Annotation of parser3/src/main/pa_common.C, revision 1.236
1.15 paf 1: /** @file
1.16 paf 2: Parser: commonly functions.
3:
1.205 paf 4: Copyright(c) 2001-2005 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:
1.236 ! misha 29: static const char * const IDENT_COMMON_C="$Date: 2008-06-16 12:46:00 $";
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"
1.98 paf 38:
1.93 paf 39: // some maybe-undefined constants
40:
1.82 paf 41: #ifndef _O_TEXT
42: # define _O_TEXT 0
43: #endif
44: #ifndef _O_BINARY
45: # define _O_BINARY 0
1.47 paf 46: #endif
1.80 paf 47:
1.138 paf 48: #ifdef HAVE_FTRUNCATE
49: # define PA_O_TRUNC 0
50: #else
51: # ifdef _O_TRUNC
52: # define PA_O_TRUNC _O_TRUNC
53: # else
54: # error you must have either ftruncate function or _O_TRUNC bit declared
55: # endif
1.154 paf 56: #endif
1.176 paf 57:
1.154 paf 58: // defines for globals
59:
60: #define FILE_STATUS_NAME "status"
61:
62: // globals
63:
64: const String file_status_name(FILE_STATUS_NAME);
65:
66: // functions
1.127 paf 67:
1.154 paf 68: void fix_line_breaks(char *str, size_t& length) {
1.87 paf 69: //_asm int 3;
1.154 paf 70: const char* const eob=str+length;
71: char* dest=str;
1.72 parser 72: // fix DOS: \r\n -> \n
73: // fix Macintosh: \r -> \n
1.154 paf 74: char* bol=str;
1.137 paf 75: while(char* eol=(char*)memchr(bol, '\r', eob -bol)) {
1.72 parser 76: size_t len=eol-bol;
77: if(dest!=bol)
1.126 paf 78: memcpy(dest, bol, len);
1.72 parser 79: dest+=len;
1.126 paf 80: *dest++='\n';
1.72 parser 81:
1.126 paf 82: if(&eol[1]<eob && eol[1]=='\n') { // \r, \n = DOS
1.72 parser 83: bol=eol+2;
1.154 paf 84: length--;
1.126 paf 85: } else // \r, not \n = Macintosh
1.72 parser 86: bol=eol+1;
87: }
1.154 paf 88: // last piece without \r
1.72 parser 89: if(dest!=bol)
1.126 paf 90: memcpy(dest, bol, eob-bol);
1.154 paf 91: str[length]=0; // terminating
1.72 parser 92: }
1.18 paf 93:
1.154 paf 94: char* file_read_text(Request_charsets& charsets,
1.229 misha 95: const String& file_spec,
96: bool fail_on_read_problem,
1.234 misha 97: HashStringValue* params,
98: bool transcode_result) {
1.154 paf 99: File_read_result file=
1.234 misha 100: file_read(charsets, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result);
1.154 paf 101: return file.success?file.str:0;
1.126 paf 102: }
103:
1.206 paf 104: /// these options were handled but not checked elsewhere, now check them
1.214 paf 105: int pa_get_valid_file_options_count(HashStringValue& options)
1.206 paf 106: {
107: int result=0;
108: if(options.get(PA_SQL_LIMIT_NAME))
109: result++;
110: if(options.get(PA_SQL_OFFSET_NAME))
111: result++;
112: if(options.get(PA_COLUMN_SEPARATOR_NAME))
113: result++;
114: if(options.get(PA_COLUMN_ENCLOSER_NAME))
115: result++;
1.223 misha 116: if(options.get(PA_CHARSET_NAME))
117: result++;
1.206 paf 118: return result;
119: }
120:
1.123 paf 121: #ifndef DOXYGEN
122: struct File_read_action_info {
1.154 paf 123: char **data; size_t *data_size;
1.188 paf 124: char* buf; size_t offset; size_t count;
1.126 paf 125: };
1.123 paf 126: #endif
1.154 paf 127: static void file_read_action(
1.229 misha 128: struct stat& finfo,
129: int f,
130: const String& file_spec, const char* /*fname*/, bool as_text,
131: void *context) {
1.126 paf 132: File_read_action_info& info=*static_cast<File_read_action_info *>(context);
1.188 paf 133: size_t to_read_size=info.count;
134: if(!to_read_size)
135: to_read_size=(size_t)finfo.st_size;
136: assert( !(info.buf && as_text) );
137: if(to_read_size) {
138: if(info.offset)
139: lseek(f, info.offset, SEEK_SET);
140: *info.data=info.buf
141: ? info.buf
142: : new(PointerFreeGC) char[to_read_size+(as_text?1:0)];
1.126 paf 143: *info.data_size=(size_t)read(f, *info.data, to_read_size);
1.123 paf 144:
145: if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size)
1.126 paf 146: throw Exception(0,
1.123 paf 147: &file_spec,
1.173 paf 148: "read failed: actually read %u bytes count not in [0..%u] valid range",
1.126 paf 149: *info.data_size, to_read_size);
1.123 paf 150: } else { // empty file
1.209 paf 151: // for both, text and binary: for text we need that terminator, for binary we need nonzero pointer to be able to save such files
152: *info.data=new(PointerFreeGC) char[1];
153: *(char*)(*info.data)=0;
1.123 paf 154: *info.data_size=0;
155: return;
156: }
1.126 paf 157: }
1.154 paf 158: File_read_result file_read(Request_charsets& charsets, const String& file_spec,
1.229 misha 159: bool as_text, HashStringValue *params,
160: bool fail_on_read_problem,
1.234 misha 161: char* buf, size_t offset, size_t count, bool transcode_text_result) {
1.167 paf 162: File_read_result result={false, 0, 0, 0};
1.154 paf 163: if(file_spec.starts_with("http://")) {
1.203 paf 164: if(offset || count)
1.224 misha 165: throw Exception(PARSER_RUNTIME,
1.203 paf 166: 0,
167: "offset and load options are not supported for HTTP:// file load");
168:
1.126 paf 169: // fail on read problem
1.234 misha 170: File_read_http_result http=pa_internal_file_read_http(charsets, file_spec, as_text, params, transcode_text_result);
1.154 paf 171: result.success=true;
172: result.str=http.str;
173: result.length=http.length;
174: result.headers=http.headers;
1.126 paf 175: } else {
1.236 ! misha 176: if(params){
1.214 paf 177: int valid_options=pa_get_valid_file_options_count(*params);
1.206 paf 178: if(valid_options!=params->count())
1.224 misha 179: throw Exception(PARSER_RUNTIME,
1.206 paf 180: 0,
181: "invalid option passed");
182: }
1.161 paf 183:
1.188 paf 184: File_read_action_info info={&result.str, &result.length,
185: buf, offset, count};
1.154 paf 186: result.success=file_read_action_under_lock(file_spec,
1.126 paf 187: "read", file_read_action, &info,
188: as_text, fail_on_read_problem);
1.223 misha 189:
1.236 ! misha 190: if(as_text && result.success){
! 191: if(result.length>=3 && strncmp(result.str, "\xEF\xBB\xBF", 3)==0){
! 192: // skip UTF-8 signature: EF BB BF (BOM code)
! 193: result.str+=3;
! 194: result.length-=3;
! 195: }
! 196:
! 197: 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
! 198: if( Value* vcharset_name=params->get(PA_CHARSET_NAME) ){
! 199: Charset asked_charset=::charsets.get(vcharset_name->as_string().
