Annotation of parser3/src/main/pa_common.C, revision 1.238
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:
1.238 ! misha 29: static const char * const IDENT_COMMON_C="$Date: 2008-08-21 15:57:48 $";
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"
1.98 paf 39:
1.93 paf 40: // some maybe-undefined constants
41:
1.82 paf 42: #ifndef _O_TEXT
43: # define _O_TEXT 0
44: #endif
45: #ifndef _O_BINARY
46: # define _O_BINARY 0
1.47 paf 47: #endif
1.80 paf 48:
1.138 paf 49: #ifdef HAVE_FTRUNCATE
50: # define PA_O_TRUNC 0
51: #else
52: # ifdef _O_TRUNC
53: # define PA_O_TRUNC _O_TRUNC
54: # else
55: # error you must have either ftruncate function or _O_TRUNC bit declared
56: # endif
1.154 paf 57: #endif
1.176 paf 58:
1.154 paf 59: // defines for globals
60:
61: #define FILE_STATUS_NAME "status"
62:
63: // globals
64:
65: const String file_status_name(FILE_STATUS_NAME);
66:
67: // functions
1.127 paf 68:
1.154 paf 69: void fix_line_breaks(char *str, size_t& length) {
1.87 paf 70: //_asm int 3;
1.154 paf 71: const char* const eob=str+length;
72: char* dest=str;
1.72 parser 73: // fix DOS: \r\n -> \n
74: // fix Macintosh: \r -> \n
1.154 paf 75: char* bol=str;
1.137 paf 76: while(char* eol=(char*)memchr(bol, '\r', eob -bol)) {
1.72 parser 77: size_t len=eol-bol;
78: if(dest!=bol)
1.126 paf 79: memcpy(dest, bol, len);
1.72 parser 80: dest+=len;
1.126 paf 81: *dest++='\n';
1.72 parser 82:
1.126 paf 83: if(&eol[1]<eob && eol[1]=='\n') { // \r, \n = DOS
1.72 parser 84: bol=eol+2;
1.154 paf 85: length--;
1.126 paf 86: } else // \r, not \n = Macintosh
1.72 parser 87: bol=eol+1;
88: }
1.154 paf 89: // last piece without \r
1.72 parser 90: if(dest!=bol)
1.126 paf 91: memcpy(dest, bol, eob-bol);
1.154 paf 92: str[length]=0; // terminating
1.72 parser 93: }
1.18 paf 94:
1.154 paf 95: char* file_read_text(Request_charsets& charsets,
1.229 misha 96: const String& file_spec,
97: bool fail_on_read_problem,
1.234 misha 98: HashStringValue* params,
99: bool transcode_result) {
1.154 paf 100: File_read_result file=
1.234 misha 101: file_read(charsets, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result);
1.154 paf 102: return file.success?file.str:0;
1.126 paf 103: }
104:
1.206 paf 105: /// these options were handled but not checked elsewhere, now check them
1.214 paf 106: int pa_get_valid_file_options_count(HashStringValue& options)
1.206 paf 107: {
108: int result=0;
109: if(options.get(PA_SQL_LIMIT_NAME))
110: result++;
111: if(options.get(PA_SQL_OFFSET_NAME))
112: result++;
113: if(options.get(PA_COLUMN_SEPARATOR_NAME))
114: result++;
115: if(options.get(PA_COLUMN_ENCLOSER_NAME))
116: result++;
1.223 misha 117: if(options.get(PA_CHARSET_NAME))
118: result++;
1.206 paf 119: return result;
120: }
121:
1.123 paf 122: #ifndef DOXYGEN
123: struct File_read_action_info {
1.154 paf 124: char **data; size_t *data_size;
1.188 paf 125: char* buf; size_t offset; size_t count;
1.126 paf 126: };
1.123 paf 127: #endif
1.154 paf 128: static void file_read_action(
1.229 misha 129: struct stat& finfo,
130: int f,
131: const String& file_spec, const char* /*fname*/, bool as_text,
132: void *context) {
1.126 paf 133: File_read_action_info& info=*static_cast<File_read_action_info *>(context);
1.188 paf 134: size_t to_read_size=info.count;
135: if(!to_read_size)
136: to_read_size=(size_t)finfo.st_size;
137: assert( !(info.buf && as_text) );
138: if(to_read_size) {
139: if(info.offset)
140: lseek(f, info.offset, SEEK_SET);
141: *info.data=info.buf
142: ? info.buf
143: : new(PointerFreeGC) char[to_read_size+(as_text?1:0)];
1.126 paf 144: *info.data_size=(size_t)read(f, *info.data, to_read_size);
1.123 paf 145:
146: if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size)
1.238 ! misha 147: throw Exception("file.read",
1.123 paf 148: &file_spec,
1.173 paf 149: "read failed: actually read %u bytes count not in [0..%u] valid range",
1.126 paf 150: *info.data_size, to_read_size);
1.123 paf 151: } else { // empty file
1.209 paf 152: // for both, text and binary: for text we need that terminator, for binary we need nonzero pointer to be able to save such files
153: *info.data=new(PointerFreeGC) char[1];
154: *(char*)(*info.data)=0;
1.123 paf 155: *info.data_size=0;
156: return;
157: }
1.126 paf 158: }
1.154 paf 159: File_read_result file_read(Request_charsets& charsets, const String& file_spec,
1.229 misha 160: bool as_text, HashStringValue *params,
161: bool fail_on_read_problem,
1.234 misha 162: char* buf, size_t offset, size_t count, bool transcode_text_result) {
1.167 paf 163: File_read_result result={false, 0, 0, 0};
1.154 paf 164: if(file_spec.starts_with("http://")) {
1.203 paf 165: if(offset || count)
1.224 misha 166: throw Exception(PARSER_RUNTIME,
1.203 paf 167: 0,
168: "offset and load options are not supported for HTTP:// file load");
169:
1.126 paf 170: // fail on read problem
1.234 misha 171: File_read_http_result http=pa_internal_file_read_http(charsets, file_spec, as_text, params, transcode_text_result);
1.154 paf 172: result.success=true;
173: result.str=http.str;
174: result.length=http.length;
175: result.headers=http.headers;
1.126 paf 176: } else {
1.236 misha 177: if(params){
1.214 paf 178: int valid_options=pa_get_valid_file_options_count(*params);
1.206 paf 179: if(valid_options!=params->count())
1.224 misha 180: throw Exception(PARSER_RUNTIME,
1.206 paf 181: 0,
182: "invalid option passed");
183: }
1.161 paf 184:
1.188 paf 185: File_read_action_info info={&result.str, &result.length,
186: buf, offset, count};
1.154 paf 187: result.success=file_read_action_under_lock(file_spec,
1.126 paf 188: "read", file_read_action, &info,
189: as_text, fail_on_read_problem);
1.223 misha 190:
1.236 misha 191: if(as_text && result.success){
192: if(result.length>=3 && strncmp(result.str, "\xEF\xBB\xBF", 3)==0){
193: // skip UTF-8 signature: EF BB BF (BOM code)
194: result.str+=3;
195: result.length-=3;
196: }
197:
198: 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
199: if( Value* vcharset_name=params->get(PA_CHARSET_NAME) ){
200: Charset asked_charset=::charsets.get(vcharset_name->as_string().
