Annotation of parser3/src/main/pa_common.C, revision 1.227
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:
1.227 ! misha 29: static const char * const IDENT_COMMON_C="$Date: 2007/11/27 09:57:08 $";
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)
1.224 misha 164: throw Exception(PARSER_RUNTIME,
1.203 paf 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())
1.224 misha 178: throw Exception(PARSER_RUNTIME,
1.206 paf 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())
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
1.98 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,
1.225 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,
1.58 parser 471: size_t& rsize,
1.126 paf 472: time_t& ratime,
473: time_t& rmtime,
474: time_t& rctime,
1.64 parser 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) {
494: char* result=*row_ref;
1.8 paf 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;
501: }
502: return result;
503: }
504:
1.126 paf 505: char* lsplit(char* string, char delim) {
1.23 paf 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: }
512: }
513: return 0;
514: }
515:
1.126 paf 516: char* lsplit(char* *string_ref, char delim) {
517: char* result=*string_ref;
518: char* next=lsplit(*string_ref, delim);
1.8 paf 519: *string_ref=next;
520: return result;
1.9 paf 521: }
522:
1.126 paf 523: char* rsplit(char* string, char delim) {
1.18 paf 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: }
530: }
531: return NULL;
1.10 paf 532: }
533:
1.37 paf 534: /// @todo less stupid type detection
1.154 paf 535: const char* format(double value, char* fmt) {
1.126 paf 536: char local_buf[MAX_NUMBER];
1.108 paf 537: size_t size;
538:
1.10 paf 539: if(fmt)
540: if(strpbrk(fmt, "diouxX"))
541: if(strpbrk(fmt, "ouxX"))
1.126 paf 542: size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value);
1.10 paf 543: else
1.126 paf 544: size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value);
1.10 paf 545: else
1.126 paf 546: size=snprintf(local_buf, sizeof(local_buf), fmt, value);
1.10 paf 547: else
1.126 paf 548: size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value);
1.10 paf 549:
1.154 paf 550: return pa_strdup(local_buf, size);
1.12 paf 551: }
552:
1.36 paf 553: size_t stdout_write(const void *buf, size_t size) {
1.12 paf 554: #ifdef WIN32
1.187 paf 555: size_t to_write = size;
1.12 paf 556: do{
1.154 paf 557: int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout);
1.12 paf 558: if(chunk_written<=0)
559: break;
560: size-=chunk_written;
1.36 paf 561: buf=((const char*)buf)+chunk_written;
1.126 paf 562: } while(size>0);
1.12 paf 563:
1.187 paf 564: return to_write-size;
1.12 paf 565: #else
1.126 paf 566: return fwrite(buf, 1, size, stdout);
1.12 paf 567: #endif
1.2 paf 568: }
1.14 paf 569:
1.154 paf 570: char* unescape_chars(const char* cp, int len) {
571: char* s=new(PointerFreeGC) char[len + 1];
1.14 paf 572: enum EscapeState {
1.33 paf 573: EscapeRest,
574: EscapeFirst,
1.14 paf 575: EscapeSecond
576: } escapeState=EscapeRest;
1.216 paf 577: uint escapedValue=0;
1.14 paf 578: int srcPos=0;
579: int dstPos=0;
580: while(srcPos < len) {
1.193 paf 581: uchar ch=(uchar)cp[srcPos];
1.14 paf 582: switch(escapeState) {
583: case EscapeRest:
584: if(ch=='%') {
585: escapeState=EscapeFirst;
586: } else if(ch=='+') {
1.126 paf 587: s[dstPos++]=' ';
1.14 paf 588: } else {
589: s[dstPos++]=ch;
590: }
591: break;
