf 245: struct sockaddr_in dest;
1.154 paf 246:
247: if(!set_addr(&dest, host, port))
1.126 paf 248: throw Exception("http.host",
1.154 paf 249: 0,
1.127 paf 250: "can not resolve hostname \"%s\"", host);
1.126 paf 251:
1.195 paf 252: if((sock=socket(AF_INET, SOCK_STREAM, IPPROTO_TCP/*0*/))<0) {
253: int no=pa_socks_errno();
1.126 paf 254: throw Exception("http.connect",
1.154 paf 255: 0,
1.195 paf 256: "can not make socket: %s (%d)", pa_socks_strerr(no), no);
257: }
1.196 paf 258:
259: #ifdef SO_DONTLINGER
260: int dont_linger = 0;
261: setsockopt(sock, SOL_SOCKET, SO_DONTLINGER, (const char *)&dont_linger, sizeof(dont_linger));
262: #else
263: // To enable SO_DONTLINGER (that is, disable SO_LINGER)
264: // l_onoff should be set to zero and setsockopt should be called
265: linger dont_linger={0,0};
266: setsockopt(sock, SOL_SOCKET, SO_LINGER, (const char *)&dont_linger, sizeof(dont_linger));
267: #endif
268:
1.198 paf 269: #if defined(SO_SNDTIMEO) || defined(SO_RCVTIMEO)
1.196 paf 270: int timeout_ms=timeout_secs*1000;
1.198 paf 271: #endif
272: #ifdef SO_SNDTIMEO
1.196 paf 273: setsockopt(sock, SOL_SOCKET, SO_SNDTIMEO, (const char*)&timeout_ms, sizeof(timeout_ms));
1.198 paf 274: #endif
275: #ifdef SO_RCVTIMEO
1.196 paf 276: setsockopt(sock, SOL_SOCKET, SO_RCVTIMEO, (const char*)&timeout_ms, sizeof(timeout_ms));
277: #endif
278:
1.195 paf 279: if(connect(sock, (struct sockaddr *)&dest, sizeof(dest))) {
280: int no=pa_socks_errno();
1.126 paf 281: throw Exception("http.connect",
1.154 paf 282: 0,
1.195 paf 283: "can not connect to host \"%s\": %s (%d)", host, pa_socks_strerr(no), no);
284: }
1.126 paf 285: size_t request_size=strlen(request);
1.195 paf 286: if(send(sock, request, request_size, 0)!=(ssize_t)request_size) {
287: int no=pa_socks_errno();
1.197 paf 288: throw Exception("http.timeout",
1.154 paf 289: 0,
1.195 paf 290: "error sending request: %s (%d)", pa_socks_strerr(no), no);
291: }
1.126 paf 292:
1.171 paf 293: result=http_read_response(response, response_size, sock, fail_on_status_ne_200);
1.142 paf 294: closesocket(sock);
1.145 paf 295: #ifdef PA_USE_ALARM
1.142 paf 296: alarm(0);
1.126 paf 297: #endif
1.147 paf 298: return result;
1.146 paf 299: } catch(...) {
1.145 paf 300: #ifdef PA_USE_ALARM
1.146 paf 301: alarm(0);
1.126 paf 302: #endif
1.146 paf 303: if(sock>=0)
304: closesocket(sock);
1.154 paf 305: rethrow;
1.146 paf 306: }
1.148 paf 307: #ifdef PA_USE_ALARM
1.126 paf 308: }
1.148 paf 309: #endif
1.126 paf 310: }
311:
1.127 paf 312: #ifndef DOXYGEN
313: struct Http_pass_header_info {
1.154 paf 314: Request_charsets* charsets;
1.127 paf 315: String* request;
316: bool user_agent_specified;
317: };
318: #endif
1.154 paf 319: static void http_pass_header(HashStringValue::key_type key,
320: HashStringValue::value_type value,
321: Http_pass_header_info *info) {
322: *info->request <<key<<": "
323: << attributed_meaning_to_string(*value, String::L_HTTP_HEADER, false)
324: << CRLF;
1.135 paf 325:
1.154 paf 326: if(String(key, String::L_TAINTED).change_case(info->charsets->source(), String::CC_UPPER)=="USER-AGENT")
327: info->user_agent_specified=true;
1.126 paf 328: }
1.154 paf 329:
330:
331: static Charset* detect_charset(Charset& source_charset, const String& content_type_value) {
1.156 paf 332: const String::Body CONTENT_TYPE_VALUE=
1.154 paf 333: content_type_value.change_case(source_charset, String::CC_UPPER);
334: // content-type: xxx/xxx; source_charset=WE-NEED-THIS
335: // content-type: xxx/xxx; source_charset="WE-NEED-THIS"
336: // content-type: xxx/xxx; source_charset="WE-NEED-THIS";
337: size_t before_charseteq_pos=CONTENT_TYPE_VALUE.pos("CHARSET=");
338: if(before_charseteq_pos!=STRING_NOT_FOUND) {
339: size_t charset_begin=before_charseteq_pos+8/*CHARSET="*/;
340: size_t open_quote_pos=CONTENT_TYPE_VALUE.pos('"', charset_begin);
341: bool quoted=open_quote_pos==charset_begin;
342: if(quoted)
343: charset_begin++; // skip opening '"'
344: size_t charset_end=CONTENT_TYPE_VALUE.length();
345: if(quoted) {
346: size_t close_quote_pos=CONTENT_TYPE_VALUE.pos('"', charset_begin);
347: if(close_quote_pos!=STRING_NOT_FOUND)
348: charset_end=close_quote_pos;
349: } else {
350: size_t delim_pos=CONTENT_TYPE_VALUE.pos(';', charset_begin);
351: if(delim_pos!=STRING_NOT_FOUND)
352: charset_end=delim_pos;
353: }
1.156 paf 354: const String::Body CHARSET_NAME_BODY=
1.154 paf 355: CONTENT_TYPE_VALUE.mid(charset_begin, charset_end);
356:
357: return &charsets.get(CHARSET_NAME_BODY);
358: }
359:
360: return 0;
361: }
362:
1.178 paf 363: static const String* basic_authorization_field(const char* user, const char* pass) {
364: if(!user&& !pass)
365: return 0;
366:
367: String combined;
368: if(user)
369: combined<<user;
370: combined<<":";
371: if(pass)
372: combined<<pass;
373:
374: String* result=new String("Basic "); *result<<pa_base64(combined.cstr(), combined.length());
375: return result;
376: }
1.154 paf 377:
1.181 paf 378: static void form_string_value2string(
379: HashStringValue::key_type key,
380: const String& value,
381: String& result)
382: {
383: result << String(key, String::L_TAINTED) << "=";
384: result.append(value, String::L_URI, true);
385: result<< "&";
386: }
387: #ifndef DOXYGEN
388: struct Form_table_value2string_info {
389: HashStringValue::key_type key;
390: String& result;
391:
392: Form_table_value2string_info(HashStringValue::key_type akey, String& aresult):
393: key(akey), result(aresult) {}
394: };
395: #endif
396: static void form_table_value2string(Table::element_type row, Form_table_value2string_info* info) {
397: form_string_value2string(info->key, *row->get(0), info->result);
398: }
1.180 paf 399: static void form_value2string(
400: HashStringValue::key_type key,
401: HashStringValue::value_type value,
402: String* result)
403: {
1.181 paf 404: if(const String* svalue=value->get_string())
405: form_string_value2string(key, *svalue, *result);
406: else if(Table* tvalue=value->get_table()) {
407: Form_table_value2string_info info(key, *result);
408: tvalue->for_each(form_table_value2string, &info);
409: } else
410: throw Exception(0,
411: new String(key, String::L_TAINTED),
412: "is %s, "HTTP_FORM_NAME" option value must either string or table", value->type());
1.180 paf 413: }
414: static const char* form2string(HashStringValue& form) {
415: String string;
416: form.for_each(form_value2string, &string);
1.181 paf 417: return string.cstr(String::L_UNSPECIFIED);
1.180 paf 418: }
1.154 paf 419: #ifndef DOXYGEN
420: struct File_read_http_result {
421: char *str; size_t length;
422: HashStringValue* headers;
423: };
424: #endif
1.194 paf 425: static void find_headers_end(char* p,
426: char*& headers_end_at,
427: char*& raw_body)
428: {
429: raw_body=p;
430: // \n\n
431: // \r\n\r\n
432: while((p=strchr(p, '\n'))) {
433: headers_end_at=++p; // \n>.<
434: if(*p=='\r') // \r\n>\r?<\n
435: p++;
436: if(*p=='\n') { // \r\n\r>\n?<
437: raw_body=p+1;
438: return;
439: }
440: }
441: headers_end_at=0;
442: }
443:
1.155 paf 444: /// @todo build .cookies field. use ^file.tables.SET-COOKIES.menu{ for now
1.154 paf 445: static File_read_http_result file_read_http(Request_charsets& charsets,
446: const String& file_spec,
1.169 paf 447: bool as_text,
448: HashStringValue *options=0) {
1.154 paf 449: File_read_http_result result;
1.126 paf 450: char host[MAX_STRING];
1.129 paf 451: const char* uri;
1.193 paf 452: short port;
1.180 paf 453: const char* method="GET"; bool method_is_get;
454: HashStringValue* form=0;
455: const char* body_cstr=0;
1.196 paf 456: int timeout_secs=2;
1.142 paf 457: bool fail_on_status_ne_200=true;
1.154 paf 458: Value* vheaders=0;
459: Charset *asked_remote_charset=0;
1.178 paf 460: const char* user_cstr=0;
461: const char* password_cstr=0;
1.126 paf 462:
1.127 paf 463: if(options) {
464: int valid_options=0;
1.177 paf 465: if(Value* vmethod=options->get(HTTP_METHOD_NAME)) {
1.127 paf 466: valid_options++;
1.154 paf 467: method=vmethod->as_string().cstr();
1.181 paf 468: }
469: if(Value* vform=options->get(HTTP_FORM_NAME)) {
1.180 paf 470: valid_options++;
471: form=vform->get_hash();
1.181 paf 472: }
473: if(Value* vbody=options->get(HTTP_BODY_NAME)) {
1.180 paf 474: valid_options++;
475: body_cstr=vbody->as_string().cstr(String::L_UNSPECIFIED);
1.181 paf 476: }
477: if(Value* vtimeout=options->get(HTTP_TIMEOUT_NAME)) {
1.127 paf 478: valid_options++;
1.196 paf 479: timeout_secs=vtimeout->as_int();
1.181 paf 480: }
481: if((vheaders=options->get(HTTP_HEADERS_NAME))) {
1.127 paf 482: valid_options++;
1.181 paf 483: }
484: if(Value* vany_status=options->get(HTTP_ANY_STATUS_NAME)) {
1.142 paf 485: valid_options++;
486: fail_on_status_ne_200=!vany_status->as_bool();
1.181 paf 487: }
488: if(Value* vcharset_name=options->get(HTTP_CHARSET_NAME)) {
1.154 paf 489: valid_options++;
490: asked_remote_charset=&::charsets.get(vcharset_name->as_string().