201: change_case(charsets.source(), String::CC_UPPER));
202:
203: String::C body=String::C(result.str, result.length);
204: body=Charset::transcode(body, asked_charset, charsets.source());
1.123 paf 205:
1.236 misha 206: result.str=const_cast<char*>(body.str); // hacking a little
207: result.length=body.length;
208: }
1.131 paf 209: }
210: }
1.123 paf 211: }
1.236 misha 212:
213: if(as_text && result.length)
214: fix_line_breaks(result.str, result.length);
1.126 paf 215:
216: return result;
1.123 paf 217: }
218:
1.154 paf 219: #ifdef PA_SAFE_MODE
220: void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) {
221: if(finfo.st_uid/*foreign?*/!=geteuid()
222: && finfo.st_gid/*foreign?*/!=getegid())
1.224 misha 223: throw Exception(PARSER_RUNTIME,
1.154 paf 224: &file_spec,
225: "parser is in safe mode: "
226: "reading files of foreign group and user disabled "
227: "[recompile parser with --disable-safe-mode configure option], "
228: "actual filename '%s', "
229: "fuid(%d)!=euid(%d) or fgid(%d)!=egid(%d)",
230: fname,
231: finfo.st_uid, geteuid(),
232: finfo.st_gid, getegid());
233: }
234: #endif
1.149 paf 235:
1.154 paf 236: bool file_read_action_under_lock(const String& file_spec,
1.126 paf 237: const char* action_name, File_read_action action, void *context,
238: bool as_text,
1.123 paf 239: bool fail_on_read_problem) {
1.154 paf 240: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.33 paf 241: int f;
242:
243: // first open, next stat:
1.45 paf 244: // directory update of NTFS hard links performed on open.
1.33 paf 245: // ex:
246: // a.html:^test[] and b.html hardlink to a.html
247: // user inserts ! before ^test in a.html
1.126 paf 248: // directory entry of b.html in NTFS not updated at once,
1.35 paf 249: // they delay update till open, so we would receive "!^test[" string
250: // if would do stat, next open.
1.123 paf 251: // later: it seems, even this does not help sometimes
1.229 misha 252: if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) {
1.123 paf 253: try {
1.162 paf 254: if(pa_lock_shared_blocking(f)!=0)
1.126 paf 255: throw Exception("file.lock",
1.123 paf 256: &file_spec,
257: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 258: strerror(errno), errno, fname);
1.123 paf 259:
1.124 paf 260: struct stat finfo;
261: if(stat(fname, &finfo)!=0)
262: throw Exception("file.missing", // hardly possible: we just opened it OK
263: &file_spec,
264: "stat failed: %s (%d), actual filename '%s'",
1.154 paf 265: strerror(errno), errno, fname);
1.124 paf 266:
1.140 paf 267: #ifdef PA_SAFE_MODE
1.149 paf 268: check_safe_mode(finfo, file_spec, fname);
1.105 paf 269: #endif
1.32 paf 270:
1.154 paf 271: action(finfo, f, file_spec, fname, as_text, context);
1.123 paf 272: } catch(...) {
1.162 paf 273: pa_unlock(f);close(f);
1.123 paf 274: if(fail_on_read_problem)
1.154 paf 275: rethrow;
1.123 paf 276: return false;
277: }
1.87 paf 278:
1.162 paf 279: pa_unlock(f);close(f);
1.72 parser 280: return true;
1.229 misha 281: } else {
1.118 paf 282: if(fail_on_read_problem)
1.126 paf 283: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.118 paf 284: &file_spec,
1.123 paf 285: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 286: action_name, strerror(errno), errno, fname);
1.118 paf 287: return false;
288: }
1.8 paf 289: }
290:
1.202 paf 291: void create_dir_for_file(const String& file_spec) {
1.63 parser 292: size_t pos_after=1;
1.154 paf 293: size_t pos_before;
294: while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) {
295: mkdir(file_spec.mid(0, pos_before).cstr(String::L_FILE_SPEC), 0775);
1.63 parser 296: pos_after=pos_before+1;
297: }
298: }
299:
1.98 paf 300: bool file_write_action_under_lock(
1.28 paf 301: const String& file_spec,
1.225 misha 302: const char* action_name,
303: File_write_action action,
304: void *context,
1.126 paf 305: bool as_text,
306: bool do_append,
307: bool do_block,
1.110 paf 308: bool fail_on_lock_problem) {
1.154 paf 309: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.28 paf 310: int f;
1.80 paf 311: if(access(fname, W_OK)!=0) // no
1.126 paf 312: create_dir_for_file(file_spec);
1.50 paf 313:
1.80 paf 314: if((f=open(fname,
315: O_CREAT|O_RDWR
316: |(as_text?_O_TEXT:_O_BINARY)
1.138 paf 317: |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) {
1.162 paf 318: if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) {
1.126 paf 319: Exception e("file.lock",
1.110 paf 320: &file_spec,
321: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 322: strerror(errno), errno, fname);
1.126 paf 323: close(f);
1.110 paf 324: if(fail_on_lock_problem)
325: throw e;
1.98 paf 326: return false;
327: }
