Annotation of parser3/src/main/pa_common.C, revision 1.223
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:
1.223 ! misha 29: static const char * const IDENT_COMMON_C="$Date: 2007/02/07 15:50:04 $";
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,
95: const String& file_spec,
96: bool fail_on_read_problem,
97: HashStringValue* params/*, HashStringValue* * out_fields*/) {
98: File_read_result file=
99: file_read(charsets, file_spec, true, params, fail_on_read_problem);
100: return file.success?file.str:0;
1.126 paf 101: }
102:
1.206 paf 103: /// these options were handled but not checked elsewhere, now check them
1.214 paf 104: int pa_get_valid_file_options_count(HashStringValue& options)
1.206 paf 105: {
106: int result=0;
107: if(options.get(PA_SQL_LIMIT_NAME))
108: result++;
109: if(options.get(PA_SQL_OFFSET_NAME))
110: result++;
111: if(options.get(PA_COLUMN_SEPARATOR_NAME))
112: result++;
113: if(options.get(PA_COLUMN_ENCLOSER_NAME))
114: result++;
1.223 ! misha 115: if(options.get(PA_CHARSET_NAME))
! 116: result++;
1.206 paf 117: return result;
118: }
119:
1.123 paf 120: #ifndef DOXYGEN
121: struct File_read_action_info {
1.154 paf 122: char **data; size_t *data_size;
1.188 paf 123: char* buf; size_t offset; size_t count;
1.126 paf 124: };
1.123 paf 125: #endif
1.154 paf 126: static void file_read_action(
127: struct stat& finfo,
128: int f,
1.166 paf 129: const String& file_spec, const char* /*fname*/, bool as_text,
1.154 paf 130: void *context) {
1.126 paf 131: File_read_action_info& info=*static_cast<File_read_action_info *>(context);
1.188 paf 132: size_t to_read_size=info.count;
133: if(!to_read_size)
134: to_read_size=(size_t)finfo.st_size;
135: assert( !(info.buf && as_text) );
136: if(to_read_size) {
137: if(info.offset)
138: lseek(f, info.offset, SEEK_SET);
139: *info.data=info.buf
140: ? info.buf
141: : new(PointerFreeGC) char[to_read_size+(as_text?1:0)];
1.126 paf 142: *info.data_size=(size_t)read(f, *info.data, to_read_size);
1.123 paf 143:
144: if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size)
1.126 paf 145: throw Exception(0,
1.123 paf 146: &file_spec,
1.173 paf 147: "read failed: actually read %u bytes count not in [0..%u] valid range",
1.126 paf 148: *info.data_size, to_read_size);
1.123 paf 149: } else { // empty file
1.209 paf 150: // for both, text and binary: for text we need that terminator, for binary we need nonzero pointer to be able to save such files
151: *info.data=new(PointerFreeGC) char[1];
152: *(char*)(*info.data)=0;
1.123 paf 153: *info.data_size=0;
154: return;
155: }
1.126 paf 156: }
1.154 paf 157: File_read_result file_read(Request_charsets& charsets, const String& file_spec,
158: bool as_text, HashStringValue *params,
1.188 paf 159: bool fail_on_read_problem,
160: char* buf, size_t offset, size_t count) {
1.167 paf 161: File_read_result result={false, 0, 0, 0};
1.154 paf 162: if(file_spec.starts_with("http://")) {
1.203 paf 163: if(offset || count)
164: throw Exception("parser.runtime",
165: 0,
166: "offset and load options are not supported for HTTP:// file load");
167:
1.126 paf 168: // fail on read problem
1.214 paf 169: File_read_http_result http=pa_internal_file_read_http(charsets, file_spec, as_text, params);
1.154 paf 170: result.success=true;
171: result.str=http.str;
172: result.length=http.length;
173: result.headers=http.headers;
1.126 paf 174: } else {
1.206 paf 175: if(params) {
1.214 paf 176: int valid_options=pa_get_valid_file_options_count(*params);
1.206 paf 177: if(valid_options!=params->count())
178: throw Exception("parser.runtime",
179: 0,
180: "invalid option passed");
181: }
1.161 paf 182:
1.188 paf 183: File_read_action_info info={&result.str, &result.length,
184: buf, offset, count};
1.154 paf 185: result.success=file_read_action_under_lock(file_spec,
1.126 paf 186: "read", file_read_action, &info,
187: as_text, fail_on_read_problem);
1.223 ! misha 188:
! 189: if(result.length && as_text && params) {
! 190: if( Value* vcharset_name=params->get(PA_CHARSET_NAME) ) {
! 191: Charset asked_charset=::charsets.get(vcharset_name->as_string().