491: change_case(charsets.source(), String::CC_UPPER));
1.181 paf 492: }
493: if(Value* vuser=options->get(HTTP_USER)) {
1.178 paf 494: valid_options++;
495: user_cstr=vuser->as_string().cstr();
1.181 paf 496: }
497: if(Value* vpassword=options->get(HTTP_PASSWORD)) {
1.178 paf 498: valid_options++;
499: password_cstr=vpassword->as_string().cstr();
500: }
1.142 paf 501:
1.154 paf 502: if(valid_options!=options->count())
1.127 paf 503: throw Exception("parser.runtime",
504: 0,
505: "invalid option passed");
1.154 paf 506: }
507: if(!asked_remote_charset) // defaulting to $request:charset
508: asked_remote_charset=&charsets.source();
509:
1.180 paf 510: method_is_get=strcmp(method, "GET")==0;
511: if(method_is_get && body_cstr)
512: throw Exception("parser.runtime",
513: 0,
514: "you can not use $."HTTP_BODY_NAME" option with method GET");
515:
1.154 paf 516: //preparing request
517: String& connect_string=*new String;
518: // not in ^sql{... L_SQL ...} spirit, but closer to ^file::load one
519: connect_string.append(file_spec, String::L_URI); // tainted pieces -> URI pieces
520:
1.180 paf 521: String request_head_and_body;
1.154 paf 522: {
523: // influence URLencoding of tainted pieces to String::L_URI lang
524: Temp_client_charset temp(charsets, *asked_remote_charset);
525:
526: const char* connect_string_cstr=connect_string.cstr(String::L_UNSPECIFIED);
1.126 paf 527:
1.154 paf 528: const char* current=connect_string_cstr;
529: if(strncmp(current, "http://", 7)!=0)
530: throw Exception(0,
531: &connect_string,
532: "does not start with http://"); //never
533: current+=7;
534:
535: strncpy(host, current, sizeof(host)-1); host[sizeof(host)-1]=0;
536: char* host_uri=lsplit(host, '/');
537: uri=host_uri?current+(host_uri-1-host):"/";
538: char* port_cstr=lsplit(host, ':');
539: char* error_pos=0;
1.193 paf 540: port=port_cstr?(short)strtol(port_cstr, &error_pos, 0):80;
1.154 paf 541:
1.180 paf 542: if(strchr(uri, '?') && form)
543: throw Exception("parser.runtime",
544: 0,
545: "use either uri with ?params or $."HTTP_FORM_NAME" option");
546:
547: //making request head
548: String head;
549: head << method;
550: head << " " << uri;
551: if(form)
552: if(method_is_get)
1.181 paf 553: head << "?" << form2string(*form);
1.180 paf 554: head <<" HTTP/1.0" CRLF
1.154 paf 555: "host: "<< host << CRLF;
1.181 paf 556: if(form && !method_is_get) {
557: head << "content-type: application/x-www-form-urlencoded" CRLF;
558: body_cstr = form2string(*form);
559: }
1.178 paf 560:
1.179 paf 561: // http://www.ietf.org/rfc/rfc2617.txt
1.178 paf 562: if(const String* authorization_field_value=basic_authorization_field(user_cstr, password_cstr))
1.180 paf 563: head<<"authorization: "<<*authorization_field_value<<CRLF;
1.178 paf 564:
1.154 paf 565: bool user_agent_specified=false;
566: if(vheaders && !vheaders->is_string()) { // allow empty
567: if(HashStringValue *headers=vheaders->get_hash()) {
1.180 paf 568: Http_pass_header_info info={&charsets, &head, false};
1.154 paf 569: headers->for_each(http_pass_header, &info);
570: user_agent_specified=info.user_agent_specified;
571: } else
572: throw Exception("parser.runtime",
573: &connect_string,
574: "headers param must be hash");
575: };
576: if(!user_agent_specified) // defaulting
1.180 paf 577: head << "user-agent: " DEFAULT_USER_AGENT CRLF;
578:
579: if(body_cstr) {
580: // recode those pieces which are not in String::L_URI lang
581: // [those violating HTTP standard, but widly used]
582: body_cstr=Charset::transcode(
583: String::C(body_cstr, strlen(body_cstr)),
584: charsets.source(),
585: *asked_remote_charset);
1.181 paf 586:
587: head << "content-length: " << format(strlen(body_cstr), "%u") << CRLF;
1.180 paf 588: }
1.154 paf 589:
1.181 paf 590: const char* head_cstr=head.cstr(String::L_UNSPECIFIED);
591:
592: // recode those pieces which are not in String::L_URI lang
593: // [those violating HTTP standard, but widly used]
594: head_cstr=Charset::transcode(
595: String::C(head_cstr, strlen(head_cstr)),
596: charsets.source(),
597: *asked_remote_charset);
1.180 paf 598:
1.181 paf 599: // head + end of header
600: request_head_and_body << head_cstr << CRLF;
1.180 paf 601: // body
602: if(body_cstr)
603: request_head_and_body << body_cstr;
1.154 paf 604: }
1.126 paf 605:
606: //sending request
1.171 paf 607: char* response;
608: size_t response_size;
609: int status_code=http_request(response, response_size,
1.180 paf 610: host, port, request_head_and_body.cstr(),
1.196 paf 611: timeout_secs, fail_on_status_ne_200);
1.126 paf 612:
613: //processing results
1.171 paf 614: char* raw_body; size_t raw_body_size;
1.194 paf 615: char* headers_end_at;
616: find_headers_end(response,
617: headers_end_at,
618: raw_body);
1.191 paf 619: raw_body_size=response_size-(raw_body-response);
1.171 paf 620:
1.154 paf 621: result.headers=new HashStringValue;
1.155 paf 622: VHash* vtables=new VHash;
1.177 paf 623: result.headers->put(HTTP_TABLES_NAME, vtables);
1.194 paf 624: Charset* real_remote_charset=0; // undetected, yet
625:
626: if(headers_end_at) {
627: *headers_end_at=0;
628: const String header_block(String::C(response, headers_end_at-response), true);
629:
630: ArrayString aheaders;
631: HashStringValue& tables=vtables->hash();
1.155 paf 632:
1.194 paf 633: size_t pos_after=0;
634: header_block.split(aheaders, pos_after, "\n");
635:
636: //processing headers
637: size_t aheaders_count=aheaders.count();
638: for(size_t i=1; i<aheaders_count; i++) {
639: const String& line=*aheaders.get(i);
640: size_t pos=line.pos(':');
641: if(pos==STRING_NOT_FOUND || pos<1)
642: throw Exception("http.response",
643: &connect_string,
644: "bad response from host - bad header \"%s\"", line.cstr());
645: const String::Body HEADER_NAME=
646: line.mid(0, pos).change_case(charsets.source(), String::CC_UPPER);
647: const String& header_value=line.mid(pos+1, line.length()).trim(String::TRIM_BOTH, " \t\r");
648: if(as_text && HEADER_NAME=="CONTENT-TYPE")
649: real_remote_charset=detect_charset(charsets.source(), header_value);
650:
651: // tables
652: {
653: Value *valready=(Value *)tables.get(HEADER_NAME);
654: bool existed=valready!=0;
655: Table *table;
656: if(existed) {
657: // second+ appearence
658: table=valready->get_table();
659: } else {
660: // first appearence
661: Table::columns_type columns =new ArrayString(1);
662: *columns+=new String("value");
663: table=new Table(columns);
664: }
665: // this string becomes next row
666: ArrayString& row=*new ArrayString(1);
667: row+=&header_value;
668: *table+=&row;
669: // not existed before? add it
670: if(!existed)
671: tables.put(HEADER_NAME, new VTable(table));
1.155 paf 672: }
1.194 paf 673:
674: result.headers->put(HEADER_NAME, new VString(header_value));
1.155 paf 675: }
1.126 paf 676: }
677:
678: // output response
1.171 paf 679: String::C real_body=String::C(raw_body, raw_body_size);
1.192 paf 680: if(as_text && raw_body_size) { // must be checked because transcode returns CONST string in case length==0, which contradicts hacking few lines below
681: // defaulting to used-asked charset [it's never empty!]
