Annotation of parser3/src/main/pa_common.C, revision 1.258
1.15 paf 1: /** @file
1.16 paf 2: Parser: commonly functions.
3:
1.242 misha 4: Copyright(c) 2001-2009 ArtLebedev Group (http://www.artlebedev.com)
1.101 paf 5: Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
1.16 paf 6:
1.210 paf 7: * BASE64 part
8: * Authors: Michael Zucchi <notzed@ximian.com>
9: * Jeffrey Stedfast <fejj@ximian.com>
10: *
11: * Copyright 2000-2004 Ximian, Inc. (www.ximian.com)
12: *
13: * This program is free software; you can redistribute it and/or modify
14: * it under the terms of the GNU General Public License as published by
15: * the Free Software Foundation; either version 2 of the License, or
16: * (at your option) any later version.
17: *
18: * This program is distributed in the hope that it will be useful,
19: * but WITHOUT ANY WARRANTY; without even the implied warranty of
20: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
21: * GNU General Public License for more details.
22: *
23: * You should have received a copy of the GNU General Public License
24: * along with this program; if not, write to the Free Software
25: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
26: *
27: */
28:
1.258 ! misha 29: static const char * const IDENT_COMMON_C="$Date: 2010-04-19 19:35:55 $";
1.1 paf 30:
31: #include "pa_common.h"
1.4 paf 32: #include "pa_exception.h"
1.154 paf 33: #include "pa_hash.h"
1.14 paf 34: #include "pa_globals.h"
1.154 paf 35: #include "pa_charsets.h"
1.214 paf 36: #include "pa_http.h"
1.223 misha 37: #include "pa_request_charsets.h"
1.237 misha 38: #include "pcre.h"
1.241 misha 39: #include "pa_request.h"
1.98 paf 40:
1.93 paf 41: // some maybe-undefined constants
42:
1.82 paf 43: #ifndef _O_TEXT
44: # define _O_TEXT 0
45: #endif
46: #ifndef _O_BINARY
47: # define _O_BINARY 0
1.47 paf 48: #endif
1.80 paf 49:
1.138 paf 50: #ifdef HAVE_FTRUNCATE
51: # define PA_O_TRUNC 0
52: #else
53: # ifdef _O_TRUNC
54: # define PA_O_TRUNC _O_TRUNC
55: # else
56: # error you must have either ftruncate function or _O_TRUNC bit declared
57: # endif
1.154 paf 58: #endif
1.176 paf 59:
1.154 paf 60: // defines for globals
61:
62: #define FILE_STATUS_NAME "status"
63:
64: // globals
65:
66: const String file_status_name(FILE_STATUS_NAME);
67:
68: // functions
1.127 paf 69:
1.255 misha 70: bool capitalized(const char* s){
71: bool upper=true;
72: for(const char* c=s; *c; c++){
73: if(*c != (upper ? toupper((unsigned char)*c) : tolower((unsigned char)*c)))
74: return false;
75: upper=strchr("-_ ", *c) != 0;
76: }
77: return true;
78: }
79:
1.249 misha 80: const char* capitalize(const char* s){
1.255 misha 81: if(!s || capitalized(s))
82: return s;
83:
1.249 misha 84: char* result=pa_strdup(s);
85: if(result){
86: bool upper=true;
87: for(char* c=result; *c; c++){
88: *c=upper ? (char)toupper((unsigned char)*c) : (char)tolower((unsigned char)*c);
89: upper=strchr("-_ ", *c) != 0;
90: }
91: }
92: return (const char*)result;
93: }
94:
1.154 paf 95: void fix_line_breaks(char *str, size_t& length) {
1.87 paf 96: //_asm int 3;
1.154 paf 97: const char* const eob=str+length;
98: char* dest=str;
1.72 parser 99: // fix DOS: \r\n -> \n
100: // fix Macintosh: \r -> \n
1.154 paf 101: char* bol=str;
1.137 paf 102: while(char* eol=(char*)memchr(bol, '\r', eob -bol)) {
1.72 parser 103: size_t len=eol-bol;
104: if(dest!=bol)
1.126 paf 105: memcpy(dest, bol, len);
1.72 parser 106: dest+=len;
1.126 paf 107: *dest++='\n';
1.72 parser 108:
1.126 paf 109: if(&eol[1]<eob && eol[1]=='\n') { // \r, \n = DOS
1.72 parser 110: bol=eol+2;
1.154 paf 111: length--;
1.126 paf 112: } else // \r, not \n = Macintosh
1.72 parser 113: bol=eol+1;
114: }
1.154 paf 115: // last piece without \r
1.72 parser 116: if(dest!=bol)
1.126 paf 117: memcpy(dest, bol, eob-bol);
1.154 paf 118: str[length]=0; // terminating
1.72 parser 119: }
1.18 paf 120:
1.154 paf 121: char* file_read_text(Request_charsets& charsets,
1.229 misha 122: const String& file_spec,
123: bool fail_on_read_problem,
1.234 misha 124: HashStringValue* params,
125: bool transcode_result) {
1.154 paf 126: File_read_result file=
1.234 misha 127: file_read(charsets, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result);
1.154 paf 128: return file.success?file.str:0;
1.126 paf 129: }
130:
1.241 misha 131: char* file_load_text(Request& r,
132: const String& file_spec,
133: bool fail_on_read_problem,
134: HashStringValue* params,
135: bool transcode_result) {
136: File_read_result file=
137: file_load(r, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result);
138: return file.success?file.str:0;
139: }
140:
1.206 paf 141: /// these options were handled but not checked elsewhere, now check them
1.239 misha 142: int pa_get_valid_file_options_count(HashStringValue& options) {
1.206 paf 143: int result=0;
144: if(options.get(PA_SQL_LIMIT_NAME))
145: result++;
146: if(options.get(PA_SQL_OFFSET_NAME))
147: result++;
148: if(options.get(PA_COLUMN_SEPARATOR_NAME))
149: result++;
150: if(options.get(PA_COLUMN_ENCLOSER_NAME))
151: result++;
1.223 misha 152: if(options.get(PA_CHARSET_NAME))
153: result++;
1.206 paf 154: return result;
155: }
156:
1.123 paf 157: #ifndef DOXYGEN
158: struct File_read_action_info {
1.154 paf 159: char **data; size_t *data_size;
1.188 paf 160: char* buf; size_t offset; size_t count;
1.126 paf 161: };
1.123 paf 162: #endif
1.154 paf 163: static void file_read_action(
1.229 misha 164: struct stat& finfo,
165: int f,
166: const String& file_spec, const char* /*fname*/, bool as_text,
