Annotation of parser3/src/main/pa_common.C, revision 1.262
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:
1.262 ! misha 29: static const char * const IDENT_COMMON_C="$Date: 2010-07-05 01:08:24 $";
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)
1.260 misha 105: memmove(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)
1.260 misha 117: memmove(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())
1.262 ! misha 203: throw Exception(PARSER_RUNTIME, 0, CALLED_WITH_INVALID_OPTION);
1.241 misha 204: }
1.203 paf 205:
1.241 misha 206: File_read_action_info info={&result.str, &result.length, buf, offset, count};
1.161 paf 207:
1.241 misha 208: result.success=file_read_action_under_lock(file_spec,
209: "read", file_read_action, &info,
210: as_text, fail_on_read_problem);
1.223 misha 211:
1.241 misha 212: if(as_text){
213: if(result.success){
1.236 misha 214: if(result.length>=3 && strncmp(result.str, "\xEF\xBB\xBF", 3)==0){
1.240 misha 215: // skip UTF-8 signature (BOM code)
1.236 misha 216: result.str+=3;
217: result.length-=3;
218: }
219:
220: 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 221: if(Value* vcharset_name=params->get(PA_CHARSET_NAME)){
1.236 misha 222: Charset asked_charset=::charsets.get(vcharset_name->as_string().
223: change_case(charsets.source(), String::CC_UPPER));
224:
225: String::C body=String::C(result.str, result.length);
226: body=Charset::transcode(body, asked_charset, charsets.source());
1.123 paf 227:
1.236 misha 228: result.str=const_cast<char*>(body.str); // hacking a little
229: result.length=body.length;
230: }
1.131 paf 231: }
232: }
1.241 misha 233: if(result.length)
234: fix_line_breaks(result.str, result.length);
1.123 paf 235: }
1.241 misha 236:
237: return result;
238: }
239:
240: File_read_result file_load(Request& r, const String& file_spec,
241: bool as_text, HashStringValue *params,
242: bool fail_on_read_problem,
243: char* buf, size_t offset, size_t count, bool transcode_text_result) {
244:
245: File_read_result result={false, 0, 0, 0};
246: if(file_spec.starts_with("http://")) {
247: if(offset || count)
248: throw Exception(PARSER_RUNTIME,
249: 0,
250: "offset and load options are not supported for HTTP:// file load");
251:
252: // fail on read problem
253: File_read_http_result http=pa_internal_file_read_http(r, file_spec, as_text, params, transcode_text_result);
254: result.success=true;
255: result.str=http.str;
256: result.length=http.length;
257: result.headers=http.headers;
258: } else
259: result=
260: file_read(r.charsets, file_spec, as_text, params, fail_on_read_problem, buf, offset, count, transcode_text_result);
1.126 paf 261:
262: return result;
1.123 paf 263: }
264:
1.257 pretende 265:
1.154 paf 266: #ifdef PA_SAFE_MODE
1.259 misha 267: void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) {
1.154 paf 268: if(finfo.st_uid/*foreign?*/!=geteuid()
269: && finfo.st_gid/*foreign?*/!=getegid())
1.224 misha 270: throw Exception(PARSER_RUNTIME,
1.154 paf 271: &file_spec,
272: "parser is in safe mode: "
273: "reading files of foreign group and user disabled "
274: "[recompile parser with --disable-safe-mode configure option], "
275: "actual filename '%s', "
276: "fuid(%d)!=euid(%d) or fgid(%d)!=egid(%d)",
277: fname,
278: finfo.st_uid, geteuid(),
1.259 misha 279: finfo.st_gid, getegid());
280: }
281: #else
282: void check_safe_mode(struct stat, const String&, const char*) {
283: }
1.257 pretende 284: #endif
1.259 misha 285:
1.257 pretende 286:
1.149 paf 287:
1.154 paf 288: bool file_read_action_under_lock(const String& file_spec,
1.126 paf 289: const char* action_name, File_read_action action, void *context,
290: bool as_text,
1.123 paf 291: bool fail_on_read_problem) {
1.247 misha 292: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.33 paf 293: int f;
294:
295: // first open, next stat:
1.45 paf 296: // directory update of NTFS hard links performed on open.
1.33 paf 297: // ex:
298: // a.html:^test[] and b.html hardlink to a.html
299: // user inserts ! before ^test in a.html
1.126 paf 300: // directory entry of b.html in NTFS not updated at once,
1.35 paf 301: // they delay update till open, so we would receive "!^test[" string
302: // if would do stat, next open.