682: if(!real_remote_charset)
683: real_remote_charset=asked_remote_charset;
1.169 paf 684: real_body=Charset::transcode(real_body, *real_remote_charset, charsets.source());
1.192 paf 685: }
1.154 paf 686:
687: result.str=const_cast<char *>(real_body.str); // hacking a little
688: result.length=real_body.length;
689: result.headers->put(file_status_name, new VInt(status_code));
690: return result;
1.34 paf 691: }
1.123 paf 692:
1.154 paf 693: #endif
694:
1.123 paf 695: #ifndef DOXYGEN
696: struct File_read_action_info {
1.154 paf 697: char **data; size_t *data_size;
1.188 paf 698: char* buf; size_t offset; size_t count;
1.126 paf 699: };
1.123 paf 700: #endif
1.154 paf 701: static void file_read_action(
702: struct stat& finfo,
703: int f,
1.166 paf 704: const String& file_spec, const char* /*fname*/, bool as_text,
1.154 paf 705: void *context) {
1.126 paf 706: File_read_action_info& info=*static_cast<File_read_action_info *>(context);
1.188 paf 707: size_t to_read_size=info.count;
708: if(!to_read_size)
709: to_read_size=(size_t)finfo.st_size;
710: assert( !(info.buf && as_text) );
711: if(to_read_size) {
712: if(info.offset)
713: lseek(f, info.offset, SEEK_SET);
714: *info.data=info.buf
715: ? info.buf
716: : new(PointerFreeGC) char[to_read_size+(as_text?1:0)];
1.126 paf 717: *info.data_size=(size_t)read(f, *info.data, to_read_size);
1.123 paf 718:
719: if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size)
1.126 paf 720: throw Exception(0,
1.123 paf 721: &file_spec,
1.173 paf 722: "read failed: actually read %u bytes count not in [0..%u] valid range",
1.126 paf 723: *info.data_size, to_read_size);
1.123 paf 724: } else { // empty file
725: if(as_text) {
1.154 paf 726: *info.data=new(PointerFreeGC) char[1];
1.123 paf 727: *(char*)(*info.data)=0;
728: } else
729: *info.data=0;
730: *info.data_size=0;
731: return;
732: }
1.126 paf 733: }
1.154 paf 734: File_read_result file_read(Request_charsets& charsets, const String& file_spec,
735: bool as_text, HashStringValue *params,
1.188 paf 736: bool fail_on_read_problem,
737: char* buf, size_t offset, size_t count) {
1.167 paf 738: File_read_result result={false, 0, 0, 0};
1.154 paf 739: #ifdef PA_HTTP
740: if(file_spec.starts_with("http://")) {
1.126 paf 741: // fail on read problem
1.169 paf 742: File_read_http_result http=file_read_http(charsets, file_spec, as_text, params);
1.154 paf 743: result.success=true;
744: result.str=http.str;
745: result.length=http.length;
746: result.headers=http.headers;
1.126 paf 747: } else {
1.154 paf 748: #endif
1.161 paf 749: if(params && params->count())
750: throw Exception("parser.runtime",
751: 0,
752: "invalid option passed");
753:
1.188 paf 754: File_read_action_info info={&result.str, &result.length,
755: buf, offset, count};
1.154 paf 756: result.success=file_read_action_under_lock(file_spec,
1.126 paf 757: "read", file_read_action, &info,
758: as_text, fail_on_read_problem);
1.154 paf 759: #ifdef PA_HTTP
1.126 paf 760: }
1.154 paf 761: #endif
1.123 paf 762:
1.154 paf 763: if(result.success && as_text) {
1.131 paf 764: // UTF-8 signature: EF BB BF
1.154 paf 765: if(result.length>=3) {
766: char *in=(char *)result.str;
1.159 paf 767: if(strncmp(in, "\xEF\xBB\xBF", 3)==0) {
1.154 paf 768: result.str=in+3; result.length-=3;// skip prefix
1.131 paf 769: }
770: }
771:
1.154 paf 772: fix_line_breaks((char *)(result.str), result.length);
1.123 paf 773: }
1.126 paf 774:
775: return result;
1.123 paf 776: }
777:
1.154 paf 778: #ifdef PA_SAFE_MODE
779: void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) {
780: if(finfo.st_uid/*foreign?*/!=geteuid()
781: && finfo.st_gid/*foreign?*/!=getegid())
782: throw Exception("parser.runtime",
783: &file_spec,
784: "parser is in safe mode: "
785: "reading files of foreign group and user disabled "
786: "[recompile parser with --disable-safe-mode configure option], "
787: "actual filename '%s', "
788: "fuid(%d)!=euid(%d) or fgid(%d)!=egid(%d)",
789: fname,
790: finfo.st_uid, geteuid(),
791: finfo.st_gid, getegid());
792: }
793: #endif
1.149 paf 794:
1.154 paf 795: bool file_read_action_under_lock(const String& file_spec,
1.126 paf 796: const char* action_name, File_read_action action, void *context,
797: bool as_text,
1.123 paf 798: bool fail_on_read_problem) {
1.154 paf 799: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.33 paf 800: int f;
801:
802: // first open, next stat:
1.45 paf 803: // directory update of NTFS hard links performed on open.
1.33 paf 804: // ex:
805: // a.html:^test[] and b.html hardlink to a.html
806: // user inserts ! before ^test in a.html
1.126 paf 807: // directory entry of b.html in NTFS not updated at once,
1.35 paf 808: // they delay update till open, so we would receive "!^test[" string
809: // if would do stat, next open.
