Annotation of parser3/src/main/pa_common.C, revision 1.195
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:
1.195 ! paf 8: static const char * const IDENT_COMMON_C="$Date: 2004/09/09 13:57:27 $";
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);
168: if(received_size<=0)
1.126 paf 169: break;
1.171 paf 170: response=(char*)pa_realloc(response, response_size+received_size+1/*terminator*/);
171: memcpy(response+response_size, buf, received_size);
172: response_size+=received_size;
173: response[response_size]=0;
174:
1.192 paf 175: if(!result && (EOLat=strstr(response, "\n"))) { // checking status in first response
1.171 paf 176: const String status_line(pa_strdup(response, EOLat-response));
1.154 paf 177: ArrayString astatus;
178: size_t pos_after=0;
179: status_line.split(astatus, pos_after, " ");
1.182 paf 180: const String& status_code=*astatus.get(astatus.count()>1?1:0);
1.154 paf 181: result=status_code.as_int();
1.142 paf 182:
1.154 paf 183: if(fail_on_status_ne_200 && result!=200)
1.142 paf 184: throw Exception("http.status",
1.154 paf 185: &status_code,
1.142 paf 186: "invalid HTTP response status");
187: }
188: }
1.154 paf 189: if(result)
190: return result;
1.142 paf 191: else
192: throw Exception("http.response",
193: 0,
1.173 paf 194: "bad response from host - no status found (size=%u)", response_size);
1.126 paf 195: }
196:
197: /* ********************** request *************************** */
198:
199: #if defined(SIGALRM) && defined(HAVE_SIGSETJMP) && defined(HAVE_SIGLONGJMP)
1.145 paf 200: # define PA_USE_ALARM
1.126 paf 201: #endif
202:
1.145 paf 203: #ifdef PA_USE_ALARM
1.126 paf 204: static sigjmp_buf timeout_env;
205: static void timeout_handler(int sig){
206: siglongjmp(timeout_env, 1);
207: }
208: #endif
209:
1.171 paf 210: static int http_request(char*& response, size_t& response_size,
1.193 paf 211: const char* host, short port,
1.152 paf 212: const char* request,
1.166 paf 213: int
214: #ifdef PA_USE_ALARM
215: timeout
216: #endif
217: ,
1.152 paf 218: bool fail_on_status_ne_200) {
1.126 paf 219: if(!host)
220: throw Exception("http.host",
1.154 paf 221: 0,
1.126 paf 222: "zero hostname"); //never
223:
1.145 paf 224: #ifdef PA_USE_ALARM
1.146 paf 225: signal(SIGALRM, timeout_handler);
1.126 paf 226: #endif
227: int sock=-1;
1.145 paf 228: #ifdef PA_USE_ALARM
229: if(sigsetjmp(timeout_env, 1)) {
230: // stupid gcc [2.95.4] generated bad code
231: // which failed to handle sigsetjmp+throw: crashed inside of pre-throw code.
232: // rewritten simplier [though duplicating closesocket code]
233: if(sock>=0)
234: closesocket(sock);
235: throw Exception("http.timeout",
236: origin_string,
237: "timeout occured while retrieving document");
1.146 paf 238: return 0; // never
1.145 paf 239: } else {
240: alarm(timeout);
241: #endif
1.146 paf 242: try {
243: int result;
1.126 paf 244: struct sockaddr_in dest;
1.154 paf 245:
246: if(!set_addr(&dest, host, port))
1.126 paf 247: throw Exception("http.host",
1.154 paf 248: 0,
1.127 paf 249: "can not resolve hostname \"%s\"", host);
1.126 paf 250:
1.195 ! paf 251: if((sock=socket(AF_INET, SOCK_STREAM, IPPROTO_TCP/*0*/))<0) {
! 252: int no=pa_socks_errno();
1.126 paf 253: throw Exception("http.connect",
1.154 paf 254: 0,
1.195 ! paf 255: "can not make socket: %s (%d)", pa_socks_strerr(no), no);
! 256: }
! 257: if(connect(sock, (struct sockaddr *)&dest, sizeof(dest))) {
! 258: int no=pa_socks_errno();
1.126 paf 259: throw Exception("http.connect",
1.154 paf 260: 0,
1.195 ! paf 261: "can not connect to host \"%s\": %s (%d)", host, pa_socks_strerr(no), no);
! 262: }
1.126 paf 263: size_t request_size=strlen(request);
1.195 ! paf 264: if(send(sock, request, request_size, 0)!=(ssize_t)request_size) {
! 265: int no=pa_socks_errno();
1.126 paf 266: throw Exception("http.connect",
1.154 paf 267: 0,
1.195 ! paf 268: "error sending request: %s (%d)", pa_socks_strerr(no), no);
! 269: }
1.126 paf 270:
1.171 paf 271: result=http_read_response(response, response_size, sock, fail_on_status_ne_200);
1.142 paf 272: closesocket(sock);
1.145 paf 273: #ifdef PA_USE_ALARM
1.142 paf 274: alarm(0);
1.126 paf 275: #endif
1.147 paf 276: return result;
1.146 paf 277: } catch(...) {
1.145 paf 278: #ifdef PA_USE_ALARM
1.146 paf 279: alarm(0);
1.126 paf 280: #endif
1.146 paf 281: if(sock>=0)
282: closesocket(sock);
1.154 paf 283: rethrow;
1.146 paf 284: }
1.148 paf 285: #ifdef PA_USE_ALARM
1.126 paf 286: }
1.148 paf 287: #endif
1.126 paf 288: }
289:
1.127 paf 290: #ifndef DOXYGEN
291: struct Http_pass_header_info {
1.154 paf 292: Request_charsets* charsets;
1.127 paf 293: String* request;
294: bool user_agent_specified;
295: };
296: #endif
1.154 paf 297: static void http_pass_header(HashStringValue::key_type key,
298: HashStringValue::value_type value,
299: Http_pass_header_info *info) {
300: *info->request <<key<<": "
301: << attributed_meaning_to_string(*value, String::L_HTTP_HEADER, false)
302: << CRLF;
1.135 paf 303:
1.154 paf 304: if(String(key, String::L_TAINTED).change_case(info->charsets->source(), String::CC_UPPER)=="USER-AGENT")
305: info->user_agent_specified=true;
1.126 paf 306: }
1.154 paf 307:
308:
309: static Charset* detect_charset(Charset& source_charset, const String& content_type_value) {
1.156 paf 310: const String::Body CONTENT_TYPE_VALUE=
1.154 paf 311: content_type_value.change_case(source_charset, String::CC_UPPER);
312: // content-type: xxx/xxx; source_charset=WE-NEED-THIS
