apt-pkg/contrib/strutl.cc

   1 // -*- mode: cpp; mode: fold -*-
   2 // Description                                                          /*{{{*/
   3 // $Id: strutl.cc,v 1.48 2003/07/18 14:15:11 mdz Exp $
   4 /* ######################################################################
   5
   6    String Util - Some useful string functions.
   7
   8    These have been collected from here and there to do all sorts of useful
   9    things to strings. They are useful in file parsers, URI handlers and
  10    especially in APT methods.
  11
  12    This source is placed in the Public Domain, do with it what you will
  13    It was originally written by Jason Gunthorpe <jgg@gpu.srv.ualberta.ca>
  14
  15    ##################################################################### */
  16                                                                         /*}}}*/
  17 // Includes                                                             /*{{{*/
  18 #include <config.h>
  19
  20 #include <apt-pkg/strutl.h>
  21 #include <apt-pkg/fileutl.h>
  22 #include <apt-pkg/error.h>
  23
  24 #include <algorithm>
  25 #include <iomanip>
  26 #include <locale>
  27 #include <sstream>
  28 #include <string>
  29 #include <vector>
  30
  31 #include <stddef.h>
  32 #include <stdlib.h>
  33 #include <time.h>
  34 #include <ctype.h>
  35 #include <string.h>
  36 #include <stdio.h>
  37 #include <unistd.h>
  38 #include <regex.h>
  39 #include <errno.h>
  40 #include <stdarg.h>
  41 #include <iconv.h>
  42
  43 #include <apti18n.h>
  44                                                                         /*}}}*/
  45 using namespace std;
  46
  47 // Strip - Remove white space from the front and back of a string       /*{{{*/
  48 // ---------------------------------------------------------------------
  49 namespace APT {
  50    namespace String {
  51 std::string Strip(const std::string &str)
  52 {
  53    // ensure we have at least one character
  54    if (str.empty() == true)
  55       return str;
  56
  57    char const * const s = str.c_str();
  58    size_t start = 0;
  59    for (; isspace(s[start]) != 0; ++start)
  60       ; // find the first not-space
  61
  62    // string contains only whitespaces
  63    if (s[start] == '\0')
  64       return "";
  65
  66    size_t end = str.length() - 1;
  67    for (; isspace(s[end]) != 0; --end)
  68       ; // find the last not-space
  69
  70    return str.substr(start, end - start + 1);
  71 }
  72
  73 bool Endswith(const std::string &s, const std::string &end)
  74 {
  75    if (end.size() > s.size())
  76       return false;
  77    return (s.compare(s.size() - end.size(), end.size(), end) == 0);
  78 }
  79
  80 bool Startswith(const std::string &s, const std::string &start)
  81 {
  82    if (start.size() > s.size())
  83       return false;
  84    return (s.compare(0, start.size(), start) == 0);
  85 }
  86
  87 }
  88 }
  89                                                                         /*}}}*/
  90 // UTF8ToCodeset - Convert some UTF-8 string for some codeset           /*{{{*/
  91 // ---------------------------------------------------------------------
  92 /* This is handy to use before display some information for enduser  */
  93 bool UTF8ToCodeset(const char *codeset, const string &orig, string *dest)
  94 {
  95   iconv_t cd;
  96   const char *inbuf;
  97   char *inptr, *outbuf;
  98   size_t insize, bufsize;
  99   dest->clear();
 100
 101   cd = iconv_open(codeset, "UTF-8");
 102   if (cd == (iconv_t)(-1)) {
 103      // Something went wrong
 104      if (errno == EINVAL)
 105         _error->Error("conversion from 'UTF-8' to '%s' not available",
 106                codeset);
 107      else
 108         perror("iconv_open");
 109
 110      return false;
 111   }
 112
 113   insize = bufsize = orig.size();
 114   inbuf = orig.data();
 115   inptr = (char *)inbuf;
 116   outbuf = new char[bufsize];
 117   size_t lastError = -1;
 118
 119   while (insize != 0)
 120   {
 121      char *outptr = outbuf;
 122      size_t outsize = bufsize;
 123      size_t const err = iconv(cd, &inptr, &insize, &outptr, &outsize);
 124      dest->append(outbuf, outptr - outbuf);
 125      if (err == (size_t)(-1))
 126      {
 127         switch (errno)
 128         {
 129         case EILSEQ:
 130            insize--;
 131            inptr++;
 132            // replace a series of unknown multibytes with a single "?"