1.123 paf 810: // later: it seems, even this does not help sometimes
1.98 paf 811: if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) {
1.123 paf 812: try {
1.162 paf 813: if(pa_lock_shared_blocking(f)!=0)
1.126 paf 814: throw Exception("file.lock",
1.123 paf 815: &file_spec,
816: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 817: strerror(errno), errno, fname);
1.123 paf 818:
1.124 paf 819: struct stat finfo;
820: if(stat(fname, &finfo)!=0)
821: throw Exception("file.missing", // hardly possible: we just opened it OK
822: &file_spec,
823: "stat failed: %s (%d), actual filename '%s'",
1.154 paf 824: strerror(errno), errno, fname);
1.124 paf 825:
1.140 paf 826: #ifdef PA_SAFE_MODE
1.149 paf 827: check_safe_mode(finfo, file_spec, fname);
1.105 paf 828: #endif
1.32 paf 829:
1.154 paf 830: action(finfo, f, file_spec, fname, as_text, context);
1.123 paf 831: } catch(...) {
1.162 paf 832: pa_unlock(f);close(f);
1.123 paf 833: if(fail_on_read_problem)
1.154 paf 834: rethrow;
1.123 paf 835: return false;
836: }
1.87 paf 837:
1.162 paf 838: pa_unlock(f);close(f);
1.72 parser 839: return true;
1.118 paf 840: } else {
841: if(fail_on_read_problem)
1.126 paf 842: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.118 paf 843: &file_spec,
1.123 paf 844: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 845: action_name, strerror(errno), errno, fname);
1.118 paf 846: return false;
847: }
1.8 paf 848: }
849:
1.63 parser 850: static void create_dir_for_file(const String& file_spec) {
851: size_t pos_after=1;
1.154 paf 852: size_t pos_before;
853: while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) {
854: mkdir(file_spec.mid(0, pos_before).cstr(String::L_FILE_SPEC), 0775);
1.63 parser 855: pos_after=pos_before+1;
856: }
857: }
858:
1.98 paf 859: bool file_write_action_under_lock(
1.28 paf 860: const String& file_spec,
1.126 paf 861: const char* action_name, File_write_action action, void *context,
862: bool as_text,
863: bool do_append,
864: bool do_block,
1.110 paf 865: bool fail_on_lock_problem) {
1.154 paf 866: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.28 paf 867: int f;
1.80 paf 868: if(access(fname, W_OK)!=0) // no
1.126 paf 869: create_dir_for_file(file_spec);
1.50 paf 870:
1.80 paf 871: if((f=open(fname,
872: O_CREAT|O_RDWR
873: |(as_text?_O_TEXT:_O_BINARY)
1.138 paf 874: |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) {
1.162 paf 875: if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) {
1.126 paf 876: Exception e("file.lock",
1.110 paf 877: &file_spec,
878: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 879: strerror(errno), errno, fname);
1.126 paf 880: close(f);
1.110 paf 881: if(fail_on_lock_problem)
882: throw e;
1.98 paf 883: return false;
884: }
1.96 paf 885:
1.158 paf 886: try {
1.126 paf 887: action(f, context);
1.158 paf 888: } catch(...) {
1.138 paf 889: #ifdef HAVE_FTRUNCATE
1.104 paf 890: if(!do_append)
1.125 paf 891: ftruncate(f, lseek(f, 0, SEEK_CUR)); // one can not use O_TRUNC, read lower
1.138 paf 892: #endif
1.162 paf 893: pa_unlock(f);close(f);
1.154 paf 894: rethrow;
1.158 paf 895: }
1.80 paf 896:
1.138 paf 897: #ifdef HAVE_FTRUNCATE
1.104 paf 898: if(!do_append)
1.125 paf 899: ftruncate(f, lseek(f, 0, SEEK_CUR)); // O_TRUNC truncates even exclusevely write-locked file [thanks to Igor Milyakov <virtan@rotabanner.com> for discovering]
1.138 paf 900: #endif
1.162 paf 901: pa_unlock(f);close(f);
1.98 paf 902: return true;
1.80 paf 903: } else
1.126 paf 904: throw Exception(errno==EACCES?"file.access":0,
1.80 paf 905: &file_spec,
1.96 paf 906: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 907: action_name, strerror(errno), errno, fname);
1.96 paf 908: // here should be nothing, see rethrow above
909: }
910:
911: #ifndef DOXYGEN
912: struct File_write_action_info {
1.154 paf 913: const char* str; size_t length;
1.126 paf 914: };
1.96 paf 915: #endif
916: static void file_write_action(int f, void *context) {
1.126 paf 917: File_write_action_info& info=*static_cast<File_write_action_info *>(context);
1.154 paf 918: if(info.length) {
919: int written=write(f, info.str, info.length);
1.116 paf 920: if(written<0)
1.126 paf 921: throw Exception(0,
922: 0,
923: "write failed: %s (%d)", strerror(errno), errno);
1.113 paf 924: }
1.96 paf 925: }
926: void file_write(
927: const String& file_spec,
1.154 paf 928: const char* data, size_t size,
1.126 paf 929: bool as_text,
1.96 paf 930: bool do_append) {
1.126 paf 931: File_write_action_info info={data, size};
1.98 paf 932: file_write_action_under_lock(
1.154 paf 933: file_spec,
934: "write", file_write_action, &info,
935: as_text,
936: do_append);
1.30 paf 937: }
938:
1.63 parser 939: // throws nothing! [this is required in file_move & file_delete]
1.50 paf 940: static void rmdir(const String& file_spec, size_t pos_after) {
1.154 paf 941: size_t pos_before;
942: if((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND)
1.126 paf 943: rmdir(file_spec, pos_before+1);
1.50 paf 944:
1.154 paf 945: rmdir(file_spec.mid(0, pos_after-1/* / */).cstr(String::L_FILE_SPEC));
1.50 paf 946: }
1.164 paf 947: bool file_delete(const String& file_spec, bool fail_on_problem) {
1.154 paf 948: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.54 parser 949: if(unlink(fname)!=0)
1.164 paf 950: if(fail_on_problem)