! 192: change_case(charsets.source(), String::CC_UPPER));
! 193:
! 194: String::C body=String::C(result.str, result.length);
! 195: body=Charset::transcode(body, asked_charset, charsets.source());
! 196:
! 197: result.str=const_cast<char *>(body.str); // hacking a little
! 198: result.length=body.length;
! 199: }
! 200: }
1.126 paf 201: }
1.123 paf 202:
1.154 paf 203: if(result.success && as_text) {
1.131 paf 204: // UTF-8 signature: EF BB BF
1.154 paf 205: if(result.length>=3) {
206: char *in=(char *)result.str;
1.159 paf 207: if(strncmp(in, "\xEF\xBB\xBF", 3)==0) {
1.154 paf 208: result.str=in+3; result.length-=3;// skip prefix
1.131 paf 209: }
210: }
211:
1.154 paf 212: fix_line_breaks((char *)(result.str), result.length);
1.123 paf 213: }
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())
222: throw Exception("parser.runtime",
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
1.98 paf 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;
1.118 paf 280: } else {
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,
1.126 paf 301: const char* action_name, File_write_action action, void *context,
302: bool as_text,
303: bool do_append,
304: bool do_block,
1.110 paf 305: bool fail_on_lock_problem) {
1.154 paf 306: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.28 paf 307: int f;
1.80 paf 308: if(access(fname, W_OK)!=0) // no
1.126 paf 309: create_dir_for_file(file_spec);
1.50 paf 310:
1.80 paf 311: if((f=open(fname,
312: O_CREAT|O_RDWR
313: |(as_text?_O_TEXT:_O_BINARY)
1.138 paf 314: |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) {
1.162 paf 315: if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) {
1.126 paf 316: Exception e("file.lock",
1.110 paf 317: &file_spec,
318: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 319: strerror(errno), errno, fname);
1.126 paf 320: close(f);
1.110 paf 321: if(fail_on_lock_problem)
322: throw e;
1.98 paf 323: return false;
324: }
1.96 paf 325:
1.158 paf 326: try {
1.126 paf 327: action(f, context);
1.158 paf 328: } catch(...) {
1.138 paf 329: #ifdef HAVE_FTRUNCATE
1.104 paf 330: if(!do_append)
1.125 paf 331: ftruncate(f, lseek(f, 0, SEEK_CUR)); // one can not use O_TRUNC, read lower
1.138 paf 332: #endif
1.162 paf 333: pa_unlock(f);close(f);
1.154 paf 334: rethrow;
1.158 paf 335: }
1.80 paf 336:
1.138 paf 337: #ifdef HAVE_FTRUNCATE
1.104 paf 338: if(!do_append)
1.125 paf 339: 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 340: #endif
1.162 paf 341: pa_unlock(f);close(f);
1.98 paf 342: return true;
1.80 paf 343: } else
1.126 paf 344: throw Exception(errno==EACCES?"file.access":0,
1.80 paf 345: &file_spec,
1.96 paf 346: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 347: action_name, strerror(errno), errno, fname);
1.96 paf 348: // here should be nothing, see rethrow above
349: }
350:
351: #ifndef DOXYGEN
352: struct File_write_action_info {
1.154 paf 353: const char* str; size_t length;
1.126 paf 354: };
1.96 paf 355: #endif
356: static void file_write_action(int f, void *context) {
1.126 paf 357: File_write_action_info& info=*static_cast<File_write_action_info *>(context);
1.154 paf 358: if(info.length) {
359: int written=write(f, info.str, info.length);
1.116 paf 360: if(written<0)
1.126 paf 361: throw Exception(0,
362: 0,
363: "write failed: %s (%d)", strerror(errno), errno);
1.113 paf 364: }
1.96 paf 365: }
366: void file_write(
367: const String& file_spec,
1.154 paf 368: const char* data, size_t size,
1.126 paf 369: bool as_text,