167: void *context) {
1.126 paf 168: File_read_action_info& info=*static_cast<File_read_action_info *>(context);
1.188 paf 169: size_t to_read_size=info.count;
170: if(!to_read_size)
171: to_read_size=(size_t)finfo.st_size;
172: assert( !(info.buf && as_text) );
173: if(to_read_size) {
174: if(info.offset)
175: lseek(f, info.offset, SEEK_SET);
176: *info.data=info.buf
177: ? info.buf
1.253 misha 178: : (char *)pa_malloc_atomic(to_read_size+1);
1.126 paf 179: *info.data_size=(size_t)read(f, *info.data, to_read_size);
1.123 paf 180:
181: if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size)
1.238 misha 182: throw Exception("file.read",
1.123 paf 183: &file_spec,
1.173 paf 184: "read failed: actually read %u bytes count not in [0..%u] valid range",
1.126 paf 185: *info.data_size, to_read_size);
1.123 paf 186: } else { // empty file
1.209 paf 187: // for both, text and binary: for text we need that terminator, for binary we need nonzero pointer to be able to save such files
1.253 misha 188: *info.data=(char *)pa_malloc_atomic(1);
1.209 paf 189: *(char*)(*info.data)=0;
1.123 paf 190: *info.data_size=0;
191: return;
192: }
1.126 paf 193: }
1.241 misha 194:
1.154 paf 195: File_read_result file_read(Request_charsets& charsets, const String& file_spec,
1.229 misha 196: bool as_text, HashStringValue *params,
197: bool fail_on_read_problem,
1.234 misha 198: char* buf, size_t offset, size_t count, bool transcode_text_result) {
1.167 paf 199: File_read_result result={false, 0, 0, 0};
1.241 misha 200: if(params){
201: int valid_options=pa_get_valid_file_options_count(*params);
202: if(valid_options!=params->count())
1.224 misha 203: throw Exception(PARSER_RUNTIME,
1.203 paf 204: 0,
1.250 misha 205: INVALID_OPTION_PASSED);
1.241 misha 206: }
1.203 paf 207:
1.241 misha 208: File_read_action_info info={&result.str, &result.length, buf, offset, count};
1.161 paf 209:
1.241 misha 210: result.success=file_read_action_under_lock(file_spec,
211: "read", file_read_action, &info,
212: as_text, fail_on_read_problem);
1.223 misha 213:
1.241 misha 214: if(as_text){
215: if(result.success){
1.236 misha 216: if(result.length>=3 && strncmp(result.str, "\xEF\xBB\xBF", 3)==0){
1.240 misha 217: // skip UTF-8 signature (BOM code)
1.236 misha 218: result.str+=3;
219: result.length-=3;
220: }
221:
222: if(result.length && transcode_text_result && params){ // must be checked because transcode returns CONST string in case length==0, which contradicts hacking few lines below
1.240 misha 223: if(Value* vcharset_name=params->get(PA_CHARSET_NAME)){
1.236 misha 224: Charset asked_charset=::charsets.get(vcharset_name->as_string().
225: change_case(charsets.source(), String::CC_UPPER));
226:
227: String::C body=String::C(result.str, result.length);
228: body=Charset::transcode(body, asked_charset, charsets.source());
1.123 paf 229:
1.236 misha 230: result.str=const_cast<char*>(body.str); // hacking a little
231: result.length=body.length;
232: }
1.131 paf 233: }
234: }
1.241 misha 235: if(result.length)
236: fix_line_breaks(result.str, result.length);
1.123 paf 237: }
1.241 misha 238:
239: return result;
240: }
241:
242: File_read_result file_load(Request& r, const String& file_spec,
243: bool as_text, HashStringValue *params,
244: bool fail_on_read_problem,
245: char* buf, size_t offset, size_t count, bool transcode_text_result) {
246:
247: File_read_result result={false, 0, 0, 0};
248: if(file_spec.starts_with("http://")) {
249: if(offset || count)
250: throw Exception(PARSER_RUNTIME,
251: 0,
252: "offset and load options are not supported for HTTP:// file load");
253:
254: // fail on read problem
255: File_read_http_result http=pa_internal_file_read_http(r, file_spec, as_text, params, transcode_text_result);
256: result.success=true;
257: result.str=http.str;
258: result.length=http.length;
259: result.headers=http.headers;
260: } else
261: result=
262: file_read(r.charsets, file_spec, as_text, params, fail_on_read_problem, buf, offset, count, transcode_text_result);
1.126 paf 263:
264: return result;
1.123 paf 265: }
266:
1.257 pretende 267:
268: void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) {
1.154 paf 269: #ifdef PA_SAFE_MODE
270: if(finfo.st_uid/*foreign?*/!=geteuid()
271: && finfo.st_gid/*foreign?*/!=getegid())
1.224 misha 272: throw Exception(PARSER_RUNTIME,
1.154 paf 273: &file_spec,
274: "parser is in safe mode: "
275: "reading files of foreign group and user disabled "
276: "[recompile parser with --disable-safe-mode configure option], "
277: "actual filename '%s', "
278: "fuid(%d)!=euid(%d) or fgid(%d)!=egid(%d)",
279: fname,
280: finfo.st_uid, geteuid(),
281: finfo.st_gid, getegid());
1.257 pretende 282: #endif
1.154 paf 283: }
1.257 pretende 284:
1.149 paf 285:
1.154 paf 286: bool file_read_action_under_lock(const String& file_spec,
1.126 paf 287: const char* action_name, File_read_action action, void *context,
288: bool as_text,
1.123 paf 289: bool fail_on_read_problem) {
1.247 misha 290: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.33 paf 291: int f;
292:
293: // first open, next stat:
1.45 paf 294: // directory update of NTFS hard links performed on open.
1.33 paf 295: // ex:
296: // a.html:^test[] and b.html hardlink to a.html
297: // user inserts ! before ^test in a.html
1.126 paf 298: // directory entry of b.html in NTFS not updated at once,
1.35 paf 299: // they delay update till open, so we would receive "!^test[" string
300: // if would do stat, next open.