313: // content-type: xxx/xxx; source_charset="WE-NEED-THIS"
314: // content-type: xxx/xxx; source_charset="WE-NEED-THIS";
315: size_t before_charseteq_pos=CONTENT_TYPE_VALUE.pos("CHARSET=");
316: if(before_charseteq_pos!=STRING_NOT_FOUND) {
317: size_t charset_begin=before_charseteq_pos+8/*CHARSET="*/;
318: size_t open_quote_pos=CONTENT_TYPE_VALUE.pos('"', charset_begin);
319: bool quoted=open_quote_pos==charset_begin;
320: if(quoted)
321: charset_begin++; // skip opening '"'
322: size_t charset_end=CONTENT_TYPE_VALUE.length();
323: if(quoted) {
324: size_t close_quote_pos=CONTENT_TYPE_VALUE.pos('"', charset_begin);
325: if(close_quote_pos!=STRING_NOT_FOUND)
326: charset_end=close_quote_pos;
327: } else {
328: size_t delim_pos=CONTENT_TYPE_VALUE.pos(';', charset_begin);
329: if(delim_pos!=STRING_NOT_FOUND)
330: charset_end=delim_pos;
331: }
1.156 paf 332: const String::Body CHARSET_NAME_BODY=
1.154 paf 333: CONTENT_TYPE_VALUE.mid(charset_begin, charset_end);
334:
335: return &charsets.get(CHARSET_NAME_BODY);
336: }
337:
338: return 0;
339: }
340:
1.178 paf 341: static const String* basic_authorization_field(const char* user, const char* pass) {
342: if(!user&& !pass)
343: return 0;
344:
345: String combined;
346: if(user)
347: combined<<user;
348: combined<<":";
349: if(pass)
350: combined<<pass;
351:
352: String* result=new String("Basic "); *result<<pa_base64(combined.cstr(), combined.length());
353: return result;
354: }
1.154 paf 355:
1.181 paf 356: static void form_string_value2string(
357: HashStringValue::key_type key,
358: const String& value,
359: String& result)
360: {
361: result << String(key, String::L_TAINTED) << "=";
362: result.append(value, String::L_URI, true);
363: result<< "&";
364: }
365: #ifndef DOXYGEN
366: struct Form_table_value2string_info {
367: HashStringValue::key_type key;
368: String& result;
369:
370: Form_table_value2string_info(HashStringValue::key_type akey, String& aresult):
371: key(akey), result(aresult) {}
372: };
373: #endif
374: static void form_table_value2string(Table::element_type row, Form_table_value2string_info* info) {
375: form_string_value2string(info->key, *row->get(0), info->result);
376: }
1.180 paf 377: static void form_value2string(
378: HashStringValue::key_type key,
379: HashStringValue::value_type value,
380: String* result)
381: {
1.181 paf 382: if(const String* svalue=value->get_string())
383: form_string_value2string(key, *svalue, *result);
384: else if(Table* tvalue=value->get_table()) {
385: Form_table_value2string_info info(key, *result);
386: tvalue->for_each(form_table_value2string, &info);
387: } else
388: throw Exception(0,
389: new String(key, String::L_TAINTED),
390: "is %s, "HTTP_FORM_NAME" option value must either string or table", value->type());
1.180 paf 391: }
392: static const char* form2string(HashStringValue& form) {
393: String string;
394: form.for_each(form_value2string, &string);
1.181 paf 395: return string.cstr(String::L_UNSPECIFIED);
1.180 paf 396: }
1.154 paf 397: #ifndef DOXYGEN
398: struct File_read_http_result {
399: char *str; size_t length;
400: HashStringValue* headers;
401: };
402: #endif
1.194 paf 403: static void find_headers_end(char* p,
404: char*& headers_end_at,
405: char*& raw_body)
406: {
407: raw_body=p;
408: // \n\n
409: // \r\n\r\n
410: while((p=strchr(p, '\n'))) {
411: headers_end_at=++p; // \n>.<
412: if(*p=='\r') // \r\n>\r?<\n
413: p++;
414: if(*p=='\n') { // \r\n\r>\n?<
415: raw_body=p+1;
416: return;
417: }
418: }
419: headers_end_at=0;
420: }
421:
1.155 paf 422: /// @todo build .cookies field. use ^file.tables.SET-COOKIES.menu{ for now
1.154 paf 423: static File_read_http_result file_read_http(Request_charsets& charsets,
424: const String& file_spec,
1.169 paf 425: bool as_text,
426: HashStringValue *options=0) {
1.154 paf 427: File_read_http_result result;
1.126 paf 428: char host[MAX_STRING];
1.129 paf 429: const char* uri;
1.193 paf 430: short port;
1.180 paf 431: const char* method="GET"; bool method_is_get;
432: HashStringValue* form=0;
433: const char* body_cstr=0;
1.126 paf 434: int timeout=2;
1.142 paf 435: bool fail_on_status_ne_200=true;
1.154 paf 436: Value* vheaders=0;
437: Charset *asked_remote_charset=0;
1.178 paf 438: const char* user_cstr=0;
439: const char* password_cstr=0;
1.126 paf 440:
1.127 paf 441: if(options) {
442: int valid_options=0;
1.177 paf 443: if(Value* vmethod=options->get(HTTP_METHOD_NAME)) {
1.127 paf 444: valid_options++;
1.154 paf 445: method=vmethod->as_string().cstr();
1.181 paf 446: }
447: if(Value* vform=options->get(HTTP_FORM_NAME)) {
1.180 paf 448: valid_options++;
449: form=vform->get_hash();
1.181 paf 450: }
451: if(Value* vbody=options->get(HTTP_BODY_NAME)) {
1.180 paf 452: valid_options++;
453: body_cstr=vbody->as_string().cstr(String::L_UNSPECIFIED);
1.181 paf 454: }
455: if(Value* vtimeout=options->get(HTTP_TIMEOUT_NAME)) {
1.127 paf 456: valid_options++;
457: timeout=vtimeout->as_int();
1.181 paf 458: }
459: if((vheaders=options->get(HTTP_HEADERS_NAME))) {
1.127 paf 460: valid_options++;
1.181 paf 461: }
462: if(Value* vany_status=options->get(HTTP_ANY_STATUS_NAME)) {
1.142 paf 463: valid_options++;
464: fail_on_status_ne_200=!vany_status->as_bool();
1.181 paf 465: }
466: if(Value* vcharset_name=options->get(HTTP_CHARSET_NAME)) {
1.154 paf 467: valid_options++;
468: asked_remote_charset=&::charsets.get(vcharset_name->as_string().