 133            if (lastError != insize) {
 134               lastError = insize - 1;
 135               dest->append("?");
 136            }
 137            break;
 138         case EINVAL:
 139            insize = 0;
 140            break;
 141         case E2BIG:
 142            if (outptr == outbuf)
 143            {
 144               bufsize *= 2;
 145               delete[] outbuf;
 146               outbuf = new char[bufsize];
 147            }
 148            break;
 149         }
 150      }
 151   }
 152
 153   delete[] outbuf;
 154
 155   iconv_close(cd);
 156
 157   return true;
 158 }
 159                                                                         /*}}}*/
 160 // strstrip - Remove white space from the front and back of a string    /*{{{*/
 161 // ---------------------------------------------------------------------
 162 /* This is handy to use when parsing a file. It also removes \n's left
 163    over from fgets and company */
 164 char *_strstrip(char *String)
 165 {
 166    for (;*String != 0 && (*String == ' ' || *String == '\t'); String++);
 167
 168    if (*String == 0)
 169       return String;
 170    return _strrstrip(String);
 171 }
 172                                                                         /*}}}*/
 173 // strrstrip - Remove white space from the back of a string     /*{{{*/
 174 // ---------------------------------------------------------------------
 175 char *_strrstrip(char *String)
 176 {
 177    char *End = String + strlen(String) - 1;
 178    for (;End != String - 1 && (*End == ' ' || *End == '\t' || *End == '\n' ||
 179                                *End == '\r'); End--);
 180    End++;
 181    *End = 0;
 182    return String;
 183 }
 184                                                                         /*}}}*/
 185 // strtabexpand - Converts tabs into 8 spaces                           /*{{{*/
 186 // ---------------------------------------------------------------------
 187 /* */
 188 char *_strtabexpand(char *String,size_t Len)
 189 {
 190    for (char *I = String; I != I + Len && *I != 0; I++)
 191    {
 192       if (*I != '\t')
 193          continue;
 194       if (I + 8 > String + Len)
 195       {
 196          *I = 0;
 197          return String;
 198       }
 199
 200       /* Assume the start of the string is 0 and find the next 8 char
 201          division */
 202       int Len;
 203       if (String == I)
 204          Len = 1;
 205       else
 206          Len = 8 - ((String - I) % 8);
 207       Len -= 2;
 208       if (Len <= 0)
 209       {
 210          *I = ' ';
 211          continue;
 212       }
 213
 214       memmove(I + Len,I + 1,strlen(I) + 1);
 215       for (char *J = I; J + Len != I; *I = ' ', I++);
 216    }
 217    return String;
 218 }
 219                                                                         /*}}}*/
 220 // ParseQuoteWord - Parse a single word out of a string                 /*{{{*/
 221 // ---------------------------------------------------------------------
 222 /* This grabs a single word, converts any % escaped characters to their
 223    proper values and advances the pointer. Double quotes are understood
 224    and striped out as well. This is for URI/URL parsing. It also can
 225    understand [] brackets.*/
 226 bool ParseQuoteWord(const char *&String,string &Res)
 227 {
 228    // Skip leading whitespace
 229    const char *C = String;
 230    for (;*C != 0 && *C == ' '; C++);
 231    if (*C == 0)
 232       return false;
 233
 234    // Jump to the next word
 235    for (;*C != 0 && isspace(*C) == 0; C++)
 236    {
 237       if (*C == '"')
 238       {
 239          C = strchr(C + 1, '"');
 240          if (C == NULL)
 241             return false;
 242       }
 243       if (*C == '[')
 244       {
 245          C = strchr(C + 1, ']');
 246          if (C == NULL)
 247             return false;
 248       }
 249    }
 250
 251    // Now de-quote characters
 252    char Buffer[1024];
 253    char Tmp[3];
 254    const char *Start = String;
 255    char *I;
 256    for (I = Buffer; I < Buffer + sizeof(Buffer) && Start != C; I++)
 257    {
 258       if (*Start == '%' && Start + 2 < C &&
 259           isxdigit(Start[1]) && isxdigit(Start[2]))
 260       {
 261          Tmp[0] = Start[1];
 262          Tmp[1] = Start[2];
 263          Tmp[2] = 0;
 264          *I = (char)strtol(Tmp,0,16);
 265          Start += 3;
 266          continue;
 267       }
 268       if (*Start != '"')
 269          *I = *Start;
 270       else
 271          I--;
 272       Start++;
 273    }
 274    *I = 0;
 275    Res = Buffer;
 276
 277    // Skip ending white space
 278    for (;*C != 0 && isspace(*C) != 0; C++);
 279    String = C;
 280    return true;
 281 }
 282                                                                         /*}}}*/
 283 // ParseCWord - Parses a string like a C "" expression                  /*{{{*/
 284 // ---------------------------------------------------------------------
 285 /* This expects a series of space separated strings enclosed in ""'s.