1.123 paf 301: // later: it seems, even this does not help sometimes
1.229 misha 302: if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) {
1.123 paf 303: try {
1.162 paf 304: if(pa_lock_shared_blocking(f)!=0)
1.126 paf 305: throw Exception("file.lock",
1.123 paf 306: &file_spec,
307: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 308: strerror(errno), errno, fname);
1.123 paf 309:
1.124 paf 310: struct stat finfo;
1.254 misha 311: if(fstat(f, &finfo)!=0)
1.124 paf 312: throw Exception("file.missing", // hardly possible: we just opened it OK
313: &file_spec,
314: "stat failed: %s (%d), actual filename '%s'",
1.154 paf 315: strerror(errno), errno, fname);
1.124 paf 316:
1.149 paf 317: check_safe_mode(finfo, file_spec, fname);
1.32 paf 318:
1.154 paf 319: action(finfo, f, file_spec, fname, as_text, context);
1.123 paf 320: } catch(...) {
1.162 paf 321: pa_unlock(f);close(f);
1.123 paf 322: if(fail_on_read_problem)
1.154 paf 323: rethrow;
1.123 paf 324: return false;
325: }
1.87 paf 326:
1.162 paf 327: pa_unlock(f);close(f);
1.72 parser 328: return true;
1.229 misha 329: } else {
1.118 paf 330: if(fail_on_read_problem)
1.256 misha 331: throw Exception(errno==EACCES?"file.access"
332: :(errno==ENOENT || errno==ENOTDIR || errno==ENODEV)?"file.missing":0,
1.118 paf 333: &file_spec,
1.123 paf 334: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 335: action_name, strerror(errno), errno, fname);
1.118 paf 336: return false;
337: }
1.8 paf 338: }
339:
1.202 paf 340: void create_dir_for_file(const String& file_spec) {
1.63 parser 341: size_t pos_after=1;
1.154 paf 342: size_t pos_before;
343: while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) {
1.247 misha 344: mkdir(file_spec.mid(0, pos_before).taint_cstr(String::L_FILE_SPEC), 0775);
1.63 parser 345: pos_after=pos_before+1;
346: }
347: }
348:
1.98 paf 349: bool file_write_action_under_lock(
1.28 paf 350: const String& file_spec,
1.225 misha 351: const char* action_name,
352: File_write_action action,
353: void *context,
1.126 paf 354: bool as_text,
355: bool do_append,
356: bool do_block,
1.110 paf 357: bool fail_on_lock_problem) {
1.247 misha 358: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.28 paf 359: int f;
1.80 paf 360: if(access(fname, W_OK)!=0) // no
1.126 paf 361: create_dir_for_file(file_spec);
1.50 paf 362:
1.80 paf 363: if((f=open(fname,
364: O_CREAT|O_RDWR
365: |(as_text?_O_TEXT:_O_BINARY)
1.138 paf 366: |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) {
1.162 paf 367: if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) {
1.126 paf 368: Exception e("file.lock",
1.110 paf 369: &file_spec,
370: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 371: strerror(errno), errno, fname);
1.126 paf 372: close(f);
1.110 paf 373: if(fail_on_lock_problem)
374: throw e;
1.98 paf 375: return false;
376: }
1.96 paf 377:
1.158 paf 378: try {
1.254 misha 379: #if (defined(HAVE_FCHMOD) && defined(PA_SAFE_MODE))
380: struct stat finfo;
381: if(fstat(f, &finfo)==0 && finfo.st_mode & 0111)
382: fchmod(f, finfo.st_mode & 0666/*clear executable bits*/); // backward: ignore errors if any
383: #endif
384: action(f, context);
1.158 paf 385: } catch(...) {
1.138 paf 386: #ifdef HAVE_FTRUNCATE
1.104 paf 387: if(!do_append)
1.125 paf 388: ftruncate(f, lseek(f, 0, SEEK_CUR)); // one can not use O_TRUNC, read lower
1.138 paf 389: #endif
1.162 paf 390: pa_unlock(f);close(f);
1.154 paf 391: rethrow;
1.158 paf 392: }
1.80 paf 393:
1.138 paf 394: #ifdef HAVE_FTRUNCATE
1.104 paf 395: if(!do_append)
1.125 paf 396: 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 397: #endif
1.162 paf 398: pa_unlock(f);close(f);
1.98 paf 399: return true;
1.80 paf 400: } else
1.126 paf 401: throw Exception(errno==EACCES?"file.access":0,
1.80 paf 402: &file_spec,
1.96 paf 403: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 404: action_name, strerror(errno), errno, fname);
1.96 paf 405: // here should be nothing, see rethrow above
406: }
407:
408: #ifndef DOXYGEN
409: struct File_write_action_info {
1.250 misha 410: const char* str;
411: size_t length;
1.126 paf 412: };
1.96 paf 413: #endif
414: static void file_write_action(int f, void *context) {
1.126 paf 415: File_write_action_info& info=*static_cast<File_write_action_info *>(context);
1.154 paf 416: if(info.length) {
417: int written=write(f, info.str, info.length);
1.116 paf 418: if(written<0)
1.238 misha 419: throw Exception("file.access",
1.126 paf 420: 0,
421: "write failed: %s (%d)", strerror(errno), errno);
1.113 paf 422: }
1.96 paf 423: }
424: void file_write(
1.250 misha 425: Request_charsets& charsets,
426: const String& file_spec,
427: const char* data,
428: size_t size,
1.126 paf 429: bool as_text,
1.250 misha 430: bool do_append,
431: Charset* asked_charset) {
432:
433: if(as_text && asked_charset){
434: String::C body=String::C(data, size);
435: body=Charset::transcode(body, charsets.source(), *asked_charset);
436: data=body.str;
437: size=body.length;
438: };
439:
1.126 paf 440: File_write_action_info info={data, size};
1.225 misha 441:
1.98 paf 442: file_write_action_under_lock(
1.154 paf 443: file_spec,
1.225 misha 444: "write",
445: file_write_action,
446: &info,
1.154 paf 447: as_text,
448: do_append);
1.30 paf 449: }