1.123 paf 303: // later: it seems, even this does not help sometimes
1.229 misha 304: if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) {
1.123 paf 305: try {
1.162 paf 306: if(pa_lock_shared_blocking(f)!=0)
1.126 paf 307: throw Exception("file.lock",
1.123 paf 308: &file_spec,
309: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 310: strerror(errno), errno, fname);
1.123 paf 311:
1.124 paf 312: struct stat finfo;
1.254 misha 313: if(fstat(f, &finfo)!=0)
1.124 paf 314: throw Exception("file.missing", // hardly possible: we just opened it OK
315: &file_spec,
316: "stat failed: %s (%d), actual filename '%s'",
1.154 paf 317: strerror(errno), errno, fname);
1.124 paf 318:
1.149 paf 319: check_safe_mode(finfo, file_spec, fname);
1.32 paf 320:
1.154 paf 321: action(finfo, f, file_spec, fname, as_text, context);
1.123 paf 322: } catch(...) {
1.162 paf 323: pa_unlock(f);close(f);
1.123 paf 324: if(fail_on_read_problem)
1.154 paf 325: rethrow;
1.123 paf 326: return false;
327: }
1.87 paf 328:
1.162 paf 329: pa_unlock(f);close(f);
1.72 parser 330: return true;
1.229 misha 331: } else {
1.118 paf 332: if(fail_on_read_problem)
1.256 misha 333: throw Exception(errno==EACCES?"file.access"
334: :(errno==ENOENT || errno==ENOTDIR || errno==ENODEV)?"file.missing":0,
1.118 paf 335: &file_spec,
1.123 paf 336: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 337: action_name, strerror(errno), errno, fname);
1.118 paf 338: return false;
339: }
1.8 paf 340: }
341:
1.202 paf 342: void create_dir_for_file(const String& file_spec) {
1.63 parser 343: size_t pos_after=1;
1.154 paf 344: size_t pos_before;
345: while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) {
1.247 misha 346: mkdir(file_spec.mid(0, pos_before).taint_cstr(String::L_FILE_SPEC), 0775);
1.63 parser 347: pos_after=pos_before+1;
348: }
349: }
350:
1.98 paf 351: bool file_write_action_under_lock(
1.28 paf 352: const String& file_spec,
1.225 misha 353: const char* action_name,
354: File_write_action action,
355: void *context,
1.126 paf 356: bool as_text,
357: bool do_append,
358: bool do_block,
1.110 paf 359: bool fail_on_lock_problem) {
1.247 misha 360: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.28 paf 361: int f;
1.80 paf 362: if(access(fname, W_OK)!=0) // no
1.126 paf 363: create_dir_for_file(file_spec);
1.50 paf 364:
1.80 paf 365: if((f=open(fname,
366: O_CREAT|O_RDWR
367: |(as_text?_O_TEXT:_O_BINARY)
1.138 paf 368: |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) {
1.162 paf 369: if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) {
1.126 paf 370: Exception e("file.lock",
1.110 paf 371: &file_spec,
372: "shared lock failed: %s (%d), actual filename '%s'",
1.154 paf 373: strerror(errno), errno, fname);
1.126 paf 374: close(f);
1.110 paf 375: if(fail_on_lock_problem)
376: throw e;
1.98 paf 377: return false;
378: }
1.96 paf 379:
1.158 paf 380: try {
1.254 misha 381: #if (defined(HAVE_FCHMOD) && defined(PA_SAFE_MODE))
382: struct stat finfo;
383: if(fstat(f, &finfo)==0 && finfo.st_mode & 0111)
384: fchmod(f, finfo.st_mode & 0666/*clear executable bits*/); // backward: ignore errors if any
385: #endif
386: action(f, context);
1.158 paf 387: } catch(...) {
1.138 paf 388: #ifdef HAVE_FTRUNCATE
1.104 paf 389: if(!do_append)
1.125 paf 390: ftruncate(f, lseek(f, 0, SEEK_CUR)); // one can not use O_TRUNC, read lower
1.138 paf 391: #endif
1.162 paf 392: pa_unlock(f);close(f);
1.154 paf 393: rethrow;
1.158 paf 394: }
1.80 paf 395:
1.138 paf 396: #ifdef HAVE_FTRUNCATE
1.104 paf 397: if(!do_append)
1.125 paf 398: 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 399: #endif
1.162 paf 400: pa_unlock(f);close(f);
1.98 paf 401: return true;
1.80 paf 402: } else
1.126 paf 403: throw Exception(errno==EACCES?"file.access":0,
1.80 paf 404: &file_spec,
1.96 paf 405: "%s failed: %s (%d), actual filename '%s'",
1.154 paf 406: action_name, strerror(errno), errno, fname);
1.96 paf 407: // here should be nothing, see rethrow above
408: }
409:
410: #ifndef DOXYGEN
411: struct File_write_action_info {
1.250 misha 412: const char* str;
413: size_t length;
1.126 paf 414: };
1.96 paf 415: #endif
416: static void file_write_action(int f, void *context) {
1.126 paf 417: File_write_action_info& info=*static_cast<File_write_action_info *>(context);
1.154 paf 418: if(info.length) {
419: int written=write(f, info.str, info.length);
1.116 paf 420: if(written<0)
1.238 misha 421: throw Exception("file.access",
1.126 paf 422: 0,
423: "write failed: %s (%d)", strerror(errno), errno);
1.113 paf 424: }