1.96 paf 370: bool do_append) {
1.126 paf 371: File_write_action_info info={data, size};
1.98 paf 372: file_write_action_under_lock(
1.154 paf 373: file_spec,
374: "write", file_write_action, &info,
375: as_text,
376: do_append);
1.30 paf 377: }
378:
1.63 parser 379: // throws nothing! [this is required in file_move & file_delete]
1.50 paf 380: static void rmdir(const String& file_spec, size_t pos_after) {
1.154 paf 381: size_t pos_before;
382: if((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND)
1.126 paf 383: rmdir(file_spec, pos_before+1);
1.50 paf 384:
1.154 paf 385: rmdir(file_spec.mid(0, pos_after-1/* / */).cstr(String::L_FILE_SPEC));
1.50 paf 386: }
1.164 paf 387: bool file_delete(const String& file_spec, bool fail_on_problem) {
1.154 paf 388: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.54 parser 389: if(unlink(fname)!=0)
1.164 paf 390: if(fail_on_problem)
1.126 paf 391: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.93 paf 392: &file_spec,
393: "unlink failed: %s (%d), actual filename '%s'",
1.154 paf 394: strerror(errno), errno, fname);
1.93 paf 395: else
396: return false;
1.50 paf 397:
1.126 paf 398: rmdir(file_spec, 1);
1.93 paf 399: return true;
1.60 parser 400: }
1.95 paf 401: void file_move(const String& old_spec, const String& new_spec) {
1.154 paf 402: const char* old_spec_cstr=old_spec.cstr(String::L_FILE_SPEC);
403: const char* new_spec_cstr=new_spec.cstr(String::L_FILE_SPEC);
1.63 parser 404:
1.126 paf 405: create_dir_for_file(new_spec);
1.63 parser 406:
1.60 parser 407: if(rename(old_spec_cstr, new_spec_cstr)!=0)
1.126 paf 408: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.60 parser 409: &old_spec,
410: "rename failed: %s (%d), actual filename '%s' to '%s'",
1.154 paf 411: strerror(errno), errno, old_spec_cstr, new_spec_cstr);
1.63 parser 412:
1.126 paf 413: rmdir(old_spec, 1);
1.31 paf 414: }
415:
1.51 paf 416:
1.126 paf 417: bool entry_exists(const char* fname, struct stat *afinfo) {
1.118 paf 418: struct stat lfinfo;
419: bool result=stat(fname, &lfinfo)==0;
420: if(afinfo)
421: *afinfo=lfinfo;
422: return result;
1.119 paf 423: }
424:
425: bool entry_exists(const String& file_spec) {
1.154 paf 426: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
1.126 paf 427: return entry_exists(fname, 0);
1.118 paf 428: }
429:
1.51 paf 430: static bool entry_readable(const String& file_spec, bool need_dir) {
1.154 paf 431: char* fname=file_spec.cstrm(String::L_FILE_SPEC);
1.120 paf 432: if(need_dir) {
1.126 paf 433: size_t size=strlen(fname);
1.120 paf 434: while(size) {
1.126 paf 435: char c=fname[size-1];
1.120 paf 436: if(c=='/' || c=='\\')
437: fname[--size]=0;
438: else
439: break;
440: }
441: }
1.51 paf 442: struct stat finfo;
1.118 paf 443: if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) {
1.109 paf 444: bool is_dir=(finfo.st_mode&S_IFDIR) != 0;
1.51 paf 445: return is_dir==need_dir;
446: }
447: return false;
448: }
1.215 paf 449: bool file_exist(const String& file_spec) {
1.126 paf 450: return entry_readable(file_spec, false);
1.51 paf 451: }
1.215 paf 452: bool dir_exists(const String& file_spec) {
1.126 paf 453: return entry_readable(file_spec, true);
1.65 parser 454: }
1.215 paf 455: const String* file_exist(const String& path, const String& name) {
1.154 paf 456: String& result=*new String(path);
457: result << "/";
458: result << name;
1.215 paf 459: return file_exist(result)?&result:0;
1.43 paf 460: }