1.126 paf 951: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.93 paf 952: &file_spec,
953: "unlink failed: %s (%d), actual filename '%s'",
1.154 paf 954: strerror(errno), errno, fname);
1.93 paf 955: else
956: return false;
1.50 paf 957:
1.126 paf 958: rmdir(file_spec, 1);
1.93 paf 959: return true;
1.60 parser 960: }
1.95 paf 961: void file_move(const String& old_spec, const String& new_spec) {
1.154 paf 962: const char* old_spec_cstr=old_spec.cstr(String::L_FILE_SPEC);
963: const char* new_spec_cstr=new_spec.cstr(String::L_FILE_SPEC);
1.63 parser 964:
1.126 paf 965: create_dir_for_file(new_spec);
1.63 parser 966:
1.60 parser 967: if(rename(old_spec_cstr, new_spec_cstr)!=0)
1.126 paf 968: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.60 parser 969: &old_spec,
970: "rename failed: %s (%d), actual filename '%s' to '%s'",
1.154 paf 971: strerror(errno), errno, old_spec_cstr, new_spec_cstr);
1.63 parser 972:
1.126 paf 973: rmdir(old_spec, 1);
1.31 paf 974: }
975:
1.51 paf 976:
1.126 paf 977: bool entry_exists(const char* fname, struct stat *afinfo) {
1.118 paf 978: struct stat lfinfo;
979: bool result=stat(fname, &lfinfo)==0;
980: if(afinfo)
981: *afinfo=lfinfo;
982: return result;
1.119 paf 983: }
984:
985: bool entry_exists(const String& file_spec) {
1.154 paf 986: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.126 paf 987: return entry_exists(fname, 0);
1.118 paf 988: }
989:
1.51 paf 990: static bool entry_readable(const String& file_spec, bool need_dir) {
1.154 paf 991: char* fname=file_spec.cstrm(String::L_FILE_SPEC);
1.120 paf 992: if(need_dir) {
1.126 paf 993: size_t size=strlen(fname);
1.120 paf 994: while(size) {
1.126 paf 995: char c=fname[size-1];
1.120 paf 996: if(c=='/' || c=='\\')
997: fname[--size]=0;
998: else
999: break;
1000: }
1001: }
1.51 paf 1002: struct stat finfo;
1.118 paf 1003: if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) {
1.109 paf 1004: bool is_dir=(finfo.st_mode&S_IFDIR) != 0;
1.51 paf 1005: return is_dir==need_dir;
1006: }
1007: return false;
1008: }
1.31 paf 1009: bool file_readable(const String& file_spec) {
1.126 paf 1010: return entry_readable(file_spec, false);
1.51 paf 1011: }
1012: bool dir_readable(const String& file_spec) {
1.126 paf 1013: return entry_readable(file_spec, true);
1.65 parser 1014: }
1.154 paf 1015: const String* file_readable(const String& path, const String& name) {
1016: String& result=*new String(path);
1017: result << "/";
1018: result << name;
1019: return file_readable(result)?&result:0;
1.43 paf 1020: }
1021: bool file_executable(const String& file_spec) {
1.154 paf 1022: return access(file_spec.cstr(String::L_FILE_SPEC), X_OK)==0;
1.44 paf 1023: }
1024:
1.64 parser 1025: bool file_stat(const String& file_spec,
1.58 parser 1026: size_t& rsize,
1.126 paf 1027: time_t& ratime,
1028: time_t& rmtime,
1029: time_t& rctime,
1.64 parser 1030: bool fail_on_read_problem) {
1.154 paf 1031: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1032: struct stat finfo;
1.44 paf 1033: if(stat(fname, &finfo)!=0)
1.64 parser 1034: if(fail_on_read_problem)
1.126 paf 1035: throw Exception("file.missing",
1.67 parser 1036: &file_spec,
1037: "getting file size failed: %s (%d), real filename '%s'",
1.154 paf 1038: strerror(errno), errno, fname);
1.64 parser 1039: else
1040: return false;
1.58 parser 1041: rsize=finfo.st_size;
1042: ratime=finfo.st_atime;
1043: rmtime=finfo.st_mtime;
1044: rctime=finfo.st_ctime;
1.64 parser 1045: return true;
1.18 paf 1046: }
1047:
1.126 paf 1048: char* getrow(char* *row_ref, char delim) {
1049: char* result=*row_ref;
1.8 paf 1050: if(result) {
1.126 paf 1051: *row_ref=strchr(result, delim);
1.8 paf 1052: if(*row_ref)
1053: *((*row_ref)++)=0;
1054: else if(!*result)
1055: return 0;
1056: }
1057: return result;
1058: }
1059:
1.126 paf 1060: char* lsplit(char* string, char delim) {
1.23 paf 1061: if(string) {
1.126 paf 1062: char* v=strchr(string, delim);
1.8 paf 1063: if(v) {
1064: *v=0;
1065: return v+1;
1066: }
1067: }
1068: return 0;
1069: }
1070:
1.126 paf 1071: char* lsplit(char* *string_ref, char delim) {
1072: char* result=*string_ref;
1073: char* next=lsplit(*string_ref, delim);
1.8 paf 1074: *string_ref=next;
1075: return result;
1.9 paf 1076: }
1077:
1.126 paf 1078: char* rsplit(char* string, char delim) {
1.18 paf 1079: if(string) {
1.126 paf 1080: char* v=strrchr(string, delim);
1.18 paf 1081: if(v) {
1.9 paf 1082: *v=0;
1083: return v+1;
1084: }
1085: }
1086: return NULL;
1.10 paf 1087: }
1088:
1.37 paf 1089: /// @todo less stupid type detection
1.154 paf 1090: const char* format(double value, char* fmt) {
1.126 paf 1091: char local_buf[MAX_NUMBER];
1.108 paf 1092: size_t size;
1093:
1.10 paf 1094: if(fmt)
1095: if(strpbrk(fmt, "diouxX"))
1096: if(strpbrk(fmt, "ouxX"))
1.126 paf 1097: size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value);
1.10 paf 1098: else
1.126 paf 1099: size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value);
1.10 paf 1100: else
1.126 paf 1101: size=snprintf(local_buf, sizeof(local_buf), fmt, value);
1.10 paf 1102: else
1.126 paf 1103: size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value);
1.10 paf 1104:
1.154 paf 1105: return pa_strdup(local_buf, size);
1.12 paf 1106: }
1107:
1.36 paf 1108: size_t stdout_write(const void *buf, size_t size) {