592: case EscapeFirst:
1.216 paf 593: escapedValue=hex_value[ch] << 4;
1.14 paf 594: escapeState=EscapeSecond;
595: break;
596: case EscapeSecond:
1.126 paf 597: escapedValue +=hex_value[ch];
1.216 paf 598: s[dstPos++]=(char)escapedValue;
1.14 paf 599: escapeState=EscapeRest;
600: break;
601: }
1.126 paf 602: srcPos++;
1.14 paf 603: }
604: s[dstPos]=0;
605: return s;
1.24 paf 606: }
607:
608: #ifdef WIN32
1.126 paf 609: void back_slashes_to_slashes(char* s) {
1.24 paf 610: if(s)
611: for(; *s; s++)
612: if(*s=='\\')
1.126 paf 613: *s='/';
1.24 paf 614: }
1.42 paf 615: /*
1.126 paf 616: void slashes_to_back_slashes(char* s) {
1.42 paf 617: if(s)
618: for(; *s; s++)
619: if(*s=='/')
1.126 paf 620: *s='\\';
1.42 paf 621: }
622: */
1.24 paf 623: #endif
1.41 paf 624:
1.126 paf 625: bool StrEqNc(const char* s1, const char* s2, bool strict) {
1.41 paf 626: while(true) {
627: if(!(*s1)) {
628: if(!(*s2))
629: return true;
630: else
631: return !strict;
632: } else if(!(*s2))
633: return !strict;
1.189 paf 634: if(isalpha((unsigned char)*s1)) {
1.190 paf 635: if(tolower((unsigned char)*s1) !=tolower((unsigned char)*s2))
1.41 paf 636: return false;
637: } else if((*s1) !=(*s2))
638: return false;
1.126 paf 639: s1++;
640: s2++;
1.41 paf 641: }
1.57 parser 642: }
643:
1.84 paf 644: static bool isLeap(int year) {
1.57 parser 645: return !(
646: (year % 4) || ((year % 400) && !(year % 100))
1.126 paf 647: );
1.57 parser 648: }
649:
650: int getMonthDays(int year, int month) {
1.220 misha 651: static int monthDays[]={
1.126 paf 652: 31,
1.220 misha 653: 28,
1.126 paf 654: 31,
655: 30,
656: 31,
657: 30,
658: 31,
659: 31,
660: 30,
661: 31,
662: 30,
1.57 parser 663: 31
1.126 paf 664: };
1.227 ! misha 665: return (month == 1 && isLeap(year)) ? 29 : monthDays[month];
1.41 paf 666: }
1.69 parser 667:
1.226 misha 668: int remove_crlf(char* start, char* end) {
669: char* from=start;
670: char* to=start;
671: bool skip=false;
672: while(from < end){
673: switch(*from){
674: case '\n':
675: case '\r':
676: case '\t':
677: case ' ':
678: if(!skip){
679: *to=' ';
680: to++;
681: skip=true;
682: }
683: break;
684: default:
685: if(from != to)
686: *to=*from;
687: to++;
688: skip=false;
1.69 parser 689: }
1.226 misha 690: from++;
691: }
692: return to-start;
1.91 paf 693: }
694:
695:
696: /// must be last in this file
697: #undef vsnprintf
1.126 paf 698: int __vsnprintf(char* b, size_t s, const char* f, va_list l) {
1.91 paf 699: if(!s)
700: return 0;
701:
702: int r;
703: // note: on win32& maybe somewhere else
704: // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing
705: --s;
1.172 paf 706:
707: // clients do not check for negative 's', feature: ignore such prints
708: if((ssize_t)s<0)
709: return 0;
710:
1.91 paf 711: #if _MSC_VER
712: /*
713: win32:
714: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm
715:
1.154 paf 716: if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 is returned
1.91 paf 717: */
1.126 paf 718: r=_vsnprintf(b, s, f, l);
1.91 paf 719: if(r<0)
720: r=s;
721: #else
1.126 paf 722: r=vsnprintf(b, s, f, l);
1.91 paf 723: /*
724: solaris:
725: man vsnprintf
726:
727: The snprintf() function returns the number of characters
728: formatted, that is, the number of characters that would have
729: been written to the buffer if it were large enough. If the
730: value of n is 0 on a call to snprintf(), an unspecified
731: value less than 1 is returned.