461: bool file_executable(const String& file_spec) {
1.154 paf 462: return access(file_spec.cstr(String::L_FILE_SPEC), X_OK)==0;
1.44 paf 463: }
464:
1.64 parser 465: bool file_stat(const String& file_spec,
1.58 parser 466: size_t& rsize,
1.126 paf 467: time_t& ratime,
468: time_t& rmtime,
469: time_t& rctime,
1.64 parser 470: bool fail_on_read_problem) {
1.154 paf 471: const char* fname=file_spec.cstr(String::L_FILE_SPEC);
472: struct stat finfo;
1.44 paf 473: if(stat(fname, &finfo)!=0)
1.64 parser 474: if(fail_on_read_problem)
1.126 paf 475: throw Exception("file.missing",
1.67 parser 476: &file_spec,
477: "getting file size failed: %s (%d), real filename '%s'",
1.154 paf 478: strerror(errno), errno, fname);
1.64 parser 479: else
480: return false;
1.58 parser 481: rsize=finfo.st_size;
482: ratime=finfo.st_atime;
483: rmtime=finfo.st_mtime;
484: rctime=finfo.st_ctime;
1.64 parser 485: return true;
1.18 paf 486: }
487:
1.126 paf 488: char* getrow(char* *row_ref, char delim) {
489: char* result=*row_ref;
1.8 paf 490: if(result) {
1.126 paf 491: *row_ref=strchr(result, delim);
1.8 paf 492: if(*row_ref)
493: *((*row_ref)++)=0;
494: else if(!*result)
495: return 0;
496: }
497: return result;
498: }
499:
1.126 paf 500: char* lsplit(char* string, char delim) {
1.23 paf 501: if(string) {
1.126 paf 502: char* v=strchr(string, delim);
1.8 paf 503: if(v) {
504: *v=0;
505: return v+1;
506: }
507: }
508: return 0;
509: }
510:
1.126 paf 511: char* lsplit(char* *string_ref, char delim) {
512: char* result=*string_ref;
513: char* next=lsplit(*string_ref, delim);
1.8 paf 514: *string_ref=next;
515: return result;
1.9 paf 516: }
517:
1.126 paf 518: char* rsplit(char* string, char delim) {
1.18 paf 519: if(string) {
1.126 paf 520: char* v=strrchr(string, delim);
1.18 paf 521: if(v) {
1.9 paf 522: *v=0;
523: return v+1;
524: }
525: }
526: return NULL;
1.10 paf 527: }
528:
1.37 paf 529: /// @todo less stupid type detection
1.154 paf 530: const char* format(double value, char* fmt) {
1.126 paf 531: char local_buf[MAX_NUMBER];
1.108 paf 532: size_t size;
533:
1.10 paf 534: if(fmt)
535: if(strpbrk(fmt, "diouxX"))
536: if(strpbrk(fmt, "ouxX"))
1.126 paf 537: size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value);
1.10 paf 538: else
1.126 paf 539: size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value);
1.10 paf 540: else
1.126 paf 541: size=snprintf(local_buf, sizeof(local_buf), fmt, value);
1.10 paf 542: else
1.126 paf 543: size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value);
1.10 paf 544:
1.154 paf 545: return pa_strdup(local_buf, size);
1.12 paf 546: }
547:
1.36 paf 548: size_t stdout_write(const void *buf, size_t size) {
1.12 paf 549: #ifdef WIN32
1.187 paf 550: size_t to_write = size;
1.12 paf 551: do{
1.154 paf 552: int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout);
1.12 paf 553: if(chunk_written<=0)
554: break;
555: size-=chunk_written;
1.36 paf 556: buf=((const char*)buf)+chunk_written;
1.126 paf 557: } while(size>0);
1.12 paf 558:
1.187 paf 559: return to_write-size;
1.12 paf 560: #else
1.126 paf 561: return fwrite(buf, 1, size, stdout);
1.12 paf 562: #endif
1.2 paf 563: }
1.14 paf 564:
1.154 paf 565: char* unescape_chars(const char* cp, int len) {
566: char* s=new(PointerFreeGC) char[len + 1];
1.14 paf 567: enum EscapeState {
1.33 paf 568: EscapeRest,
569: EscapeFirst,
1.14 paf 570: EscapeSecond
571: } escapeState=EscapeRest;
1.216 paf 572: uint escapedValue=0;
1.14 paf 573: int srcPos=0;
574: int dstPos=0;
