Annotation of parser3/src/main/pa_common.C, revision 1.260
1.15 paf 1: /** @file
1.16 paf 2: Parser: commonly functions.
3:
1.242 misha 4: Copyright(c) 2001-2009 ArtLebedev Group (http://www.artlebedev.com)
1.101 paf 5: Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
1.16 paf 6:
1.210 paf 7: * BASE64 part
8: * Authors: Michael Zucchi <notzed@ximian.com>
9: * Jeffrey Stedfast <fejj@ximian.com>
10: *
11: * Copyright 2000-2004 Ximian, Inc. (www.ximian.com)
12: *
13: * This program is free software; you can redistribute it and/or modify
14: * it under the terms of the GNU General Public License as published by
15: * the Free Software Foundation; either version 2 of the License, or
16: * (at your option) any later version.
17: *
18: * This program is distributed in the hope that it will be useful,
19: * but WITHOUT ANY WARRANTY; without even the implied warranty of
20: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
21: * GNU General Public License for more details.
22: *
23: * You should have received a copy of the GNU General Public License
24: * along with this program; if not, write to the Free Software
25: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
26: *
27: */
28:
1.260 ! misha 29: static const char * const IDENT_COMMON_C="$Date: 2010-05-25 09:30:30 $";
1.1 paf 30:
31: #include "pa_common.h"
1.4 paf 32: #include "pa_exception.h"
1.154 paf 33: #include "pa_hash.h"
1.14 paf 34: #include "pa_globals.h"
1.154 paf 35: #include "pa_charsets.h"
1.214 paf 36: #include "pa_http.h"
1.223 misha 37: #include "pa_request_charsets.h"
1.237 misha 38: #include "pcre.h"
1.241 misha 39: #include "pa_request.h"
1.98 paf 40:
1.93 paf 41: // some maybe-undefined constants
42:
1.82 paf 43: #ifndef _O_TEXT
44: # define _O_TEXT 0
45: #endif
46: #ifndef _O_BINARY
47: # define _O_BINARY 0
1.47 paf 48: #endif
1.80 paf 49:
1.138 paf 50: #ifdef HAVE_FTRUNCATE
51: # define PA_O_TRUNC 0
52: #else
53: # ifdef _O_TRUNC
54: # define PA_O_TRUNC _O_TRUNC
55: # else
56: # error you must have either ftruncate function or _O_TRUNC bit declared
57: # endif
1.154 paf 58: #endif
1.176 paf 59:
1.154 paf 60: // defines for globals
61:
62: #define FILE_STATUS_NAME "status"
63:
64: // globals
65:
66: const String file_status_name(FILE_STATUS_NAME);
67:
68: // functions
1.127 paf 69:
1.255 misha 70: bool capitalized(const char* s){
71: bool upper=true;
72: for(const char* c=s; *c; c++){
73: if(*c != (upper ? toupper((unsigned char)*c) : tolower((unsigned char)*c)))
74: return false;
75: upper=strchr("-_ ", *c) != 0;
76: }
77: return true;
78: }
79:
1.249 misha 80: const char* capitalize(const char* s){
1.255 misha 81: if(!s || capitalized(s))
82: return s;
83:
1.249 misha 84: char* result=pa_strdup(s);
85: if(result){
86: bool upper=true;
87: for(char* c=result; *c; c++){
88: *c=upper ? (char)toupper((unsigned char)*c) : (char)tolower((unsigned char)*c);
89: upper=strchr("-_ ", *c) != 0;
90: }
91: }
92: return (const char*)result;
93: }
94:
1.154 paf 95: void fix_line_breaks(char *str, size_t& length) {
1.87 paf 96: //_asm int 3;
1.154 paf 97: const char* const eob=str+length;
98: char* dest=str;
1.72 parser 99: // fix DOS: \r\n -> \n
100: // fix Macintosh: \r -> \n
1.154 paf 101: char* bol=str;
1.137 paf 102: while(char* eol=(char*)memchr(bol, '\r', eob -bol)) {
1.72 parser 103: size_t len=eol-bol;
104: if(dest!=bol)
1.260 ! misha 105: memmove(dest, bol, len);
1.72 parser 106: dest+=len;
1.126 paf 107: *dest++='\n';
1.72 parser 108:
1.126 paf 109: if(&eol[1]<eob && eol[1]=='\n') { // \r, \n = DOS
1.72 parser 110: bol=eol+2;
1.154 paf 111: length--;
1.126 paf 112: } else // \r, not \n = Macintosh
1.72 parser 113: bol=eol+1;
114: }
1.154 paf 115: // last piece without \r
1.72 parser 116: if(dest!=bol)
1.260 ! misha 117: memmove(dest, bol, eob-bol);
1.154 paf 118: str[length]=0; // terminating
1.72 parser 119: }
1.18 paf 120:
1.154 paf 121: char* file_read_text(Request_charsets& charsets,
1.229 misha 122: const String& file_spec,
123: bool fail_on_read_problem,
1.234 misha 124: HashStringValue* params,
125: bool transcode_result) {
1.154 paf 126: File_read_result file=
1.234 misha 127: file_read(charsets, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result);
1.154 paf 128: return file.success?file.str:0;
1.126 paf 129: }
130:
1.241 misha 131: char* file_load_text(Request& r,
132: const String& file_spec,
133: bool fail_on_read_problem,
134: HashStringValue* params,
135: bool transcode_result) {
136: File_read_result file=
137: file_load(r, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result);
138: return file.success?file.str:0;
139: }
140:
1.206 paf 141: /// these options were handled but not checked elsewhere, now check them
1.239 misha 142: int pa_get_valid_file_options_count(HashStringValue& options) {
1.206 paf 143: int result=0;
144: if(options.get(PA_SQL_LIMIT_NAME))
145: result++;
146: if(options.get(PA_SQL_OFFSET_NAME))
147: result++;
148: if(options.get(PA_COLUMN_SEPARATOR_NAME))
149: result++;
150: if(options.get(PA_COLUMN_ENCLOSER_NAME))
151: result++;
1.223 misha 152: if(options.get(PA_CHARSET_NAME))
153: result++;
1.206 paf 154: return result;
155: }
156:
1.123 paf 157: #ifndef DOXYGEN
158: struct File_read_action_info {
1.154 paf 159: char **data; size_t *data_size;
1.188 paf 160: char* buf; size_t offset; size_t count;
1.126 paf 161: };
1.123 paf 162: #endif
1.154 paf 163: static void file_read_action(