450:
1.63 parser 451: // throws nothing! [this is required in file_move & file_delete]
1.50 paf 452: static void rmdir(const String& file_spec, size_t pos_after) {
1.154 paf 453: size_t pos_before;
454: if((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND)
1.126 paf 455: rmdir(file_spec, pos_before+1);
1.50 paf 456:
1.247 misha 457: rmdir(file_spec.mid(0, pos_after-1/* / */).taint_cstr(String::L_FILE_SPEC));
1.50 paf 458: }
1.239 misha 459:
1.164 paf 460: bool file_delete(const String& file_spec, bool fail_on_problem) {
1.247 misha 461: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.54 parser 462: if(unlink(fname)!=0)
1.164 paf 463: if(fail_on_problem)
1.126 paf 464: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.93 paf 465: &file_spec,
466: "unlink failed: %s (%d), actual filename '%s'",
1.154 paf 467: strerror(errno), errno, fname);
1.93 paf 468: else
469: return false;
1.50 paf 470:
1.126 paf 471: rmdir(file_spec, 1);
1.93 paf 472: return true;
1.60 parser 473: }
1.239 misha 474:
1.95 paf 475: void file_move(const String& old_spec, const String& new_spec) {
1.247 misha 476: const char* old_spec_cstr=old_spec.taint_cstr(String::L_FILE_SPEC);
477: const char* new_spec_cstr=new_spec.taint_cstr(String::L_FILE_SPEC);
1.63 parser 478:
1.126 paf 479: create_dir_for_file(new_spec);
1.63 parser 480:
1.60 parser 481: if(rename(old_spec_cstr, new_spec_cstr)!=0)
1.126 paf 482: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.60 parser 483: &old_spec,
484: "rename failed: %s (%d), actual filename '%s' to '%s'",
1.154 paf 485: strerror(errno), errno, old_spec_cstr, new_spec_cstr);
1.63 parser 486:
1.126 paf 487: rmdir(old_spec, 1);
1.31 paf 488: }
489:
1.51 paf 490:
1.126 paf 491: bool entry_exists(const char* fname, struct stat *afinfo) {
1.118 paf 492: struct stat lfinfo;
493: bool result=stat(fname, &lfinfo)==0;
494: if(afinfo)
495: *afinfo=lfinfo;
496: return result;
1.119 paf 497: }
498:
499: bool entry_exists(const String& file_spec) {
1.247 misha 500: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.126 paf 501: return entry_exists(fname, 0);
1.118 paf 502: }
503:
1.51 paf 504: static bool entry_readable(const String& file_spec, bool need_dir) {
1.247 misha 505: char* fname=file_spec.taint_cstrm(String::L_FILE_SPEC);
1.120 paf 506: if(need_dir) {
1.126 paf 507: size_t size=strlen(fname);
1.120 paf 508: while(size) {
1.126 paf 509: char c=fname[size-1];
1.120 paf 510: if(c=='/' || c=='\\')
511: fname[--size]=0;
512: else
513: break;
514: }
515: }
1.51 paf 516: struct stat finfo;
1.118 paf 517: if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) {
1.109 paf 518: bool is_dir=(finfo.st_mode&S_IFDIR) != 0;
1.51 paf 519: return is_dir==need_dir;
520: }
521: return false;
522: }
1.239 misha 523:
1.215 paf 524: bool file_exist(const String& file_spec) {
1.126 paf 525: return entry_readable(file_spec, false);
1.51 paf 526: }
1.239 misha 527:
1.215 paf 528: bool dir_exists(const String& file_spec) {
1.126 paf 529: return entry_readable(file_spec, true);
1.65 parser 530: }
1.239 misha 531:
1.215 paf 532: const String* file_exist(const String& path, const String& name) {
1.154 paf 533: String& result=*new String(path);
534: result << "/";
535: result << name;
1.215 paf 536: return file_exist(result)?&result:0;
1.43 paf 537: }
1.239 misha 538:
1.43 paf 539: bool file_executable(const String& file_spec) {
1.247 misha 540: return access(file_spec.taint_cstr(String::L_FILE_SPEC), X_OK)==0;
1.44 paf 541: }
542:
1.64 parser 543: bool file_stat(const String& file_spec,
1.229 misha 544: size_t& rsize,
545: time_t& ratime,
546: time_t& rmtime,
547: time_t& rctime,
548: bool fail_on_read_problem) {
1.247 misha 549: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.154 paf 550: struct stat finfo;
1.44 paf 551: if(stat(fname, &finfo)!=0)
1.64 parser 552: if(fail_on_read_problem)
1.126 paf 553: throw Exception("file.missing",
1.67 parser 554: &file_spec,
555: "getting file size failed: %s (%d), real filename '%s'",
1.154 paf 556: strerror(errno), errno, fname);
1.64 parser 557: else
558: return false;
1.58 parser 559: rsize=finfo.st_size;
560: ratime=finfo.st_atime;
561: rmtime=finfo.st_mtime;
562: rctime=finfo.st_ctime;
1.64 parser 563: return true;
1.18 paf 564: }
565:
1.126 paf 566: char* getrow(char* *row_ref, char delim) {
1.229 misha 567: char* result=*row_ref;
568: if(result) {
1.126 paf 569: *row_ref=strchr(result, delim);
1.8 paf 570: if(*row_ref)
571: *((*row_ref)++)=0;
572: else if(!*result)
573: return 0;
1.229 misha 574: }
575: return result;
1.8 paf 576: }
577:
1.126 paf 578: char* lsplit(char* string, char delim) {
1.229 misha 579: if(string) {
1.126 paf 580: char* v=strchr(string, delim);
1.8 paf 581: if(v) {
582: *v=0;
583: return v+1;
584: }
1.229 misha 585: }
586: return 0;
1.8 paf 587: }
588:
1.126 paf 589: char* lsplit(char* *string_ref, char delim) {
1.229 misha 590: char* result=*string_ref;
1.126 paf 591: char* next=lsplit(*string_ref, delim);
1.229 misha 592: *string_ref=next;
593: return result;
1.9 paf 594: }
595:
1.126 paf 596: char* rsplit(char* string, char delim) {
1.229 misha 597: if(string) {
1.126 paf 598: char* v=strrchr(string, delim);
1.18 paf 599: if(v) {
1.9 paf 600: *v=0;
601: return v+1;