575: while(srcPos < len) {
1.193 paf 576: uchar ch=(uchar)cp[srcPos];
1.14 paf 577: switch(escapeState) {
578: case EscapeRest:
579: if(ch=='%') {
580: escapeState=EscapeFirst;
581: } else if(ch=='+') {
1.126 paf 582: s[dstPos++]=' ';
1.14 paf 583: } else {
584: s[dstPos++]=ch;
585: }
586: break;
587: case EscapeFirst:
1.216 paf 588: escapedValue=hex_value[ch] << 4;
1.14 paf 589: escapeState=EscapeSecond;
590: break;
591: case EscapeSecond:
1.126 paf 592: escapedValue +=hex_value[ch];
1.216 paf 593: s[dstPos++]=(char)escapedValue;
1.14 paf 594: escapeState=EscapeRest;
595: break;
596: }
1.126 paf 597: srcPos++;
1.14 paf 598: }
599: s[dstPos]=0;
600: return s;
1.24 paf 601: }
602:
603: #ifdef WIN32
1.126 paf 604: void back_slashes_to_slashes(char* s) {
1.24 paf 605: if(s)
606: for(; *s; s++)
607: if(*s=='\\')
1.126 paf 608: *s='/';
1.24 paf 609: }
1.42 paf 610: /*
1.126 paf 611: void slashes_to_back_slashes(char* s) {
1.42 paf 612: if(s)
613: for(; *s; s++)
614: if(*s=='/')
1.126 paf 615: *s='\\';
1.42 paf 616: }
617: */
1.24 paf 618: #endif
1.41 paf 619:
1.126 paf 620: bool StrEqNc(const char* s1, const char* s2, bool strict) {
1.41 paf 621: while(true) {
622: if(!(*s1)) {
623: if(!(*s2))
624: return true;
625: else
626: return !strict;
627: } else if(!(*s2))
628: return !strict;
1.189 paf 629: if(isalpha((unsigned char)*s1)) {
1.190 paf 630: if(tolower((unsigned char)*s1) !=tolower((unsigned char)*s2))
1.41 paf 631: return false;
632: } else if((*s1) !=(*s2))
633: return false;
1.126 paf 634: s1++;
635: s2++;
1.41 paf 636: }
1.57 parser 637: }
638:
1.84 paf 639: static bool isLeap(int year) {
1.57 parser 640: return !(
641: (year % 4) || ((year % 400) && !(year % 100))
1.126 paf 642: );
1.57 parser 643: }
644:
645: int getMonthDays(int year, int month) {
1.220 misha 646: static int monthDays[]={
1.126 paf 647: 31,
1.220 misha 648: 28,
1.126 paf 649: 31,
650: 30,
651: 31,
652: 30,
653: 31,
654: 31,
655: 30,
656: 31,
657: 30,
1.57 parser 658: 31
1.126 paf 659: };
1.220 misha 660: return (month == 2 && isLeap(year)) ? 29 : monthDays[month];
1.41 paf 661: }
1.69 parser 662:
1.126 paf 663: void remove_crlf(char* start, char* end) {
664: for(char* p=start; p<end; p++)
1.69 parser 665: switch(*p) {
1.126 paf 666: case '\n': *p='|'; break;
667: case '\r': *p=' '; break;
1.69 parser 668: }
1.91 paf 669: }
670:
671:
672: /// must be last in this file
673: #undef vsnprintf
1.126 paf 674: int __vsnprintf(char* b, size_t s, const char* f, va_list l) {
1.91 paf 675: if(!s)
676: return 0;
677:
678: int r;
679: // note: on win32& maybe somewhere else
680: // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing
681: --s;
1.172 paf 682:
683: // clients do not check for negative 's', feature: ignore such prints
684: if((ssize_t)s<0)
685: return 0;
686:
1.91 paf 687: #if _MSC_VER
688: /*
689: win32:
690: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm
691:
1.154 paf 692: if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 is returned
1.91 paf 693: */
1.126 paf 694: r=_vsnprintf(b, s, f, l);
1.91 paf 695: if(r<0)
696: r=s;
697: #else
1.126 paf 698: r=vsnprintf(b, s, f, l);
1.91 paf 699: /*
700: solaris:
701: man vsnprintf
702:
703: The snprintf() function returns the number of characters
704: formatted, that is, the number of characters that would have
705: been written to the buffer if it were large enough. If the
706: value of n is 0 on a call to snprintf(), an unspecified
707: value less than 1 is returned.