1.96 paf 425: }
426: void file_write(
1.250 misha 427: Request_charsets& charsets,
428: const String& file_spec,
429: const char* data,
430: size_t size,
1.126 paf 431: bool as_text,
1.250 misha 432: bool do_append,
433: Charset* asked_charset) {
434:
435: if(as_text && asked_charset){
436: String::C body=String::C(data, size);
437: body=Charset::transcode(body, charsets.source(), *asked_charset);
438: data=body.str;
439: size=body.length;
440: };
441:
1.126 paf 442: File_write_action_info info={data, size};
1.225 misha 443:
1.98 paf 444: file_write_action_under_lock(
1.154 paf 445: file_spec,
1.225 misha 446: "write",
447: file_write_action,
448: &info,
1.154 paf 449: as_text,
450: do_append);
1.30 paf 451: }
452:
1.261 misha 453: static size_t get_dir(char* fname, size_t helper_length){
454: bool dir=false;
455: size_t pos=0;
456: for(pos=helper_length; pos; pos--){
457: char c=fname[pos-1];
458: if(c=='/' || c=='\\'){
459: fname[pos-1]=0;
460: dir=true;
461: } else if(dir) break;
462: }
463: return pos;
464: }
465:
466: static bool entry_readable(char* fname, bool need_dir) {
467: if(need_dir){
468: size_t size=strlen(fname);
469: while(size) {
470: char c=fname[size-1];
471: if(c=='/' || c=='\\')
472: fname[--size]=0;
473: else
474: break;
475: }
476: }
477:
478: struct stat finfo;
479: if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) {
480: bool is_dir=(finfo.st_mode&S_IFDIR) != 0;
481: return is_dir==need_dir;
482: }
483: return false;
484: }
485:
486: static bool entry_readable(const String& file_spec, bool need_dir) {
487: return entry_readable(file_spec.taint_cstrm(String::L_FILE_SPEC), need_dir);
488: }
489:
1.63 parser 490: // throws nothing! [this is required in file_move & file_delete]
1.261 misha 491: static void rmdir(const String& file_spec, size_t pos_after=0) {
492: char* dir_spec=file_spec.taint_cstrm(String::L_FILE_SPEC);
493: size_t length=strlen(dir_spec);
494: while( (length=get_dir(dir_spec, length)) && (length > pos_after) ){
495: #ifdef WIN32
496: if(!entry_readable(dir_spec, true))
497: break;
498: DWORD attrs=GetFileAttributes(dir_spec);
499: if(
500: (attrs==INVALID_FILE_ATTRIBUTES)
501: || !(attrs & FILE_ATTRIBUTE_DIRECTORY)
502: || (attrs & FILE_ATTRIBUTE_REPARSE_POINT)
503: )
504: break;
505: #endif
506: if( rmdir(dir_spec) )
507: break;
508: };
1.50 paf 509: }
1.239 misha 510:
1.164 paf 511: bool file_delete(const String& file_spec, bool fail_on_problem) {
1.247 misha 512: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.54 parser 513: if(unlink(fname)!=0)
1.164 paf 514: if(fail_on_problem)
1.126 paf 515: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.93 paf 516: &file_spec,
517: "unlink failed: %s (%d), actual filename '%s'",
1.154 paf 518: strerror(errno), errno, fname);
1.93 paf 519: else
520: return false;
1.50 paf 521:
1.126 paf 522: rmdir(file_spec, 1);
1.93 paf 523: return true;
1.60 parser 524: }
1.239 misha 525:
1.95 paf 526: void file_move(const String& old_spec, const String& new_spec) {
1.247 misha 527: const char* old_spec_cstr=old_spec.taint_cstr(String::L_FILE_SPEC);
528: const char* new_spec_cstr=new_spec.taint_cstr(String::L_FILE_SPEC);
1.63 parser 529:
1.126 paf 530: create_dir_for_file(new_spec);
1.63 parser 531:
1.60 parser 532: if(rename(old_spec_cstr, new_spec_cstr)!=0)
1.126 paf 533: throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0,
1.60 parser 534: &old_spec,
535: "rename failed: %s (%d), actual filename '%s' to '%s'",
1.154 paf 536: strerror(errno), errno, old_spec_cstr, new_spec_cstr);
1.63 parser 537:
1.126 paf 538: rmdir(old_spec, 1);
1.31 paf 539: }
540:
1.51 paf 541:
1.126 paf 542: bool entry_exists(const char* fname, struct stat *afinfo) {
1.118 paf 543: struct stat lfinfo;
544: bool result=stat(fname, &lfinfo)==0;
545: if(afinfo)
546: *afinfo=lfinfo;
547: return result;
1.119 paf 548: }
549:
550: bool entry_exists(const String& file_spec) {
1.247 misha 551: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.126 paf 552: return entry_exists(fname, 0);
1.118 paf 553: }
554:
1.215 paf 555: bool file_exist(const String& file_spec) {
1.126 paf 556: return entry_readable(file_spec, false);
1.51 paf 557: }
1.239 misha 558:
1.215 paf 559: bool dir_exists(const String& file_spec) {
1.126 paf 560: return entry_readable(file_spec, true);
1.65 parser 561: }
1.239 misha 562:
1.215 paf 563: const String* file_exist(const String& path, const String& name) {
1.154 paf 564: String& result=*new String(path);
565: result << "/";
566: result << name;
1.215 paf 567: return file_exist(result)?&result:0;