469: change_case(charsets.source(), String::CC_UPPER));
1.181 paf 470: }
471: if(Value* vuser=options->get(HTTP_USER)) {
1.178 paf 472: valid_options++;
473: user_cstr=vuser->as_string().cstr();
1.181 paf 474: }
475: if(Value* vpassword=options->get(HTTP_PASSWORD)) {
1.178 paf 476: valid_options++;
477: password_cstr=vpassword->as_string().cstr();
478: }
1.142 paf 479:
1.154 paf 480: if(valid_options!=options->count())
1.127 paf 481: throw Exception("parser.runtime",
482: 0,
483: "invalid option passed");
1.154 paf 484: }
485: if(!asked_remote_charset) // defaulting to $request:charset
486: asked_remote_charset=&charsets.source();
487:
1.180 paf 488: method_is_get=strcmp(method, "GET")==0;
489: if(method_is_get && body_cstr)
490: throw Exception("parser.runtime",
491: 0,
492: "you can not use $."HTTP_BODY_NAME" option with method GET");
493:
1.154 paf 494: //preparing request
495: String& connect_string=*new String;
496: // not in ^sql{... L_SQL ...} spirit, but closer to ^file::load one
497: connect_string.append(file_spec, String::L_URI); // tainted pieces -> URI pieces
498:
1.180 paf 499: String request_head_and_body;
1.154 paf 500: {
501: // influence URLencoding of tainted pieces to String::L_URI lang
502: Temp_client_charset temp(charsets, *asked_remote_charset);
503:
504: const char* connect_string_cstr=connect_string.cstr(String::L_UNSPECIFIED);
1.126 paf 505:
1.154 paf 506: const char* current=connect_string_cstr;
507: if(strncmp(current, "http://", 7)!=0)
508: throw Exception(0,
509: &connect_string,
510: "does not start with http://"); //never
511: current+=7;
512:
513: strncpy(host, current, sizeof(host)-1); host[sizeof(host)-1]=0;
514: char* host_uri=lsplit(host, '/');
515: uri=host_uri?current+(host_uri-1-host):"/";
516: char* port_cstr=lsplit(host, ':');
517: char* error_pos=0;
1.193 paf 518: port=port_cstr?(short)strtol(port_cstr, &error_pos, 0):80;
1.154 paf 519:
1.180 paf 520: if(strchr(uri, '?') && form)
521: throw Exception("parser.runtime",
522: 0,
523: "use either uri with ?params or $."HTTP_FORM_NAME" option");
524:
525: //making request head
526: String head;
527: head << method;
528: head << " " << uri;
529: if(form)
530: if(method_is_get)
1.181 paf 531: head << "?" << form2string(*form);
1.180 paf 532: head <<" HTTP/1.0" CRLF
1.154 paf 533: "host: "<< host << CRLF;
1.181 paf 534: if(form && !method_is_get) {
535: head << "content-type: application/x-www-form-urlencoded" CRLF;
536: body_cstr = form2string(*form);
537: }
1.178 paf 538:
1.179 paf 539: // http://www.ietf.org/rfc/rfc2617.txt
1.178 paf 540: if(const String* authorization_field_value=basic_authorization_field(user_cstr, password_cstr))
1.180 paf 541: head<<"authorization: "<<*authorization_field_value<<CRLF;
1.178 paf 542:
1.154 paf 543: bool user_agent_specified=false;
544: if(vheaders && !vheaders->is_string()) { // allow empty
545: if(HashStringValue *headers=vheaders->get_hash()) {
1.180 paf 546: Http_pass_header_info info={&charsets, &head, false};
1.154 paf 547: headers->for_each(http_pass_header, &info);
548: user_agent_specified=info.user_agent_specified;
549: } else
550: throw Exception("parser.runtime",
551: &connect_string,
552: "headers param must be hash");
553: };
554: if(!user_agent_specified) // defaulting
1.180 paf 555: head << "user-agent: " DEFAULT_USER_AGENT CRLF;
556:
557: if(body_cstr) {
558: // recode those pieces which are not in String::L_URI lang
559: // [those violating HTTP standard, but widly used]
560: body_cstr=Charset::transcode(
561: String::C(body_cstr, strlen(body_cstr)),
562: charsets.source(),
563: *asked_remote_charset);
1.181 paf 564:
565: head << "content-length: " << format(strlen(body_cstr), "%u") << CRLF;
1.180 paf 566: }
1.154 paf 567:
1.181 paf 568: const char* head_cstr=head.cstr(String::L_UNSPECIFIED);
569:
570: // recode those pieces which are not in String::L_URI lang
571: // [those violating HTTP standard, but widly used]
572: head_cstr=Charset::transcode(
573: String::C(head_cstr, strlen(head_cstr)),
574: charsets.source(),
575: *asked_remote_charset);
1.180 paf 576:
1.181 paf 577: // head + end of header
578: request_head_and_body << head_cstr << CRLF;
1.180 paf 579: // body
580: if(body_cstr)
581: request_head_and_body << body_cstr;
1.154 paf 582: }
1.126 paf 583:
584: //sending request
1.171 paf 585: char* response;
586: size_t response_size;
587: int status_code=http_request(response, response_size,
1.180 paf 588: host, port, request_head_and_body.cstr(),
1.142 paf 589: timeout, fail_on_status_ne_200);
1.126 paf 590:
591: //processing results
1.171 paf 592: char* raw_body; size_t raw_body_size;
1.194 paf 593: char* headers_end_at;
594: find_headers_end(response,
595: headers_end_at,
596: raw_body);
1.191 paf 597: raw_body_size=response_size-(raw_body-response);
1.171 paf 598:
1.154 paf 599: result.headers=new HashStringValue;
1.155 paf 600: VHash* vtables=new VHash;
1.177 paf 601: result.headers->put(HTTP_TABLES_NAME, vtables);
1.194 paf 602: Charset* real_remote_charset=0; // undetected, yet
603:
604: if(headers_end_at) {
605: *headers_end_at=0;
606: const String header_block(String::C(response, headers_end_at-response), true);
607:
608: ArrayString aheaders;
609: HashStringValue& tables=vtables->hash();
1.155 paf 610:
1.194 paf 611: size_t pos_after=0;
612: header_block.split(aheaders, pos_after, "\n");
613:
614: //processing headers
615: size_t aheaders_count=aheaders.count();
616: for(size_t i=1; i<aheaders_count; i++) {
617: const String& line=*aheaders.get(i);
618: size_t pos=line.pos(':');
619: if(pos==STRING_NOT_FOUND || pos<1)
620: throw Exception("http.response",
621: &connect_string,
622: "bad response from host - bad header \"%s\"", line.cstr());
623: const String::Body HEADER_NAME=
624: line.mid(0, pos).change_case(charsets.source(), String::CC_UPPER);