708: */
709:
710: if(r<0)
711: r=0;
1.167 paf 712: else if((size_t)r>s)
1.91 paf 713: r=s;
714: #endif
715: b[r]=0;
716: return r;
717: }
718:
1.126 paf 719: int __snprintf(char* b, size_t s, const char* f, ...) {
1.91 paf 720: va_list l;
1.126 paf 721: va_start(l, f);
722: int r=__vsnprintf(b, s, f, l);
723: va_end(l);
1.91 paf 724: return r;
1.178 paf 725: }
726:
727: /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */
728: /*
729: * Authors: Michael Zucchi <notzed@helixcode.com>
730: * Jeffrey Stedfast <fejj@helixcode.com>
731: *
732: * Copyright 2000 Helix Code, Inc. (www.helixcode.com)
733: *
734: * This program is free software; you can redistribute it and/or modify
735: * it under the terms of the GNU General Public License as published by
736: * the Free Software Foundation; either version 2 of the License, or
737: * (at your option) any later version.
738: *
739: * This program is distributed in the hope that it will be useful,
740: * but WITHOUT ANY WARRANTY; without even the implied warranty of
741: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
742: * GNU General Public License for more details.
743: *
744: * You should have received a copy of the GNU General Public License
745: * along with this program; if not, write to the Free Software
746: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
747: *
748: */
749: static char *base64_alphabet =
750: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
751:
752: /**
753: * g_mime_utils_base64_encode_step:
754: * @in: input stream
755: * @inlen: length of the input
756: * @out: output string
757: * @state: holds the number of bits that are stored in @save
758: * @save: leftover bits that have not yet been encoded
759: *
760: * Base64 encodes a chunk of data. Performs an 'encode step', only
761: * encodes blocks of 3 characters to the output at a time, saves
762: * left-over state in state and save (initialise to 0 on first
763: * invocation).
764: *
765: * Returns the number of bytes encoded.
766: **/
767: static size_t
768: g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
769: {
1.186 paf 770: register const unsigned char *inptr;
1.178 paf 771: register unsigned char *outptr;
772:
773: if (inlen <= 0)
774: return 0;
775:
776: inptr = in;
777: outptr = out;
778:
779: if (inlen + ((unsigned char *)save)[0] > 2) {
780: const unsigned char *inend = in + inlen - 2;
781: register int c1 = 0, c2 = 0, c3 = 0;
782: register int already;
783:
784: already = *state;
785:
786: switch (((char *)save)[0]) {
787: case 1: c1 = ((unsigned char *)save)[1]; goto skip1;
788: case 2: c1 = ((unsigned char *)save)[1];
789: c2 = ((unsigned char *)save)[2]; goto skip2;
790: }
791:
792: /* yes, we jump into the loop, no i'm not going to change it, its beautiful! */
793: while (inptr < inend) {
794: c1 = *inptr++;
795: skip1:
796: c2 = *inptr++;
797: skip2:
798: c3 = *inptr++;
799: *outptr++ = base64_alphabet [c1 >> 2];
800: *outptr++ = base64_alphabet [(c2 >> 4) | ((c1 & 0x3) << 4)];
801: *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)];
802: *outptr++ = base64_alphabet [c3 & 0x3f];
803: /* this is a bit ugly ... */
804: if ((++already) >= 19) {
805: *outptr++ = '\n';
806: already = 0;
807: }
808: }
809:
810: ((unsigned char *)save)[0] = 0;
811: inlen = 2 - (inptr - inend);
812: *state = already;
813: }
814:
815: //d(printf ("state = %d, inlen = %d\n", (int)((char *)save)[0], inlen));
816:
817: if (inlen > 0) {
818: register char *saveout;
819:
820: /* points to the slot for the next char to save */
821: saveout = & (((char *)save)[1]) + ((char *)save)[0];
822:
823: /* inlen can only be 0 1 or 2 */
824: switch (inlen) {
825: case 2: *saveout++ = *inptr++;
826: case 1: *saveout++ = *inptr++;
827: }
1.216 paf 828: *(char *)save = *(char *)save+(char)inlen;
1.178 paf 829: }
830:
831: /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n",
832: (int)((char *)save)[0],
833: (int)((char *)save)[1],
834: (int)((char *)save)[2]));*/
835:
836: return (outptr - out);
837: }
838:
839: /**
840: * g_mime_utils_base64_encode_close:
841: * @in: input stream
842: * @inlen: length of the input
843: * @out: output string
844: * @state: holds the number of bits that are stored in @save
845: * @save: leftover bits that have not yet been encoded
846: *
847: * Base64 encodes the input stream to the output stream. Call this
848: * when finished encoding data with g_mime_utils_base64_encode_step to
849: * flush off the last little bit.