! 200: change_case(charsets.source(), String::CC_UPPER));
! 201:
! 202: String::C body=String::C(result.str, result.length);
! 203: body=Charset::transcode(body, asked_charset, charsets.source());
1.123 paf 204:
1.236 ! misha 205: result.str=const_cast<char*>(body.str); // hacking a little
! 206: result.length=body.length;
! 207: }
1.131 paf 208: }
209: }
1.123 paf 210: }
1.236 ! misha 211:
! 212: if(as_text && result.length)
! 213: fix_line_breaks(result.str, result.length);
1.126 paf 214:
215: return result;
1.123 paf 216: }
217:
1.154 paf 218: #ifdef PA_SAFE_MODE
219: void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) {
220: if(finfo.st_uid/*foreign?*/!=geteuid()
221: && finfo.st_gid/*foreign?*/!=getegid())
1.224 misha 222: throw Exception(PARSER_RUNTIME,
1.154 paf 223: &file_spec,
224: "parser is in safe mode: "
225: "reading files of foreign group and user disabled "
226: "[recompile parser with --disable-safe-mode configure option], "
227: "actual filename '%s', "
228: "fuid(%d)!=euid(%d) or fgid(%d)!=egid(%d)",
229: fname,
230: finfo.st_uid, geteuid(),
231: finfo.st_gid, getegid());
232: }
233: #endif
1.149 paf 234:
1.154 paf 235: bool file_read_action_under_lock(const String& file_spec,
1.126 paf 236: const char* action_name, File_read_action action, void *context,
237: bool as_text,
1.123 paf 238: bool fail_on_read_problem) {
1.154 paf 239: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.33 paf 240: int f;
241:
242: // first open, next stat:
1.45 paf 243: // directory update of NTFS hard links performed on open.
1.33 paf 244: // ex:
245: // a.html:^test[] and b.html hardlink to a.html
246: // user inserts ! before ^test in a.html
1.126 paf 247: // directory entry of b.html in NTFS not updated at once,
1.35 paf 248: // they delay update till open, so we would receive "!^test[" string
249: // if would do stat, next open.
1.123 paf 250: // later: it seems, even this does not help sometimes
1.229 misha 251: if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) {
1.123 paf 252: try {
1.162 paf 253: if(pa_lock_shared_blocking(f)!=0)
1.126 paf 254: throw Exception("file.lock",
1.123 paf 255: &file_spec,
256: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 257: strerror(errno), errno, fname);
1.123 paf 258:
1.124 paf 259: struct stat finfo;
260: if(stat(fname, &finfo)!=0)
261: throw Exception("file.missing", // hardly possible: we just opened it OK
262: &file_spec,
263: "stat failed: %s (%d), actual filename '%s'",
1.154 paf 264: strerror(errno), errno, fname);
1.124 paf 265:
1.140 paf 266: #ifdef PA_SAFE_MODE
1.149 paf 267: check_safe_mode(finfo, file_spec, fname);
1.105 paf 268: #endif
1.32 paf 269:
1.154 paf 270: action(finfo, f, file_spec, fname, as_text, context);
1.123 paf 271: } catch(...) {
1.162 paf 272: pa_unlock(f);close(f);
1.123 paf 273: if(fail_on_read_problem)
1.154 paf 274: rethrow;
1.123 paf 275: return false;
276: }
1.87 paf 277:
1.162 paf 278: pa_unlock(f);close(f);
1.72 parser 279: return true;
1.229 misha 280: } else {
1.118 paf 281: if(fail_on_read_problem)
1.126 paf 282: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.118 paf 283: &file_spec,
1.123 paf 284: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 285: action_name, strerror(errno), errno, fname);
1.118 paf 286: return false;
287: }
1.8 paf 288: }
289:
1.202 paf 290: void create_dir_for_file(const String& file_spec) {
1.63 parser 291: size_t pos_after=1;
1.154 paf 292: size_t pos_before;
293: while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) {
294: mkdir(file_spec.mid(0, pos_before).cstr(String::L_FILE_SPEC), 0775);
1.63 parser 295: pos_after=pos_before+1;
296: }
297: }
298:
1.98 paf 299: bool file_write_action_under_lock(
1.28 paf 300: const String& file_spec,
1.225 misha 301: const char* action_name,
302: File_write_action action,
303: void *context,
1.126 paf 304: bool as_text,
305: bool do_append,
306: bool do_block,
1.110 paf 307: bool fail_on_lock_problem) {
1.154 paf 308: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.28 paf 309: int f;
1.80 paf 310: if(access(fname, W_OK)!=0) // no
1.126 paf 311: create_dir_for_file(file_spec);
1.50 paf 312:
1.80 paf 313: if((f=open(fname,
314: O_CREAT|O_RDWR
315: |(as_text?_O_TEXT:_O_BINARY)
1.138 paf 316: |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) {
1.162 paf 317: if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) {
1.126 paf 318: Exception e("file.lock",
1.110 paf 319: &file_spec,
320: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 321: strerror(errno), errno, fname);
1.126 paf 322: close(f);
1.110 paf 323: if(fail_on_lock_problem)