1.229 misha 164: struct stat& finfo,
165: int f,
166: const String& file_spec, const char* /*fname*/, bool as_text,
167: void *context) {
1.126 paf 168: File_read_action_info& info=*static_cast<File_read_action_info *>(context);
1.188 paf 169: size_t to_read_size=info.count;
170: if(!to_read_size)
171: to_read_size=(size_t)finfo.st_size;
172: assert( !(info.buf && as_text) );
173: if(to_read_size) {
174: if(info.offset)
175: lseek(f, info.offset, SEEK_SET);
176: *info.data=info.buf
177: ? info.buf
1.253 misha 178: : (char *)pa_malloc_atomic(to_read_size+1);
1.126 paf 179: *info.data_size=(size_t)read(f, *info.data, to_read_size);
1.123 paf 180:
181: if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size)
1.238 misha 182: throw Exception("file.read",
1.123 paf 183: &file_spec,
1.173 paf 184: "read failed: actually read %u bytes count not in [0..%u] valid range",
1.126 paf 185: *info.data_size, to_read_size);
1.123 paf 186: } else { // empty file
1.209 paf 187: // for both, text and binary: for text we need that terminator, for binary we need nonzero pointer to be able to save such files
1.253 misha 188: *info.data=(char *)pa_malloc_atomic(1);
1.209 paf 189: *(char*)(*info.data)=0;
1.123 paf 190: *info.data_size=0;
191: return;
192: }
1.126 paf 193: }
1.241 misha 194:
1.154 paf 195: File_read_result file_read(Request_charsets& charsets, const String& file_spec,
1.229 misha 196: bool as_text, HashStringValue *params,
197: bool fail_on_read_problem,
1.234 misha 198: char* buf, size_t offset, size_t count, bool transcode_text_result) {
1.167 paf 199: File_read_result result={false, 0, 0, 0};
1.241 misha 200: if(params){
201: int valid_options=pa_get_valid_file_options_count(*params);
202: if(valid_options!=params->count())
1.224 misha 203: throw Exception(PARSER_RUNTIME,
1.203 paf 204: 0,
1.250 misha 205: INVALID_OPTION_PASSED);
1.241 misha 206: }
1.203 paf 207:
1.241 misha 208: File_read_action_info info={&result.str, &result.length, buf, offset, count};
1.161 paf 209:
1.241 misha 210: result.success=file_read_action_under_lock(file_spec,
211: "read", file_read_action, &info,
212: as_text, fail_on_read_problem);
1.223 misha 213:
1.241 misha 214: if(as_text){
215: if(result.success){
1.236 misha 216: if(result.length>=3 && strncmp(result.str, "\xEF\xBB\xBF", 3)==0){
1.240 misha 217: // skip UTF-8 signature (BOM code)
1.236 misha 218: result.str+=3;
219: result.length-=3;
220: }
221:
222: if(result.length && transcode_text_result && params){ // must be checked because transcode returns CONST string in case length==0, which contradicts hacking few lines below
1.240 misha 223: if(Value* vcharset_name=params->get(PA_CHARSET_NAME)){
1.236 misha 224: Charset asked_charset=::charsets.get(vcharset_name->as_string().
225: change_case(charsets.source(), String::CC_UPPER));
226:
227: String::C body=String::C(result.str, result.length);
228: body=Charset::transcode(body, asked_charset, charsets.source());
1.123 paf 229:
1.236 misha 230: result.str=const_cast<char*>(body.str); // hacking a little
231: result.length=body.length;
232: }
1.131 paf 233: }
234: }
1.241 misha 235: if(result.length)
236: fix_line_breaks(result.str, result.length);
1.123 paf 237: }
1.241 misha 238:
239: return result;
240: }
241:
242: File_read_result file_load(Request& r, const String& file_spec,
243: bool as_text, HashStringValue *params,
244: bool fail_on_read_problem,
245: char* buf, size_t offset, size_t count, bool transcode_text_result) {
246:
247: File_read_result result={false, 0, 0, 0};
248: if(file_spec.starts_with("http://")) {
249: if(offset || count)
250: throw Exception(PARSER_RUNTIME,
251: 0,
252: "offset and load options are not supported for HTTP:// file load");
253:
254: // fail on read problem
255: File_read_http_result http=pa_internal_file_read_http(r, file_spec, as_text, params, transcode_text_result);
256: result.success=true;
257: result.str=http.str;
258: result.length=http.length;
259: result.headers=http.headers;
260: } else
261: result=
262: file_read(r.charsets, file_spec, as_text, params, fail_on_read_problem, buf, offset, count, transcode_text_result);
1.126 paf 263:
264: return result;
1.123 paf 265: }
266:
1.257 pretende 267:
1.154 paf 268: #ifdef PA_SAFE_MODE
1.259 misha 269: void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) {
1.154 paf 270: if(finfo.st_uid/*foreign?*/!=geteuid()
271: && finfo.st_gid/*foreign?*/!=getegid())
1.224 misha 272: throw Exception(PARSER_RUNTIME,
1.154 paf 273: &file_spec,
274: "parser is in safe mode: "
275: "reading files of foreign group and user disabled "
276: "[recompile parser with --disable-safe-mode configure option], "
277: "actual filename '%s', "
278: "fuid(%d)!=euid(%d) or fgid(%d)!=egid(%d)",
279: fname,
280: finfo.st_uid, geteuid(),
1.259 misha 281: finfo.st_gid, getegid());
282: }
283: #else
284: void check_safe_mode(struct stat, const String&, const char*) {
285: }
1.257 pretende 286: #endif
1.259 misha 287:
1.257 pretende 288:
1.149 paf 289:
1.154 paf 290: bool file_read_action_under_lock(const String& file_spec,
1.126 paf 291: const char* action_name, File_read_action action, void *context,
292: bool as_text,
1.123 paf 293: bool fail_on_read_problem) {
1.247 misha 294: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.33 paf 295: int f;
296:
297: // first open, next stat:
1.45 paf 298: // directory update of NTFS hard links performed on open.