625: const String& header_value=line.mid(pos+1, line.length()).trim(String::TRIM_BOTH, " \t\r");
626: if(as_text && HEADER_NAME=="CONTENT-TYPE")
627: real_remote_charset=detect_charset(charsets.source(), header_value);
628:
629: // tables
630: {
631: Value *valready=(Value *)tables.get(HEADER_NAME);
632: bool existed=valready!=0;
633: Table *table;
634: if(existed) {
635: // second+ appearence
636: table=valready->get_table();
637: } else {
638: // first appearence
639: Table::columns_type columns =new ArrayString(1);
640: *columns+=new String("value");
641: table=new Table(columns);
642: }
643: // this string becomes next row
644: ArrayString& row=*new ArrayString(1);
645: row+=&header_value;
646: *table+=&row;
647: // not existed before? add it
648: if(!existed)
649: tables.put(HEADER_NAME, new VTable(table));
1.155 paf 650: }
1.194 paf 651:
652: result.headers->put(HEADER_NAME, new VString(header_value));
1.155 paf 653: }
1.126 paf 654: }
655:
656: // output response
1.171 paf 657: String::C real_body=String::C(raw_body, raw_body_size);
1.192 paf 658: if(as_text && raw_body_size) { // must be checked because transcode returns CONST string in case length==0, which contradicts hacking few lines below
659: // defaulting to used-asked charset [it's never empty!]
660: if(!real_remote_charset)
661: real_remote_charset=asked_remote_charset;
1.169 paf 662: real_body=Charset::transcode(real_body, *real_remote_charset, charsets.source());
1.192 paf 663: }
1.154 paf 664:
665: result.str=const_cast<char *>(real_body.str); // hacking a little
666: result.length=real_body.length;
667: result.headers->put(file_status_name, new VInt(status_code));
668: return result;
1.34 paf 669: }
1.123 paf 670:
1.154 paf 671: #endif
672:
1.123 paf 673: #ifndef DOXYGEN
674: struct File_read_action_info {
1.154 paf 675: char **data; size_t *data_size;
1.188 paf 676: char* buf; size_t offset; size_t count;
1.126 paf 677: };
1.123 paf 678: #endif
1.154 paf 679: static void file_read_action(
680: struct stat& finfo,
681: int f,
1.166 paf 682: const String& file_spec, const char* /*fname*/, bool as_text,
1.154 paf 683: void *context) {
1.126 paf 684: File_read_action_info& info=*static_cast<File_read_action_info *>(context);
1.188 paf 685: size_t to_read_size=info.count;
686: if(!to_read_size)
687: to_read_size=(size_t)finfo.st_size;
688: assert( !(info.buf && as_text) );
689: if(to_read_size) {
690: if(info.offset)
691: lseek(f, info.offset, SEEK_SET);
692: *info.data=info.buf
693: ? info.buf
694: : new(PointerFreeGC) char[to_read_size+(as_text?1:0)];
1.126 paf 695: *info.data_size=(size_t)read(f, *info.data, to_read_size);
1.123 paf 696:
697: if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size)
1.126 paf 698: throw Exception(0,
1.123 paf 699: &file_spec,
1.173 paf 700: "read failed: actually read %u bytes count not in [0..%u] valid range",
1.126 paf 701: *info.data_size, to_read_size);
1.123 paf 702: } else { // empty file
703: if(as_text) {
1.154 paf 704: *info.data=new(PointerFreeGC) char[1];
1.123 paf 705: *(char*)(*info.data)=0;
706: } else
707: *info.data=0;
708: *info.data_size=0;
709: return;
710: }
1.126 paf 711: }
1.154 paf 712: File_read_result file_read(Request_charsets& charsets, const String& file_spec,
713: bool as_text, HashStringValue *params,
1.188 paf 714: bool fail_on_read_problem,
715: char* buf, size_t offset, size_t count) {
1.167 paf 716: File_read_result result={false, 0, 0, 0};
1.154 paf 717: #ifdef PA_HTTP
718: if(file_spec.starts_with("http://")) {
1.126 paf 719: // fail on read problem
1.169 paf 720: File_read_http_result http=file_read_http(charsets, file_spec, as_text, params);
1.154 paf 721: result.success=true;
722: result.str=http.str;
723: result.length=http.length;
724: result.headers=http.headers;
1.126 paf 725: } else {
1.154 paf 726: #endif
1.161 paf 727: if(params && params->count())
728: throw Exception("parser.runtime",
729: 0,
730: "invalid option passed");
731:
1.188 paf 732: File_read_action_info info={&result.str, &result.length,
733: buf, offset, count};
1.154 paf 734: result.success=file_read_action_under_lock(file_spec,
1.126 paf 735: "read", file_read_action, &info,
736: as_text, fail_on_read_problem);
1.154 paf 737: #ifdef PA_HTTP
1.126 paf 738: }
1.154 paf 739: #endif
1.123 paf 740:
1.154 paf 741: if(result.success && as_text) {
1.131 paf 742: // UTF-8 signature: EF BB BF
1.154 paf 743: if(result.length>=3) {
744: char *in=(char *)result.str;
1.159 paf 745: if(strncmp(in, "\xEF\xBB\xBF", 3)==0) {
1.154 paf 746: result.str=in+3; result.length-=3;// skip prefix
1.131 paf 747: }
748: }
749:
1.154 paf 750: fix_line_breaks((char *)(result.str), result.length);
1.123 paf 751: }
1.126 paf 752:
753: return result;
1.123 paf 754: }
755:
1.154 paf 756: #ifdef PA_SAFE_MODE
757: void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) {
758: if(finfo.st_uid/*foreign?*/!=geteuid()
759: && finfo.st_gid/*foreign?*/!=getegid())
760: throw Exception("parser.runtime",
761: &file_spec,
762: "parser is in safe mode: "
763: "reading files of foreign group and user disabled "
764: "[recompile parser with --disable-safe-mode configure option], "
765: "actual filename '%s', "
766: "fuid(%d)!=euid(%d) or fgid(%d)!=egid(%d)",
767: fname,
768: finfo.st_uid, geteuid(),
769: finfo.st_gid, getegid());
770: }
771: #endif
1.149 paf 772:
1.154 paf 773: bool file_read_action_under_lock(const String& file_spec,
1.126 paf 774: const char* action_name, File_read_action action, void *context,
775: bool as_text,
1.123 paf 776: bool fail_on_read_problem) {
1.154 paf 777: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.33 paf 778: int f;
779:
780: // first open, next stat:
1.45 paf 781: // directory update of NTFS hard links performed on open.