602: }
1.229 misha 603: }
604: return NULL;
1.10 paf 605: }
606:
1.229 misha 607:
608: // format: %[flags][width][.precision]type http://msdn.microsoft.com/ru-ru/library/56e442dc(en-us,VS.80).aspx
609: // flags: '-', '+', ' ', '#', '0' http://msdn.microsoft.com/ru-ru/library/8aky45ct(en-us,VS.80).aspx
610: // width, precision: non negative decimal number
611: enum FormatType {
612: FormatInvalid,
613: FormatInt,
614: FormatUInt,
615: FormatDouble
616: };
617: FormatType format_type(char* fmt){
618: enum FormatState {
619: Percent,
620: Flags,
621: Width,
622: Precision,
623: Done
624: } state=Percent;
625:
626: FormatType result=FormatInvalid;
627:
628: char* pos=fmt;
629: while(char c=*(pos++)){
630: switch(state){
631: case Percent:
632: if(c=='%'){
633: state=Flags;
634: } else {
635: return FormatInvalid; // 1st char must be '%' only
636: }
637: break;
638: case Flags:
639: if(strchr("-+ #0", c)!=0){
640: break;
641: }
642: // go to the next step
643: case Width:
644: if(c=='.'){
645: state=Precision;
646: break;
647: }
648: // go to the next step
649: case Precision:
650: if(c>='0' && c<='9'){
651: if(state == Flags) state=Width; // no more flags
652: break;
653: } else if(c=='d' || c=='i'){
654: result=FormatInt;
655: } else if(strchr("feEgG", c)!=0){
656: result=FormatDouble;
657: } else if(strchr("uoxX", c)!=0){
658: result=FormatUInt;
659: } else {
660: return FormatInvalid; // invalid char
661: }
662: state=Done;
663: break;
664: case Done:
665: return FormatInvalid; // no chars allowed after 'type'
666: }
667: }
668: return result;
669: }
670:
671:
1.154 paf 672: const char* format(double value, char* fmt) {
1.229 misha 673: char local_buf[MAX_NUMBER];
1.235 misha 674: int size=-1;
1.229 misha 675:
676: if(fmt && strlen(fmt)){
677: switch(format_type(fmt)){
678: case FormatDouble:
679: size=snprintf(local_buf, sizeof(local_buf), fmt, value);
680: break;
681: case FormatInt:
682: size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value);
683: break;
684: case FormatUInt:
1.126 paf 685: size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value);
1.229 misha 686: break;
687: case FormatInvalid:
688: throw Exception(PARSER_RUNTIME,
689: 0,
690: "Incorrect format string '%s' was specified.", fmt);
691: }
692: } else
693: size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value);
694:
695: if(size < 0 || size >= MAX_NUMBER-1){ // on win32 we manually reduce max size while printing
696: throw Exception(PARSER_RUNTIME,
697: 0,
698: "Error occure white executing snprintf with format string '%s'.", fmt);
699: }
700:
1.235 misha 701: return pa_strdup(local_buf, (size_t)size);
1.12 paf 702: }
703:
1.36 paf 704: size_t stdout_write(const void *buf, size_t size) {
1.12 paf 705: #ifdef WIN32
1.187 paf 706: size_t to_write = size;
1.12 paf 707: do{
1.154 paf 708: int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout);
1.12 paf 709: if(chunk_written<=0)
710: break;
711: size-=chunk_written;
1.36 paf 712: buf=((const char*)buf)+chunk_written;
1.126 paf 713: } while(size>0);
1.12 paf 714:
1.187 paf 715: return to_write-size;
1.12 paf 716: #else
1.126 paf 717: return fwrite(buf, 1, size, stdout);
1.12 paf 718: #endif
1.2 paf 719: }
1.14 paf 720:
1.229 misha 721: enum EscapeState {
722: EscapeRest,
723: EscapeFirst,
724: EscapeSecond,
725: EscapeUnicode
726: };
727:
1.236 misha 728: // @todo prescan for reduce required size (unescaped sting in 1 byte charset requires less memory usually)
1.258 ! misha 729: char* unescape_chars(const char* cp, int len, Charset* charset, bool js){
1.236 misha 730: char* s=new(PointerFreeGC) char[len+1]; // must be enough (%uXXXX==6 bytes, max utf-8 char length==6 bytes)
1.230 misha 731: char* dst=s;
1.229 misha 732: EscapeState escapeState=EscapeRest;
733: uint escapedValue=0;
734: int srcPos=0;
1.230 misha 735: short int jsCnt=0;
1.236 misha 736: while(srcPos<len){
1.229 misha 737: uchar c=(uchar)cp[srcPos];
1.258 ! misha 738: if(c=='%' || (c=='\\' && js)){
1.229 misha 739: escapeState=EscapeFirst;
740: } else {
741: switch(escapeState) {
742: case EscapeRest:
1.258 ! misha 743: if(!js && c=='+'){
1.230 misha 744: *dst++=' ';
1.229 misha 745: } else {
1.230 misha 746: *dst++=c;
1.229 misha 747: }
748: break;
749: case EscapeFirst:
1.232 misha 750: if(charset && c=='u'){
1.229 misha 751: // escaped unicode value: %u0430
752: jsCnt=0;
753: escapedValue=0;
754: escapeState=EscapeUnicode;
755: } else {
1.231 misha 756: if(isxdigit(c)){
1.229 misha 757: escapedValue=hex_value[c] << 4;
758: escapeState=EscapeSecond;
759: } else {
1.230 misha 760: *dst++=c;
1.229 misha 761: escapeState=EscapeRest;
762: }
763: }
764: break;
765: case EscapeSecond:
1.231 misha 766: if(isxdigit(c)){
1.229 misha 767: escapedValue+=hex_value[c];
1.230 misha 768: *dst++=(char)escapedValue;
1.229 misha 769: }
770: escapeState=EscapeRest;
771: break;
772: case EscapeUnicode:
1.231 misha 773: if(isxdigit(c)){
1.229 misha 774: escapedValue=(escapedValue << 4) + hex_value[c];
775: if(++jsCnt==4){
1.230 misha 776: // transcode utf8 char to client charset (we can lost some chars here)
1.232 misha 777: charset->store_Char((XMLByte*&)dst, (XMLCh)escapedValue, '?');
1.229 misha 778: escapeState=EscapeRest;
779: }
780: } else {