1.96 paf 328:
1.158 paf 329: try {
1.126 paf 330: action(f, context);
1.158 paf 331: } catch(...) {
1.138 paf 332: #ifdef HAVE_FTRUNCATE
1.104 paf 333: if(!do_append)
1.125 paf 334: ftruncate(f, lseek(f, 0, SEEK_CUR)); // one can not use O_TRUNC, read lower
1.138 paf 335: #endif
1.162 paf 336: pa_unlock(f);close(f);
1.154 paf 337: rethrow;
1.158 paf 338: }
1.80 paf 339:
1.138 paf 340: #ifdef HAVE_FTRUNCATE
1.104 paf 341: if(!do_append)
1.125 paf 342: 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 343: #endif
1.162 paf 344: pa_unlock(f);close(f);
1.98 paf 345: return true;
1.80 paf 346: } else
1.126 paf 347: throw Exception(errno==EACCES?"file.access":0,
1.80 paf 348: &file_spec,
1.96 paf 349: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 350: action_name, strerror(errno), errno, fname);
1.96 paf 351: // here should be nothing, see rethrow above
352: }
353:
354: #ifndef DOXYGEN
355: struct File_write_action_info {
1.154 paf 356: const char* str; size_t length;
1.126 paf 357: };
1.96 paf 358: #endif
359: static void file_write_action(int f, void *context) {
1.126 paf 360: File_write_action_info& info=*static_cast<File_write_action_info *>(context);
1.154 paf 361: if(info.length) {
362: int written=write(f, info.str, info.length);
1.116 paf 363: if(written<0)
1.238 ! misha 364: throw Exception("file.access",
1.126 paf 365: 0,
366: "write failed: %s (%d)", strerror(errno), errno);
1.113 paf 367: }
1.96 paf 368: }
369: void file_write(
370: const String& file_spec,
1.154 paf 371: const char* data, size_t size,
1.126 paf 372: bool as_text,
1.96 paf 373: bool do_append) {
1.126 paf 374: File_write_action_info info={data, size};
1.225 misha 375:
1.98 paf 376: file_write_action_under_lock(
1.154 paf 377: file_spec,
1.225 misha 378: "write",
379: file_write_action,
380: &info,
1.154 paf 381: as_text,
382: do_append);
1.30 paf 383: }
384:
1.63 parser 385: // throws nothing! [this is required in file_move & file_delete]
1.50 paf 386: static void rmdir(const String& file_spec, size_t pos_after) {
1.154 paf 387: size_t pos_before;
388: if((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND)
1.126 paf 389: rmdir(file_spec, pos_before+1);
1.50 paf 390:
1.154 paf 391: rmdir(file_spec.mid(0, pos_after-1/* / */).cstr(String::L_FILE_SPEC));
1.50 paf 392: }
1.164 paf 393: bool file_delete(const String& file_spec, bool fail_on_problem) {
1.154 paf 394: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.54 parser 395: if(unlink(fname)!=0)
1.164 paf 396: if(fail_on_problem)
1.126 paf 397: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.93 paf 398: &file_spec,
399: "unlink failed: %s (%d), actual filename '%s'",
1.154 paf 400: strerror(errno), errno, fname);
1.93 paf 401: else
402: return false;
1.50 paf 403:
1.126 paf 404: rmdir(file_spec, 1);
1.93 paf 405: return true;
1.60 parser 406: }
1.95 paf 407: void file_move(const String& old_spec, const String& new_spec) {
1.154 paf 408: const char* old_spec_cstr=old_spec.cstr(String::L_FILE_SPEC);
409: const char* new_spec_cstr=new_spec.cstr(String::L_FILE_SPEC);
1.63 parser 410:
1.126 paf 411: create_dir_for_file(new_spec);
1.63 parser 412:
1.60 parser 413: if(rename(old_spec_cstr, new_spec_cstr)!=0)
1.126 paf 414: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.60 parser 415: &old_spec,
416: "rename failed: %s (%d), actual filename '%s' to '%s'",
1.154 paf 417: strerror(errno), errno, old_spec_cstr, new_spec_cstr);
1.63 parser 418:
1.126 paf 419: rmdir(old_spec, 1);
1.31 paf 420: }
421:
1.51 paf 422:
1.126 paf 423: bool entry_exists(const char* fname, struct stat *afinfo) {
1.118 paf 424: struct stat lfinfo;
425: bool result=stat(fname, &lfinfo)==0;
426: if(afinfo)
427: *afinfo=lfinfo;
428: return result;
1.119 paf 429: }
430:
431: bool entry_exists(const String& file_spec) {
1.154 paf 432: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.126 paf 433: return entry_exists(fname, 0);
1.118 paf 434: }
435:
1.51 paf 436: static bool entry_readable(const String& file_spec, bool need_dir) {
1.154 paf 437: char* fname=file_spec.cstrm(String::L_FILE_SPEC);
1.120 paf 438: if(need_dir) {
1.126 paf 439: size_t size=strlen(fname);
1.120 paf 440: while(size) {
1.126 paf 441: char c=fname[size-1];
1.120 paf 442: if(c=='/' || c=='\\')
443: fname[--size]=0;
444: else
445: break;
446: }
447: }
1.51 paf 448: struct stat finfo;
1.118 paf 449: if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) {
1.109 paf 450: bool is_dir=(finfo.st_mode&S_IFDIR) != 0;
1.51 paf 451: return is_dir==need_dir;
452: }
453: return false;
454: }
1.215 paf 455: bool file_exist(const String& file_spec) {
1.126 paf 456: return entry_readable(file_spec, false);