1.33 paf 782: // ex:
783: // a.html:^test[] and b.html hardlink to a.html
784: // user inserts ! before ^test in a.html
1.126 paf 785: // directory entry of b.html in NTFS not updated at once,
1.35 paf 786: // they delay update till open, so we would receive "!^test[" string
787: // if would do stat, next open.
1.123 paf 788: // later: it seems, even this does not help sometimes
1.98 paf 789: if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) {
1.123 paf 790: try {
1.162 paf 791: if(pa_lock_shared_blocking(f)!=0)
1.126 paf 792: throw Exception("file.lock",
1.123 paf 793: &file_spec,
794: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 795: strerror(errno), errno, fname);
1.123 paf 796:
1.124 paf 797: struct stat finfo;
798: if(stat(fname, &finfo)!=0)
799: throw Exception("file.missing", // hardly possible: we just opened it OK
800: &file_spec,
801: "stat failed: %s (%d), actual filename '%s'",
1.154 paf 802: strerror(errno), errno, fname);
1.124 paf 803:
1.140 paf 804: #ifdef PA_SAFE_MODE
1.149 paf 805: check_safe_mode(finfo, file_spec, fname);
1.105 paf 806: #endif
1.32 paf 807:
1.154 paf 808: action(finfo, f, file_spec, fname, as_text, context);
1.123 paf 809: } catch(...) {
1.162 paf 810: pa_unlock(f);close(f);
1.123 paf 811: if(fail_on_read_problem)
1.154 paf 812: rethrow;
1.123 paf 813: return false;
814: }
1.87 paf 815:
1.162 paf 816: pa_unlock(f);close(f);
1.72 parser 817: return true;
1.118 paf 818: } else {
819: if(fail_on_read_problem)
1.126 paf 820: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.118 paf 821: &file_spec,
1.123 paf 822: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 823: action_name, strerror(errno), errno, fname);
1.118 paf 824: return false;
825: }
1.8 paf 826: }
827:
1.63 parser 828: static void create_dir_for_file(const String& file_spec) {
829: size_t pos_after=1;
1.154 paf 830: size_t pos_before;
831: while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) {
832: mkdir(file_spec.mid(0, pos_before).cstr(String::L_FILE_SPEC), 0775);
1.63 parser 833: pos_after=pos_before+1;
834: }
835: }
836:
1.98 paf 837: bool file_write_action_under_lock(
1.28 paf 838: const String& file_spec,
1.126 paf 839: const char* action_name, File_write_action action, void *context,
840: bool as_text,
841: bool do_append,
842: bool do_block,
1.110 paf 843: bool fail_on_lock_problem) {
1.154 paf 844: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.28 paf 845: int f;
1.80 paf 846: if(access(fname, W_OK)!=0) // no
1.126 paf 847: create_dir_for_file(file_spec);
1.50 paf 848:
1.80 paf 849: if((f=open(fname,
850: O_CREAT|O_RDWR
851: |(as_text?_O_TEXT:_O_BINARY)
1.138 paf 852: |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) {
1.162 paf 853: if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) {
1.126 paf 854: Exception e("file.lock",
1.110 paf 855: &file_spec,
856: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 857: strerror(errno), errno, fname);
1.126 paf 858: close(f);
1.110 paf 859: if(fail_on_lock_problem)
860: throw e;
1.98 paf 861: return false;
862: }
1.96 paf 863:
1.158 paf 864: try {
1.126 paf 865: action(f, context);
1.158 paf 866: } catch(...) {
1.138 paf 867: #ifdef HAVE_FTRUNCATE
1.104 paf 868: if(!do_append)
1.125 paf 869: ftruncate(f, lseek(f, 0, SEEK_CUR)); // one can not use O_TRUNC, read lower
1.138 paf 870: #endif
1.162 paf 871: pa_unlock(f);close(f);
1.154 paf 872: rethrow;
1.158 paf 873: }
1.80 paf 874:
1.138 paf 875: #ifdef HAVE_FTRUNCATE
1.104 paf 876: if(!do_append)
1.125 paf 877: 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 878: #endif
1.162 paf 879: pa_unlock(f);close(f);
1.98 paf 880: return true;
1.80 paf 881: } else
1.126 paf 882: throw Exception(errno==EACCES?"file.access":0,
1.80 paf 883: &file_spec,
1.96 paf 884: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 885: action_name, strerror(errno), errno, fname);
1.96 paf 886: // here should be nothing, see rethrow above
887: }
888:
889: #ifndef DOXYGEN
890: struct File_write_action_info {
1.154 paf 891: const char* str; size_t length;
1.126 paf 892: };
1.96 paf 893: #endif
894: static void file_write_action(int f, void *context) {
1.126 paf 895: File_write_action_info& info=*static_cast<File_write_action_info *>(context);
1.154 paf 896: if(info.length) {
897: int written=write(f, info.str, info.length);
1.116 paf 898: if(written<0)
1.126 paf 899: throw Exception(0,
900: 0,
901: "write failed: %s (%d)", strerror(errno), errno);
1.113 paf 902: }
1.96 paf 903: }
904: void file_write(
905: const String& file_spec,
1.154 paf 906: const char* data, size_t size,
1.126 paf 907: bool as_text,
1.96 paf 908: bool do_append) {
1.126 paf 909: File_write_action_info info={data, size};
1.98 paf 910: file_write_action_under_lock(
1.154 paf 911: file_spec,
912: "write", file_write_action, &info,
913: as_text,
914: do_append);
1.30 paf 915: }
916:
1.63 parser 917: // throws nothing! [this is required in file_move & file_delete]
1.50 paf 918: static void rmdir(const String& file_spec, size_t pos_after) {
1.154 paf 919: size_t pos_before;