324: throw e;
1.98 paf 325: return false;
326: }
1.96 paf 327:
1.158 paf 328: try {
1.126 paf 329: action(f, context);
1.158 paf 330: } catch(...) {
1.138 paf 331: #ifdef HAVE_FTRUNCATE
1.104 paf 332: if(!do_append)
1.125 paf 333: ftruncate(f, lseek(f, 0, SEEK_CUR)); // one can not use O_TRUNC, read lower
1.138 paf 334: #endif
1.162 paf 335: pa_unlock(f);close(f);
1.154 paf 336: rethrow;
1.158 paf 337: }
1.80 paf 338:
1.138 paf 339: #ifdef HAVE_FTRUNCATE
1.104 paf 340: if(!do_append)
1.125 paf 341: 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 342: #endif
1.162 paf 343: pa_unlock(f);close(f);
1.98 paf 344: return true;
1.80 paf 345: } else
1.126 paf 346: throw Exception(errno==EACCES?"file.access":0,
1.80 paf 347: &file_spec,
1.96 paf 348: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 349: action_name, strerror(errno), errno, fname);
1.96 paf 350: // here should be nothing, see rethrow above
351: }
352:
353: #ifndef DOXYGEN
354: struct File_write_action_info {
1.154 paf 355: const char* str; size_t length;
1.126 paf 356: };
1.96 paf 357: #endif
358: static void file_write_action(int f, void *context) {
1.126 paf 359: File_write_action_info& info=*static_cast<File_write_action_info *>(context);
1.154 paf 360: if(info.length) {
361: int written=write(f, info.str, info.length);
1.116 paf 362: if(written<0)
1.126 paf 363: throw Exception(0,
364: 0,
365: "write failed: %s (%d)", strerror(errno), errno);
1.113 paf 366: }
1.96 paf 367: }
368: void file_write(
369: const String& file_spec,
1.154 paf 370: const char* data, size_t size,
1.126 paf 371: bool as_text,
1.96 paf 372: bool do_append) {
1.126 paf 373: File_write_action_info info={data, size};
1.225 misha 374:
1.98 paf 375: file_write_action_under_lock(
1.154 paf 376: file_spec,
1.225 misha 377: "write",
378: file_write_action,
379: &info,
1.154 paf 380: as_text,
381: do_append);
1.30 paf 382: }
383:
1.63 parser 384: // throws nothing! [this is required in file_move & file_delete]
1.50 paf 385: static void rmdir(const String& file_spec, size_t pos_after) {
1.154 paf 386: size_t pos_before;
387: if((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND)
1.126 paf 388: rmdir(file_spec, pos_before+1);
1.50 paf 389:
1.154 paf 390: rmdir(file_spec.mid(0, pos_after-1/* / */).cstr(String::L_FILE_SPEC));
1.50 paf 391: }
1.164 paf 392: bool file_delete(const String& file_spec, bool fail_on_problem) {
1.154 paf 393: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.54 parser 394: if(unlink(fname)!=0)
1.164 paf 395: if(fail_on_problem)
1.126 paf 396: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.93 paf 397: &file_spec,
398: "unlink failed: %s (%d), actual filename '%s'",
1.154 paf 399: strerror(errno), errno, fname);
1.93 paf 400: else
401: return false;
1.50 paf 402:
1.126 paf 403: rmdir(file_spec, 1);
1.93 paf 404: return true;
1.60 parser 405: }
1.95 paf 406: void file_move(const String& old_spec, const String& new_spec) {
1.154 paf 407: const char* old_spec_cstr=old_spec.cstr(String::L_FILE_SPEC);
408: const char* new_spec_cstr=new_spec.cstr(String::L_FILE_SPEC);
1.63 parser 409:
1.126 paf 410: create_dir_for_file(new_spec);
1.63 parser 411:
1.60 parser 412: if(rename(old_spec_cstr, new_spec_cstr)!=0)
1.126 paf 413: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.60 parser 414: &old_spec,
415: "rename failed: %s (%d), actual filename '%s' to '%s'",
1.154 paf 416: strerror(errno), errno, old_spec_cstr, new_spec_cstr);
1.63 parser 417:
1.126 paf 418: rmdir(old_spec, 1);
1.31 paf 419: }
420:
1.51 paf 421:
1.126 paf 422: bool entry_exists(const char* fname, struct stat *afinfo) {
1.118 paf 423: struct stat lfinfo;
424: bool result=stat(fname, &lfinfo)==0;
425: if(afinfo)
426: *afinfo=lfinfo;
427: return result;
1.119 paf 428: }
429:
430: bool entry_exists(const String& file_spec) {
1.154 paf 431: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.126 paf 432: return entry_exists(fname, 0);
1.118 paf 433: }
434:
1.51 paf 435: static bool entry_readable(const String& file_spec, bool need_dir) {
1.154 paf 436: char* fname=file_spec.cstrm(String::L_FILE_SPEC);
1.120 paf 437: if(need_dir) {
1.126 paf 438: size_t size=strlen(fname);
1.120 paf 439: while(size) {
1.126 paf 440: char c=fname[size-1];
1.120 paf 441: if(c=='/' || c=='\\')
442: fname[--size]=0;
443: else
444: break;
445: }
446: }
1.51 paf 447: struct stat finfo;
1.118 paf 448: if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) {
1.109 paf 449: bool is_dir=(finfo.st_mode&S_IFDIR) != 0;
1.51 paf 450: return is_dir==need_dir;
451: }
452: return false;
453: }