1.51 paf 457: }
1.215 paf 458: bool dir_exists(const String& file_spec) {
1.126 paf 459: return entry_readable(file_spec, true);
1.65 parser 460: }
1.215 paf 461: const String* file_exist(const String& path, const String& name) {
1.154 paf 462: String& result=*new String(path);
463: result << "/";
464: result << name;
1.215 paf 465: return file_exist(result)?&result:0;
1.43 paf 466: }
467: bool file_executable(const String& file_spec) {
1.154 paf 468: return access(file_spec.cstr(String::L_FILE_SPEC), X_OK)==0;
1.44 paf 469: }
470:
1.64 parser 471: bool file_stat(const String& file_spec,
1.229 misha 472: size_t& rsize,
473: time_t& ratime,
474: time_t& rmtime,
475: time_t& rctime,
476: bool fail_on_read_problem) {
1.154 paf 477: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
478: struct stat finfo;
1.44 paf 479: if(stat(fname, &finfo)!=0)
1.64 parser 480: if(fail_on_read_problem)
1.126 paf 481: throw Exception("file.missing",
1.67 parser 482: &file_spec,
483: "getting file size failed: %s (%d), real filename '%s'",
1.154 paf 484: strerror(errno), errno, fname);
1.64 parser 485: else
486: return false;
1.58 parser 487: rsize=finfo.st_size;
488: ratime=finfo.st_atime;
489: rmtime=finfo.st_mtime;
490: rctime=finfo.st_ctime;
1.64 parser 491: return true;
1.18 paf 492: }
493:
1.126 paf 494: char* getrow(char* *row_ref, char delim) {
1.229 misha 495: char* result=*row_ref;
496: if(result) {
1.126 paf 497: *row_ref=strchr(result, delim);
1.8 paf 498: if(*row_ref)
499: *((*row_ref)++)=0;
500: else if(!*result)
501: return 0;
1.229 misha 502: }
503: return result;
1.8 paf 504: }
505:
1.126 paf 506: char* lsplit(char* string, char delim) {
1.229 misha 507: if(string) {
1.126 paf 508: char* v=strchr(string, delim);
1.8 paf 509: if(v) {
510: *v=0;
511: return v+1;
512: }
1.229 misha 513: }
514: return 0;
1.8 paf 515: }
516:
1.126 paf 517: char* lsplit(char* *string_ref, char delim) {
1.229 misha 518: char* result=*string_ref;
1.126 paf 519: char* next=lsplit(*string_ref, delim);
1.229 misha 520: *string_ref=next;
521: return result;
1.9 paf 522: }
523:
1.126 paf 524: char* rsplit(char* string, char delim) {
1.229 misha 525: if(string) {
1.126 paf 526: char* v=strrchr(string, delim);
1.18 paf 527: if(v) {
1.9 paf 528: *v=0;
529: return v+1;
530: }
1.229 misha 531: }
532: return NULL;
1.10 paf 533: }
534:
1.229 misha 535:
536: // format: %[flags][width][.precision]type http://msdn.microsoft.com/ru-ru/library/56e442dc(en-us,VS.80).aspx
537: // flags: '-', '+', ' ', '#', '0' http://msdn.microsoft.com/ru-ru/library/8aky45ct(en-us,VS.80).aspx
538: // width, precision: non negative decimal number
539: enum FormatType {
540: FormatInvalid,
541: FormatInt,
542: FormatUInt,
543: FormatDouble
544: };
545: FormatType format_type(char* fmt){
546: enum FormatState {
547: Percent,
548: Flags,
549: Width,
550: Precision,
551: Done
552: } state=Percent;
553:
554: FormatType result=FormatInvalid;
555:
556: char* pos=fmt;
557: while(char c=*(pos++)){
558: switch(state){
559: case Percent:
560: if(c=='%'){
561: state=Flags;
562: } else {
563: return FormatInvalid; // 1st char must be '%' only
564: }
565: break;
566: case Flags:
567: if(strchr("-+ #0", c)!=0){
568: break;
569: }
570: // go to the next step
571: case Width:
572: if(c=='.'){
573: state=Precision;
574: break;
575: }
576: // go to the next step
577: case Precision:
578: if(c>='0' && c<='9'){
579: if(state == Flags) state=Width; // no more flags
580: break;
581: } else if(c=='d' || c=='i'){
582: result=FormatInt;
583: } else if(strchr("feEgG", c)!=0){
584: result=FormatDouble;
585: } else if(strchr("uoxX", c)!=0){
586: result=FormatUInt;
587: } else {
588: return FormatInvalid; // invalid char
589: }
590: state=Done;
591: break;
592: case Done:
593: return FormatInvalid; // no chars allowed after 'type'
594: }
595: }
596: return result;
597: }
598:
599:
1.154 paf 600: const char* format(double value, char* fmt) {
1.229 misha 601: char local_buf[MAX_NUMBER];
1.235 misha 602: int size=-1;
1.229 misha 603:
604: if(fmt && strlen(fmt)){
605: switch(format_type(fmt)){
606: case FormatDouble:
607: size=snprintf(local_buf, sizeof(local_buf), fmt, value);
608: break;
609: case FormatInt:
610: size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value);
611: break;
612: case FormatUInt:
1.126 paf 613: size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value);
1.229 misha 614: break;
615: case FormatInvalid:
616: throw Exception(PARSER_RUNTIME,
617: 0,
618: "Incorrect format string '%s' was specified.", fmt);
619: }
620: } else
621: size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value);
622:
623: if(size < 0 || size >= MAX_NUMBER-1){ // on win32 we manually reduce max size while printing
624: throw Exception(PARSER_RUNTIME,
625: 0,
626: "Error occure white executing snprintf with format string '%s'.", fmt);
627: }
628:
1.235 misha 629: return pa_strdup(local_buf, (size_t)size);
1.12 paf 630: }
631:
1.36 paf 632: size_t stdout_write(const void *buf, size_t size) {
1.12 paf 633: #ifdef WIN32
1.187 paf 634: size_t to_write = size;
1.12 paf 635: do{
1.154 paf 636: int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout);
1.12 paf 637: if(chunk_written<=0)
638: break;
639: size-=chunk_written;
1.36 paf 640: buf=((const char*)buf)+chunk_written;
1.126 paf 641: } while(size>0);
1.12 paf 642:
1.187 paf 643: return to_write-size;
1.12 paf 644: #else
1.126 paf 645: return fwrite(buf, 1, size, stdout);
1.12 paf 646: #endif
1.2 paf 647: }
1.14 paf 648:
1.229 misha 649: enum EscapeState {
650: EscapeRest,
651: EscapeFirst,
652: EscapeSecond,
653: EscapeUnicode
654: };
655:
1.236 misha 656: // @todo prescan for reduce required size (unescaped sting in 1 byte charset requires less memory usually)
657: char* unescape_chars(const char* cp, int len, Charset* charset, bool ignore_plus){
658: char* s=new(PointerFreeGC) char[len+1]; // must be enough (%uXXXX==6 bytes, max utf-8 char length==6 bytes)
1.230 misha 659: char* dst=s;
1.229 misha 660: EscapeState escapeState=EscapeRest;
661: uint escapedValue=0;
662: int srcPos=0;
1.230 misha 663: short int jsCnt=0;
1.236 misha 664: while(srcPos<len){
1.229 misha 665: uchar c=(uchar)cp[srcPos];
666: if(c=='%'){
667: escapeState=EscapeFirst;
668: } else {
669: switch(escapeState) {
670: case EscapeRest:
1.236 misha 671: if(!ignore_plus && c=='+'){
1.230 misha 672: *dst++=' ';
1.229 misha 673: } else {
1.230 misha 674: *dst++=c;
1.229 misha 675: }
676: break;
677: case EscapeFirst:
1.232 misha 678: if(charset && c=='u'){
1.229 misha 679: // escaped unicode value: %u0430
680: jsCnt=0;
681: escapedValue=0;
682: escapeState=EscapeUnicode;
683: } else {
1.231 misha 684: if(isxdigit(c)){
1.229 misha 685: escapedValue=hex_value[c] << 4;
686: escapeState=EscapeSecond;
687: } else {
1.230 misha 688: *dst++=c;
1.229 misha 689: escapeState=EscapeRest;
690: }
691: }
692: break;
693: case EscapeSecond:
1.231 misha 694: if(isxdigit(c)){
1.229 misha 695: escapedValue+=hex_value[c];
1.230 misha 696: *dst++=(char)escapedValue;
1.229 misha 697: }
698: escapeState=EscapeRest;
699: break;
700: case EscapeUnicode:
1.231 misha 701: if(isxdigit(c)){
1.229 misha 702: escapedValue=(escapedValue << 4) + hex_value[c];
703: if(++jsCnt==4){
1.230 misha 704: // transcode utf8 char to client charset (we can lost some chars here)
1.232 misha 705: charset->store_Char((XMLByte*&)dst, (XMLCh)escapedValue, '?');
1.229 misha 706: escapeState=EscapeRest;
707: }
708: } else {
709: // not full unicode value
710: escapeState=EscapeRest;
711: }
712: break;
713: }
714: }
715:
716: srcPos++;
717: }
718:
1.230 misha 719: *dst=0; // zero-termination
1.229 misha 720: return s;
721: }
1.24 paf 722:
723: #ifdef WIN32
1.126 paf 724: void back_slashes_to_slashes(char* s) {
1.24 paf 725: if(s)
726: for(; *s; s++)
727: if(*s=='\\')
1.126 paf 728: *s='/';
1.24 paf 729: }
1.42 paf 730: /*
1.126 paf 731: void slashes_to_back_slashes(char* s) {
1.42 paf 732: if(s)
733: for(; *s; s++)
734: if(*s=='/')
1.126 paf 735: *s='\\';
1.42 paf 736: }
737: */
1.24 paf 738: #endif
1.41 paf 739:
1.231 misha 740: bool StrStartFromNC(const char* str, const char* substr, bool equal){
1.41 paf 741: while(true) {
1.231 misha 742: if(!(*substr)){
743: if(!(*str))
1.41 paf 744: return true;
745: else
1.231 misha 746: return !equal;
747: }
748: if(!(*str))
749: return false;
750: if(isalpha((unsigned char)*str)) {
751: if(tolower((unsigned char)*str)!=tolower((unsigned char)*substr))
1.41 paf 752: return false;
1.231 misha 753: } else if((*str) != (*substr))
1.41 paf 754: return false;
1.231 misha 755: str++;
756: substr++;
1.41 paf 757: }
1.57 parser 758: }
759:
1.232 misha 760: size_t strpos(const char *str, const char *substr) {
761: const char *p = strstr(str, substr);
762: return (p==0)?STRING_NOT_FOUND:p-str;
763: }
764:
765: // content-type: xxx; charset=WE-NEED-THIS
766: // content-type: xxx; charset="WE-NEED-THIS"
767: // content-type: xxx; charset="WE-NEED-THIS";
1.233 misha 768: Charset* detect_charset(const char* content_type){
769: if(content_type){
770: size_t len=strlen(content_type);
771: char* CONTENT_TYPE=new(PointerFreeGC) char[len+1];
772: memcpy(CONTENT_TYPE, content_type, len);
773: for(char *p=CONTENT_TYPE; *p; p++)
774: *p=(char)toupper((unsigned char)*p);
775:
776: if(const char* begin=strstr(CONTENT_TYPE, "CHARSET=")){
777: begin+=8; // skip "CHARSET="
778: char* end=0;
779: if(*begin && (*begin=='"' || *begin =='\'')){
780: char quote=*begin;
781: begin++;
782: end=(char*)strchr(begin, quote);
783: }
784: if(!end)
785: end=(char*)strchr(begin, ';');
786:
787: size_t len;
788: if(end){
789: *end=0; // terminator
790: len=end-begin;
791: } else {
792: len=strlen(begin);
793: }
794:
795: String::Body NAME=String::Body(begin, len);
796: return &charsets.get(NAME);
1.232 misha 797: }
798: }
799: return 0;
800: }
801:
1.233 misha 802: Charset* detect_charset(Charset& source_charset, const String& content_type){
803: const String& CONTENT_TYPE=content_type.change_case(source_charset, String::CC_UPPER);
804: size_t begin=CONTENT_TYPE.pos("CHARSET=");
805: if(begin!=STRING_NOT_FOUND) {
806: begin+=8; // skip "CHARSET="
807: size_t end=STRING_NOT_FOUND;
808:
809: if(CONTENT_TYPE.pos('"', begin)==begin){
810: begin++;
811: end=CONTENT_TYPE.pos('"', begin);
812: } else if(CONTENT_TYPE.pos('\'', begin)==begin){
813: begin++;
814: end=CONTENT_TYPE.pos('\'', begin);
815: }
816:
817: if(end==STRING_NOT_FOUND)
818: end=CONTENT_TYPE.pos(';', begin);
819:
820: if(end==STRING_NOT_FOUND)
821: end=CONTENT_TYPE.length();
1.232 misha 822:
1.233 misha 823: const String::Body NAME=CONTENT_TYPE.mid(begin, end);
824: return &charsets.get(NAME);
825: }
826: return 0;
1.232 misha 827: }
828:
829:
1.84 paf 830: static bool isLeap(int year) {
1.229 misha 831: return !(
832: (year % 4) || ((year % 400) && !(year % 100))
833: );
1.57 parser 834: }
835:
836: int getMonthDays(int year, int month) {
1.220 misha 837: static int monthDays[]={
1.229 misha 838: 31,
839: 28,
840: 31,
841: 30,
842: 31,
843: 30,
844: 31,
845: 31,
846: 30,
847: 31,
848: 30,
849: 31
850: };
1.228 misha 851: return (month == 1 /* january -- 0 */ && isLeap(year)) ? 29 : monthDays[month];
1.41 paf 852: }
1.69 parser 853:
1.226 misha 854: int remove_crlf(char* start, char* end) {
855: char* from=start;
856: char* to=start;
857: bool skip=false;
858: while(from < end){
859: switch(*from){
860: case '\n':
861: case '\r':
862: case '\t':
863: case ' ':
864: if(!skip){
865: *to=' ';
866: to++;
867: skip=true;
868: }
869: break;
870: default:
871: if(from != to)
872: *to=*from;
873: to++;
874: skip=false;
1.69 parser 875: }
1.226 misha 876: from++;
877: }
878: return to-start;
1.91 paf 879: }
880:
881:
882: /// must be last in this file
883: #undef vsnprintf
1.126 paf 884: int __vsnprintf(char* b, size_t s, const char* f, va_list l) {
1.91 paf 885: if(!s)
886: return 0;
887:
888: int r;
889: // note: on win32& maybe somewhere else
890: // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing
891: --s;
1.172 paf 892:
893: // clients do not check for negative 's', feature: ignore such prints
894: if((ssize_t)s<0)
895: return 0;
896:
1.91 paf 897: #if _MSC_VER
898: /*
899: win32:
900: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm
901:
1.154 paf 902: if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 is returned
1.91 paf 903: */
1.126 paf 904: r=_vsnprintf(b, s, f, l);
1.91 paf 905: if(r<0)
906: r=s;
907: #else
1.126 paf 908: r=vsnprintf(b, s, f, l);
1.91 paf 909: /*
910: solaris:
911: man vsnprintf
912:
913: The snprintf() function returns the number of characters
914: formatted, that is, the number of characters that would have
915: been written to the buffer if it were large enough. If the
916: value of n is 0 on a call to snprintf(), an unspecified
917: value less than 1 is returned.
918: */
919:
920: if(r<0)
921: r=0;
1.167 paf 922: else if((size_t)r>s)
1.91 paf 923: r=s;
924: #endif
925: b[r]=0;
926: return r;
927: }
928:
1.126 paf 929: int __snprintf(char* b, size_t s, const char* f, ...) {
1.91 paf 930: va_list l;
1.126 paf 931: va_start(l, f);
932: int r=__vsnprintf(b, s, f, l);
933: va_end(l);
1.91 paf 934: return r;
1.178 paf 935: }
936:
937: /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */
938: /*
939: * Authors: Michael Zucchi <notzed@helixcode.com>
940: * Jeffrey Stedfast <fejj@helixcode.com>
941: *
942: * Copyright 2000 Helix Code, Inc. (www.helixcode.com)
943: *
944: * This program is free software; you can redistribute it and/or modify
945: * it under the terms of the GNU General Public License as published by
946: * the Free Software Foundation; either version 2 of the License, or
947: * (at your option) any later version.
948: *
949: * This program is distributed in the hope that it will be useful,
950: * but WITHOUT ANY WARRANTY; without even the implied warranty of
951: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
952: * GNU General Public License for more details.