920: if((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND)
1.126 paf 921: rmdir(file_spec, pos_before+1);
1.50 paf 922:
1.154 paf 923: rmdir(file_spec.mid(0, pos_after-1/* / */).cstr(String::L_FILE_SPEC));
1.50 paf 924: }
1.164 paf 925: bool file_delete(const String& file_spec, bool fail_on_problem) {
1.154 paf 926: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.54 parser 927: if(unlink(fname)!=0)
1.164 paf 928: if(fail_on_problem)
1.126 paf 929: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.93 paf 930: &file_spec,
931: "unlink failed: %s (%d), actual filename '%s'",
1.154 paf 932: strerror(errno), errno, fname);
1.93 paf 933: else
934: return false;
1.50 paf 935:
1.126 paf 936: rmdir(file_spec, 1);
1.93 paf 937: return true;
1.60 parser 938: }
1.95 paf 939: void file_move(const String& old_spec, const String& new_spec) {
1.154 paf 940: const char* old_spec_cstr=old_spec.cstr(String::L_FILE_SPEC);
941: const char* new_spec_cstr=new_spec.cstr(String::L_FILE_SPEC);
1.63 parser 942:
1.126 paf 943: create_dir_for_file(new_spec);
1.63 parser 944:
1.60 parser 945: if(rename(old_spec_cstr, new_spec_cstr)!=0)
1.126 paf 946: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.60 parser 947: &old_spec,
948: "rename failed: %s (%d), actual filename '%s' to '%s'",
1.154 paf 949: strerror(errno), errno, old_spec_cstr, new_spec_cstr);
1.63 parser 950:
1.126 paf 951: rmdir(old_spec, 1);
1.31 paf 952: }
953:
1.51 paf 954:
1.126 paf 955: bool entry_exists(const char* fname, struct stat *afinfo) {
1.118 paf 956: struct stat lfinfo;
957: bool result=stat(fname, &lfinfo)==0;
958: if(afinfo)
959: *afinfo=lfinfo;
960: return result;
1.119 paf 961: }
962:
963: bool entry_exists(const String& file_spec) {
1.154 paf 964: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.126 paf 965: return entry_exists(fname, 0);
1.118 paf 966: }
967:
1.51 paf 968: static bool entry_readable(const String& file_spec, bool need_dir) {
1.154 paf 969: char* fname=file_spec.cstrm(String::L_FILE_SPEC);
1.120 paf 970: if(need_dir) {
1.126 paf 971: size_t size=strlen(fname);
1.120 paf 972: while(size) {
1.126 paf 973: char c=fname[size-1];
1.120 paf 974: if(c=='/' || c=='\\')
975: fname[--size]=0;
976: else
977: break;
978: }
979: }
1.51 paf 980: struct stat finfo;
1.118 paf 981: if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) {
1.109 paf 982: bool is_dir=(finfo.st_mode&S_IFDIR) != 0;
1.51 paf 983: return is_dir==need_dir;
984: }
985: return false;
986: }
1.31 paf 987: bool file_readable(const String& file_spec) {
1.126 paf 988: return entry_readable(file_spec, false);
1.51 paf 989: }
990: bool dir_readable(const String& file_spec) {
1.126 paf 991: return entry_readable(file_spec, true);
1.65 parser 992: }
1.154 paf 993: const String* file_readable(const String& path, const String& name) {
994: String& result=*new String(path);
995: result << "/";
996: result << name;
997: return file_readable(result)?&result:0;
1.43 paf 998: }
999: bool file_executable(const String& file_spec) {
1.154 paf 1000: return access(file_spec.cstr(String::L_FILE_SPEC), X_OK)==0;
1.44 paf 1001: }
1002:
1.64 parser 1003: bool file_stat(const String& file_spec,
1.58 parser 1004: size_t& rsize,
1.126 paf 1005: time_t& ratime,
1006: time_t& rmtime,
1007: time_t& rctime,
1.64 parser 1008: bool fail_on_read_problem) {
1.154 paf 1009: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1010: struct stat finfo;
1.44 paf 1011: if(stat(fname, &finfo)!=0)
1.64 parser 1012: if(fail_on_read_problem)
1.126 paf 1013: throw Exception("file.missing",
1.67 parser 1014: &file_spec,
1015: "getting file size failed: %s (%d), real filename '%s'",
1.154 paf 1016: strerror(errno), errno, fname);
1.64 parser 1017: else
1018: return false;
1.58 parser 1019: rsize=finfo.st_size;
1020: ratime=finfo.st_atime;
1021: rmtime=finfo.st_mtime;
1022: rctime=finfo.st_ctime;
1.64 parser 1023: return true;
1.18 paf 1024: }
1025:
1.126 paf 1026: char* getrow(char* *row_ref, char delim) {
1027: char* result=*row_ref;
1.8 paf 1028: if(result) {
1.126 paf 1029: *row_ref=strchr(result, delim);
1.8 paf 1030: if(*row_ref)
1031: *((*row_ref)++)=0;
1032: else if(!*result)
1033: return 0;
1034: }
1035: return result;
1036: }
1037:
1.126 paf 1038: char* lsplit(char* string, char delim) {
1.23 paf 1039: if(string) {
1.126 paf 1040: char* v=strchr(string, delim);
1.8 paf 1041: if(v) {
1042: *v=0;
1043: return v+1;
1044: }
1045: }
1046: return 0;
1047: }
1048:
1.126 paf 1049: char* lsplit(char* *string_ref, char delim) {
1050: char* result=*string_ref;
1051: char* next=lsplit(*string_ref, delim);
1.8 paf 1052: *string_ref=next;
1053: return result;
1.9 paf 1054: }
1055:
1.126 paf 1056: char* rsplit(char* string, char delim) {
1.18 paf 1057: if(string) {
1.126 paf 1058: char* v=strrchr(string, delim);
1.18 paf 1059: if(v) {
1.9 paf 1060: *v=0;
1061: return v+1;
1062: }
1063: }
1064: return NULL;
1.10 paf 1065: }
1066:
1.37 paf 1067: /// @todo less stupid type detection
1.154 paf 1068: const char* format(double value, char* fmt) {