781: // not full unicode value
782: escapeState=EscapeRest;
783: }
784: break;
785: }
786: }
787:
788: srcPos++;
789: }
790:
1.230 misha 791: *dst=0; // zero-termination
1.229 misha 792: return s;
793: }
1.24 paf 794:
795: #ifdef WIN32
1.126 paf 796: void back_slashes_to_slashes(char* s) {
1.24 paf 797: if(s)
798: for(; *s; s++)
799: if(*s=='\\')
1.126 paf 800: *s='/';
1.24 paf 801: }
1.42 paf 802: /*
1.126 paf 803: void slashes_to_back_slashes(char* s) {
1.42 paf 804: if(s)
805: for(; *s; s++)
806: if(*s=='/')
1.126 paf 807: *s='\\';
1.42 paf 808: }
809: */
1.24 paf 810: #endif
1.41 paf 811:
1.231 misha 812: bool StrStartFromNC(const char* str, const char* substr, bool equal){
1.41 paf 813: while(true) {
1.231 misha 814: if(!(*substr)){
815: if(!(*str))
1.41 paf 816: return true;
817: else
1.231 misha 818: return !equal;
819: }
820: if(!(*str))
821: return false;
822: if(isalpha((unsigned char)*str)) {
823: if(tolower((unsigned char)*str)!=tolower((unsigned char)*substr))
1.41 paf 824: return false;
1.231 misha 825: } else if((*str) != (*substr))
1.41 paf 826: return false;
1.231 misha 827: str++;
828: substr++;
1.41 paf 829: }
1.57 parser 830: }
831:
1.232 misha 832: size_t strpos(const char *str, const char *substr) {
833: const char *p = strstr(str, substr);
834: return (p==0)?STRING_NOT_FOUND:p-str;
835: }
836:
837: // content-type: xxx; charset=WE-NEED-THIS
838: // content-type: xxx; charset="WE-NEED-THIS"
839: // content-type: xxx; charset="WE-NEED-THIS";
1.248 misha 840: Charset* detect_charset(const char* content_type){
1.233 misha 841: if(content_type){
1.245 misha 842: char* CONTENT_TYPE=pa_strdup(content_type);
1.248 misha 843:
844: for(char *p=CONTENT_TYPE; *p; p++)
845: *p=(char)toupper((unsigned char)*p);
1.233 misha 846:
847: if(const char* begin=strstr(CONTENT_TYPE, "CHARSET=")){
848: begin+=8; // skip "CHARSET="
849: char* end=0;
850: if(*begin && (*begin=='"' || *begin =='\'')){
851: char quote=*begin;
852: begin++;
853: end=(char*)strchr(begin, quote);
854: }
855: if(!end)
856: end=(char*)strchr(begin, ';');
857:
1.244 misha 858: if(end)
1.233 misha 859: *end=0; // terminator
860:
1.245 misha 861: return *begin?&charsets.get(begin):0;
1.232 misha 862: }
863: }
864: return 0;
865: }
866:
867:
1.84 paf 868: static bool isLeap(int year) {
1.229 misha 869: return !(
870: (year % 4) || ((year % 400) && !(year % 100))
871: );
1.57 parser 872: }
873:
874: int getMonthDays(int year, int month) {
1.220 misha 875: static int monthDays[]={
1.229 misha 876: 31,
877: 28,
878: 31,
879: 30,
880: 31,
881: 30,
882: 31,
883: 31,
884: 30,
885: 31,
886: 30,
887: 31
888: };
1.228 misha 889: return (month == 1 /* january -- 0 */ && isLeap(year)) ? 29 : monthDays[month];
1.41 paf 890: }
1.69 parser 891:
1.226 misha 892: int remove_crlf(char* start, char* end) {
893: char* from=start;
894: char* to=start;
895: bool skip=false;
896: while(from < end){
897: switch(*from){
898: case '\n':
899: case '\r':
900: case '\t':
901: case ' ':
902: if(!skip){
903: *to=' ';
904: to++;
905: skip=true;
906: }
907: break;
908: default:
909: if(from != to)
910: *to=*from;
911: to++;
912: skip=false;
1.69 parser 913: }
1.226 misha 914: from++;
915: }
916: return to-start;
1.91 paf 917: }
918:
919:
920: /// must be last in this file
921: #undef vsnprintf
1.126 paf 922: int __vsnprintf(char* b, size_t s, const char* f, va_list l) {
1.91 paf 923: if(!s)
924: return 0;
925:
926: int r;
927: // note: on win32& maybe somewhere else
928: // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing
929: --s;
1.172 paf 930:
931: // clients do not check for negative 's', feature: ignore such prints
932: if((ssize_t)s<0)
933: return 0;
934:
1.91 paf 935: #if _MSC_VER
936: /*
937: win32:
938: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm
939:
1.154 paf 940: if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 is returned
1.91 paf 941: */
1.126 paf 942: r=_vsnprintf(b, s, f, l);
1.91 paf 943: if(r<0)
944: r=s;
945: #else
1.126 paf 946: r=vsnprintf(b, s, f, l);
1.91 paf 947: /*
948: solaris:
949: man vsnprintf
950:
951: The snprintf() function returns the number of characters
952: formatted, that is, the number of characters that would have
953: been written to the buffer if it were large enough. If the
954: value of n is 0 on a call to snprintf(), an unspecified
955: value less than 1 is returned.
956: */
957:
958: if(r<0)
959: r=0;
1.167 paf 960: else if((size_t)r>s)
1.91 paf 961: r=s;
962: #endif
963: b[r]=0;
964: return r;
965: }
966:
1.126 paf 967: int __snprintf(char* b, size_t s, const char* f, ...) {
1.91 paf 968: va_list l;
1.241 misha 969: va_start(l, f);
970: int r=__vsnprintf(b, s, f, l);
971: va_end(l);
1.91 paf 972: return r;
1.178 paf 973: }
974:
975: /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */
976: /*
977: * Authors: Michael Zucchi <notzed@helixcode.com>
978: * Jeffrey Stedfast <fejj@helixcode.com>
979: *
980: * Copyright 2000 Helix Code, Inc. (www.helixcode.com)
981: *
982: * This program is free software; you can redistribute it and/or modify
983: * it under the terms of the GNU General Public License as published by
984: * the Free Software Foundation; either version 2 of the License, or
985: * (at your option) any later version.