1.215 paf 454: bool file_exist(const String& file_spec) {
1.126 paf 455: return entry_readable(file_spec, false);
1.51 paf 456: }
1.215 paf 457: bool dir_exists(const String& file_spec) {
1.126 paf 458: return entry_readable(file_spec, true);
1.65 parser 459: }
1.215 paf 460: const String* file_exist(const String& path, const String& name) {
1.154 paf 461: String& result=*new String(path);
462: result << "/";
463: result << name;
1.215 paf 464: return file_exist(result)?&result:0;
1.43 paf 465: }
466: bool file_executable(const String& file_spec) {
1.154 paf 467: return access(file_spec.cstr(String::L_FILE_SPEC), X_OK)==0;
1.44 paf 468: }
469:
1.64 parser 470: bool file_stat(const String& file_spec,
1.229 misha 471: size_t& rsize,
472: time_t& ratime,
473: time_t& rmtime,
474: time_t& rctime,
475: bool fail_on_read_problem) {
1.154 paf 476: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
477: struct stat finfo;
1.44 paf 478: if(stat(fname, &finfo)!=0)
1.64 parser 479: if(fail_on_read_problem)
1.126 paf 480: throw Exception("file.missing",
1.67 parser 481: &file_spec,
482: "getting file size failed: %s (%d), real filename '%s'",
1.154 paf 483: strerror(errno), errno, fname);
1.64 parser 484: else
485: return false;
1.58 parser 486: rsize=finfo.st_size;
487: ratime=finfo.st_atime;
488: rmtime=finfo.st_mtime;
489: rctime=finfo.st_ctime;
1.64 parser 490: return true;
1.18 paf 491: }
492:
1.126 paf 493: char* getrow(char* *row_ref, char delim) {
1.229 misha 494: char* result=*row_ref;
495: if(result) {
1.126 paf 496: *row_ref=strchr(result, delim);
1.8 paf 497: if(*row_ref)
498: *((*row_ref)++)=0;
499: else if(!*result)
500: return 0;
1.229 misha 501: }
502: return result;
1.8 paf 503: }
504:
1.126 paf 505: char* lsplit(char* string, char delim) {
1.229 misha 506: if(string) {
1.126 paf 507: char* v=strchr(string, delim);
1.8 paf 508: if(v) {
509: *v=0;
510: return v+1;
511: }
1.229 misha 512: }
513: return 0;
1.8 paf 514: }
515:
1.126 paf 516: char* lsplit(char* *string_ref, char delim) {
1.229 misha 517: char* result=*string_ref;
1.126 paf 518: char* next=lsplit(*string_ref, delim);
1.229 misha 519: *string_ref=next;
520: return result;
1.9 paf 521: }
522:
1.126 paf 523: char* rsplit(char* string, char delim) {
1.229 misha 524: if(string) {
1.126 paf 525: char* v=strrchr(string, delim);
1.18 paf 526: if(v) {
1.9 paf 527: *v=0;
528: return v+1;
529: }
1.229 misha 530: }
531: return NULL;
1.10 paf 532: }
533:
1.229 misha 534:
535: // format: %[flags][width][.precision]type http://msdn.microsoft.com/ru-ru/library/56e442dc(en-us,VS.80).aspx
536: // flags: '-', '+', ' ', '#', '0' http://msdn.microsoft.com/ru-ru/library/8aky45ct(en-us,VS.80).aspx
537: // width, precision: non negative decimal number
538: enum FormatType {
539: FormatInvalid,
540: FormatInt,
541: FormatUInt,
542: FormatDouble
543: };
544: FormatType format_type(char* fmt){
545: enum FormatState {
546: Percent,
547: Flags,
548: Width,
549: Precision,
550: Done
551: } state=Percent;
552:
553: FormatType result=FormatInvalid;
554:
555: char* pos=fmt;
556: while(char c=*(pos++)){
557: switch(state){
558: case Percent:
559: if(c=='%'){
560: state=Flags;
561: } else {
562: return FormatInvalid; // 1st char must be '%' only
563: }
564: break;
565: case Flags:
566: if(strchr("-+ #0", c)!=0){
567: break;
568: }
569: // go to the next step
570: case Width:
571: if(c=='.'){
572: state=Precision;
573: break;
574: }
575: // go to the next step
576: case Precision:
577: if(c>='0' && c<='9'){
578: if(state == Flags) state=Width; // no more flags
579: break;
580: } else if(c=='d' || c=='i'){
581: result=FormatInt;
582: } else if(strchr("feEgG", c)!=0){
583: result=FormatDouble;
584: } else if(strchr("uoxX", c)!=0){
585: result=FormatUInt;
586: } else {
587: return FormatInvalid; // invalid char
588: }
589: state=Done;
590: break;
591: case Done:
592: return FormatInvalid; // no chars allowed after 'type'
593: }
594: }
595: return result;
596: }
597:
598:
1.154 paf 599: const char* format(double value, char* fmt) {
1.229 misha 600: char local_buf[MAX_NUMBER];
1.235 misha 601: int size=-1;
1.229 misha 602:
603: if(fmt && strlen(fmt)){
604: switch(format_type(fmt)){
605: case FormatDouble:
606: size=snprintf(local_buf, sizeof(local_buf), fmt, value);
607: break;
608: case FormatInt:
609: size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value);
610: break;
611: case FormatUInt:
1.126 paf 612: size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value);
1.229 misha 613: break;
614: case FormatInvalid:
615: throw Exception(PARSER_RUNTIME,
616: 0,