1.33 paf 299: // ex:
300: // a.html:^test[] and b.html hardlink to a.html
301: // user inserts ! before ^test in a.html
1.126 paf 302: // directory entry of b.html in NTFS not updated at once,
1.35 paf 303: // they delay update till open, so we would receive "!^test[" string
304: // if would do stat, next open.
1.123 paf 305: // later: it seems, even this does not help sometimes
1.229 misha 306: if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) {
1.123 paf 307: try {
1.162 paf 308: if(pa_lock_shared_blocking(f)!=0)
1.126 paf 309: throw Exception("file.lock",
1.123 paf 310: &file_spec,
311: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 312: strerror(errno), errno, fname);
1.123 paf 313:
1.124 paf 314: struct stat finfo;
1.254 misha 315: if(fstat(f, &finfo)!=0)
1.124 paf 316: throw Exception("file.missing", // hardly possible: we just opened it OK
317: &file_spec,
318: "stat failed: %s (%d), actual filename '%s'",
1.154 paf 319: strerror(errno), errno, fname);
1.124 paf 320:
1.149 paf 321: check_safe_mode(finfo, file_spec, fname);
1.32 paf 322:
1.154 paf 323: action(finfo, f, file_spec, fname, as_text, context);
1.123 paf 324: } catch(...) {
1.162 paf 325: pa_unlock(f);close(f);
1.123 paf 326: if(fail_on_read_problem)
1.154 paf 327: rethrow;
1.123 paf 328: return false;
329: }
1.87 paf 330:
1.162 paf 331: pa_unlock(f);close(f);
1.72 parser 332: return true;
1.229 misha 333: } else {
1.118 paf 334: if(fail_on_read_problem)
1.256 misha 335: throw Exception(errno==EACCES?"file.access"
336: :(errno==ENOENT || errno==ENOTDIR || errno==ENODEV)?"file.missing":0,
1.118 paf 337: &file_spec,
1.123 paf 338: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 339: action_name, strerror(errno), errno, fname);
1.118 paf 340: return false;
341: }
1.8 paf 342: }
343:
1.202 paf 344: void create_dir_for_file(const String& file_spec) {
1.63 parser 345: size_t pos_after=1;
1.154 paf 346: size_t pos_before;
347: while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) {
1.247 misha 348: mkdir(file_spec.mid(0, pos_before).taint_cstr(String::L_FILE_SPEC), 0775);
1.63 parser 349: pos_after=pos_before+1;
350: }
351: }
352:
1.98 paf 353: bool file_write_action_under_lock(
1.28 paf 354: const String& file_spec,
1.225 misha 355: const char* action_name,
356: File_write_action action,
357: void *context,
1.126 paf 358: bool as_text,
359: bool do_append,
360: bool do_block,
1.110 paf 361: bool fail_on_lock_problem) {
1.247 misha 362: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.28 paf 363: int f;
1.80 paf 364: if(access(fname, W_OK)!=0) // no
1.126 paf 365: create_dir_for_file(file_spec);
1.50 paf 366:
1.80 paf 367: if((f=open(fname,
368: O_CREAT|O_RDWR
369: |(as_text?_O_TEXT:_O_BINARY)
1.138 paf 370: |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) {
1.162 paf 371: if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) {
1.126 paf 372: Exception e("file.lock",
1.110 paf 373: &file_spec,
374: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 375: strerror(errno), errno, fname);
1.126 paf 376: close(f);
1.110 paf 377: if(fail_on_lock_problem)
378: throw e;
1.98 paf 379: return false;
380: }
1.96 paf 381:
1.158 paf 382: try {
1.254 misha 383: #if (defined(HAVE_FCHMOD) && defined(PA_SAFE_MODE))
384: struct stat finfo;
385: if(fstat(f, &finfo)==0 && finfo.st_mode & 0111)
386: fchmod(f, finfo.st_mode & 0666/*clear executable bits*/); // backward: ignore errors if any
387: #endif
388: action(f, context);
1.158 paf 389: } catch(...) {
1.138 paf 390: #ifdef HAVE_FTRUNCATE
1.104 paf 391: if(!do_append)
1.125 paf 392: ftruncate(f, lseek(f, 0, SEEK_CUR)); // one can not use O_TRUNC, read lower
1.138 paf 393: #endif
1.162 paf 394: pa_unlock(f);close(f);
1.154 paf 395: rethrow;
1.158 paf 396: }
1.80 paf 397:
1.138 paf 398: #ifdef HAVE_FTRUNCATE
1.104 paf 399: if(!do_append)
1.125 paf 400: 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 401: #endif
1.162 paf 402: pa_unlock(f);close(f);
1.98 paf 403: return true;
1.80 paf 404: } else
1.126 paf 405: throw Exception(errno==EACCES?"file.access":0,
1.80 paf 406: &file_spec,
1.96 paf 407: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 408: action_name, strerror(errno), errno, fname);
1.96 paf 409: // here should be nothing, see rethrow above
410: }
411:
412: #ifndef DOXYGEN
413: struct File_write_action_info {
1.250 misha 414: const char* str;
415: size_t length;
1.126 paf 416: };
1.96 paf 417: #endif
418: static void file_write_action(int f, void *context) {
1.126 paf 419: File_write_action_info& info=*static_cast<File_write_action_info *>(context);
1.154 paf 420: if(info.length) {
421: int written=write(f, info.str, info.length);
1.116 paf 422: if(written<0)
1.238 misha 423: throw Exception("file.access",
1.126 paf 424: 0,
425: "write failed: %s (%d)", strerror(errno), errno);
1.113 paf 426: }
1.96 paf 427: }
428: void file_write(
1.250 misha 429: Request_charsets& charsets,
430: const String& file_spec,
431: const char* data,
432: size_t size,
1.126 paf 433: bool as_text,
1.250 misha 434: bool do_append,
435: Charset* asked_charset) {
436:
437: if(as_text && asked_charset){
438: String::C body=String::C(data, size);
439: body=Charset::transcode(body, charsets.source(), *asked_charset);
440: data=body.str;
441: size=body.length;
442: };
443:
1.126 paf 444: File_write_action_info info={data, size};
1.225 misha 445:
1.98 paf 446: file_write_action_under_lock(
1.154 paf 447: file_spec,
1.225 misha 448: "write",
449: file_write_action,
450: &info,
1.154 paf 451: as_text,
452: do_append);
1.30 paf 453: }
454:
1.63 parser 455: // throws nothing! [this is required in file_move & file_delete]
1.50 paf 456: static void rmdir(const String& file_spec, size_t pos_after) {
1.154 paf 457: size_t pos_before;
458: if((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND)
1.126 paf 459: rmdir(file_spec, pos_before+1);
1.50 paf 460:
1.247 misha 461: rmdir(file_spec.mid(0, pos_after-1/* / */).taint_cstr(String::L_FILE_SPEC));
1.50 paf 462: }
1.239 misha 463:
1.164 paf 464: bool file_delete(const String& file_spec, bool fail_on_problem) {
1.247 misha 465: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.54 parser 466: if(unlink(fname)!=0)
1.164 paf 467: if(fail_on_problem)
1.126 paf 468: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.93 paf 469: &file_spec,
470: "unlink failed: %s (%d), actual filename '%s'",
1.154 paf 471: strerror(errno), errno, fname);
1.93 paf 472: else
473: return false;
1.50 paf 474:
1.126 paf 475: rmdir(file_spec, 1);
1.93 paf 476: return true;
1.60 parser 477: }
1.239 misha 478:
1.95 paf 479: void file_move(const String& old_spec, const String& new_spec) {
1.247 misha 480: const char* old_spec_cstr=old_spec.taint_cstr(String::L_FILE_SPEC);
481: const char* new_spec_cstr=new_spec.taint_cstr(String::L_FILE_SPEC);
1.63 parser 482:
1.126 paf 483: create_dir_for_file(new_spec);
1.63 parser 484:
1.60 parser 485: if(rename(old_spec_cstr, new_spec_cstr)!=0)
1.126 paf 486: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.60 parser 487: &old_spec,
488: "rename failed: %s (%d), actual filename '%s' to '%s'",
1.154 paf 489: strerror(errno), errno, old_spec_cstr, new_spec_cstr);
1.63 parser 490:
1.126 paf 491: rmdir(old_spec, 1);
1.31 paf 492: }
493:
1.51 paf 494:
1.126 paf 495: bool entry_exists(const char* fname, struct stat *afinfo) {
1.118 paf 496: struct stat lfinfo;
497: bool result=stat(fname, &lfinfo)==0;
498: if(afinfo)
499: *afinfo=lfinfo;
500: return result;
1.119 paf 501: }
502:
503: bool entry_exists(const String& file_spec) {
1.247 misha 504: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.126 paf 505: return entry_exists(fname, 0);
1.118 paf 506: }
507:
1.51 paf 508: static bool entry_readable(const String& file_spec, bool need_dir) {
1.247 misha 509: char* fname=file_spec.taint_cstrm(String::L_FILE_SPEC);
1.120 paf 510: if(need_dir) {
1.126 paf 511: size_t size=strlen(fname);
1.120 paf 512: while(size) {
1.126 paf 513: char c=fname[size-1];
1.120 paf 514: if(c=='/' || c=='\\')
515: fname[--size]=0;
516: else
517: break;
518: }
519: }
1.51 paf 520: struct stat finfo;
1.118 paf 521: if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) {
1.109 paf 522: bool is_dir=(finfo.st_mode&S_IFDIR) != 0;
1.51 paf 523: return is_dir==need_dir;
524: }
525: return false;
526: }
1.239 misha 527:
1.215 paf 528: bool file_exist(const String& file_spec) {
1.126 paf 529: return entry_readable(file_spec, false);
1.51 paf 530: }
1.239 misha 531:
1.215 paf 532: bool dir_exists(const String& file_spec) {
1.126 paf 533: return entry_readable(file_spec, true);
1.65 parser 534: }
1.239 misha 535:
1.215 paf 536: const String* file_exist(const String& path, const String& name) {
1.154 paf 537: String& result=*new String(path);
538: result << "/";
539: result << name;
1.215 paf 540: return file_exist(result)?&result:0;
1.43 paf 541: }
1.239 misha 542:
1.43 paf 543: bool file_executable(const String& file_spec) {
1.247 misha 544: return access(file_spec.taint_cstr(String::L_FILE_SPEC), X_OK)==0;
1.44 paf 545: }
546:
1.64 parser 547: bool file_stat(const String& file_spec,
1.229 misha 548: size_t& rsize,
549: time_t& ratime,
550: time_t& rmtime,
551: time_t& rctime,
552: bool fail_on_read_problem) {
1.247 misha 553: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.154 paf 554: struct stat finfo;
1.44 paf 555: if(stat(fname, &finfo)!=0)
1.64 parser 556: if(fail_on_read_problem)
1.126 paf 557: throw Exception("file.missing",
1.67 parser 558: &file_spec,
559: "getting file size failed: %s (%d), real filename '%s'",
1.154 paf 560: strerror(errno), errno, fname);
1.64 parser 561: else
562: return false;
1.58 parser 563: rsize=finfo.st_size;
564: ratime=finfo.st_atime;
565: rmtime=finfo.st_mtime;
566: rctime=finfo.st_ctime;
1.64 parser 567: return true;
1.18 paf 568: }
569:
1.126 paf 570: char* getrow(char* *row_ref, char delim) {
1.229 misha 571: char* result=*row_ref;
572: if(result) {
1.126 paf 573: *row_ref=strchr(result, delim);
1.8 paf 574: if(*row_ref)
575: *((*row_ref)++)=0;
576: else if(!*result)
577: return 0;
1.229 misha 578: }
579: return result;
1.8 paf 580: }
581:
1.126 paf 582: char* lsplit(char* string, char delim) {
1.229 misha 583: if(string) {
1.126 paf 584: char* v=strchr(string, delim);
1.8 paf 585: if(v) {
586: *v=0;
587: return v+1;
588: }
1.229 misha 589: }
590: return 0;
1.8 paf 591: }
592:
1.126 paf 593: char* lsplit(char* *string_ref, char delim) {
1.229 misha 594: char* result=*string_ref;
1.126 paf 595: char* next=lsplit(*string_ref, delim);
1.229 misha 596: *string_ref=next;
597: return result;
1.9 paf 598: }
599:
1.126 paf 600: char* rsplit(char* string, char delim) {
1.229 misha 601: if(string) {
1.126 paf 602: char* v=strrchr(string, delim);
1.18 paf 603: if(v) {
1.9 paf 604: *v=0;
605: return v+1;
606: }
1.229 misha 607: }
608: return NULL;
1.10 paf 609: }
610:
1.229 misha 611:
612: // format: %[flags][width][.precision]type http://msdn.microsoft.com/ru-ru/library/56e442dc(en-us,VS.80).aspx
613: // flags: '-', '+', ' ', '#', '0' http://msdn.microsoft.com/ru-ru/library/8aky45ct(en-us,VS.80).aspx
614: // width, precision: non negative decimal number
615: enum FormatType {
616: FormatInvalid,
617: FormatInt,
618: FormatUInt,
619: FormatDouble
620: };
621: FormatType format_type(char* fmt){
622: enum FormatState {
623: Percent,
624: Flags,
625: Width,
626: Precision,
627: Done
628: } state=Percent;
629:
630: FormatType result=FormatInvalid;
631:
632: char* pos=fmt;
633: while(char c=*(pos++)){
634: switch(state){
635: case Percent:
636: if(c=='%'){
637: state=Flags;
638: } else {
639: return FormatInvalid; // 1st char must be '%' only
640: }
641: break;
642: case Flags:
643: if(strchr("-+ #0", c)!=0){
644: break;
645: }
646: // go to the next step
647: case Width:
648: if(c=='.'){
649: state=Precision;
650: break;
651: }
652: // go to the next step
653: case Precision:
654: if(c>='0' && c<='9'){
655: if(state == Flags) state=Width; // no more flags
656: break;
657: } else if(c=='d' || c=='i'){
658: result=FormatInt;
659: } else if(strchr("feEgG", c)!=0){
660: result=FormatDouble;
661: } else if(strchr("uoxX", c)!=0){
662: result=FormatUInt;
663: } else {
664: return FormatInvalid; // invalid char
665: }
666: state=Done;
667: break;
668: case Done:
669: return FormatInvalid; // no chars allowed after 'type'
670: }
671: }
672: return result;
673: }
674:
675:
1.154 paf 676: const char* format(double value, char* fmt) {
1.229 misha 677: char local_buf[MAX_NUMBER];
1.235 misha 678: int size=-1;
1.229 misha 679:
680: if(fmt && strlen(fmt)){
681: switch(format_type(fmt)){
682: case FormatDouble:
683: size=snprintf(local_buf, sizeof(local_buf), fmt, value);
684: break;
685: case FormatInt:
686: size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value);
687: break;
688: case FormatUInt:
1.126 paf 689: size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value);
1.229 misha 690: break;
691: case FormatInvalid:
692: throw Exception(PARSER_RUNTIME,
693: 0,
694: "Incorrect format string '%s' was specified.", fmt);
695: }
696: } else
697: size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value);
698:
699: if(size < 0 || size >= MAX_NUMBER-1){ // on win32 we manually reduce max size while printing
700: throw Exception(PARSER_RUNTIME,
701: 0,
702: "Error occure white executing snprintf with format string '%s'.", fmt);
703: }
704:
1.235 misha 705: return pa_strdup(local_buf, (size_t)size);
1.12 paf 706: }
707:
1.36 paf 708: size_t stdout_write(const void *buf, size_t size) {
1.12 paf 709: #ifdef WIN32
1.187 paf 710: size_t to_write = size;
1.12 paf 711: do{
1.154 paf 712: int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout);
1.12 paf 713: if(chunk_written<=0)
714: break;
715: size-=chunk_written;
1.36 paf 716: buf=((const char*)buf)+chunk_written;
1.126 paf 717: } while(size>0);
1.12 paf 718:
1.187 paf 719: return to_write-size;
1.12 paf 720: #else
1.126 paf 721: return fwrite(buf, 1, size, stdout);
1.12 paf 722: #endif
1.2 paf 723: }
1.14 paf 724:
1.229 misha 725: enum EscapeState {
726: EscapeRest,
727: EscapeFirst,
728: EscapeSecond,
729: EscapeUnicode
730: };
731:
1.236 misha 732: // @todo prescan for reduce required size (unescaped sting in 1 byte charset requires less memory usually)
1.258 misha 733: char* unescape_chars(const char* cp, int len, Charset* charset, bool js){
1.236 misha 734: char* s=new(PointerFreeGC) char[len+1]; // must be enough (%uXXXX==6 bytes, max utf-8 char length==6 bytes)
1.230 misha 735: char* dst=s;
1.229 misha 736: EscapeState escapeState=EscapeRest;
737: uint escapedValue=0;
738: int srcPos=0;
1.230 misha 739: short int jsCnt=0;
1.236 misha 740: while(srcPos<len){
1.229 misha 741: uchar c=(uchar)cp[srcPos];
1.258 misha 742: if(c=='%' || (c=='\\' && js)){
1.229 misha 743: escapeState=EscapeFirst;
744: } else {
745: switch(escapeState) {
746: case EscapeRest:
1.258 misha 747: if(!js && c=='+'){
1.230 misha 748: *dst++=' ';
1.229 misha 749: } else {
1.230 misha 750: *dst++=c;
1.229 misha 751: }
752: break;
753: case EscapeFirst:
1.232 misha 754: if(charset && c=='u'){
1.229 misha 755: // escaped unicode value: %u0430
756: jsCnt=0;
757: escapedValue=0;
758: escapeState=EscapeUnicode;
759: } else {
1.231 misha 760: if(isxdigit(c)){
1.229 misha 761: escapedValue=hex_value[c] << 4;
762: escapeState=EscapeSecond;
763: } else {
1.230 misha 764: *dst++=c;
1.229 misha 765: escapeState=EscapeRest;
766: }
767: }
768: break;
769: case EscapeSecond:
1.231 misha 770: if(isxdigit(c)){
1.229 misha 771: escapedValue+=hex_value[c];
1.230 misha 772: *dst++=(char)escapedValue;
1.229 misha 773: }
774: escapeState=EscapeRest;
775: break;
776: case EscapeUnicode:
1.231 misha 777: if(isxdigit(c)){
1.229 misha 778: escapedValue=(escapedValue << 4) + hex_value[c];
779: if(++jsCnt==4){
1.230 misha 780: // transcode utf8 char to client charset (we can lost some chars here)
1.232 misha 781: charset->store_Char((XMLByte*&)dst, (XMLCh)escapedValue, '?');