 286    It concatenates the ""'s into a single string. */
 287 bool ParseCWord(const char *&String,string &Res)
 288 {
 289    // Skip leading whitespace
 290    const char *C = String;
 291    for (;*C != 0 && *C == ' '; C++);
 292    if (*C == 0)
 293       return false;
 294
 295    char Buffer[1024];
 296    char *Buf = Buffer;
 297    if (strlen(String) >= sizeof(Buffer))
 298        return false;
 299
 300    for (; *C != 0; C++)
 301    {
 302       if (*C == '"')
 303       {
 304          for (C++; *C != 0 && *C != '"'; C++)
 305             *Buf++ = *C;
 306
 307          if (*C == 0)
 308             return false;
 309
 310          continue;
 311       }
 312
 313       if (C != String && isspace(*C) != 0 && isspace(C[-1]) != 0)
 314          continue;
 315       if (isspace(*C) == 0)
 316          return false;
 317       *Buf++ = ' ';
 318    }
 319    *Buf = 0;
 320    Res = Buffer;
 321    String = C;
 322    return true;
 323 }
 324                                                                         /*}}}*/
 325 // QuoteString - Convert a string into quoted from                      /*{{{*/
 326 // ---------------------------------------------------------------------
 327 /* */
 328 string QuoteString(const string &Str, const char *Bad)
 329 {
 330    std::stringstream Res;
 331    for (string::const_iterator I = Str.begin(); I != Str.end(); ++I)
 332    {
 333       if (strchr(Bad,*I) != 0 || isprint(*I) == 0 ||
 334           *I == 0x25 || // percent '%' char
 335           *I <= 0x20 || *I >= 0x7F) // control chars
 336       {
 337          ioprintf(Res, "%%%02hhx", *I);
 338       }
 339       else
 340          Res << *I;
 341    }
 342    return Res.str();
 343 }
 344                                                                         /*}}}*/
 345 // DeQuoteString - Convert a string from quoted from                    /*{{{*/
 346 // ---------------------------------------------------------------------
 347 /* This undoes QuoteString */
 348 string DeQuoteString(const string &Str)
 349 {
 350    return DeQuoteString(Str.begin(),Str.end());
 351 }
 352 string DeQuoteString(string::const_iterator const &begin,
 353                         string::const_iterator const &end)
 354 {
 355    string Res;
 356    for (string::const_iterator I = begin; I != end; ++I)
 357    {
 358       if (*I == '%' && I + 2 < end &&
 359           isxdigit(I[1]) && isxdigit(I[2]))
 360       {
 361          char Tmp[3];
 362          Tmp[0] = I[1];
 363          Tmp[1] = I[2];
 364          Tmp[2] = 0;
 365          Res += (char)strtol(Tmp,0,16);
 366          I += 2;
 367          continue;
 368       }
 369       else
 370          Res += *I;
 371    }
 372    return Res;
 373 }
 374
 375                                                                         /*}}}*/
 376 // SizeToStr - Convert a long into a human readable size                /*{{{*/
 377 // ---------------------------------------------------------------------
 378 /* A max of 4 digits are shown before conversion to the next highest unit.