986: *
987: * This program is distributed in the hope that it will be useful,
988: * but WITHOUT ANY WARRANTY; without even the implied warranty of
989: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
990: * GNU General Public License for more details.
991: *
992: * You should have received a copy of the GNU General Public License
993: * along with this program; if not, write to the Free Software
994: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
995: *
996: */
997: static char *base64_alphabet =
998: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
999:
1000: /**
1001: * g_mime_utils_base64_encode_step:
1002: * @in: input stream
1003: * @inlen: length of the input
1004: * @out: output string
1005: * @state: holds the number of bits that are stored in @save
1006: * @save: leftover bits that have not yet been encoded
1007: *
1008: * Base64 encodes a chunk of data. Performs an 'encode step', only
1009: * encodes blocks of 3 characters to the output at a time, saves
1010: * left-over state in state and save (initialise to 0 on first
1011: * invocation).
1012: *
1013: * Returns the number of bytes encoded.
1014: **/
1.252 misha 1015:
1016: #define BASE64_GROUPS_IN_LINE 19
1017:
1.178 paf 1018: static size_t
1019: g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1020: {
1.186 paf 1021: register const unsigned char *inptr;
1.178 paf 1022: register unsigned char *outptr;
1023:
1024: if (inlen <= 0)
1025: return 0;
1026:
1027: inptr = in;
1028: outptr = out;
1029:
1030: if (inlen + ((unsigned char *)save)[0] > 2) {
1031: const unsigned char *inend = in + inlen - 2;
1032: register int c1 = 0, c2 = 0, c3 = 0;
1033: register int already;
1034:
1035: already = *state;
1036:
1037: switch (((char *)save)[0]) {
1038: case 1: c1 = ((unsigned char *)save)[1]; goto skip1;
1039: case 2: c1 = ((unsigned char *)save)[1];
1040: c2 = ((unsigned char *)save)[2]; goto skip2;
1041: }
1042:
1043: /* yes, we jump into the loop, no i'm not going to change it, its beautiful! */
1044: while (inptr < inend) {
1045: c1 = *inptr++;
1046: skip1:
1047: c2 = *inptr++;
1048: skip2:
1049: c3 = *inptr++;
1050: *outptr++ = base64_alphabet [c1 >> 2];
1051: *outptr++ = base64_alphabet [(c2 >> 4) | ((c1 & 0x3) << 4)];
1052: *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)];
1053: *outptr++ = base64_alphabet [c3 & 0x3f];
1054: /* this is a bit ugly ... */
1.252 misha 1055: if ((++already) >= BASE64_GROUPS_IN_LINE) {
1.178 paf 1056: *outptr++ = '\n';
1057: already = 0;
1058: }
1059: }
1060:
1061: ((unsigned char *)save)[0] = 0;
1062: inlen = 2 - (inptr - inend);
1063: *state = already;
1064: }
1065:
1066: //d(printf ("state = %d, inlen = %d\n", (int)((char *)save)[0], inlen));
1067:
1068: if (inlen > 0) {
1069: register char *saveout;
1070:
1071: /* points to the slot for the next char to save */
1072: saveout = & (((char *)save)[1]) + ((char *)save)[0];
1073:
1074: /* inlen can only be 0 1 or 2 */
1075: switch (inlen) {
1076: case 2: *saveout++ = *inptr++;
1077: case 1: *saveout++ = *inptr++;
1078: }
1.216 paf 1079: *(char *)save = *(char *)save+(char)inlen;
1.178 paf 1080: }
1081:
1082: /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n",
1083: (int)((char *)save)[0],
1084: (int)((char *)save)[1],
1085: (int)((char *)save)[2]));*/
1086:
1087: return (outptr - out);
1088: }
1089:
1090: /**
1091: * g_mime_utils_base64_encode_close:
1092: * @in: input stream
1093: * @inlen: length of the input
1094: * @out: output string
1095: * @state: holds the number of bits that are stored in @save
1096: * @save: leftover bits that have not yet been encoded
1097: *
1098: * Base64 encodes the input stream to the output stream. Call this
1099: * when finished encoding data with g_mime_utils_base64_encode_step to
1100: * flush off the last little bit.
1101: *
1102: * Returns the number of bytes encoded.
1103: **/
1104: static size_t
1105: g_mime_utils_base64_encode_close (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1106: {
1107: unsigned char *outptr = out;
1108: int c1, c2;
1109:
1110: if (inlen > 0)
1111: outptr += g_mime_utils_base64_encode_step (in, inlen, outptr, state, save);
1112:
1113: c1 = ((unsigned char *)save)[1];
1114: c2 = ((unsigned char *)save)[2];
1115:
1116: switch (((unsigned char *)save)[0]) {
1117: case 2:
1118: outptr[2] = base64_alphabet [(c2 & 0x0f) << 2];
1119: goto skip;
1120: case 1:
1121: outptr[2] = '=';
1122: skip:
1123: outptr[0] = base64_alphabet [c1 >> 2];
1124: outptr[1] = base64_alphabet [c2 >> 4 | ((c1 & 0x3) << 4)];
1125: outptr[3] = '=';
1126: outptr += 4;
1127: break;
1128: }
1129:
1130: *outptr++ = 0;
1131:
1132: *save = 0;
1133: *state = 0;
1134:
1135: return (outptr - out);
1136: }
1137:
1.210 paf 1138: static unsigned char gmime_base64_rank[256] = {
1139: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1140: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1141: 255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63,
1142: 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255, 0,255,255,
1143: 255, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
1144: 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255,
1145: 255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
1146: 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255,
1147: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1148: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1149: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1150: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1151: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1152: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1153: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1154: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1155: };
1156:
1157: /**
1158: * g_mime_utils_base64_decode_step:
1159: * @in: input stream
1160: * @inlen: max length of data to decode
1161: * @out: output stream
1162: * @state: holds the number of bits that are stored in @save
1163: * @save: leftover bits that have not yet been decoded
1164: *
1165: * Decodes a chunk of base64 encoded data.
1166: *
1167: * Returns the number of bytes decoded (which have been dumped in @out).