1.12 paf 1109: #ifdef WIN32
1.187 paf 1110: size_t to_write = size;
1.12 paf 1111: do{
1.154 paf 1112: int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout);
1.12 paf 1113: if(chunk_written<=0)
1114: break;
1115: size-=chunk_written;
1.36 paf 1116: buf=((const char*)buf)+chunk_written;
1.126 paf 1117: } while(size>0);
1.12 paf 1118:
1.187 paf 1119: return to_write-size;
1.12 paf 1120: #else
1.126 paf 1121: return fwrite(buf, 1, size, stdout);
1.12 paf 1122: #endif
1.2 paf 1123: }
1.14 paf 1124:
1.154 paf 1125: char* unescape_chars(const char* cp, int len) {
1126: char* s=new(PointerFreeGC) char[len + 1];
1.14 paf 1127: enum EscapeState {
1.33 paf 1128: EscapeRest,
1129: EscapeFirst,
1.14 paf 1130: EscapeSecond
1131: } escapeState=EscapeRest;
1.193 paf 1132: uchar escapedValue=0;
1.14 paf 1133: int srcPos=0;
1134: int dstPos=0;
1135: while(srcPos < len) {
1.193 paf 1136: uchar ch=(uchar)cp[srcPos];
1.14 paf 1137: switch(escapeState) {
1138: case EscapeRest:
1139: if(ch=='%') {
1140: escapeState=EscapeFirst;
1141: } else if(ch=='+') {
1.126 paf 1142: s[dstPos++]=' ';
1.14 paf 1143: } else {
1144: s[dstPos++]=ch;
1145: }
1146: break;
1147: case EscapeFirst:
1.193 paf 1148: escapedValue=(uchar)(hex_value[ch] << 4);
1.14 paf 1149: escapeState=EscapeSecond;
1150: break;
1151: case EscapeSecond:
1.126 paf 1152: escapedValue +=hex_value[ch];
1.14 paf 1153: s[dstPos++]=escapedValue;
1154: escapeState=EscapeRest;
1155: break;
1156: }
1.126 paf 1157: srcPos++;
1.14 paf 1158: }
1159: s[dstPos]=0;
1160: return s;
1.24 paf 1161: }
1162:
1163: #ifdef WIN32
1.126 paf 1164: void back_slashes_to_slashes(char* s) {
1.24 paf 1165: if(s)
1166: for(; *s; s++)
1167: if(*s=='\\')
1.126 paf 1168: *s='/';
1.24 paf 1169: }
1.42 paf 1170: /*
1.126 paf 1171: void slashes_to_back_slashes(char* s) {
1.42 paf 1172: if(s)
1173: for(; *s; s++)
1174: if(*s=='/')
1.126 paf 1175: *s='\\';
1.42 paf 1176: }
1177: */
1.24 paf 1178: #endif
1.41 paf 1179:
1.126 paf 1180: bool StrEqNc(const char* s1, const char* s2, bool strict) {
1.41 paf 1181: while(true) {
1182: if(!(*s1)) {
1183: if(!(*s2))
1184: return true;
1185: else
1186: return !strict;
1187: } else if(!(*s2))
1188: return !strict;
1.189 paf 1189: if(isalpha((unsigned char)*s1)) {
1.190 paf 1190: if(tolower((unsigned char)*s1) !=tolower((unsigned char)*s2))
1.41 paf 1191: return false;
1192: } else if((*s1) !=(*s2))
1193: return false;
1.126 paf 1194: s1++;
1195: s2++;
1.41 paf 1196: }
1.57 parser 1197: }
1198:
1.84 paf 1199: static bool isLeap(int year) {
1.57 parser 1200: return !(
1201: (year % 4) || ((year % 400) && !(year % 100))
1.126 paf 1202: );
1.57 parser 1203: }
1204:
1205: int getMonthDays(int year, int month) {
1206: int monthDays[]={
1.126 paf 1207: 31,
1208: isLeap(year) ? 29 : 28,
1209: 31,
1210: 30,
1211: 31,
1212: 30,
1213: 31,
1214: 31,
1215: 30,
1216: 31,
1217: 30,
1.57 parser 1218: 31
1.126 paf 1219: };
1220: return monthDays[month];
1.41 paf 1221: }
1.69 parser 1222:
1.126 paf 1223: void remove_crlf(char* start, char* end) {
1224: for(char* p=start; p<end; p++)
1.69 parser 1225: switch(*p) {
1.126 paf 1226: case '\n': *p='|'; break;
1227: case '\r': *p=' '; break;
1.69 parser 1228: }
1.91 paf 1229: }
1230:
1231:
1232: /// must be last in this file
1233: #undef vsnprintf
1.126 paf 1234: int __vsnprintf(char* b, size_t s, const char* f, va_list l) {
1.91 paf 1235: if(!s)
1236: return 0;
1237:
1238: int r;
1239: // note: on win32& maybe somewhere else
1240: // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing
1241: --s;
1.172 paf 1242:
1243: // clients do not check for negative 's', feature: ignore such prints
1244: if((ssize_t)s<0)
1245: return 0;
1246:
1.91 paf 1247: #if _MSC_VER
1248: /*
1249: win32:
1250: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm
1251:
1.154 paf 1252: if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 is returned
1.91 paf 1253: */
1.126 paf 1254: r=_vsnprintf(b, s, f, l);
1.91 paf 1255: if(r<0)
1256: r=s;
1257: #else
1.126 paf 1258: r=vsnprintf(b, s, f, l);
1.91 paf 1259: /*
1260: solaris:
1261: man vsnprintf
1262:
1263: The snprintf() function returns the number of characters
1264: formatted, that is, the number of characters that would have
1265: been written to the buffer if it were large enough. If the
1266: value of n is 0 on a call to snprintf(), an unspecified
1267: value less than 1 is returned.
1268: */
1269:
1270: if(r<0)
1271: r=0;
1.167 paf 1272: else if((size_t)r>s)
1.91 paf 1273: r=s;
1274: #endif
1275: b[r]=0;
1276: return r;
1277: }
1278:
1.126 paf 1279: int __snprintf(char* b, size_t s, const char* f, ...) {
1.91 paf 1280: va_list l;
1.126 paf 1281: va_start(l, f);
1282: int r=__vsnprintf(b, s, f, l);
1283: va_end(l);
1.91 paf 1284: return r;
1.178 paf 1285: }
1286:
1287: /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */
1288: /*
1289: * Authors: Michael Zucchi <notzed@helixcode.com>
1290: * Jeffrey Stedfast <fejj@helixcode.com>
1291: *
1292: * Copyright 2000 Helix Code, Inc. (www.helixcode.com)
1293: *
1294: * This program is free software; you can redistribute it and/or modify
1295: * it under the terms of the GNU General Public License as published by
1296: * the Free Software Foundation; either version 2 of the License, or
1297: * (at your option) any later version.