 379    The max length of the string will be 5 chars unless the size is > 10
 380    YottaBytes (E24) */
 381 string SizeToStr(double Size)
 382 {
 383    double ASize;
 384    if (Size >= 0)
 385       ASize = Size;
 386    else
 387       ASize = -1*Size;
 388
 389    /* bytes, KiloBytes, MegaBytes, GigaBytes, TeraBytes, PetaBytes,
 390       ExaBytes, ZettaBytes, YottaBytes */
 391    char Ext[] = {'\0','k','M','G','T','P','E','Z','Y'};
 392    int I = 0;
 393    while (I <= 8)
 394    {
 395       if (ASize < 100 && I != 0)
 396       {
 397          std::string S;
 398          strprintf(S, "%'.1f %c", ASize, Ext[I]);
 399          return S;
 400       }
 401
 402       if (ASize < 10000)
 403       {
 404          std::string S;
 405          strprintf(S, "%'.0f %c", ASize, Ext[I]);
 406          return S;
 407       }
 408       ASize /= 1000.0;
 409       I++;
 410    }
 411    return "";
 412 }
 413                                                                         /*}}}*/
 414 // TimeToStr - Convert the time into a string                           /*{{{*/
 415 // ---------------------------------------------------------------------
 416 /* Converts a number of seconds to a hms format */
 417 string TimeToStr(unsigned long Sec)
 418 {
 419    std::string S;
 420    if (Sec > 60*60*24)
 421    {
 422       //TRANSLATOR: d means days, h means hours, min means minutes, s means seconds
 423       strprintf(S,_("%lid %lih %limin %lis"),Sec/60/60/24,(Sec/60/60) % 24,(Sec/60) % 60,Sec % 60);
 424    }
 425    else if (Sec > 60*60)
 426    {
 427       //TRANSLATOR: h means hours, min means minutes, s means seconds
 428       strprintf(S,_("%lih %limin %lis"),Sec/60/60,(Sec/60) % 60,Sec % 60);
 429    }
 430    else if (Sec > 60)
 431    {
 432       //TRANSLATOR: min means minutes, s means seconds
 433       strprintf(S,_("%limin %lis"),Sec/60,Sec % 60);
 434    }
 435    else
 436    {
 437       //TRANSLATOR: s means seconds
 438       strprintf(S,_("%lis"),Sec);
 439    }
 440    return S;
 441 }
 442                                                                         /*}}}*/
 443 // SubstVar - Substitute a string for another string                    /*{{{*/
 444 // ---------------------------------------------------------------------
 445 /* This replaces all occurrences of Subst with Contents in Str. */
 446 string SubstVar(const string &Str,const string &Subst,const string &Contents)
 447 {
 448    if (Subst.empty() == true)
 449       return Str;
 450
 451    string::size_type Pos = 0;
 452    string::size_type OldPos = 0;
 453    string Temp;
 454
 455    while (OldPos < Str.length() &&
 456           (Pos = Str.find(Subst,OldPos)) != string::npos)
 457    {
 458       if (OldPos != Pos)
 459          Temp.append(Str, OldPos, Pos - OldPos);
 460       if (Contents.empty() == false)
 461          Temp.append(Contents);
 462       OldPos = Pos + Subst.length();
 463    }
 464
 465    if (OldPos == 0)
 466       return Str;
 467
 468    if (OldPos >= Str.length())
 469       return Temp;
 470
 471    Temp.append(Str, OldPos, string::npos);
 472    return Temp;
 473 }
 474 string SubstVar(string Str,const struct SubstVar *Vars)
 475 {
 476    for (; Vars->Subst != 0; Vars++)
 477       Str = SubstVar(Str,Vars->Subst,*Vars->Contents);
 478    return Str;
 479 }
 480                                                                         /*}}}*/
 481 // OutputInDepth - return a string with separator multiplied with depth /*{{{*/
 482 // ---------------------------------------------------------------------
 483 /* Returns a string with the supplied separator depth + 1 times in it */
 484 std::string OutputInDepth(const unsigned long Depth, const char* Separator)
 485 {
 486    std::string output = "";
 487    for(unsigned long d=Depth+1; d > 0; d--)
 488       output.append(Separator);
 489    return output;
 490 }
 491                                                                         /*}}}*/
 492 // URItoFileName - Convert the uri into a unique file name              /*{{{*/
 493 // ---------------------------------------------------------------------
 494 /* This converts a URI into a safe filename. It quotes all unsafe characters
 495    and converts / to _ and removes the scheme identifier. The resulting
 496    file name should be unique and never occur again for a different file */
 497 string URItoFileName(const string &URI)
 498 {
 499    // Nuke 'sensitive' items
 500    ::URI U(URI);
 501    U.User.clear();
 502    U.Password.clear();
 503    U.Access.clear();
 504
 505    // "\x00-\x20{}|\\\\^\\[\\]<>\"\x7F-\xFF";
 506    string NewURI = QuoteString(U,"\\|{}[]<>\"^~_=!@#$%^&*");
 507    replace(NewURI.begin(),NewURI.end(),'/','_');
 508    return NewURI;
 509 }
 510                                                                         /*}}}*/
 511 // Base64Encode - Base64 Encoding routine for short strings             /*{{{*/
 512 // ---------------------------------------------------------------------
 513 /* This routine performs a base64 transformation on a string. It was ripped
 514    from wget and then patched and bug fixed.