1.43 paf 568: }
1.239 misha 569:
1.43 paf 570: bool file_executable(const String& file_spec) {
1.247 misha 571: return access(file_spec.taint_cstr(String::L_FILE_SPEC), X_OK)==0;
1.44 paf 572: }
573:
1.64 parser 574: bool file_stat(const String& file_spec,
1.229 misha 575: size_t& rsize,
576: time_t& ratime,
577: time_t& rmtime,
578: time_t& rctime,
579: bool fail_on_read_problem) {
1.247 misha 580: const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC);
1.154 paf 581: struct stat finfo;
1.44 paf 582: if(stat(fname, &finfo)!=0)
1.64 parser 583: if(fail_on_read_problem)
1.126 paf 584: throw Exception("file.missing",
1.67 parser 585: &file_spec,
586: "getting file size failed: %s (%d), real filename '%s'",
1.154 paf 587: strerror(errno), errno, fname);
1.64 parser 588: else
589: return false;
1.58 parser 590: rsize=finfo.st_size;
591: ratime=finfo.st_atime;
592: rmtime=finfo.st_mtime;
593: rctime=finfo.st_ctime;
1.64 parser 594: return true;
1.18 paf 595: }
596:
1.126 paf 597: char* getrow(char* *row_ref, char delim) {
1.229 misha 598: char* result=*row_ref;
599: if(result) {
1.126 paf 600: *row_ref=strchr(result, delim);
1.8 paf 601: if(*row_ref)
602: *((*row_ref)++)=0;
603: else if(!*result)
604: return 0;
1.229 misha 605: }
606: return result;
1.8 paf 607: }
608:
1.126 paf 609: char* lsplit(char* string, char delim) {
1.229 misha 610: if(string) {
1.126 paf 611: char* v=strchr(string, delim);
1.8 paf 612: if(v) {
613: *v=0;
614: return v+1;
615: }
1.229 misha 616: }
617: return 0;
1.8 paf 618: }
619:
1.126 paf 620: char* lsplit(char* *string_ref, char delim) {
1.229 misha 621: char* result=*string_ref;
1.126 paf 622: char* next=lsplit(*string_ref, delim);
1.229 misha 623: *string_ref=next;
624: return result;
1.9 paf 625: }
626:
1.126 paf 627: char* rsplit(char* string, char delim) {
1.229 misha 628: if(string) {
1.126 paf 629: char* v=strrchr(string, delim);
1.18 paf 630: if(v) {
1.9 paf 631: *v=0;
632: return v+1;
633: }
1.229 misha 634: }
635: return NULL;
1.10 paf 636: }
637:
1.229 misha 638:
639: // format: %[flags][width][.precision]type http://msdn.microsoft.com/ru-ru/library/56e442dc(en-us,VS.80).aspx
640: // flags: '-', '+', ' ', '#', '0' http://msdn.microsoft.com/ru-ru/library/8aky45ct(en-us,VS.80).aspx
641: // width, precision: non negative decimal number
642: enum FormatType {
643: FormatInvalid,
644: FormatInt,
645: FormatUInt,
646: FormatDouble
647: };
648: FormatType format_type(char* fmt){
649: enum FormatState {
650: Percent,
651: Flags,
652: Width,
653: Precision,
654: Done
655: } state=Percent;
656:
657: FormatType result=FormatInvalid;
658:
659: char* pos=fmt;
660: while(char c=*(pos++)){
661: switch(state){
662: case Percent:
663: if(c=='%'){
664: state=Flags;
665: } else {
666: return FormatInvalid; // 1st char must be '%' only
667: }
668: break;
669: case Flags:
670: if(strchr("-+ #0", c)!=0){
671: break;
672: }
673: // go to the next step
674: case Width:
675: if(c=='.'){
676: state=Precision;
677: break;
678: }
679: // go to the next step
680: case Precision:
681: if(c>='0' && c<='9'){
682: if(state == Flags) state=Width; // no more flags
683: break;
684: } else if(c=='d' || c=='i'){
685: result=FormatInt;
686: } else if(strchr("feEgG", c)!=0){
687: result=FormatDouble;
688: } else if(strchr("uoxX", c)!=0){
689: result=FormatUInt;
690: } else {
691: return FormatInvalid; // invalid char
692: }
693: state=Done;
694: break;
695: case Done:
696: return FormatInvalid; // no chars allowed after 'type'
697: }
698: }
699: return result;
700: }
701:
702:
1.154 paf 703: const char* format(double value, char* fmt) {
1.229 misha 704: char local_buf[MAX_NUMBER];
1.235 misha 705: int size=-1;
1.229 misha 706:
707: if(fmt && strlen(fmt)){
708: switch(format_type(fmt)){
709: case FormatDouble:
710: size=snprintf(local_buf, sizeof(local_buf), fmt, value);
711: break;
712: case FormatInt:
713: size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value);
714: break;
715: case FormatUInt:
1.126 paf 716: size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value);
1.229 misha 717: break;
718: case FormatInvalid:
719: throw Exception(PARSER_RUNTIME,
720: 0,
721: "Incorrect format string '%s' was specified.", fmt);
722: }
723: } else
724: size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value);
725:
726: if(size < 0 || size >= MAX_NUMBER-1){ // on win32 we manually reduce max size while printing
727: throw Exception(PARSER_RUNTIME,
728: 0,
729: "Error occure white executing snprintf with format string '%s'.", fmt);