617: "Incorrect format string '%s' was specified.", fmt);
618: }
619: } else
620: size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value);
621:
622: if(size < 0 || size >= MAX_NUMBER-1){ // on win32 we manually reduce max size while printing
623: throw Exception(PARSER_RUNTIME,
624: 0,
625: "Error occure white executing snprintf with format string '%s'.", fmt);
626: }
627:
1.235 misha 628: return pa_strdup(local_buf, (size_t)size);
1.12 paf 629: }
630:
1.36 paf 631: size_t stdout_write(const void *buf, size_t size) {
1.12 paf 632: #ifdef WIN32
1.187 paf 633: size_t to_write = size;
1.12 paf 634: do{
1.154 paf 635: int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout);
1.12 paf 636: if(chunk_written<=0)
637: break;
638: size-=chunk_written;
1.36 paf 639: buf=((const char*)buf)+chunk_written;
1.126 paf 640: } while(size>0);
1.12 paf 641:
1.187 paf 642: return to_write-size;
1.12 paf 643: #else
1.126 paf 644: return fwrite(buf, 1, size, stdout);
1.12 paf 645: #endif
1.2 paf 646: }
1.14 paf 647:
1.229 misha 648: enum EscapeState {
649: EscapeRest,
650: EscapeFirst,
651: EscapeSecond,
652: EscapeUnicode
653: };
654:
1.236 ! misha 655: // @todo prescan for reduce required size (unescaped sting in 1 byte charset requires less memory usually)
! 656: char* unescape_chars(const char* cp, int len, Charset* charset, bool ignore_plus){
! 657: char* s=new(PointerFreeGC) char[len+1]; // must be enough (%uXXXX==6 bytes, max utf-8 char length==6 bytes)
1.230 misha 658: char* dst=s;
1.229 misha 659: EscapeState escapeState=EscapeRest;
660: uint escapedValue=0;
661: int srcPos=0;
1.230 misha 662: short int jsCnt=0;
1.236 ! misha 663: while(srcPos<len){
1.229 misha 664: uchar c=(uchar)cp[srcPos];
665: if(c=='%'){
666: escapeState=EscapeFirst;
667: } else {
668: switch(escapeState) {
669: case EscapeRest:
1.236 ! misha 670: if(!ignore_plus && c=='+'){
1.230 misha 671: *dst++=' ';
1.229 misha 672: } else {
1.230 misha 673: *dst++=c;
1.229 misha 674: }
675: break;
676: case EscapeFirst:
1.232 misha 677: if(charset && c=='u'){
1.229 misha 678: // escaped unicode value: %u0430
679: jsCnt=0;
680: escapedValue=0;
681: escapeState=EscapeUnicode;
682: } else {
1.231 misha 683: if(isxdigit(c)){
1.229 misha 684: escapedValue=hex_value[c] << 4;
685: escapeState=EscapeSecond;
686: } else {
1.230 misha 687: *dst++=c;
1.229 misha 688: escapeState=EscapeRest;
689: }
690: }
691: break;
692: case EscapeSecond:
1.231 misha 693: if(isxdigit(c)){
1.229 misha 694: escapedValue+=hex_value[c];
1.230 misha 695: *dst++=(char)escapedValue;
1.229 misha 696: }
697: escapeState=EscapeRest;
698: break;
699: case EscapeUnicode:
1.231 misha 700: if(isxdigit(c)){
1.229 misha 701: escapedValue=(escapedValue << 4) + hex_value[c];
702: if(++jsCnt==4){
1.230 misha 703: // transcode utf8 char to client charset (we can lost some chars here)
1.232 misha 704: charset->store_Char((XMLByte*&)dst, (XMLCh)escapedValue, '?');
1.229 misha 705: escapeState=EscapeRest;
706: }
707: } else {
708: // not full unicode value
709: escapeState=EscapeRest;
710: }
711: break;
712: }
713: }
714:
715: srcPos++;
716: }
717:
1.230 misha 718: *dst=0; // zero-termination
1.229 misha 719: return s;
720: }
1.24 paf 721:
722: #ifdef WIN32
1.126 paf 723: void back_slashes_to_slashes(char* s) {
1.24 paf 724: if(s)
725: for(; *s; s++)
726: if(*s=='\\')
1.126 paf 727: *s='/';
1.24 paf 728: }
1.42 paf 729: /*
1.126 paf 730: void slashes_to_back_slashes(char* s) {
1.42 paf 731: if(s)
732: for(; *s; s++)
733: if(*s=='/')
1.126 paf 734: *s='\\';
1.42 paf 735: }
736: */
1.24 paf 737: #endif
1.41 paf 738:
1.231 misha 739: bool StrStartFromNC(const char* str, const char* substr, bool equal){
1.41 paf 740: while(true) {
1.231 misha 741: if(!(*substr)){
742: if(!(*str))
1.41 paf 743: return true;
744: else
1.231 misha 745: return !equal;
746: }
747: if(!(*str))
748: return false;
749: if(isalpha((unsigned char)*str)) {
750: if(tolower((unsigned char)*str)!=tolower((unsigned char)*substr))
1.41 paf 751: return false;
1.231 misha 752: } else if((*str) != (*substr))
1.41 paf 753: return false;
1.231 misha 754: str++;
755: substr++;
1.41 paf 756: }
1.57 parser 757: }
758:
1.232 misha 759: size_t strpos(const char *str, const char *substr) {
760: const char *p = strstr(str, substr);
761: return (p==0)?STRING_NOT_FOUND:p-str;
762: }
763:
764: // content-type: xxx; charset=WE-NEED-THIS
765: // content-type: xxx; charset="WE-NEED-THIS"
766: // content-type: xxx; charset="WE-NEED-THIS";
1.233 misha 767: Charset* detect_charset(const char* content_type){
768: if(content_type){
769: size_t len=strlen(content_type);