1.126 paf 1069: char local_buf[MAX_NUMBER];
1.108 paf 1070: size_t size;
1071:
1.10 paf 1072: if(fmt)
1073: if(strpbrk(fmt, "diouxX"))
1074: if(strpbrk(fmt, "ouxX"))
1.126 paf 1075: size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value);
1.10 paf 1076: else
1.126 paf 1077: size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value);
1.10 paf 1078: else
1.126 paf 1079: size=snprintf(local_buf, sizeof(local_buf), fmt, value);
1.10 paf 1080: else
1.126 paf 1081: size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value);
1.10 paf 1082:
1.154 paf 1083: return pa_strdup(local_buf, size);
1.12 paf 1084: }
1085:
1.36 paf 1086: size_t stdout_write(const void *buf, size_t size) {
1.12 paf 1087: #ifdef WIN32
1.187 paf 1088: size_t to_write = size;
1.12 paf 1089: do{
1.154 paf 1090: int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout);
1.12 paf 1091: if(chunk_written<=0)
1092: break;
1093: size-=chunk_written;
1.36 paf 1094: buf=((const char*)buf)+chunk_written;
1.126 paf 1095: } while(size>0);
1.12 paf 1096:
1.187 paf 1097: return to_write-size;
1.12 paf 1098: #else
1.126 paf 1099: return fwrite(buf, 1, size, stdout);
1.12 paf 1100: #endif
1.2 paf 1101: }
1.14 paf 1102:
1.154 paf 1103: char* unescape_chars(const char* cp, int len) {
1104: char* s=new(PointerFreeGC) char[len + 1];
1.14 paf 1105: enum EscapeState {
1.33 paf 1106: EscapeRest,
1107: EscapeFirst,
1.14 paf 1108: EscapeSecond
1109: } escapeState=EscapeRest;
1.193 paf 1110: uchar escapedValue=0;
1.14 paf 1111: int srcPos=0;
1112: int dstPos=0;
1113: while(srcPos < len) {
1.193 paf 1114: uchar ch=(uchar)cp[srcPos];
1.14 paf 1115: switch(escapeState) {
1116: case EscapeRest:
1117: if(ch=='%') {
1118: escapeState=EscapeFirst;
1119: } else if(ch=='+') {
1.126 paf 1120: s[dstPos++]=' ';
1.14 paf 1121: } else {
1122: s[dstPos++]=ch;
1123: }
1124: break;
1125: case EscapeFirst:
1.193 paf 1126: escapedValue=(uchar)(hex_value[ch] << 4);
1.14 paf 1127: escapeState=EscapeSecond;
1128: break;
1129: case EscapeSecond:
1.126 paf 1130: escapedValue +=hex_value[ch];
1.14 paf 1131: s[dstPos++]=escapedValue;
1132: escapeState=EscapeRest;
1133: break;
1134: }
1.126 paf 1135: srcPos++;
1.14 paf 1136: }
1137: s[dstPos]=0;
1138: return s;
1.24 paf 1139: }
1140:
1141: #ifdef WIN32
1.126 paf 1142: void back_slashes_to_slashes(char* s) {
1.24 paf 1143: if(s)
1144: for(; *s; s++)
1145: if(*s=='\\')
1.126 paf 1146: *s='/';
1.24 paf 1147: }
1.42 paf 1148: /*
1.126 paf 1149: void slashes_to_back_slashes(char* s) {
1.42 paf 1150: if(s)
1151: for(; *s; s++)
1152: if(*s=='/')
1.126 paf 1153: *s='\\';
1.42 paf 1154: }
1155: */
1.24 paf 1156: #endif
1.41 paf 1157:
1.126 paf 1158: bool StrEqNc(const char* s1, const char* s2, bool strict) {
1.41 paf 1159: while(true) {
1160: if(!(*s1)) {
1161: if(!(*s2))
1162: return true;
1163: else
1164: return !strict;
1165: } else if(!(*s2))
1166: return !strict;
1.189 paf 1167: if(isalpha((unsigned char)*s1)) {
1.190 paf 1168: if(tolower((unsigned char)*s1) !=tolower((unsigned char)*s2))
1.41 paf 1169: return false;
1170: } else if((*s1) !=(*s2))
1171: return false;
1.126 paf 1172: s1++;
1173: s2++;
1.41 paf 1174: }
1.57 parser 1175: }
1176:
1.84 paf 1177: static bool isLeap(int year) {
1.57 parser 1178: return !(
1179: (year % 4) || ((year % 400) && !(year % 100))
1.126 paf 1180: );
1.57 parser 1181: }
1182:
1183: int getMonthDays(int year, int month) {
1184: int monthDays[]={
1.126 paf 1185: 31,
1186: isLeap(year) ? 29 : 28,
1187: 31,
1188: 30,
1189: 31,
1190: 30,
1191: 31,
1192: 31,
1193: 30,
1194: 31,
1195: 30,
1.57 parser 1196: 31
1.126 paf 1197: };
1198: return monthDays[month];
1.41 paf 1199: }
1.69 parser 1200:
1.126 paf 1201: void remove_crlf(char* start, char* end) {
1202: for(char* p=start; p<end; p++)
1.69 parser 1203: switch(*p) {
1.126 paf 1204: case '\n': *p='|'; break;
1205: case '\r': *p=' '; break;
1.69 parser 1206: }
1.91 paf 1207: }
1208:
1209:
1210: /// must be last in this file
1211: #undef vsnprintf
1.126 paf 1212: int __vsnprintf(char* b, size_t s, const char* f, va_list l) {
1.91 paf 1213: if(!s)
1214: return 0;
1215:
1216: int r;
1217: // note: on win32& maybe somewhere else
1218: // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing
1219: --s;
1.172 paf 1220:
1221: // clients do not check for negative 's', feature: ignore such prints
1222: if((ssize_t)s<0)
1223: return 0;
1224:
1.91 paf 1225: #if _MSC_VER
1226: /*
1227: win32:
1228: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm
1229:
1.154 paf 1230: if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 is returned
1.91 paf 1231: */
1.126 paf 1232: r=_vsnprintf(b, s, f, l);
1.91 paf 1233: if(r<0)
1234: r=s;
1235: #else
1.126 paf 1236: r=vsnprintf(b, s, f, l);
1.91 paf 1237: /*
1238: solaris:
1239: man vsnprintf
1240:
1241: The snprintf() function returns the number of characters
1242: formatted, that is, the number of characters that would have
1243: been written to the buffer if it were large enough. If the
1244: value of n is 0 on a call to snprintf(), an unspecified
1245: value less than 1 is returned.