730: }
731:
1.235 misha 732: return pa_strdup(local_buf, (size_t)size);
1.12 paf 733: }
734:
1.36 paf 735: size_t stdout_write(const void *buf, size_t size) {
1.12 paf 736: #ifdef WIN32
1.187 paf 737: size_t to_write = size;
1.12 paf 738: do{
1.154 paf 739: int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout);
1.12 paf 740: if(chunk_written<=0)
741: break;
742: size-=chunk_written;
1.36 paf 743: buf=((const char*)buf)+chunk_written;
1.126 paf 744: } while(size>0);
1.12 paf 745:
1.187 paf 746: return to_write-size;
1.12 paf 747: #else
1.126 paf 748: return fwrite(buf, 1, size, stdout);
1.12 paf 749: #endif
1.2 paf 750: }
1.14 paf 751:
1.229 misha 752: enum EscapeState {
753: EscapeRest,
754: EscapeFirst,
755: EscapeSecond,
756: EscapeUnicode
757: };
758:
1.236 misha 759: // @todo prescan for reduce required size (unescaped sting in 1 byte charset requires less memory usually)
1.258 misha 760: char* unescape_chars(const char* cp, int len, Charset* charset, bool js){
1.236 misha 761: char* s=new(PointerFreeGC) char[len+1]; // must be enough (%uXXXX==6 bytes, max utf-8 char length==6 bytes)
1.230 misha 762: char* dst=s;
1.229 misha 763: EscapeState escapeState=EscapeRest;
764: uint escapedValue=0;
765: int srcPos=0;
1.230 misha 766: short int jsCnt=0;
1.236 misha 767: while(srcPos<len){
1.229 misha 768: uchar c=(uchar)cp[srcPos];
1.258 misha 769: if(c=='%' || (c=='\\' && js)){
1.229 misha 770: escapeState=EscapeFirst;
771: } else {
772: switch(escapeState) {
773: case EscapeRest:
1.258 misha 774: if(!js && c=='+'){
1.230 misha 775: *dst++=' ';
1.229 misha 776: } else {
1.230 misha 777: *dst++=c;
1.229 misha 778: }
779: break;
780: case EscapeFirst:
1.232 misha 781: if(charset && c=='u'){
1.229 misha 782: // escaped unicode value: %u0430
783: jsCnt=0;
784: escapedValue=0;
785: escapeState=EscapeUnicode;
786: } else {
1.231 misha 787: if(isxdigit(c)){
1.229 misha 788: escapedValue=hex_value[c] << 4;
789: escapeState=EscapeSecond;
790: } else {
1.230 misha 791: *dst++=c;
1.229 misha 792: escapeState=EscapeRest;
793: }
794: }
795: break;
796: case EscapeSecond:
1.231 misha 797: if(isxdigit(c)){
1.229 misha 798: escapedValue+=hex_value[c];
1.230 misha 799: *dst++=(char)escapedValue;
1.229 misha 800: }
801: escapeState=EscapeRest;
802: break;
803: case EscapeUnicode:
1.231 misha 804: if(isxdigit(c)){
1.229 misha 805: escapedValue=(escapedValue << 4) + hex_value[c];
806: if(++jsCnt==4){
1.230 misha 807: // transcode utf8 char to client charset (we can lost some chars here)
1.232 misha 808: charset->store_Char((XMLByte*&)dst, (XMLCh)escapedValue, '?');
1.229 misha 809: escapeState=EscapeRest;
810: }
811: } else {
812: // not full unicode value
813: escapeState=EscapeRest;
814: }
815: break;
816: }
817: }
818:
819: srcPos++;
820: }
821:
1.230 misha 822: *dst=0; // zero-termination
1.229 misha 823: return s;
824: }
1.24 paf 825:
826: #ifdef WIN32
1.126 paf 827: void back_slashes_to_slashes(char* s) {
1.24 paf 828: if(s)
829: for(; *s; s++)
830: if(*s=='\\')
1.126 paf 831: *s='/';
1.24 paf 832: }
1.42 paf 833: /*
1.126 paf 834: void slashes_to_back_slashes(char* s) {
1.42 paf 835: if(s)
836: for(; *s; s++)
837: if(*s=='/')
1.126 paf 838: *s='\\';
1.42 paf 839: }
840: */
1.24 paf 841: #endif
1.41 paf 842:
1.231 misha 843: bool StrStartFromNC(const char* str, const char* substr, bool equal){
1.41 paf 844: while(true) {
1.231 misha 845: if(!(*substr)){
846: if(!(*str))
1.41 paf 847: return true;
848: else
1.231 misha 849: return !equal;
850: }
851: if(!(*str))
852: return false;
853: if(isalpha((unsigned char)*str)) {
854: if(tolower((unsigned char)*str)!=tolower((unsigned char)*substr))
1.41 paf 855: return false;
1.231 misha 856: } else if((*str) != (*substr))
1.41 paf 857: return false;
1.231 misha 858: str++;
859: substr++;
1.41 paf 860: }
1.57 parser 861: }
862:
1.232 misha 863: size_t strpos(const char *str, const char *substr) {
864: const char *p = strstr(str, substr);
865: return (p==0)?STRING_NOT_FOUND:p-str;
866: }
867:
868: // content-type: xxx; charset=WE-NEED-THIS
869: // content-type: xxx; charset="WE-NEED-THIS"
870: // content-type: xxx; charset="WE-NEED-THIS";
1.248 misha 871: Charset* detect_charset(const char* content_type){
1.233 misha 872: if(content_type){
1.245 misha 873: char* CONTENT_TYPE=pa_strdup(content_type);
1.248 misha 874:
875: for(char *p=CONTENT_TYPE; *p; p++)
876: *p=(char)toupper((unsigned char)*p);
1.233 misha 877:
878: if(const char* begin=strstr(CONTENT_TYPE, "CHARSET=")){
879: begin+=8; // skip "CHARSET="