1.229 misha 782: escapeState=EscapeRest;
783: }
784: } else {
785: // not full unicode value
786: escapeState=EscapeRest;
787: }
788: break;
789: }
790: }
791:
792: srcPos++;
793: }
794:
1.230 misha 795: *dst=0; // zero-termination
1.229 misha 796: return s;
797: }
1.24 paf 798:
799: #ifdef WIN32
1.126 paf 800: void back_slashes_to_slashes(char* s) {
1.24 paf 801: if(s)
802: for(; *s; s++)
803: if(*s=='\\')
1.126 paf 804: *s='/';
1.24 paf 805: }
1.42 paf 806: /*
1.126 paf 807: void slashes_to_back_slashes(char* s) {
1.42 paf 808: if(s)
809: for(; *s; s++)
810: if(*s=='/')
1.126 paf 811: *s='\\';
1.42 paf 812: }
813: */
1.24 paf 814: #endif
1.41 paf 815:
1.231 misha 816: bool StrStartFromNC(const char* str, const char* substr, bool equal){
1.41 paf 817: while(true) {
1.231 misha 818: if(!(*substr)){
819: if(!(*str))
1.41 paf 820: return true;
821: else
1.231 misha 822: return !equal;
823: }
824: if(!(*str))
825: return false;
826: if(isalpha((unsigned char)*str)) {
827: if(tolower((unsigned char)*str)!=tolower((unsigned char)*substr))
1.41 paf 828: return false;
1.231 misha 829: } else if((*str) != (*substr))
1.41 paf 830: return false;
1.231 misha 831: str++;
832: substr++;
1.41 paf 833: }
1.57 parser 834: }
835:
1.232 misha 836: size_t strpos(const char *str, const char *substr) {
837: const char *p = strstr(str, substr);
838: return (p==0)?STRING_NOT_FOUND:p-str;
839: }
840:
841: // content-type: xxx; charset=WE-NEED-THIS
842: // content-type: xxx; charset="WE-NEED-THIS"
843: // content-type: xxx; charset="WE-NEED-THIS";
1.248 misha 844: Charset* detect_charset(const char* content_type){
1.233 misha 845: if(content_type){
1.245 misha 846: char* CONTENT_TYPE=pa_strdup(content_type);
1.248 misha 847:
848: for(char *p=CONTENT_TYPE; *p; p++)
849: *p=(char)toupper((unsigned char)*p);
1.233 misha 850:
851: if(const char* begin=strstr(CONTENT_TYPE, "CHARSET=")){
852: begin+=8; // skip "CHARSET="
853: char* end=0;
854: if(*begin && (*begin=='"' || *begin =='\'')){
855: char quote=*begin;
856: begin++;
857: end=(char*)strchr(begin, quote);
858: }
859: if(!end)
860: end=(char*)strchr(begin, ';');
861:
1.244 misha 862: if(end)
1.233 misha 863: *end=0; // terminator
864:
1.245 misha 865: return *begin?&charsets.get(begin):0;
1.232 misha 866: }
867: }
868: return 0;
869: }
870:
871:
1.84 paf 872: static bool isLeap(int year) {
1.229 misha 873: return !(
874: (year % 4) || ((year % 400) && !(year % 100))
875: );
1.57 parser 876: }
877:
878: int getMonthDays(int year, int month) {
1.220 misha 879: static int monthDays[]={
1.229 misha 880: 31,
881: 28,
882: 31,
883: 30,
884: 31,
885: 30,
886: 31,
887: 31,
888: 30,
889: 31,
890: 30,
891: 31
892: };
1.228 misha 893: return (month == 1 /* january -- 0 */ && isLeap(year)) ? 29 : monthDays[month];
1.41 paf 894: }
1.69 parser 895:
1.226 misha 896: int remove_crlf(char* start, char* end) {
897: char* from=start;
898: char* to=start;
899: bool skip=false;
900: while(from < end){
901: switch(*from){
902: case '\n':
903: case '\r':
904: case '\t':
905: case ' ':
906: if(!skip){
907: *to=' ';
908: to++;
909: skip=true;
910: }
911: break;
912: default:
913: if(from != to)
914: *to=*from;
915: to++;
916: skip=false;
1.69 parser 917: }
1.226 misha 918: from++;
919: }
920: return to-start;
1.91 paf 921: }
922:
923:
924: /// must be last in this file
925: #undef vsnprintf
1.126 paf 926: int __vsnprintf(char* b, size_t s, const char* f, va_list l) {
1.91 paf 927: if(!s)
928: return 0;
929:
930: int r;
931: // note: on win32& maybe somewhere else
932: // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing
933: --s;
1.172 paf 934:
935: // clients do not check for negative 's', feature: ignore such prints
936: if((ssize_t)s<0)
937: return 0;
938:
1.91 paf 939: #if _MSC_VER
940: /*
941: win32:
942: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm
943:
1.154 paf 944: if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 is returned
1.91 paf 945: */
1.126 paf 946: r=_vsnprintf(b, s, f, l);
1.91 paf 947: if(r<0)
948: r=s;
949: #else
1.126 paf 950: r=vsnprintf(b, s, f, l);
1.91 paf 951: /*
952: solaris:
953: man vsnprintf
954:
955: The snprintf() function returns the number of characters
956: formatted, that is, the number of characters that would have
957: been written to the buffer if it were large enough. If the
958: value of n is 0 on a call to snprintf(), an unspecified
959: value less than 1 is returned.
960: */
961:
962: if(r<0)
963: r=0;
1.167 paf 964: else if((size_t)r>s)
1.91 paf 965: r=s;
966: #endif
967: b[r]=0;
968: return r;
969: }
970:
1.126 paf 971: int __snprintf(char* b, size_t s, const char* f, ...) {
1.91 paf 972: va_list l;
1.241 misha 973: va_start(l, f);
974: int r=__vsnprintf(b, s, f, l);
975: va_end(l);
1.91 paf 976: return r;
1.178 paf 977: }
978:
979: /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */
980: /*
981: * Authors: Michael Zucchi <notzed@helixcode.com>
982: * Jeffrey Stedfast <fejj@helixcode.com>
983: *
984: * Copyright 2000 Helix Code, Inc. (www.helixcode.com)
985: *
986: * This program is free software; you can redistribute it and/or modify
987: * it under the terms of the GNU General Public License as published by
988: * the Free Software Foundation; either version 2 of the License, or
989: * (at your option) any later version.