 515
 516    This spec can be found in rfc2045 */
 517 string Base64Encode(const string &S)
 518 {
 519    // Conversion table.
 520    static char tbl[64] = {'A','B','C','D','E','F','G','H',
 521                           'I','J','K','L','M','N','O','P',
 522                           'Q','R','S','T','U','V','W','X',
 523                           'Y','Z','a','b','c','d','e','f',
 524                           'g','h','i','j','k','l','m','n',
 525                           'o','p','q','r','s','t','u','v',
 526                           'w','x','y','z','0','1','2','3',
 527                           '4','5','6','7','8','9','+','/'};
 528
 529    // Pre-allocate some space
 530    string Final;
 531    Final.reserve((4*S.length() + 2)/3 + 2);
 532
 533    /* Transform the 3x8 bits to 4x6 bits, as required by
 534       base64.  */
 535    for (string::const_iterator I = S.begin(); I < S.end(); I += 3)
 536    {
 537       char Bits[3] = {0,0,0};
 538       Bits[0] = I[0];
 539       if (I + 1 < S.end())
 540          Bits[1] = I[1];
 541       if (I + 2 < S.end())
 542          Bits[2] = I[2];
 543
 544       Final += tbl[Bits[0] >> 2];
 545       Final += tbl[((Bits[0] & 3) << 4) + (Bits[1] >> 4)];
 546
 547       if (I + 1 >= S.end())
 548          break;
 549
 550       Final += tbl[((Bits[1] & 0xf) << 2) + (Bits[2] >> 6)];
 551
 552       if (I + 2 >= S.end())
 553          break;
 554
 555       Final += tbl[Bits[2] & 0x3f];
 556    }
 557
 558    /* Apply the padding elements, this tells how many bytes the remote
 559       end should discard */
 560    if (S.length() % 3 == 2)
 561       Final += '=';
 562    if (S.length() % 3 == 1)
 563       Final += "==";
 564
 565    return Final;
 566 }
 567                                                                         /*}}}*/
 568 // stringcmp - Arbitrary string compare                                 /*{{{*/
 569 // ---------------------------------------------------------------------
 570 /* This safely compares two non-null terminated strings of arbitrary
 571    length */
 572 int stringcmp(const char *A,const char *AEnd,const char *B,const char *BEnd)
 573 {
 574    for (; A != AEnd && B != BEnd; A++, B++)
 575       if (*A != *B)
 576          break;
 577
 578    if (A == AEnd && B == BEnd)
 579       return 0;
 580    if (A == AEnd)
 581       return 1;
 582    if (B == BEnd)
 583       return -1;
 584    if (*A < *B)
 585       return -1;
 586    return 1;
 587 }
 588
 589 #if __GNUC__ >= 3
 590 int stringcmp(string::const_iterator A,string::const_iterator AEnd,
 591               const char *B,const char *BEnd)
 592 {
 593    for (; A != AEnd && B != BEnd; A++, B++)
 594       if (*A != *B)
 595          break;
 596
 597    if (A == AEnd && B == BEnd)
 598       return 0;
 599    if (A == AEnd)
 600       return 1;
 601    if (B == BEnd)
 602       return -1;
 603    if (*A < *B)
 604       return -1;
 605    return 1;
 606 }
 607 int stringcmp(string::const_iterator A,string::const_iterator AEnd,
 608               string::const_iterator B,string::const_iterator BEnd)
 609 {
 610    for (; A != AEnd && B != BEnd; A++, B++)
 611       if (*A != *B)
 612          break;
 613
 614    if (A == AEnd && B == BEnd)
 615       return 0;
 616    if (A == AEnd)
 617       return 1;
 618    if (B == BEnd)
 619       return -1;
 620    if (*A < *B)
 621       return -1;
 622    return 1;
 623 }
 624 #endif
 625                                                                         /*}}}*/
 626 // stringcasecmp - Arbitrary case insensitive string compare            /*{{{*/
 627 // ---------------------------------------------------------------------
 628 /* */
 629 int stringcasecmp(const char *A,const char *AEnd,const char *B,const char *BEnd)
 630 {
 631    for (; A != AEnd && B != BEnd; A++, B++)
 632       if (tolower_ascii(*A) != tolower_ascii(*B))
 633          break;
 634
 635    if (A == AEnd && B == BEnd)
 636       return 0;
 637    if (A == AEnd)
 638       return 1;
 639    if (B == BEnd)
 640       return -1;
 641    if (tolower_ascii(*A) < tolower_ascii(*B))
 642       return -1;
 643    return 1;
 644 }
 645 #if __GNUC__ >= 3