880: char* end=0;
881: if(*begin && (*begin=='"' || *begin =='\'')){
882: char quote=*begin;
883: begin++;
884: end=(char*)strchr(begin, quote);
885: }
886: if(!end)
887: end=(char*)strchr(begin, ';');
888:
1.244 misha 889: if(end)
1.233 misha 890: *end=0; // terminator
891:
1.245 misha 892: return *begin?&charsets.get(begin):0;
1.232 misha 893: }
894: }
895: return 0;
896: }
897:
898:
1.84 paf 899: static bool isLeap(int year) {
1.229 misha 900: return !(
901: (year % 4) || ((year % 400) && !(year % 100))
902: );
1.57 parser 903: }
904:
905: int getMonthDays(int year, int month) {
1.220 misha 906: static int monthDays[]={
1.229 misha 907: 31,
908: 28,
909: 31,
910: 30,
911: 31,
912: 30,
913: 31,
914: 31,
915: 30,
916: 31,
917: 30,
918: 31
919: };
1.228 misha 920: return (month == 1 /* january -- 0 */ && isLeap(year)) ? 29 : monthDays[month];
1.41 paf 921: }
1.69 parser 922:
1.226 misha 923: int remove_crlf(char* start, char* end) {
924: char* from=start;
925: char* to=start;
926: bool skip=false;
927: while(from < end){
928: switch(*from){
929: case '\n':
930: case '\r':
931: case '\t':
932: case ' ':
933: if(!skip){
934: *to=' ';
935: to++;
936: skip=true;
937: }
938: break;
939: default:
940: if(from != to)
941: *to=*from;
942: to++;
943: skip=false;
1.69 parser 944: }
1.226 misha 945: from++;
946: }
947: return to-start;
1.91 paf 948: }
949:
950:
951: /// must be last in this file
952: #undef vsnprintf
1.126 paf 953: int __vsnprintf(char* b, size_t s, const char* f, va_list l) {
1.91 paf 954: if(!s)
955: return 0;
956:
957: int r;
958: // note: on win32& maybe somewhere else
959: // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing
960: --s;
1.172 paf 961:
962: // clients do not check for negative 's', feature: ignore such prints
963: if((ssize_t)s<0)
964: return 0;
965:
1.91 paf 966: #if _MSC_VER
967: /*
968: win32:
969: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm
970:
1.154 paf 971: if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 is returned
1.91 paf 972: */
1.126 paf 973: r=_vsnprintf(b, s, f, l);
1.91 paf 974: if(r<0)
975: r=s;
976: #else
1.126 paf 977: r=vsnprintf(b, s, f, l);
1.91 paf 978: /*
979: solaris:
980: man vsnprintf
981:
982: The snprintf() function returns the number of characters
983: formatted, that is, the number of characters that would have
984: been written to the buffer if it were large enough. If the
985: value of n is 0 on a call to snprintf(), an unspecified
986: value less than 1 is returned.
987: */
988:
989: if(r<0)
990: r=0;
1.167 paf 991: else if((size_t)r>s)
1.91 paf 992: r=s;
993: #endif
994: b[r]=0;
995: return r;
996: }
997:
1.126 paf 998: int __snprintf(char* b, size_t s, const char* f, ...) {
1.91 paf 999: va_list l;
1.241 misha 1000: va_start(l, f);
1001: int r=__vsnprintf(b, s, f, l);
1002: va_end(l);
1.91 paf 1003: return r;
1.178 paf 1004: }
1005:
1006: /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */
1007: /*
1008: * Authors: Michael Zucchi <notzed@helixcode.com>
1009: * Jeffrey Stedfast <fejj@helixcode.com>
1010: *
1011: * Copyright 2000 Helix Code, Inc. (www.helixcode.com)
1012: *
1013: * This program is free software; you can redistribute it and/or modify
1014: * it under the terms of the GNU General Public License as published by
1015: * the Free Software Foundation; either version 2 of the License, or
1016: * (at your option) any later version.
1017: *
1018: * This program is distributed in the hope that it will be useful,
1019: * but WITHOUT ANY WARRANTY; without even the implied warranty of
1020: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
1021: * GNU General Public License for more details.
1022: *
1023: * You should have received a copy of the GNU General Public License
1024: * along with this program; if not, write to the Free Software
1025: * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA.
1026: *
1027: */
1028: static char *base64_alphabet =
1029: "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
1030:
1031: /**
1032: * g_mime_utils_base64_encode_step:
1033: * @in: input stream
1034: * @inlen: length of the input
1035: * @out: output string
1036: * @state: holds the number of bits that are stored in @save
1037: * @save: leftover bits that have not yet been encoded
1038: *
1039: * Base64 encodes a chunk of data. Performs an 'encode step', only
1040: * encodes blocks of 3 characters to the output at a time, saves
1041: * left-over state in state and save (initialise to 0 on first
1042: * invocation).
1043: *
1044: * Returns the number of bytes encoded.