770: char* CONTENT_TYPE=new(PointerFreeGC) char[len+1];
771: memcpy(CONTENT_TYPE, content_type, len);
772: for(char *p=CONTENT_TYPE; *p; p++)
773: *p=(char)toupper((unsigned char)*p);
774:
775: if(const char* begin=strstr(CONTENT_TYPE, "CHARSET=")){
776: begin+=8; // skip "CHARSET="
777: char* end=0;
778: if(*begin && (*begin=='"' || *begin =='\'')){
779: char quote=*begin;
780: begin++;
781: end=(char*)strchr(begin, quote);
782: }
783: if(!end)
784: end=(char*)strchr(begin, ';');
785:
786: size_t len;
787: if(end){
788: *end=0; // terminator
789: len=end-begin;
790: } else {
791: len=strlen(begin);
792: }
793:
794: String::Body NAME=String::Body(begin, len);
795: return &charsets.get(NAME);
1.232 misha 796: }
797: }
798: return 0;
799: }
800:
1.233 misha 801: Charset* detect_charset(Charset& source_charset, const String& content_type){
802: const String& CONTENT_TYPE=content_type.change_case(source_charset, String::CC_UPPER);
803: size_t begin=CONTENT_TYPE.pos("CHARSET=");
804: if(begin!=STRING_NOT_FOUND) {
805: begin+=8; // skip "CHARSET="
806: size_t end=STRING_NOT_FOUND;
807:
808: if(CONTENT_TYPE.pos('"', begin)==begin){
809: begin++;
810: end=CONTENT_TYPE.pos('"', begin);
811: } else if(CONTENT_TYPE.pos('\'', begin)==begin){
812: begin++;
813: end=CONTENT_TYPE.pos('\'', begin);
814: }
815:
816: if(end==STRING_NOT_FOUND)
817: end=CONTENT_TYPE.pos(';', begin);
818:
819: if(end==STRING_NOT_FOUND)
820: end=CONTENT_TYPE.length();
1.232 misha 821:
1.233 misha 822: const String::Body NAME=CONTENT_TYPE.mid(begin, end);
823: return &charsets.get(NAME);
824: }
825: return 0;
1.232 misha 826: }
827:
828:
1.84 paf 829: static bool isLeap(int year) {
1.229 misha 830: return !(
831: (year % 4) || ((year % 400) && !(year % 100))
832: );
1.57 parser 833: }
834:
835: int getMonthDays(int year, int month) {
1.220 misha 836: static int monthDays[]={
1.229 misha 837: 31,
838: 28,
839: 31,
840: 30,
841: 31,
842: 30,
843: 31,
844: 31,
845: 30,
846: 31,
847: 30,
848: 31
849: };
1.228 misha 850: return (month == 1 /* january -- 0 */ && isLeap(year)) ? 29 : monthDays[month];
1.41 paf 851: }
1.69 parser 852:
1.226 misha 853: int remove_crlf(char* start, char* end) {
854: char* from=start;
855: char* to=start;
856: bool skip=false;
857: while(from < end){
858: switch(*from){
859: case '\n':
860: case '\r':
861: case '\t':
862: case ' ':
863: if(!skip){
864: *to=' ';
865: to++;
866: skip=true;
867: }
868: break;
869: default:
870: if(from != to)
871: *to=*from;
872: to++;
873: skip=false;
1.69 parser 874: }
1.226 misha 875: from++;
876: }
877: return to-start;
1.91 paf 878: }
879:
880:
881: /// must be last in this file
882: #undef vsnprintf
1.126 paf 883: int __vsnprintf(char* b, size_t s, const char* f, va_list l) {
1.91 paf 884: if(!s)
885: return 0;
886:
887: int r;
888: // note: on win32& maybe somewhere else
889: // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing
890: --s;
1.172 paf 891:
892: // clients do not check for negative 's', feature: ignore such prints
893: if((ssize_t)s<0)
894: return 0;
895:
1.91 paf 896: #if _MSC_VER
897: /*
898: win32:
899: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm
900:
1.154 paf 901: if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 is returned
1.91 paf 902: */
1.126 paf 903: r=_vsnprintf(b, s, f, l);
1.91 paf 904: if(r<0)
905: r=s;
906: #else
1.126 paf 907: r=vsnprintf(b, s, f, l);
1.91 paf 908: /*
909: solaris:
910: man vsnprintf
911:
912: The snprintf() function returns the number of characters
913: formatted, that is, the number of characters that would have
914: been written to the buffer if it were large enough. If the
915: value of n is 0 on a call to snprintf(), an unspecified
916: value less than 1 is returned.
917: */
918:
919: if(r<0)
920: r=0;
1.167 paf 921: else if((size_t)r>s)
1.91 paf 922: r=s;
923: #endif
924: b[r]=0;
925: return r;
926: }
927:
1.126 paf 928: int __snprintf(char* b, size_t s, const char* f, ...) {
1.91 paf 929: va_list l;
1.126 paf 930: va_start(l, f);
931: int r=__vsnprintf(b, s, f, l);
932: va_end(l);
1.91 paf 933: return r;
1.178 paf 934: }
935:
936: /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */
937: /*
938: * Authors: Michael Zucchi <notzed@helixcode.com>
939: * Jeffrey Stedfast <fejj@helixcode.com>
940: *
941: * Copyright 2000 Helix Code, Inc. (www.helixcode.com)
942: *
943: * This program is free software; you can redistribute it and/or modify
944: * it under the terms of the GNU General Public License as published by
945: * the Free Software Foundation; either version 2 of the License, or
946: * (at your option) any later version.