 646 int stringcasecmp(string::const_iterator A,string::const_iterator AEnd,
 647                   const char *B,const char *BEnd)
 648 {
 649    for (; A != AEnd && B != BEnd; A++, B++)
 650       if (tolower_ascii(*A) != tolower_ascii(*B))
 651          break;
 652
 653    if (A == AEnd && B == BEnd)
 654       return 0;
 655    if (A == AEnd)
 656       return 1;
 657    if (B == BEnd)
 658       return -1;
 659    if (tolower_ascii(*A) < tolower_ascii(*B))
 660       return -1;
 661    return 1;
 662 }
 663 int stringcasecmp(string::const_iterator A,string::const_iterator AEnd,
 664                   string::const_iterator B,string::const_iterator BEnd)
 665 {
 666    for (; A != AEnd && B != BEnd; A++, B++)
 667       if (tolower_ascii(*A) != tolower_ascii(*B))
 668          break;
 669
 670    if (A == AEnd && B == BEnd)
 671       return 0;
 672    if (A == AEnd)
 673       return 1;
 674    if (B == BEnd)
 675       return -1;
 676    if (tolower_ascii(*A) < tolower_ascii(*B))
 677       return -1;
 678    return 1;
 679 }
 680 #endif
 681                                                                         /*}}}*/
 682 // LookupTag - Lookup the value of a tag in a taged string              /*{{{*/
 683 // ---------------------------------------------------------------------
 684 /* The format is like those used in package files and the method
 685    communication system */
 686 string LookupTag(const string &Message,const char *Tag,const char *Default)
 687 {
 688    // Look for a matching tag.
 689    int Length = strlen(Tag);
 690    for (string::const_iterator I = Message.begin(); I + Length < Message.end(); ++I)
 691    {
 692       // Found the tag
 693       if (I[Length] == ':' && stringcasecmp(I,I+Length,Tag) == 0)
 694       {
 695          // Find the end of line and strip the leading/trailing spaces
 696          string::const_iterator J;
 697          I += Length + 1;
 698          for (; isspace_ascii(*I) != 0 && I < Message.end(); ++I);
 699          for (J = I; *J != '\n' && J < Message.end(); ++J);
 700          for (; J > I && isspace_ascii(J[-1]) != 0; --J);
 701
 702          return string(I,J);
 703       }
 704
 705       for (; *I != '\n' && I < Message.end(); ++I);
 706    }
 707
 708    // Failed to find a match
 709    if (Default == 0)
 710       return string();
 711    return Default;
 712 }
 713                                                                         /*}}}*/
 714 // StringToBool - Converts a string into a boolean                      /*{{{*/
 715 // ---------------------------------------------------------------------
 716 /* This inspects the string to see if it is true or if it is false and
 717    then returns the result. Several varients on true/false are checked. */
 718 int StringToBool(const string &Text,int Default)
 719 {
 720    char *ParseEnd;
 721    int Res = strtol(Text.c_str(),&ParseEnd,0);
 722    // ensure that the entire string was converted by strtol to avoid
 723    // failures on "apt-cache show -a 0ad" where the "0" is converted
 724    const char *TextEnd = Text.c_str()+Text.size();
 725    if (ParseEnd == TextEnd && Res >= 0 && Res <= 1)
 726       return Res;
 727
 728    // Check for positives
 729    if (strcasecmp(Text.c_str(),"no") == 0 ||
 730        strcasecmp(Text.c_str(),"false") == 0 ||
 731        strcasecmp(Text.c_str(),"without") == 0 ||
 732        strcasecmp(Text.c_str(),"off") == 0 ||
 733        strcasecmp(Text.c_str(),"disable") == 0)
 734       return 0;
 735
 736    // Check for negatives
 737    if (strcasecmp(Text.c_str(),"yes") == 0 ||
 738        strcasecmp(Text.c_str(),"true") == 0 ||
 739        strcasecmp(Text.c_str(),"with") == 0 ||
 740        strcasecmp(Text.c_str(),"on") == 0 ||
 741        strcasecmp(Text.c_str(),"enable") == 0)
 742       return 1;
 743
 744    return Default;
 745 }
 746                                                                         /*}}}*/
 747 // TimeRFC1123 - Convert a time_t into RFC1123 format                   /*{{{*/
 748 // ---------------------------------------------------------------------
 749 /* This converts a time_t into a string time representation that is