953: *
954: * You should have received a copy of the GNU General Public License
955: * along with this program; if not, write to the Free Software
956: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
957: *
958: */
959: static char *base64_alphabet =
960: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
961:
962: /**
963: * g_mime_utils_base64_encode_step:
964: * @in: input stream
965: * @inlen: length of the input
966: * @out: output string
967: * @state: holds the number of bits that are stored in @save
968: * @save: leftover bits that have not yet been encoded
969: *
970: * Base64 encodes a chunk of data. Performs an 'encode step', only
971: * encodes blocks of 3 characters to the output at a time, saves
972: * left-over state in state and save (initialise to 0 on first
973: * invocation).
974: *
975: * Returns the number of bytes encoded.
976: **/
977: static size_t
978: g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
979: {
1.186 paf 980: register const unsigned char *inptr;
1.178 paf 981: register unsigned char *outptr;
982:
983: if (inlen <= 0)
984: return 0;
985:
986: inptr = in;
987: outptr = out;
988:
989: if (inlen + ((unsigned char *)save)[0] > 2) {
990: const unsigned char *inend = in + inlen - 2;
991: register int c1 = 0, c2 = 0, c3 = 0;
992: register int already;
993:
994: already = *state;
995:
996: switch (((char *)save)[0]) {
997: case 1: c1 = ((unsigned char *)save)[1]; goto skip1;
998: case 2: c1 = ((unsigned char *)save)[1];
999: c2 = ((unsigned char *)save)[2]; goto skip2;
1000: }
1001:
1002: /* yes, we jump into the loop, no i'm not going to change it, its beautiful! */
1003: while (inptr < inend) {
1004: c1 = *inptr++;
1005: skip1:
1006: c2 = *inptr++;
1007: skip2:
1008: c3 = *inptr++;
1009: *outptr++ = base64_alphabet [c1 >> 2];
1010: *outptr++ = base64_alphabet [(c2 >> 4) | ((c1 & 0x3) << 4)];
1011: *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)];
1012: *outptr++ = base64_alphabet [c3 & 0x3f];
1013: /* this is a bit ugly ... */
1014: if ((++already) >= 19) {
1015: *outptr++ = '\n';
1016: already = 0;
1017: }
1018: }
1019:
1020: ((unsigned char *)save)[0] = 0;
1021: inlen = 2 - (inptr - inend);
1022: *state = already;
1023: }
1024:
1025: //d(printf ("state = %d, inlen = %d\n", (int)((char *)save)[0], inlen));
1026:
1027: if (inlen > 0) {
1028: register char *saveout;
1029:
1030: /* points to the slot for the next char to save */
1031: saveout = & (((char *)save)[1]) + ((char *)save)[0];
1032:
1033: /* inlen can only be 0 1 or 2 */
1034: switch (inlen) {
1035: case 2: *saveout++ = *inptr++;
1036: case 1: *saveout++ = *inptr++;
1037: }
1.216 paf 1038: *(char *)save = *(char *)save+(char)inlen;
1.178 paf 1039: }
1040:
1041: /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n",
1042: (int)((char *)save)[0],
1043: (int)((char *)save)[1],
1044: (int)((char *)save)[2]));*/
1045:
1046: return (outptr - out);
1047: }
1048:
1049: /**
1050: * g_mime_utils_base64_encode_close:
1051: * @in: input stream
1052: * @inlen: length of the input
1053: * @out: output string
1054: * @state: holds the number of bits that are stored in @save
1055: * @save: leftover bits that have not yet been encoded
1056: *
1057: * Base64 encodes the input stream to the output stream. Call this
1058: * when finished encoding data with g_mime_utils_base64_encode_step to
1059: * flush off the last little bit.
1060: *
1061: * Returns the number of bytes encoded.
1062: **/
1063: static size_t
1064: g_mime_utils_base64_encode_close (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1065: {
1066: unsigned char *outptr = out;
1067: int c1, c2;
1068:
1069: if (inlen > 0)
1070: outptr += g_mime_utils_base64_encode_step (in, inlen, outptr, state, save);
1071:
1072: c1 = ((unsigned char *)save)[1];
1073: c2 = ((unsigned char *)save)[2];
1074:
1075: switch (((unsigned char *)save)[0]) {
1076: case 2:
1077: outptr[2] = base64_alphabet [(c2 & 0x0f) << 2];
1078: goto skip;
1079: case 1:
1080: outptr[2] = '=';
1081: skip:
1082: outptr[0] = base64_alphabet [c1 >> 2];
1083: outptr[1] = base64_alphabet [c2 >> 4 | ((c1 & 0x3) << 4)];
1084: outptr[3] = '=';
1085: outptr += 4;
1086: break;
1087: }
1088:
1089: *outptr++ = 0;
1090:
1091: *save = 0;
1092: *state = 0;
1093:
1094: return (outptr - out);
1095: }
1096:
1.210 paf 1097: static unsigned char gmime_base64_rank[256] = {
1098: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1099: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1100: 255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63,
1101: 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255, 0,255,255,
1102: 255, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
1103: 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255,
1104: 255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
1105: 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255,
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: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1114: };
1115:
1116: /**
1117: * g_mime_utils_base64_decode_step:
1118: * @in: input stream
1119: * @inlen: max length of data to decode
1120: * @out: output stream
1121: * @state: holds the number of bits that are stored in @save
1122: * @save: leftover bits that have not yet been decoded
1123: *
1124: * Decodes a chunk of base64 encoded data.
1125: *
1126: * Returns the number of bytes decoded (which have been dumped in @out).