947: *
948: * This program is distributed in the hope that it will be useful,
949: * but WITHOUT ANY WARRANTY; without even the implied warranty of
950: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
951: * GNU General Public License for more details.
952: *
953: * You should have received a copy of the GNU General Public License
954: * along with this program; if not, write to the Free Software
955: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
956: *
957: */
958: static char *base64_alphabet =
959: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
960:
961: /**
962: * g_mime_utils_base64_encode_step:
963: * @in: input stream
964: * @inlen: length of the input
965: * @out: output string
966: * @state: holds the number of bits that are stored in @save
967: * @save: leftover bits that have not yet been encoded
968: *
969: * Base64 encodes a chunk of data. Performs an 'encode step', only
970: * encodes blocks of 3 characters to the output at a time, saves
971: * left-over state in state and save (initialise to 0 on first
972: * invocation).
973: *
974: * Returns the number of bytes encoded.
975: **/
976: static size_t
977: g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
978: {
1.186 paf 979: register const unsigned char *inptr;
1.178 paf 980: register unsigned char *outptr;
981:
982: if (inlen <= 0)
983: return 0;
984:
985: inptr = in;
986: outptr = out;
987:
988: if (inlen + ((unsigned char *)save)[0] > 2) {
989: const unsigned char *inend = in + inlen - 2;
990: register int c1 = 0, c2 = 0, c3 = 0;
991: register int already;
992:
993: already = *state;
994:
995: switch (((char *)save)[0]) {
996: case 1: c1 = ((unsigned char *)save)[1]; goto skip1;
997: case 2: c1 = ((unsigned char *)save)[1];
998: c2 = ((unsigned char *)save)[2]; goto skip2;
999: }
1000:
1001: /* yes, we jump into the loop, no i'm not going to change it, its beautiful! */
1002: while (inptr < inend) {
1003: c1 = *inptr++;
1004: skip1:
1005: c2 = *inptr++;
1006: skip2:
1007: c3 = *inptr++;
1008: *outptr++ = base64_alphabet [c1 >> 2];
1009: *outptr++ = base64_alphabet [(c2 >> 4) | ((c1 & 0x3) << 4)];
1010: *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)];
1011: *outptr++ = base64_alphabet [c3 & 0x3f];
1012: /* this is a bit ugly ... */
1013: if ((++already) >= 19) {
1014: *outptr++ = '\n';
1015: already = 0;
1016: }
1017: }
1018:
1019: ((unsigned char *)save)[0] = 0;
1020: inlen = 2 - (inptr - inend);
1021: *state = already;
1022: }
1023:
1024: //d(printf ("state = %d, inlen = %d\n", (int)((char *)save)[0], inlen));
1025:
1026: if (inlen > 0) {
1027: register char *saveout;
1028:
1029: /* points to the slot for the next char to save */
1030: saveout = & (((char *)save)[1]) + ((char *)save)[0];
1031:
1032: /* inlen can only be 0 1 or 2 */
1033: switch (inlen) {
1034: case 2: *saveout++ = *inptr++;
1035: case 1: *saveout++ = *inptr++;
1036: }
1.216 paf 1037: *(char *)save = *(char *)save+(char)inlen;
1.178 paf 1038: }
1039:
1040: /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n",
1041: (int)((char *)save)[0],
1042: (int)((char *)save)[1],
1043: (int)((char *)save)[2]));*/
1044:
1045: return (outptr - out);
1046: }
1047:
1048: /**
1049: * g_mime_utils_base64_encode_close:
1050: * @in: input stream
1051: * @inlen: length of the input
1052: * @out: output string
1053: * @state: holds the number of bits that are stored in @save
1054: * @save: leftover bits that have not yet been encoded
1055: *
1056: * Base64 encodes the input stream to the output stream. Call this
1057: * when finished encoding data with g_mime_utils_base64_encode_step to
1058: * flush off the last little bit.
1059: *
1060: * Returns the number of bytes encoded.
1061: **/
1062: static size_t
1063: g_mime_utils_base64_encode_close (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1064: {
1065: unsigned char *outptr = out;
1066: int c1, c2;
1067:
1068: if (inlen > 0)
1069: outptr += g_mime_utils_base64_encode_step (in, inlen, outptr, state, save);
1070:
1071: c1 = ((unsigned char *)save)[1];
1072: c2 = ((unsigned char *)save)[2];
1073:
1074: switch (((unsigned char *)save)[0]) {
1075: case 2:
1076: outptr[2] = base64_alphabet [(c2 & 0x0f) << 2];
1077: goto skip;
1078: case 1:
1079: outptr[2] = '=';
1080: skip:
1081: outptr[0] = base64_alphabet [c1 >> 2];
1082: outptr[1] = base64_alphabet [c2 >> 4 | ((c1 & 0x3) << 4)];
1083: outptr[3] = '=';
1084: outptr += 4;
1085: break;
1086: }
1087:
1088: *outptr++ = 0;
1089:
1090: *save = 0;
1091: *state = 0;
1092:
1093: return (outptr - out);
1094: }
1095:
1.210 paf 1096: static unsigned char gmime_base64_rank[256] = {
1097: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1098: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1099: 255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63,
1100: 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255, 0,255,255,
1101: 255, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
1102: 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255,
1103: 255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
1104: 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255,
1105: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1106: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1107: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1108: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1109: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1110: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1111: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1112: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1113: };
1114:
1115: /**
1116: * g_mime_utils_base64_decode_step:
1117: * @in: input stream
1118: * @inlen: max length of data to decode
1119: * @out: output stream
1120: * @state: holds the number of bits that are stored in @save
1121: * @save: leftover bits that have not yet been decoded
1122: *
1123: * Decodes a chunk of base64 encoded data.
1124: *
1125: * Returns the number of bytes decoded (which have been dumped in @out).