1246: */
1247:
1248: if(r<0)
1249: r=0;
1.167 paf 1250: else if((size_t)r>s)
1.91 paf 1251: r=s;
1252: #endif
1253: b[r]=0;
1254: return r;
1255: }
1256:
1.126 paf 1257: int __snprintf(char* b, size_t s, const char* f, ...) {
1.91 paf 1258: va_list l;
1.126 paf 1259: va_start(l, f);
1260: int r=__vsnprintf(b, s, f, l);
1261: va_end(l);
1.91 paf 1262: return r;
1.178 paf 1263: }
1264:
1265: /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */
1266: /*
1267: * Authors: Michael Zucchi <notzed@helixcode.com>
1268: * Jeffrey Stedfast <fejj@helixcode.com>
1269: *
1270: * Copyright 2000 Helix Code, Inc. (www.helixcode.com)
1271: *
1272: * This program is free software; you can redistribute it and/or modify
1273: * it under the terms of the GNU General Public License as published by
1274: * the Free Software Foundation; either version 2 of the License, or
1275: * (at your option) any later version.
1276: *
1277: * This program is distributed in the hope that it will be useful,
1278: * but WITHOUT ANY WARRANTY; without even the implied warranty of
1279: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
1280: * GNU General Public License for more details.
1281: *
1282: * You should have received a copy of the GNU General Public License
1283: * along with this program; if not, write to the Free Software
1284: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
1285: *
1286: */
1287: static char *base64_alphabet =
1288: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
1289:
1290: /**
1291: * g_mime_utils_base64_encode_step:
1292: * @in: input stream
1293: * @inlen: length of the input
1294: * @out: output string
1295: * @state: holds the number of bits that are stored in @save
1296: * @save: leftover bits that have not yet been encoded
1297: *
1298: * Base64 encodes a chunk of data. Performs an 'encode step', only
1299: * encodes blocks of 3 characters to the output at a time, saves
1300: * left-over state in state and save (initialise to 0 on first
1301: * invocation).
1302: *
1303: * Returns the number of bytes encoded.
1304: **/
1305: static size_t
1306: g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1307: {
1.186 paf 1308: register const unsigned char *inptr;
1.178 paf 1309: register unsigned char *outptr;
1310:
1311: if (inlen <= 0)
1312: return 0;
1313:
1314: inptr = in;
1315: outptr = out;
1316:
1317: if (inlen + ((unsigned char *)save)[0] > 2) {
1318: const unsigned char *inend = in + inlen - 2;
1319: register int c1 = 0, c2 = 0, c3 = 0;
1320: register int already;
1321:
1322: already = *state;
1323:
1324: switch (((char *)save)[0]) {
1325: case 1: c1 = ((unsigned char *)save)[1]; goto skip1;
1326: case 2: c1 = ((unsigned char *)save)[1];
1327: c2 = ((unsigned char *)save)[2]; goto skip2;
1328: }
1329:
1330: /* yes, we jump into the loop, no i'm not going to change it, its beautiful! */
1331: while (inptr < inend) {
1332: c1 = *inptr++;
1333: skip1:
1334: c2 = *inptr++;
1335: skip2:
1336: c3 = *inptr++;
1337: *outptr++ = base64_alphabet [c1 >> 2];
1338: *outptr++ = base64_alphabet [(c2 >> 4) | ((c1 & 0x3) << 4)];
1339: *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)];
1340: *outptr++ = base64_alphabet [c3 & 0x3f];
1341: /* this is a bit ugly ... */
1342: if ((++already) >= 19) {
1343: *outptr++ = '\n';
1344: already = 0;
1345: }
1346: }
1347:
1348: ((unsigned char *)save)[0] = 0;
1349: inlen = 2 - (inptr - inend);
1350: *state = already;
1351: }
1352:
1353: //d(printf ("state = %d, inlen = %d\n", (int)((char *)save)[0], inlen));
1354:
1355: if (inlen > 0) {
1356: register char *saveout;
1357:
1358: /* points to the slot for the next char to save */
1359: saveout = & (((char *)save)[1]) + ((char *)save)[0];
1360:
1361: /* inlen can only be 0 1 or 2 */
1362: switch (inlen) {
1363: case 2: *saveout++ = *inptr++;
1364: case 1: *saveout++ = *inptr++;
1365: }
1366: ((char *)save)[0] += inlen;
1367: }
1368:
1369: /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n",
1370: (int)((char *)save)[0],
1371: (int)((char *)save)[1],
1372: (int)((char *)save)[2]));*/
1373:
1374: return (outptr - out);
1375: }
1376:
1377: /**
1378: * g_mime_utils_base64_encode_close:
1379: * @in: input stream
1380: * @inlen: length of the input
1381: * @out: output string
1382: * @state: holds the number of bits that are stored in @save
1383: * @save: leftover bits that have not yet been encoded
1384: *
1385: * Base64 encodes the input stream to the output stream. Call this
1386: * when finished encoding data with g_mime_utils_base64_encode_step to
1387: * flush off the last little bit.
1388: *
1389: * Returns the number of bytes encoded.
1390: **/
1391: static size_t
1392: g_mime_utils_base64_encode_close (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1393: {
1394: unsigned char *outptr = out;
1395: int c1, c2;
1396:
1397: if (inlen > 0)
1398: outptr += g_mime_utils_base64_encode_step (in, inlen, outptr, state, save);
1399:
1400: c1 = ((unsigned char *)save)[1];
1401: c2 = ((unsigned char *)save)[2];
1402:
1403: switch (((unsigned char *)save)[0]) {
1404: case 2:
1405: outptr[2] = base64_alphabet [(c2 & 0x0f) << 2];
1406: goto skip;
1407: case 1:
1408: outptr[2] = '=';
1409: skip:
1410: outptr[0] = base64_alphabet [c1 >> 2];
1411: outptr[1] = base64_alphabet [c2 >> 4 | ((c1 & 0x3) << 4)];
1412: outptr[3] = '=';
1413: outptr += 4;
1414: break;
1415: }
1416:
1417: *outptr++ = 0;
1418:
1419: *save = 0;
1420: *state = 0;
1421:
1422: return (outptr - out);
1423: }
1424:
1425: char* pa_base64(const char *in, size_t len)
1426: {
1427: /* wont go to more than 2x size (overly conservative) */
1428: char* result=new(PointerFreeGC) char[len * 2 + 6];
1429: int state=0;
1430: int save=0;
1.183 paf 1431: #ifndef NDEBUG
1432: size_t filled=
1433: #endif
1434: g_mime_utils_base64_encode_close ((const unsigned char*)in, len,
1.178 paf 1435: (unsigned char*)result, &state, &save);
1436: assert(filled <= len * 2 + 6);
1437:
1438: return result;
1.98 paf 1439: }
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