1127: **/
1128: size_t
1129: g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1130: {
1.213 paf 1131: const unsigned char *inptr;
1132: unsigned char *outptr;
1.210 paf 1133: const unsigned char *inend;
1.213 paf 1134: int saved;
1.210 paf 1135: unsigned char c;
1136: int i;
1137:
1138: inend = in + inlen;
1139: outptr = out;
1140:
1141: /* convert 4 base64 bytes to 3 normal bytes */
1142: saved = *save;
1143: i = *state;
1144: inptr = in;
1145: while (inptr < inend) {
1146: c = gmime_base64_rank[*inptr++];
1147: if (c != 0xff) {
1148: saved = (saved << 6) | c;
1149: i++;
1150: if (i == 4) {
1.217 paf 1151: *outptr++ = (unsigned char)(saved >> 16);
1152: *outptr++ = (unsigned char)(saved >> 8);
1153: *outptr++ = (unsigned char)(saved);
1.210 paf 1154: i = 0;
1155: }
1156: }
1157: }
1158:
1159: *save = saved;
1160: *state = i;
1161:
1162: /* quick scan back for '=' on the end somewhere */
1163: /* fortunately we can drop 1 output char for each trailing = (upto 2) */
1164: i = 2;
1165: while (inptr > in && i) {
1166: inptr--;
1167: if (gmime_base64_rank[*inptr] != 0xff) {
1168: if (*inptr == '=' && outptr > out)
1169: outptr--;
1170: i--;
1171: }
1172: }
1173:
1174: /* if i != 0 then there is a truncation error! */
1175: return (outptr - out);
1176: }
1177:
1178:
1179: char* pa_base64_encode(const char *in, size_t in_size)
1.178 paf 1180: {
1181: /* wont go to more than 2x size (overly conservative) */
1.210 paf 1182: char* result=new(PointerFreeGC) char[in_size * 2 + 6];
1.178 paf 1183: int state=0;
1184: int save=0;
1.183 paf 1185: #ifndef NDEBUG
1186: size_t filled=
1187: #endif
1.210 paf 1188: g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size,
1.178 paf 1189: (unsigned char*)result, &state, &save);
1.210 paf 1190: assert(filled <= in_size * 2 + 6);
1.178 paf 1191:
1192: return result;
1.98 paf 1193: }
1.210 paf 1194:
1.222 misha 1195:
1196: char* pa_base64_encode(const String& file_spec)
1197: {
1198: unsigned char* base64=0;
1199: File_base64_action_info info={&base64};
1200:
1201: file_read_action_under_lock(file_spec,
1202: "pa_base64_encode", file_base64_file_action, &info);
1203:
1204: return (char*)base64;
1205: }
1206:
1207:
1208: static void file_base64_file_action(
1.229 misha 1209: struct stat& finfo,
1210: int f,
1211: const String&, const char* /*fname*/, bool,
1212: void *context) {
1.222 misha 1213:
1214: if(finfo.st_size) {
1215: File_base64_action_info& info=*static_cast<File_base64_action_info *>(context);
1216: *info.base64=new(PointerFreeGC) unsigned char[finfo.st_size * 2 + 6];
1217: unsigned char* base64 = *info.base64;
1218: int state=0;
1219: int save=0;
1220: int nCount;
1221: do {
1222: unsigned char buffer[FILE_BUFFER_SIZE];
1223: nCount = file_block_read(f, buffer, sizeof(buffer));
1224: if( nCount ){
1225: size_t filled=g_mime_utils_base64_encode_step ((const unsigned char*)buffer, nCount, base64, &state, &save);
1226: base64+=filled;
1227: }
1228: } while(nCount > 0);
1229: g_mime_utils_base64_encode_close (0, 0, base64, &state, &save);
1230: }
1231: }
1232:
1.211 paf 1233: void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size)
1.210 paf 1234: {
1235: /* wont go to more than had (overly conservative) */
1.211 paf 1236: result=new(PointerFreeGC) char[in_size+1/*terminator*/];
1.210 paf 1237: int state=0;
1238: int save=0;
1239: result_size=
1240: g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size,
1241: (unsigned char*)result, &state, &save);
1242: assert(result_size <= in_size);
1.211 paf 1243: result[result_size]=0; // for text files
1.210 paf 1244: }
1.218 misha 1245:
1246:
1.221 misha 1247: int file_block_read(const int f, unsigned char* buffer, const size_t size){
1248: int nCount = read(f, buffer, size);
1249: if (nCount < 0)
1.238 ! misha 1250: throw Exception("file.read",
1.221 misha 1251: 0,
1252: "read failed: %s (%d)", strerror(errno), errno);
1253: return nCount;
1254: }
1255:
1.218 misha 1256: const unsigned long pa_crc32(const char *in, size_t in_size)
1257: {
1258: unsigned long crc32=0xFFFFFFFF;
1.220 misha 1259:
1.218 misha 1260: InitCrc32Table();
1261: for(size_t i = 0; i < in_size; i++) CalcCrc32(in[i], crc32);
1.220 misha 1262:
1.218 misha 1263: return ~crc32;
1264: }
1265:
1266: const unsigned long pa_crc32(const String& file_spec)
1267: {
1268: unsigned long crc32=0xFFFFFFFF;
1269: file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32);
1270: return ~crc32;
1271: }
1272:
1273: static void file_crc32_file_action(
1.229 misha 1274: struct stat& finfo,
1275: int f,
1276: const String&, const char* /*fname*/, bool,
1277: void *context)
1.218 misha 1278: {
1279: unsigned long& crc32=*static_cast<unsigned long *>(context);
1280: if(finfo.st_size) {
1281: InitCrc32Table();
1.220 misha 1282: int nCount=0;
1.218 misha 1283: do {
1.221 misha 1284: unsigned char buffer[FILE_BUFFER_SIZE];
1285: nCount = file_block_read(f, buffer, sizeof(buffer));
1.220 misha 1286: for(int i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32);
1287: } while(nCount > 0);
1.218 misha 1288: }
1289: }
1290:
1.237 misha 1291: char* print_pcre_exec_error_text(int exec_result){
1292: switch(exec_result){
1293: case PCRE_ERROR_BADUTF8:
1294: return "validation of UTF-8 string failed while pcre_exec (%d).";
1295: break;
1296: default:
1297: return "regular expression execute error (%d)";
1298: }
1299: }
1300:
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