850: *
851: * Returns the number of bytes encoded.
852: **/
853: static size_t
854: g_mime_utils_base64_encode_close (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
855: {
856: unsigned char *outptr = out;
857: int c1, c2;
858:
859: if (inlen > 0)
860: outptr += g_mime_utils_base64_encode_step (in, inlen, outptr, state, save);
861:
862: c1 = ((unsigned char *)save)[1];
863: c2 = ((unsigned char *)save)[2];
864:
865: switch (((unsigned char *)save)[0]) {
866: case 2:
867: outptr[2] = base64_alphabet [(c2 & 0x0f) << 2];
868: goto skip;
869: case 1:
870: outptr[2] = '=';
871: skip:
872: outptr[0] = base64_alphabet [c1 >> 2];
873: outptr[1] = base64_alphabet [c2 >> 4 | ((c1 & 0x3) << 4)];
874: outptr[3] = '=';
875: outptr += 4;
876: break;
877: }
878:
879: *outptr++ = 0;
880:
881: *save = 0;
882: *state = 0;
883:
884: return (outptr - out);
885: }
886:
1.210 paf 887: static unsigned char gmime_base64_rank[256] = {
888: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
889: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
890: 255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63,
891: 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255, 0,255,255,
892: 255, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
893: 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255,
894: 255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
895: 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255,
896: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
897: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
898: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
899: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
900: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
901: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
902: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
903: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
904: };
905:
906: /**
907: * g_mime_utils_base64_decode_step:
908: * @in: input stream
909: * @inlen: max length of data to decode
910: * @out: output stream
911: * @state: holds the number of bits that are stored in @save
912: * @save: leftover bits that have not yet been decoded
913: *
914: * Decodes a chunk of base64 encoded data.
915: *
916: * Returns the number of bytes decoded (which have been dumped in @out).
917: **/
918: size_t
919: g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
920: {
1.213 paf 921: const unsigned char *inptr;
922: unsigned char *outptr;
1.210 paf 923: const unsigned char *inend;
1.213 paf 924: int saved;
1.210 paf 925: unsigned char c;
926: int i;
927:
928: inend = in + inlen;
929: outptr = out;
930:
931: /* convert 4 base64 bytes to 3 normal bytes */
932: saved = *save;
933: i = *state;
934: inptr = in;
935: while (inptr < inend) {
936: c = gmime_base64_rank[*inptr++];
937: if (c != 0xff) {
938: saved = (saved << 6) | c;
939: i++;
940: if (i == 4) {
1.217 paf 941: *outptr++ = (unsigned char)(saved >> 16);
942: *outptr++ = (unsigned char)(saved >> 8);
943: *outptr++ = (unsigned char)(saved);
1.210 paf 944: i = 0;
945: }
946: }
947: }
948:
949: *save = saved;
950: *state = i;
951:
952: /* quick scan back for '=' on the end somewhere */
953: /* fortunately we can drop 1 output char for each trailing = (upto 2) */
954: i = 2;
955: while (inptr > in && i) {
956: inptr--;
957: if (gmime_base64_rank[*inptr] != 0xff) {
958: if (*inptr == '=' && outptr > out)
959: outptr--;
960: i--;
961: }
962: }
963:
964: /* if i != 0 then there is a truncation error! */
965: return (outptr - out);
966: }
967:
968:
969: char* pa_base64_encode(const char *in, size_t in_size)
1.178 paf 970: {