 750    year 2000 complient and timezone neutral */
 751 string TimeRFC1123(time_t Date)
 752 {
 753    return TimeRFC1123(Date, false);
 754 }
 755 string TimeRFC1123(time_t Date, bool const NumericTimezone)
 756 {
 757    struct tm Conv;
 758    if (gmtime_r(&Date, &Conv) == NULL)
 759       return "";
 760
 761    auto const posix = std::locale("C.UTF-8");
 762    std::ostringstream datestr;
 763    datestr.imbue(posix);
 764    APT::StringView const fmt("%a, %d %b %Y %H:%M:%S");
 765    std::use_facet<std::time_put<char>>(posix).put(
 766                     std::ostreambuf_iterator<char>(datestr),
 767                     datestr, ' ', &Conv, fmt.data(), fmt.data() + fmt.size());
 768    if (NumericTimezone)
 769       datestr << " +0000";
 770    else
 771       datestr << " GMT";
 772    return datestr.str();
 773 }
 774                                                                         /*}}}*/
 775 // ReadMessages - Read messages from the FD                             /*{{{*/
 776 // ---------------------------------------------------------------------
 777 /* This pulls full messages from the input FD into the message buffer.
 778    It assumes that messages will not pause during transit so no
 779    fancy buffering is used.
 780
 781    In particular: this reads blocks from the input until it believes
 782    that it's run out of input text.  Each block is terminated by a
 783    double newline ('\n' followed by '\n').
 784  */
 785 bool ReadMessages(int Fd, vector<string> &List)
 786 {
 787    char Buffer[64000];
 788    // Represents any left-over from the previous iteration of the
 789    // parse loop.  (i.e., if a message is split across the end
 790    // of the buffer, it goes here)
 791    string PartialMessage;
 792
 793    do {
 794       int const Res = read(Fd, Buffer, sizeof(Buffer));
 795       if (Res < 0 && errno == EINTR)
 796          continue;
 797
 798       // process we read from has died
 799       if (Res == 0)
 800          return false;
 801
 802       // No data
 803 #if EAGAIN != EWOULDBLOCK
 804       if (Res < 0 && (errno == EAGAIN || errno == EWOULDBLOCK))
 805 #else
 806       if (Res < 0 && errno == EAGAIN)
 807 #endif
 808          return true;
 809       if (Res < 0)
 810          return false;
 811
 812       // extract the message(s) from the buffer
 813       char const *Start = Buffer;
 814       char const * const End = Buffer + Res;
 815
 816       char const * NL = (char const *) memchr(Start, '\n', End - Start);
 817       if (NL == NULL)
 818       {
 819          // end of buffer: store what we have so far and read new data in
 820          PartialMessage.append(Start, End - Start);
 821          Start = End;
 822       }
 823       else
 824          ++NL;
 825
 826       if (PartialMessage.empty() == false && Start < End)
 827       {
 828          // if we start with a new line, see if the partial message we have ended with one
 829          // so that we properly detect records ending between two read() runs
 830          // cases are: \n|\n  ,  \r\n|\r\n  and  \r\n\r|\n
 831          // the case \r|\n\r\n is handled by the usual double-newline handling
 832          if ((NL - Start) == 1 || ((NL - Start) == 2 && *Start == '\r'))
 833          {
 834             if (APT::String::Endswith(PartialMessage, "\n") || APT::String::Endswith(PartialMessage, "\r\n\r"))
 835             {
 836                PartialMessage.erase(PartialMessage.find_last_not_of("\r\n") + 1);
 837                List.push_back(PartialMessage);
 838                PartialMessage.clear();
 839                while (NL < End && (*NL == '\n' || *NL == '\r')) ++NL;
 840                Start = NL;
 841             }
 842          }
 843       }
 844
 845       while (Start < End) {
 846          char const * NL2 = (char const *) memchr(NL, '\n', End - NL);
 847          if (NL2 == NULL)
 848          {
 849             // end of buffer: store what we have so far and read new data in
 850             PartialMessage.append(Start, End - Start);
 851             break;
 852          }
 853          ++NL2;
 854
 855          // did we find a double newline?