990: *
991: * This program is distributed in the hope that it will be useful,
992: * but WITHOUT ANY WARRANTY; without even the implied warranty of
993: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
994: * GNU General Public License for more details.
995: *
996: * You should have received a copy of the GNU General Public License
997: * along with this program; if not, write to the Free Software
998: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
999: *
1000: */
1001: static char *base64_alphabet =
1002: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
1003:
1004: /**
1005: * g_mime_utils_base64_encode_step:
1006: * @in: input stream
1007: * @inlen: length of the input
1008: * @out: output string
1009: * @state: holds the number of bits that are stored in @save
1010: * @save: leftover bits that have not yet been encoded
1011: *
1012: * Base64 encodes a chunk of data. Performs an 'encode step', only
1013: * encodes blocks of 3 characters to the output at a time, saves
1014: * left-over state in state and save (initialise to 0 on first
1015: * invocation).
1016: *
1017: * Returns the number of bytes encoded.
1018: **/
1.252 misha 1019:
1020: #define BASE64_GROUPS_IN_LINE 19
1021:
1.178 paf 1022: static size_t
1023: g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1024: {
1.186 paf 1025: register const unsigned char *inptr;
1.178 paf 1026: register unsigned char *outptr;
1027:
1028: if (inlen <= 0)
1029: return 0;
1030:
1031: inptr = in;
1032: outptr = out;
1033:
1034: if (inlen + ((unsigned char *)save)[0] > 2) {
1035: const unsigned char *inend = in + inlen - 2;
1036: register int c1 = 0, c2 = 0, c3 = 0;
1037: register int already;
1038:
1039: already = *state;
1040:
1041: switch (((char *)save)[0]) {
1042: case 1: c1 = ((unsigned char *)save)[1]; goto skip1;
1043: case 2: c1 = ((unsigned char *)save)[1];
1044: c2 = ((unsigned char *)save)[2]; goto skip2;
1045: }
1046:
1047: /* yes, we jump into the loop, no i'm not going to change it, its beautiful! */
1048: while (inptr < inend) {
1049: c1 = *inptr++;
1050: skip1:
1051: c2 = *inptr++;
1052: skip2:
1053: c3 = *inptr++;
1054: *outptr++ = base64_alphabet [c1 >> 2];
1055: *outptr++ = base64_alphabet [(c2 >> 4) | ((c1 & 0x3) << 4)];
1056: *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)];
1057: *outptr++ = base64_alphabet [c3 & 0x3f];
1058: /* this is a bit ugly ... */
1.252 misha 1059: if ((++already) >= BASE64_GROUPS_IN_LINE) {
1.178 paf 1060: *outptr++ = '\n';
1061: already = 0;
1062: }
1063: }
1064:
1065: ((unsigned char *)save)[0] = 0;
1066: inlen = 2 - (inptr - inend);
1067: *state = already;
1068: }
1069:
1070: //d(printf ("state = %d, inlen = %d\n", (int)((char *)save)[0], inlen));
1071:
1072: if (inlen > 0) {
1073: register char *saveout;
1074:
1075: /* points to the slot for the next char to save */
1076: saveout = & (((char *)save)[1]) + ((char *)save)[0];
1077:
1078: /* inlen can only be 0 1 or 2 */
1079: switch (inlen) {
1080: case 2: *saveout++ = *inptr++;
1081: case 1: *saveout++ = *inptr++;
1082: }
1.216 paf 1083: *(char *)save = *(char *)save+(char)inlen;
1.178 paf 1084: }
1085:
1086: /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n",
1087: (int)((char *)save)[0],
1088: (int)((char *)save)[1],
1089: (int)((char *)save)[2]));*/
1090:
1091: return (outptr - out);
1092: }
1093:
1094: /**
1095: * g_mime_utils_base64_encode_close:
1096: * @in: input stream
1097: * @inlen: length of the input
1098: * @out: output string
1099: * @state: holds the number of bits that are stored in @save
1100: * @save: leftover bits that have not yet been encoded
1101: *
1102: * Base64 encodes the input stream to the output stream. Call this
1103: * when finished encoding data with g_mime_utils_base64_encode_step to
1104: * flush off the last little bit.
1105: *
1106: * Returns the number of bytes encoded.
1107: **/
1108: static size_t
1109: g_mime_utils_base64_encode_close (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1110: {
1111: unsigned char *outptr = out;
1112: int c1, c2;
1113:
1114: if (inlen > 0)
1115: outptr += g_mime_utils_base64_encode_step (in, inlen, outptr, state, save);
1116:
1117: c1 = ((unsigned char *)save)[1];
1118: c2 = ((unsigned char *)save)[2];
1119:
1120: switch (((unsigned char *)save)[0]) {
1121: case 2:
1122: outptr[2] = base64_alphabet [(c2 & 0x0f) << 2];
1123: goto skip;
1124: case 1:
1125: outptr[2] = '=';
1126: skip:
1127: outptr[0] = base64_alphabet [c1 >> 2];
1128: outptr[1] = base64_alphabet [c2 >> 4 | ((c1 & 0x3) << 4)];
1129: outptr[3] = '=';
1130: outptr += 4;
1131: break;
1132: }
1133:
1134: *outptr++ = 0;
1135:
1136: *save = 0;
1137: *state = 0;
1138:
1139: return (outptr - out);
1140: }
1141:
1.210 paf 1142: static unsigned char gmime_base64_rank[256] = {
1143: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1144: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1145: 255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63,
1146: 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255, 0,255,255,
1147: 255, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
1148: 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255,
1149: 255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
1150: 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255,
1151: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1152: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1153: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1154: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1155: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1156: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1157: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1158: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1159: };
1160:
1161: /**
1162: * g_mime_utils_base64_decode_step:
1163: * @in: input stream
1164: * @inlen: max length of data to decode
1165: * @out: output stream
1166: * @state: holds the number of bits that are stored in @save
1167: * @save: leftover bits that have not yet been decoded
1168: *
1169: * Decodes a chunk of base64 encoded data.
1170: *
1171: * Returns the number of bytes decoded (which have been dumped in @out).