732: */
733:
734: if(r<0)
735: r=0;
1.167 paf 736: else if((size_t)r>s)
1.91 paf 737: r=s;
738: #endif
739: b[r]=0;
740: return r;
741: }
742:
1.126 paf 743: int __snprintf(char* b, size_t s, const char* f, ...) {
1.91 paf 744: va_list l;
1.126 paf 745: va_start(l, f);
746: int r=__vsnprintf(b, s, f, l);
747: va_end(l);
1.91 paf 748: return r;
1.178 paf 749: }
750:
751: /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */
752: /*
753: * Authors: Michael Zucchi <notzed@helixcode.com>
754: * Jeffrey Stedfast <fejj@helixcode.com>
755: *
756: * Copyright 2000 Helix Code, Inc. (www.helixcode.com)
757: *
758: * This program is free software; you can redistribute it and/or modify
759: * it under the terms of the GNU General Public License as published by
760: * the Free Software Foundation; either version 2 of the License, or
761: * (at your option) any later version.
762: *
763: * This program is distributed in the hope that it will be useful,
764: * but WITHOUT ANY WARRANTY; without even the implied warranty of
765: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
766: * GNU General Public License for more details.
767: *
768: * You should have received a copy of the GNU General Public License
769: * along with this program; if not, write to the Free Software
770: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
771: *
772: */
773: static char *base64_alphabet =
774: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
775:
776: /**
777: * g_mime_utils_base64_encode_step:
778: * @in: input stream
779: * @inlen: length of the input
780: * @out: output string
781: * @state: holds the number of bits that are stored in @save
782: * @save: leftover bits that have not yet been encoded
783: *
784: * Base64 encodes a chunk of data. Performs an 'encode step', only
785: * encodes blocks of 3 characters to the output at a time, saves
786: * left-over state in state and save (initialise to 0 on first
787: * invocation).
788: *
789: * Returns the number of bytes encoded.
790: **/
791: static size_t
792: g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
793: {
1.186 paf 794: register const unsigned char *inptr;
1.178 paf 795: register unsigned char *outptr;
796:
797: if (inlen <= 0)
798: return 0;
799:
800: inptr = in;
801: outptr = out;
802:
803: if (inlen + ((unsigned char *)save)[0] > 2) {
804: const unsigned char *inend = in + inlen - 2;
805: register int c1 = 0, c2 = 0, c3 = 0;
806: register int already;
807:
808: already = *state;
809:
810: switch (((char *)save)[0]) {
811: case 1: c1 = ((unsigned char *)save)[1]; goto skip1;
812: case 2: c1 = ((unsigned char *)save)[1];
813: c2 = ((unsigned char *)save)[2]; goto skip2;
814: }
815:
816: /* yes, we jump into the loop, no i'm not going to change it, its beautiful! */
817: while (inptr < inend) {
818: c1 = *inptr++;
819: skip1:
820: c2 = *inptr++;
821: skip2:
822: c3 = *inptr++;
823: *outptr++ = base64_alphabet [c1 >> 2];
824: *outptr++ = base64_alphabet [(c2 >> 4) | ((c1 & 0x3) << 4)];
825: *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)];
826: *outptr++ = base64_alphabet [c3 & 0x3f];
827: /* this is a bit ugly ... */
828: if ((++already) >= 19) {
829: *outptr++ = '\n';
830: already = 0;
831: }
832: }
833:
834: ((unsigned char *)save)[0] = 0;
835: inlen = 2 - (inptr - inend);
836: *state = already;
837: }
838:
839: //d(printf ("state = %d, inlen = %d\n", (int)((char *)save)[0], inlen));
840:
841: if (inlen > 0) {
842: register char *saveout;
843:
844: /* points to the slot for the next char to save */
845: saveout = & (((char *)save)[1]) + ((char *)save)[0];
846:
847: /* inlen can only be 0 1 or 2 */
848: switch (inlen) {
849: case 2: *saveout++ = *inptr++;
850: case 1: *saveout++ = *inptr++;
851: }
1.216 paf 852: *(char *)save = *(char *)save+(char)inlen;
1.178 paf 853: }
854:
855: /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n",
856: (int)((char *)save)[0],
857: (int)((char *)save)[1],
858: (int)((char *)save)[2]));*/
859:
860: return (outptr - out);
861: }
862:
863: /**
864: * g_mime_utils_base64_encode_close:
865: * @in: input stream
866: * @inlen: length of the input
867: * @out: output string
868: * @state: holds the number of bits that are stored in @save
869: * @save: leftover bits that have not yet been encoded
870: *
871: * Base64 encodes the input stream to the output stream. Call this
872: * when finished encoding data with g_mime_utils_base64_encode_step to
873: * flush off the last little bit.