971: /* wont go to more than 2x size (overly conservative) */
1.210 paf 972: char* result=new(PointerFreeGC) char[in_size * 2 + 6];
1.178 paf 973: int state=0;
974: int save=0;
1.183 paf 975: #ifndef NDEBUG
976: size_t filled=
977: #endif
1.210 paf 978: g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size,
1.178 paf 979: (unsigned char*)result, &state, &save);
1.210 paf 980: assert(filled <= in_size * 2 + 6);
1.178 paf 981:
982: return result;
1.98 paf 983: }
1.210 paf 984:
1.222 misha 985:
986: char* pa_base64_encode(const String& file_spec)
987: {
988: unsigned char* base64=0;
989: File_base64_action_info info={&base64};
990:
991: file_read_action_under_lock(file_spec,
992: "pa_base64_encode", file_base64_file_action, &info);
993:
994: return (char*)base64;
995: }
996:
997:
998: static void file_base64_file_action(
999: struct stat& finfo,
1000: int f,
1001: const String&, const char* /*fname*/, bool,
1002: void *context) {
1003:
1004: if(finfo.st_size) {
1005: File_base64_action_info& info=*static_cast<File_base64_action_info *>(context);
1006: *info.base64=new(PointerFreeGC) unsigned char[finfo.st_size * 2 + 6];
1007: unsigned char* base64 = *info.base64;
1008: int state=0;
1009: int save=0;
1010: int nCount;
1011: do {
1012: unsigned char buffer[FILE_BUFFER_SIZE];
1013: nCount = file_block_read(f, buffer, sizeof(buffer));
1014: if( nCount ){
1015: size_t filled=g_mime_utils_base64_encode_step ((const unsigned char*)buffer, nCount, base64, &state, &save);
1016: base64+=filled;
1017: }
1018: } while(nCount > 0);
1019: g_mime_utils_base64_encode_close (0, 0, base64, &state, &save);
1020: }
1021: }
1022:
1.211 paf 1023: void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size)
1.210 paf 1024: {
1025: /* wont go to more than had (overly conservative) */
1.211 paf 1026: result=new(PointerFreeGC) char[in_size+1/*terminator*/];
1.210 paf 1027: int state=0;
1028: int save=0;
1029: result_size=
1030: g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size,
1031: (unsigned char*)result, &state, &save);
1032: assert(result_size <= in_size);
1.211 paf 1033: result[result_size]=0; // for text files
1.210 paf 1034: }
1.218 misha 1035:
1036:
1.221 misha 1037: int file_block_read(const int f, unsigned char* buffer, const size_t size){
1038: int nCount = read(f, buffer, size);
1039: if (nCount < 0)
1040: throw Exception(0,
1041: 0,
1042: "read failed: %s (%d)", strerror(errno), errno);
1043: return nCount;
1044: }
1045:
1.218 misha 1046: const unsigned long pa_crc32(const char *in, size_t in_size)
1047: {
1048: unsigned long crc32=0xFFFFFFFF;
1.220 misha 1049:
1.218 misha 1050: InitCrc32Table();
1051: for(size_t i = 0; i < in_size; i++) CalcCrc32(in[i], crc32);
1.220 misha 1052:
1.218 misha 1053: return ~crc32;
1054: }
1055:
1056: const unsigned long pa_crc32(const String& file_spec)
1057: {
1058: unsigned long crc32=0xFFFFFFFF;
1059: file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32);
1060: return ~crc32;
1061: }
1062:
1063: static void file_crc32_file_action(
1064: struct stat& finfo,
1065: int f,
1.219 misha 1066: const String&, const char* /*fname*/, bool,
1.218 misha 1067: void *context)
1068: {
1069: unsigned long& crc32=*static_cast<unsigned long *>(context);
1070: if(finfo.st_size) {
1071: InitCrc32Table();
1.220 misha 1072: int nCount=0;
1.218 misha 1073: do {
1.221 misha 1074: unsigned char buffer[FILE_BUFFER_SIZE];
1075: nCount = file_block_read(f, buffer, sizeof(buffer));
1.220 misha 1076: for(int i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32);
1077: } while(nCount > 0);
1.218 misha 1078: }
1079: }
1080:
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