1298: *
1299: * This program is distributed in the hope that it will be useful,
1300: * but WITHOUT ANY WARRANTY; without even the implied warranty of
1301: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
1302: * GNU General Public License for more details.
1303: *
1304: * You should have received a copy of the GNU General Public License
1305: * along with this program; if not, write to the Free Software
1306: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
1307: *
1308: */
1309: static char *base64_alphabet =
1310: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
1311:
1312: /**
1313: * g_mime_utils_base64_encode_step:
1314: * @in: input stream
1315: * @inlen: length of the input
1316: * @out: output string
1317: * @state: holds the number of bits that are stored in @save
1318: * @save: leftover bits that have not yet been encoded
1319: *
1320: * Base64 encodes a chunk of data. Performs an 'encode step', only
1321: * encodes blocks of 3 characters to the output at a time, saves
1322: * left-over state in state and save (initialise to 0 on first
1323: * invocation).
1324: *
1325: * Returns the number of bytes encoded.
1326: **/
1327: static size_t
1328: g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1329: {
1.186 paf 1330: register const unsigned char *inptr;
1.178 paf 1331: register unsigned char *outptr;
1332:
1333: if (inlen <= 0)
1334: return 0;
1335:
1336: inptr = in;
1337: outptr = out;
1338:
1339: if (inlen + ((unsigned char *)save)[0] > 2) {
1340: const unsigned char *inend = in + inlen - 2;
1341: register int c1 = 0, c2 = 0, c3 = 0;
1342: register int already;
1343:
1344: already = *state;
1345:
1346: switch (((char *)save)[0]) {
1347: case 1: c1 = ((unsigned char *)save)[1]; goto skip1;
1348: case 2: c1 = ((unsigned char *)save)[1];
1349: c2 = ((unsigned char *)save)[2]; goto skip2;
1350: }
1351:
1352: /* yes, we jump into the loop, no i'm not going to change it, its beautiful! */
1353: while (inptr < inend) {
1354: c1 = *inptr++;
1355: skip1:
1356: c2 = *inptr++;
1357: skip2:
1358: c3 = *inptr++;
1359: *outptr++ = base64_alphabet [c1 >> 2];
1360: *outptr++ = base64_alphabet [(c2 >> 4) | ((c1 & 0x3) << 4)];
1361: *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)];
1362: *outptr++ = base64_alphabet [c3 & 0x3f];
1363: /* this is a bit ugly ... */
1364: if ((++already) >= 19) {
1365: *outptr++ = '\n';
1366: already = 0;
1367: }
1368: }
1369:
1370: ((unsigned char *)save)[0] = 0;
1371: inlen = 2 - (inptr - inend);
1372: *state = already;
1373: }
1374:
1375: //d(printf ("state = %d, inlen = %d\n", (int)((char *)save)[0], inlen));
1376:
1377: if (inlen > 0) {
1378: register char *saveout;
1379:
1380: /* points to the slot for the next char to save */
1381: saveout = & (((char *)save)[1]) + ((char *)save)[0];
1382:
1383: /* inlen can only be 0 1 or 2 */
1384: switch (inlen) {
1385: case 2: *saveout++ = *inptr++;
1386: case 1: *saveout++ = *inptr++;
1387: }
1388: ((char *)save)[0] += inlen;
1389: }
1390:
1391: /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n",
1392: (int)((char *)save)[0],
1393: (int)((char *)save)[1],
1394: (int)((char *)save)[2]));*/
1395:
1396: return (outptr - out);
1397: }
1398:
1399: /**
1400: * g_mime_utils_base64_encode_close:
1401: * @in: input stream
1402: * @inlen: length of the input
1403: * @out: output string
1404: * @state: holds the number of bits that are stored in @save
1405: * @save: leftover bits that have not yet been encoded
1406: *
1407: * Base64 encodes the input stream to the output stream. Call this
1408: * when finished encoding data with g_mime_utils_base64_encode_step to
1409: * flush off the last little bit.
1410: *
1411: * Returns the number of bytes encoded.
1412: **/
1413: static size_t
1414: g_mime_utils_base64_encode_close (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1415: {
1416: unsigned char *outptr = out;
1417: int c1, c2;
1418:
1419: if (inlen > 0)
1420: outptr += g_mime_utils_base64_encode_step (in, inlen, outptr, state, save);
1421:
1422: c1 = ((unsigned char *)save)[1];
1423: c2 = ((unsigned char *)save)[2];
1424:
1425: switch (((unsigned char *)save)[0]) {
1426: case 2:
1427: outptr[2] = base64_alphabet [(c2 & 0x0f) << 2];
1428: goto skip;
1429: case 1:
1430: outptr[2] = '=';
1431: skip:
1432: outptr[0] = base64_alphabet [c1 >> 2];
1433: outptr[1] = base64_alphabet [c2 >> 4 | ((c1 & 0x3) << 4)];
1434: outptr[3] = '=';
1435: outptr += 4;
1436: break;
1437: }
1438:
1439: *outptr++ = 0;
1440:
1441: *save = 0;
1442: *state = 0;
1443:
1444: return (outptr - out);
1445: }
1446:
1447: char* pa_base64(const char *in, size_t len)
1448: {
1449: /* wont go to more than 2x size (overly conservative) */
1450: char* result=new(PointerFreeGC) char[len * 2 + 6];
1451: int state=0;
1452: int save=0;
1.183 paf 1453: #ifndef NDEBUG
1454: size_t filled=
1455: #endif
1456: g_mime_utils_base64_encode_close ((const unsigned char*)in, len,
1.178 paf 1457: (unsigned char*)result, &state, &save);
1458: assert(filled <= len * 2 + 6);
1459:
1460: return result;
1.98 paf 1461: }
E-mail:
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to pa_common.C CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / parser3 / src / main