874: *
875: * Returns the number of bytes encoded.
876: **/
877: static size_t
878: g_mime_utils_base64_encode_close (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
879: {
880: unsigned char *outptr = out;
881: int c1, c2;
882:
883: if (inlen > 0)
884: outptr += g_mime_utils_base64_encode_step (in, inlen, outptr, state, save);
885:
886: c1 = ((unsigned char *)save)[1];
887: c2 = ((unsigned char *)save)[2];
888:
889: switch (((unsigned char *)save)[0]) {
890: case 2:
891: outptr[2] = base64_alphabet [(c2 & 0x0f) << 2];
892: goto skip;
893: case 1:
894: outptr[2] = '=';
895: skip:
896: outptr[0] = base64_alphabet [c1 >> 2];
897: outptr[1] = base64_alphabet [c2 >> 4 | ((c1 & 0x3) << 4)];
898: outptr[3] = '=';
899: outptr += 4;
900: break;
901: }
902:
903: *outptr++ = 0;
904:
905: *save = 0;
906: *state = 0;
907:
908: return (outptr - out);
909: }
910:
1.210 paf 911: static unsigned char gmime_base64_rank[256] = {
912: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
913: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
914: 255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63,
915: 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255, 0,255,255,
916: 255, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
917: 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255,
918: 255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
919: 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255,
920: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
921: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
922: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
923: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
924: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
925: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
926: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
927: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
928: };
929:
930: /**
931: * g_mime_utils_base64_decode_step:
932: * @in: input stream
933: * @inlen: max length of data to decode
934: * @out: output stream
935: * @state: holds the number of bits that are stored in @save
936: * @save: leftover bits that have not yet been decoded
937: *
938: * Decodes a chunk of base64 encoded data.
939: *
940: * Returns the number of bytes decoded (which have been dumped in @out).
941: **/
942: size_t
943: g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
944: {
1.213 paf 945: const unsigned char *inptr;
946: unsigned char *outptr;
1.210 paf 947: const unsigned char *inend;
1.213 paf 948: int saved;
1.210 paf 949: unsigned char c;
950: int i;
951:
952: inend = in + inlen;
953: outptr = out;
954:
955: /* convert 4 base64 bytes to 3 normal bytes */
956: saved = *save;
957: i = *state;
958: inptr = in;
959: while (inptr < inend) {
960: c = gmime_base64_rank[*inptr++];
961: if (c != 0xff) {
962: saved = (saved << 6) | c;
963: i++;
964: if (i == 4) {
1.217 paf 965: *outptr++ = (unsigned char)(saved >> 16);
966: *outptr++ = (unsigned char)(saved >> 8);
967: *outptr++ = (unsigned char)(saved);
1.210 paf 968: i = 0;
969: }
970: }
971: }
972:
973: *save = saved;
974: *state = i;
975:
976: /* quick scan back for '=' on the end somewhere */
977: /* fortunately we can drop 1 output char for each trailing = (upto 2) */
978: i = 2;
979: while (inptr > in && i) {
980: inptr--;
981: if (gmime_base64_rank[*inptr] != 0xff) {
982: if (*inptr == '=' && outptr > out)
983: outptr--;
984: i--;
985: }
986: }
987:
988: /* if i != 0 then there is a truncation error! */
989: return (outptr - out);
990: }
991:
992:
993: char* pa_base64_encode(const char *in, size_t in_size)
1.178 paf 994: {