1045: **/
1.252 misha 1046:
1047: #define BASE64_GROUPS_IN_LINE 19
1048:
1.178 paf 1049: static size_t
1050: g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1051: {
1.186 paf 1052: register const unsigned char *inptr;
1.178 paf 1053: register unsigned char *outptr;
1054:
1055: if (inlen <= 0)
1056: return 0;
1057:
1058: inptr = in;
1059: outptr = out;
1060:
1061: if (inlen + ((unsigned char *)save)[0] > 2) {
1062: const unsigned char *inend = in + inlen - 2;
1063: register int c1 = 0, c2 = 0, c3 = 0;
1064: register int already;
1065:
1066: already = *state;
1067:
1068: switch (((char *)save)[0]) {
1069: case 1: c1 = ((unsigned char *)save)[1]; goto skip1;
1070: case 2: c1 = ((unsigned char *)save)[1];
1071: c2 = ((unsigned char *)save)[2]; goto skip2;
1072: }
1073:
1074: /* yes, we jump into the loop, no i'm not going to change it, its beautiful! */
1075: while (inptr < inend) {
1076: c1 = *inptr++;
1077: skip1:
1078: c2 = *inptr++;
1079: skip2:
1080: c3 = *inptr++;
1081: *outptr++ = base64_alphabet [c1 >> 2];
1082: *outptr++ = base64_alphabet [(c2 >> 4) | ((c1 & 0x3) << 4)];
1083: *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)];
1084: *outptr++ = base64_alphabet [c3 & 0x3f];
1085: /* this is a bit ugly ... */
1.252 misha 1086: if ((++already) >= BASE64_GROUPS_IN_LINE) {
1.178 paf 1087: *outptr++ = '\n';
1088: already = 0;
1089: }
1090: }
1091:
1092: ((unsigned char *)save)[0] = 0;
1093: inlen = 2 - (inptr - inend);
1094: *state = already;
1095: }
1096:
1097: //d(printf ("state = %d, inlen = %d\n", (int)((char *)save)[0], inlen));
1098:
1099: if (inlen > 0) {
1100: register char *saveout;
1101:
1102: /* points to the slot for the next char to save */
1103: saveout = & (((char *)save)[1]) + ((char *)save)[0];
1104:
1105: /* inlen can only be 0 1 or 2 */
1106: switch (inlen) {
1107: case 2: *saveout++ = *inptr++;
1108: case 1: *saveout++ = *inptr++;
1109: }
1.216 paf 1110: *(char *)save = *(char *)save+(char)inlen;
1.178 paf 1111: }
1112:
1113: /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n",
1114: (int)((char *)save)[0],
1115: (int)((char *)save)[1],
1116: (int)((char *)save)[2]));*/
1117:
1118: return (outptr - out);
1119: }
1120:
1121: /**
1122: * g_mime_utils_base64_encode_close:
1123: * @in: input stream
1124: * @inlen: length of the input
1125: * @out: output string
1126: * @state: holds the number of bits that are stored in @save
1127: * @save: leftover bits that have not yet been encoded
1128: *
1129: * Base64 encodes the input stream to the output stream. Call this
1130: * when finished encoding data with g_mime_utils_base64_encode_step to
1131: * flush off the last little bit.
1132: *
1133: * Returns the number of bytes encoded.
1134: **/
1135: static size_t
1136: g_mime_utils_base64_encode_close (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1137: {
1138: unsigned char *outptr = out;
1139: int c1, c2;
1140:
1141: if (inlen > 0)
1142: outptr += g_mime_utils_base64_encode_step (in, inlen, outptr, state, save);
1143:
1144: c1 = ((unsigned char *)save)[1];
1145: c2 = ((unsigned char *)save)[2];
1146:
1147: switch (((unsigned char *)save)[0]) {
1148: case 2:
1149: outptr[2] = base64_alphabet [(c2 & 0x0f) << 2];
1150: goto skip;
1151: case 1:
1152: outptr[2] = '=';
1153: skip:
1154: outptr[0] = base64_alphabet [c1 >> 2];
1155: outptr[1] = base64_alphabet [c2 >> 4 | ((c1 & 0x3) << 4)];
1156: outptr[3] = '=';
1157: outptr += 4;
1158: break;
1159: }
1160:
1161: *outptr++ = 0;
1162:
1163: *save = 0;
1164: *state = 0;
1165:
1166: return (outptr - out);
1167: }
1168:
1.210 paf 1169: static unsigned char gmime_base64_rank[256] = {
1170: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1171: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1172: 255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63,
1173: 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255, 0,255,255,
1174: 255, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
1175: 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255,
1176: 255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
1177: 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255,
1178: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1179: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1180: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1181: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1182: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1183: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1184: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1185: 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,
1186: };
1187:
1188: /**
1189: * g_mime_utils_base64_decode_step:
1190: * @in: input stream
1191: * @inlen: max length of data to decode
1192: * @out: output stream
1193: * @state: holds the number of bits that are stored in @save
1194: * @save: leftover bits that have not yet been decoded
1195: *
1196: * Decodes a chunk of base64 encoded data.
1197: *
1198: * Returns the number of bytes decoded (which have been dumped in @out).