1126: **/
1127: size_t
1128: g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1129: {
1.213 paf 1130: const unsigned char *inptr;
1131: unsigned char *outptr;
1.210 paf 1132: const unsigned char *inend;
1.213 paf 1133: int saved;
1.210 paf 1134: unsigned char c;
1135: int i;
1136:
1137: inend = in + inlen;
1138: outptr = out;
1139:
1140: /* convert 4 base64 bytes to 3 normal bytes */
1141: saved = *save;
1142: i = *state;
1143: inptr = in;
1144: while (inptr < inend) {
1145: c = gmime_base64_rank[*inptr++];
1146: if (c != 0xff) {
1147: saved = (saved << 6) | c;
1148: i++;
1149: if (i == 4) {
1.217 paf 1150: *outptr++ = (unsigned char)(saved >> 16);
1151: *outptr++ = (unsigned char)(saved >> 8);
1152: *outptr++ = (unsigned char)(saved);
1.210 paf 1153: i = 0;
1154: }
1155: }
1156: }
1157:
1158: *save = saved;
1159: *state = i;
1160:
1161: /* quick scan back for '=' on the end somewhere */
1162: /* fortunately we can drop 1 output char for each trailing = (upto 2) */
1163: i = 2;
1164: while (inptr > in && i) {
1165: inptr--;
1166: if (gmime_base64_rank[*inptr] != 0xff) {
1167: if (*inptr == '=' && outptr > out)
1168: outptr--;
1169: i--;
1170: }
1171: }
1172:
1173: /* if i != 0 then there is a truncation error! */
1174: return (outptr - out);
1175: }
1176:
1177:
1178: char* pa_base64_encode(const char *in, size_t in_size)
1.178 paf 1179: {
1180: /* wont go to more than 2x size (overly conservative) */
1.210 paf 1181: char* result=new(PointerFreeGC) char[in_size * 2 + 6];
1.178 paf 1182: int state=0;
1183: int save=0;
1.183 paf 1184: #ifndef NDEBUG
1185: size_t filled=
1186: #endif
1.210 paf 1187: g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size,
1.178 paf 1188: (unsigned char*)result, &state, &save);
1.210 paf 1189: assert(filled <= in_size * 2 + 6);
1.178 paf 1190:
1191: return result;
1.98 paf 1192: }
1.210 paf 1193:
1.222 misha 1194:
1195: char* pa_base64_encode(const String& file_spec)
1196: {
1197: unsigned char* base64=0;
1198: File_base64_action_info info={&base64};
1199:
1200: file_read_action_under_lock(file_spec,
1201: "pa_base64_encode", file_base64_file_action, &info);
1202:
1203: return (char*)base64;
1204: }
1205:
1206:
1207: static void file_base64_file_action(
1.229 misha 1208: struct stat& finfo,
1209: int f,
1210: const String&, const char* /*fname*/, bool,
1211: void *context) {
1.222 misha 1212:
1213: if(finfo.st_size) {
1214: File_base64_action_info& info=*static_cast<File_base64_action_info *>(context);
1215: *info.base64=new(PointerFreeGC) unsigned char[finfo.st_size * 2 + 6];
1216: unsigned char* base64 = *info.base64;
1217: int state=0;
1218: int save=0;
1219: int nCount;
1220: do {
1221: unsigned char buffer[FILE_BUFFER_SIZE];
1222: nCount = file_block_read(f, buffer, sizeof(buffer));
1223: if( nCount ){
1224: size_t filled=g_mime_utils_base64_encode_step ((const unsigned char*)buffer, nCount, base64, &state, &save);
1225: base64+=filled;
1226: }
1227: } while(nCount > 0);
1228: g_mime_utils_base64_encode_close (0, 0, base64, &state, &save);
1229: }
1230: }
1231:
1.211 paf 1232: void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size)
1.210 paf 1233: {
1234: /* wont go to more than had (overly conservative) */
1.211 paf 1235: result=new(PointerFreeGC) char[in_size+1/*terminator*/];
1.210 paf 1236: int state=0;
1237: int save=0;
1238: result_size=
1239: g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size,
1240: (unsigned char*)result, &state, &save);
1241: assert(result_size <= in_size);
1.211 paf 1242: result[result_size]=0; // for text files
1.210 paf 1243: }
1.218 misha 1244:
1245:
1.221 misha 1246: int file_block_read(const int f, unsigned char* buffer, const size_t size){
1247: int nCount = read(f, buffer, size);
1248: if (nCount < 0)
1249: throw Exception(0,
1250: 0,
1251: "read failed: %s (%d)", strerror(errno), errno);
1252: return nCount;
1253: }
1254:
1.218 misha 1255: const unsigned long pa_crc32(const char *in, size_t in_size)
1256: {
1257: unsigned long crc32=0xFFFFFFFF;
1.220 misha 1258:
1.218 misha 1259: InitCrc32Table();
1260: for(size_t i = 0; i < in_size; i++) CalcCrc32(in[i], crc32);
1.220 misha 1261:
1.218 misha 1262: return ~crc32;
1263: }
1264:
1265: const unsigned long pa_crc32(const String& file_spec)
1266: {
1267: unsigned long crc32=0xFFFFFFFF;
1268: file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32);
1269: return ~crc32;
1270: }
1271:
1272: static void file_crc32_file_action(
1.229 misha 1273: struct stat& finfo,
1274: int f,
1275: const String&, const char* /*fname*/, bool,
1276: void *context)
1.218 misha 1277: {
1278: unsigned long& crc32=*static_cast<unsigned long *>(context);
1279: if(finfo.st_size) {
1280: InitCrc32Table();
1.220 misha 1281: int nCount=0;
1.218 misha 1282: do {
1.221 misha 1283: unsigned char buffer[FILE_BUFFER_SIZE];
1284: nCount = file_block_read(f, buffer, sizeof(buffer));
1.220 misha 1285: for(int i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32);
1286: } while(nCount > 0);
1.218 misha 1287: }
1288: }
1289:
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