995: /* wont go to more than 2x size (overly conservative) */
1.210 paf 996: char* result=new(PointerFreeGC) char[in_size * 2 + 6];
1.178 paf 997: int state=0;
998: int save=0;
1.183 paf 999: #ifndef NDEBUG
1000: size_t filled=
1001: #endif
1.210 paf 1002: g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size,
1.178 paf 1003: (unsigned char*)result, &state, &save);
1.210 paf 1004: assert(filled <= in_size * 2 + 6);
1.178 paf 1005:
1006: return result;
1.98 paf 1007: }
1.210 paf 1008:
1.222 misha 1009:
1010: char* pa_base64_encode(const String& file_spec)
1011: {
1012: unsigned char* base64=0;
1013: File_base64_action_info info={&base64};
1014:
1015: file_read_action_under_lock(file_spec,
1016: "pa_base64_encode", file_base64_file_action, &info);
1017:
1018: return (char*)base64;
1019: }
1020:
1021:
1022: static void file_base64_file_action(
1023: struct stat& finfo,
1024: int f,
1025: const String&, const char* /*fname*/, bool,
1026: void *context) {
1027:
1028: if(finfo.st_size) {
1029: File_base64_action_info& info=*static_cast<File_base64_action_info *>(context);
1030: *info.base64=new(PointerFreeGC) unsigned char[finfo.st_size * 2 + 6];
1031: unsigned char* base64 = *info.base64;
1032: int state=0;
1033: int save=0;
1034: int nCount;
1035: do {
1036: unsigned char buffer[FILE_BUFFER_SIZE];
1037: nCount = file_block_read(f, buffer, sizeof(buffer));
1038: if( nCount ){
1039: size_t filled=g_mime_utils_base64_encode_step ((const unsigned char*)buffer, nCount, base64, &state, &save);
1040: base64+=filled;
1041: }
1042: } while(nCount > 0);
1043: g_mime_utils_base64_encode_close (0, 0, base64, &state, &save);
1044: }
1045: }
1046:
1.211 paf 1047: void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size)
1.210 paf 1048: {
1049: /* wont go to more than had (overly conservative) */
1.211 paf 1050: result=new(PointerFreeGC) char[in_size+1/*terminator*/];
1.210 paf 1051: int state=0;
1052: int save=0;
1053: result_size=
1054: g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size,
1055: (unsigned char*)result, &state, &save);
1056: assert(result_size <= in_size);
1.211 paf 1057: result[result_size]=0; // for text files
1.210 paf 1058: }
1.218 misha 1059:
1060:
1.221 misha 1061: int file_block_read(const int f, unsigned char* buffer, const size_t size){
1062: int nCount = read(f, buffer, size);
1063: if (nCount < 0)
1064: throw Exception(0,
1065: 0,
1066: "read failed: %s (%d)", strerror(errno), errno);
1067: return nCount;
1068: }
1069:
1.218 misha 1070: const unsigned long pa_crc32(const char *in, size_t in_size)
1071: {
1072: unsigned long crc32=0xFFFFFFFF;
1.220 misha 1073:
1.218 misha 1074: InitCrc32Table();
1075: for(size_t i = 0; i < in_size; i++) CalcCrc32(in[i], crc32);
1.220 misha 1076:
1.218 misha 1077: return ~crc32;
1078: }
1079:
1080: const unsigned long pa_crc32(const String& file_spec)
1081: {
1082: unsigned long crc32=0xFFFFFFFF;
1083: file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32);
1084: return ~crc32;
1085: }
1086:
1087: static void file_crc32_file_action(
1088: struct stat& finfo,
1089: int f,
1.219 misha 1090: const String&, const char* /*fname*/, bool,
1.218 misha 1091: void *context)
1092: {
1093: unsigned long& crc32=*static_cast<unsigned long *>(context);
1094: if(finfo.st_size) {
1095: InitCrc32Table();
1.220 misha 1096: int nCount=0;
1.218 misha 1097: do {
1.221 misha 1098: unsigned char buffer[FILE_BUFFER_SIZE];
1099: nCount = file_block_read(f, buffer, sizeof(buffer));
1.220 misha 1100: for(int i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32);
1101: } while(nCount > 0);
1.218 misha 1102: }
1103: }
1104:
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