1168: **/
1169: size_t
1170: g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1171: {
1.213 paf 1172: const unsigned char *inptr;
1173: unsigned char *outptr;
1.210 paf 1174: const unsigned char *inend;
1.213 paf 1175: int saved;
1.210 paf 1176: unsigned char c;
1177: int i;
1178:
1179: inend = in + inlen;
1180: outptr = out;
1181:
1182: /* convert 4 base64 bytes to 3 normal bytes */
1183: saved = *save;
1184: i = *state;
1185: inptr = in;
1186: while (inptr < inend) {
1187: c = gmime_base64_rank[*inptr++];
1188: if (c != 0xff) {
1189: saved = (saved << 6) | c;
1190: i++;
1191: if (i == 4) {
1.217 paf 1192: *outptr++ = (unsigned char)(saved >> 16);
1193: *outptr++ = (unsigned char)(saved >> 8);
1194: *outptr++ = (unsigned char)(saved);
1.210 paf 1195: i = 0;
1196: }
1197: }
1198: }
1199:
1200: *save = saved;
1201: *state = i;
1202:
1203: /* quick scan back for '=' on the end somewhere */
1204: /* fortunately we can drop 1 output char for each trailing = (upto 2) */
1205: i = 2;
1206: while (inptr > in && i) {
1207: inptr--;
1208: if (gmime_base64_rank[*inptr] != 0xff) {
1209: if (*inptr == '=' && outptr > out)
1210: outptr--;
1211: i--;
1212: }
1213: }
1214:
1215: /* if i != 0 then there is a truncation error! */
1216: return (outptr - out);
1217: }
1218:
1219:
1.239 misha 1220: char* pa_base64_encode(const char *in, size_t in_size){
1.252 misha 1221: size_t new_size = ((in_size / 3 + 1) * 4);
1222: new_size += new_size / (BASE64_GROUPS_IN_LINE * 4)/*new lines*/ + 1/*zero terminator*/;
1223: char* result = new(PointerFreeGC) char[new_size];
1.178 paf 1224: int state=0;
1225: int save=0;
1.183 paf 1226: #ifndef NDEBUG
1227: size_t filled=
1228: #endif
1.251 misha 1229: g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size, (unsigned char*)result, &state, &save);
1230:
1231: //throw Exception(PARSER_RUNTIME, 0, "%d %d %d", in_size, new_size, filled);
1232: assert(filled <= new_size);
1.178 paf 1233:
1234: return result;
1.98 paf 1235: }
1.210 paf 1236:
1.222 misha 1237:
1.239 misha 1238: char* pa_base64_encode(const String& file_spec){
1.222 misha 1239: unsigned char* base64=0;
1240: File_base64_action_info info={&base64};
1241:
1242: file_read_action_under_lock(file_spec,
1243: "pa_base64_encode", file_base64_file_action, &info);
1244:
1245: return (char*)base64;
1246: }
1247:
1248:
1249: static void file_base64_file_action(
1.229 misha 1250: struct stat& finfo,
1251: int f,
1252: const String&, const char* /*fname*/, bool,
1253: void *context) {
1.222 misha 1254:
1255: if(finfo.st_size) {
1256: File_base64_action_info& info=*static_cast<File_base64_action_info *>(context);
1257: *info.base64=new(PointerFreeGC) unsigned char[finfo.st_size * 2 + 6];
1258: unsigned char* base64 = *info.base64;
1259: int state=0;
1260: int save=0;
1261: int nCount;
1262: do {
1263: unsigned char buffer[FILE_BUFFER_SIZE];
1264: nCount = file_block_read(f, buffer, sizeof(buffer));
1265: if( nCount ){
1266: size_t filled=g_mime_utils_base64_encode_step ((const unsigned char*)buffer, nCount, base64, &state, &save);
1267: base64+=filled;
1268: }
1269: } while(nCount > 0);
1270: g_mime_utils_base64_encode_close (0, 0, base64, &state, &save);
1271: }
1272: }
1273:
1.239 misha 1274: void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size) {
1.210 paf 1275: /* wont go to more than had (overly conservative) */
1.211 paf 1276: result=new(PointerFreeGC) char[in_size+1/*terminator*/];
1.210 paf 1277: int state=0;
1278: int save=0;
1279: result_size=
1280: g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size,
1281: (unsigned char*)result, &state, &save);
1282: assert(result_size <= in_size);
1.211 paf 1283: result[result_size]=0; // for text files
1.210 paf 1284: }
1.218 misha 1285:
1286:
1.221 misha 1287: int file_block_read(const int f, unsigned char* buffer, const size_t size){
1288: int nCount = read(f, buffer, size);
1289: if (nCount < 0)
1.238 misha 1290: throw Exception("file.read",
1.221 misha 1291: 0,
1292: "read failed: %s (%d)", strerror(errno), errno);
1293: return nCount;
1294: }
1295:
1.239 misha 1296: const unsigned long pa_crc32(const char *in, size_t in_size){
1.218 misha 1297: unsigned long crc32=0xFFFFFFFF;
1.220 misha 1298:
1.240 misha 1299: InitCrc32Table();
1.239 misha 1300: for(size_t i = 0; i<in_size; i++)
1301: CalcCrc32(in[i], crc32);
1.220 misha 1302:
1.218 misha 1303: return ~crc32;
1304: }
1305:
1.239 misha 1306: const unsigned long pa_crc32(const String& file_spec){
1.218 misha 1307: unsigned long crc32=0xFFFFFFFF;
1308: file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32);
1309: return ~crc32;
1310: }
1311:
1312: static void file_crc32_file_action(
1.229 misha 1313: struct stat& finfo,
1314: int f,
1315: const String&, const char* /*fname*/, bool,
1.239 misha 1316: void *context) {
1.218 misha 1317: unsigned long& crc32=*static_cast<unsigned long *>(context);
1318: if(finfo.st_size) {
1319: InitCrc32Table();
1.220 misha 1320: int nCount=0;
1.218 misha 1321: do {
1.221 misha 1322: unsigned char buffer[FILE_BUFFER_SIZE];
1323: nCount = file_block_read(f, buffer, sizeof(buffer));
1.220 misha 1324: for(int i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32);
1325: } while(nCount > 0);
1.218 misha 1326: }
1327: }
1328:
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