1172: **/
1173: size_t
1174: g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1175: {
1.213 paf 1176: const unsigned char *inptr;
1177: unsigned char *outptr;
1.210 paf 1178: const unsigned char *inend;
1.213 paf 1179: int saved;
1.210 paf 1180: unsigned char c;
1181: int i;
1182:
1183: inend = in + inlen;
1184: outptr = out;
1185:
1186: /* convert 4 base64 bytes to 3 normal bytes */
1187: saved = *save;
1188: i = *state;
1189: inptr = in;
1190: while (inptr < inend) {
1191: c = gmime_base64_rank[*inptr++];
1192: if (c != 0xff) {
1193: saved = (saved << 6) | c;
1194: i++;
1195: if (i == 4) {
1.217 paf 1196: *outptr++ = (unsigned char)(saved >> 16);
1197: *outptr++ = (unsigned char)(saved >> 8);
1198: *outptr++ = (unsigned char)(saved);
1.210 paf 1199: i = 0;
1200: }
1201: }
1202: }
1203:
1204: *save = saved;
1205: *state = i;
1206:
1207: /* quick scan back for '=' on the end somewhere */
1208: /* fortunately we can drop 1 output char for each trailing = (upto 2) */
1209: i = 2;
1210: while (inptr > in && i) {
1211: inptr--;
1212: if (gmime_base64_rank[*inptr] != 0xff) {
1213: if (*inptr == '=' && outptr > out)
1214: outptr--;
1215: i--;
1216: }
1217: }
1218:
1219: /* if i != 0 then there is a truncation error! */
1220: return (outptr - out);
1221: }
1222:
1223:
1.239 misha 1224: char* pa_base64_encode(const char *in, size_t in_size){
1.252 misha 1225: size_t new_size = ((in_size / 3 + 1) * 4);
1226: new_size += new_size / (BASE64_GROUPS_IN_LINE * 4)/*new lines*/ + 1/*zero terminator*/;
1227: char* result = new(PointerFreeGC) char[new_size];
1.178 paf 1228: int state=0;
1229: int save=0;
1.183 paf 1230: #ifndef NDEBUG
1231: size_t filled=
1232: #endif
1.251 misha 1233: g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size, (unsigned char*)result, &state, &save);
1234:
1235: //throw Exception(PARSER_RUNTIME, 0, "%d %d %d", in_size, new_size, filled);
1236: assert(filled <= new_size);
1.178 paf 1237:
1238: return result;
1.98 paf 1239: }
1.210 paf 1240:
1.222 misha 1241:
1.239 misha 1242: char* pa_base64_encode(const String& file_spec){
1.222 misha 1243: unsigned char* base64=0;
1244: File_base64_action_info info={&base64};
1245:
1246: file_read_action_under_lock(file_spec,
1247: "pa_base64_encode", file_base64_file_action, &info);
1248:
1249: return (char*)base64;
1250: }
1251:
1252:
1253: static void file_base64_file_action(
1.229 misha 1254: struct stat& finfo,
1255: int f,
1256: const String&, const char* /*fname*/, bool,
1257: void *context) {
1.222 misha 1258:
1259: if(finfo.st_size) {
1260: File_base64_action_info& info=*static_cast<File_base64_action_info *>(context);
1261: *info.base64=new(PointerFreeGC) unsigned char[finfo.st_size * 2 + 6];
1262: unsigned char* base64 = *info.base64;
1263: int state=0;
1264: int save=0;
1265: int nCount;
1266: do {
1267: unsigned char buffer[FILE_BUFFER_SIZE];
1268: nCount = file_block_read(f, buffer, sizeof(buffer));
1269: if( nCount ){
1270: size_t filled=g_mime_utils_base64_encode_step ((const unsigned char*)buffer, nCount, base64, &state, &save);
1271: base64+=filled;
1272: }
1273: } while(nCount > 0);
1274: g_mime_utils_base64_encode_close (0, 0, base64, &state, &save);
1275: }
1276: }
1277:
1.239 misha 1278: void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size) {
1.210 paf 1279: /* wont go to more than had (overly conservative) */
1.211 paf 1280: result=new(PointerFreeGC) char[in_size+1/*terminator*/];
1.210 paf 1281: int state=0;
1282: int save=0;
1283: result_size=
1284: g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size,
1285: (unsigned char*)result, &state, &save);
1286: assert(result_size <= in_size);
1.211 paf 1287: result[result_size]=0; // for text files
1.210 paf 1288: }
1.218 misha 1289:
1290:
1.221 misha 1291: int file_block_read(const int f, unsigned char* buffer, const size_t size){
1292: int nCount = read(f, buffer, size);
1293: if (nCount < 0)
1.238 misha 1294: throw Exception("file.read",
1.221 misha 1295: 0,
1296: "read failed: %s (%d)", strerror(errno), errno);
1297: return nCount;
1298: }
1299:
1.239 misha 1300: const unsigned long pa_crc32(const char *in, size_t in_size){
1.218 misha 1301: unsigned long crc32=0xFFFFFFFF;
1.220 misha 1302:
1.240 misha 1303: InitCrc32Table();
1.239 misha 1304: for(size_t i = 0; i<in_size; i++)
1305: CalcCrc32(in[i], crc32);
1.220 misha 1306:
1.218 misha 1307: return ~crc32;
1308: }
1309:
1.239 misha 1310: const unsigned long pa_crc32(const String& file_spec){
1.218 misha 1311: unsigned long crc32=0xFFFFFFFF;
1312: file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32);
1313: return ~crc32;
1314: }
1315:
1316: static void file_crc32_file_action(
1.229 misha 1317: struct stat& finfo,
1318: int f,
1319: const String&, const char* /*fname*/, bool,
1.239 misha 1320: void *context) {
1.218 misha 1321: unsigned long& crc32=*static_cast<unsigned long *>(context);
1322: if(finfo.st_size) {
1323: InitCrc32Table();
1.220 misha 1324: int nCount=0;
1.218 misha 1325: do {
1.221 misha 1326: unsigned char buffer[FILE_BUFFER_SIZE];
1327: nCount = file_block_read(f, buffer, sizeof(buffer));
1.220 misha 1328: for(int i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32);
1329: } while(nCount > 0);
1.218 misha 1330: }
1331: }
1332:
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