1199: **/
1200: size_t
1201: g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save)
1202: {
1.213 paf 1203: const unsigned char *inptr;
1204: unsigned char *outptr;
1.210 paf 1205: const unsigned char *inend;
1.213 paf 1206: int saved;
1.210 paf 1207: unsigned char c;
1208: int i;
1209:
1210: inend = in + inlen;
1211: outptr = out;
1212:
1213: /* convert 4 base64 bytes to 3 normal bytes */
1214: saved = *save;
1215: i = *state;
1216: inptr = in;
1217: while (inptr < inend) {
1218: c = gmime_base64_rank[*inptr++];
1219: if (c != 0xff) {
1220: saved = (saved << 6) | c;
1221: i++;
1222: if (i == 4) {
1.217 paf 1223: *outptr++ = (unsigned char)(saved >> 16);
1224: *outptr++ = (unsigned char)(saved >> 8);
1225: *outptr++ = (unsigned char)(saved);
1.210 paf 1226: i = 0;
1227: }
1228: }
1229: }
1230:
1231: *save = saved;
1232: *state = i;
1233:
1234: /* quick scan back for '=' on the end somewhere */
1235: /* fortunately we can drop 1 output char for each trailing = (upto 2) */
1236: i = 2;
1237: while (inptr > in && i) {
1238: inptr--;
1239: if (gmime_base64_rank[*inptr] != 0xff) {
1240: if (*inptr == '=' && outptr > out)
1241: outptr--;
1242: i--;
1243: }
1244: }
1245:
1246: /* if i != 0 then there is a truncation error! */
1247: return (outptr - out);
1248: }
1249:
1250:
1.239 misha 1251: char* pa_base64_encode(const char *in, size_t in_size){
1.252 misha 1252: size_t new_size = ((in_size / 3 + 1) * 4);
1253: new_size += new_size / (BASE64_GROUPS_IN_LINE * 4)/*new lines*/ + 1/*zero terminator*/;
1254: char* result = new(PointerFreeGC) char[new_size];
1.178 paf 1255: int state=0;
1256: int save=0;
1.183 paf 1257: #ifndef NDEBUG
1258: size_t filled=
1259: #endif
1.251 misha 1260: g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size, (unsigned char*)result, &state, &save);
1261:
1262: //throw Exception(PARSER_RUNTIME, 0, "%d %d %d", in_size, new_size, filled);
1263: assert(filled <= new_size);
1.178 paf 1264:
1265: return result;
1.98 paf 1266: }
1.210 paf 1267:
1.222 misha 1268:
1.239 misha 1269: char* pa_base64_encode(const String& file_spec){
1.222 misha 1270: unsigned char* base64=0;
1271: File_base64_action_info info={&base64};
1272:
1273: file_read_action_under_lock(file_spec,
1274: "pa_base64_encode", file_base64_file_action, &info);
1275:
1276: return (char*)base64;
1277: }
1278:
1279:
1280: static void file_base64_file_action(
1.229 misha 1281: struct stat& finfo,
1282: int f,
1283: const String&, const char* /*fname*/, bool,
1284: void *context) {
1.222 misha 1285:
1286: if(finfo.st_size) {
1287: File_base64_action_info& info=*static_cast<File_base64_action_info *>(context);
1288: *info.base64=new(PointerFreeGC) unsigned char[finfo.st_size * 2 + 6];
1289: unsigned char* base64 = *info.base64;
1290: int state=0;
1291: int save=0;
1292: int nCount;
1293: do {
1294: unsigned char buffer[FILE_BUFFER_SIZE];
1295: nCount = file_block_read(f, buffer, sizeof(buffer));
1296: if( nCount ){
1297: size_t filled=g_mime_utils_base64_encode_step ((const unsigned char*)buffer, nCount, base64, &state, &save);
1298: base64+=filled;
1299: }
1300: } while(nCount > 0);
1301: g_mime_utils_base64_encode_close (0, 0, base64, &state, &save);
1302: }
1303: }
1304:
1.239 misha 1305: void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size) {
1.210 paf 1306: /* wont go to more than had (overly conservative) */
1.211 paf 1307: result=new(PointerFreeGC) char[in_size+1/*terminator*/];
1.210 paf 1308: int state=0;
1309: int save=0;
1310: result_size=
1311: g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size,
1312: (unsigned char*)result, &state, &save);
1313: assert(result_size <= in_size);
1.211 paf 1314: result[result_size]=0; // for text files
1.210 paf 1315: }
1.218 misha 1316:
1317:
1.221 misha 1318: int file_block_read(const int f, unsigned char* buffer, const size_t size){
1319: int nCount = read(f, buffer, size);
1320: if (nCount < 0)
1.238 misha 1321: throw Exception("file.read",
1.221 misha 1322: 0,
1323: "read failed: %s (%d)", strerror(errno), errno);
1324: return nCount;
1325: }
1326:
1.239 misha 1327: const unsigned long pa_crc32(const char *in, size_t in_size){
1.218 misha 1328: unsigned long crc32=0xFFFFFFFF;
1.220 misha 1329:
1.240 misha 1330: InitCrc32Table();
1.239 misha 1331: for(size_t i = 0; i<in_size; i++)
1332: CalcCrc32(in[i], crc32);
1.220 misha 1333:
1.218 misha 1334: return ~crc32;
1335: }
1336:
1.239 misha 1337: const unsigned long pa_crc32(const String& file_spec){
1.218 misha 1338: unsigned long crc32=0xFFFFFFFF;
1339: file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32);
1340: return ~crc32;
1341: }
1342:
1343: static void file_crc32_file_action(
1.229 misha 1344: struct stat& finfo,
1345: int f,
1346: const String&, const char* /*fname*/, bool,
1.239 misha 1347: void *context) {
1.218 misha 1348: unsigned long& crc32=*static_cast<unsigned long *>(context);
1349: if(finfo.st_size) {
1350: InitCrc32Table();
1.220 misha 1351: int nCount=0;
1.218 misha 1352: do {
1.221 misha 1353: unsigned char buffer[FILE_BUFFER_SIZE];
1354: nCount = file_block_read(f, buffer, sizeof(buffer));
1.220 misha 1355: for(int i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32);
1356: } while(nCount > 0);
1.218 misha 1357: }
1358: }
1359:
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