 856          if ((NL2 - NL) == 1 || ((NL2 - NL) == 2 && *NL == '\r'))
 857          {
 858             PartialMessage.append(Start, NL2 - Start);
 859             PartialMessage.erase(PartialMessage.find_last_not_of("\r\n") + 1);
 860             List.push_back(PartialMessage);
 861             PartialMessage.clear();
 862             while (NL2 < End && (*NL2 == '\n' || *NL2 == '\r')) ++NL2;
 863             Start = NL2;
 864          }
 865          NL = NL2;
 866       }
 867
 868       // we have read at least one complete message and nothing left
 869       if (PartialMessage.empty() == true)
 870          return true;
 871
 872       if (WaitFd(Fd) == false)
 873          return false;
 874    } while (true);
 875 }
 876                                                                         /*}}}*/
 877 // MonthConv - Converts a month string into a number                    /*{{{*/
 878 // ---------------------------------------------------------------------
 879 /* This was lifted from the boa webserver which lifted it from 'wn-v1.07'
 880    Made it a bit more robust with a few tolower_ascii though. */
 881 static int MonthConv(char const * const Month)
 882 {
 883    switch (tolower_ascii(*Month))
 884    {
 885       case 'a':
 886       return tolower_ascii(Month[1]) == 'p'?3:7;
 887       case 'd':
 888       return 11;
 889       case 'f':
 890       return 1;
 891       case 'j':
 892       if (tolower_ascii(Month[1]) == 'a')
 893          return 0;
 894       return tolower_ascii(Month[2]) == 'n'?5:6;
 895       case 'm':
 896       return tolower_ascii(Month[2]) == 'r'?2:4;
 897       case 'n':
 898       return 10;
 899       case 'o':
 900       return 9;
 901       case 's':
 902       return 8;
 903
 904       // Pretend it is January..
 905       default:
 906       return 0;
 907    }
 908 }
 909                                                                         /*}}}*/
 910 // timegm - Internal timegm if the gnu version is not available         /*{{{*/
 911 // ---------------------------------------------------------------------
 912 /* Converts struct tm to time_t, assuming the data in tm is UTC rather
 913    than local timezone (mktime assumes the latter).
 914
 915    This function is a nonstandard GNU extension that is also present on
 916    the BSDs and maybe other systems. For others we follow the advice of
 917    the manpage of timegm and use his portable replacement. */
 918 #ifndef HAVE_TIMEGM
 919 static time_t timegm(struct tm *t)
 920 {
 921    char *tz = getenv("TZ");
 922    setenv("TZ", "", 1);
 923    tzset();
 924    time_t ret = mktime(t);
 925    if (tz)
 926       setenv("TZ", tz, 1);
 927    else
 928       unsetenv("TZ");
 929    tzset();
 930    return ret;
 931 }
 932 #endif
 933                                                                         /*}}}*/
 934 // RFC1123StrToTime - Converts a HTTP1.1 full date strings into a time_t        /*{{{*/
 935 // ---------------------------------------------------------------------
 936 /* tries to parses a full date as specified in RFC7231 §7.1.1.1
 937    with one exception: HTTP/1.1 valid dates need to have GMT as timezone.
 938    As we encounter dates from UTC or with a numeric timezone in other places,
 939    we allow them here to to be able to reuse the method. Either way, a date
 940    must be in UTC or parsing will fail. Previous implementations of this
 941    method used to ignore the timezone and assume always UTC. */
 942 bool RFC1123StrToTime(const char* const str,time_t &time)
 943 {
 944    unsigned short day = 0;
 945    signed int year = 0; // yes, Y23K problem – we gonna worry then…
 946    std::string weekday, month, datespec, timespec, zone;
 947    std::istringstream ss(str);
 948    auto const &posix = std::locale("C.UTF-8");
 949    ss.imbue(posix);
 950    ss >> weekday;
 951    // we only superficially check weekday, mostly to avoid accepting localized
 952    // weekdays here and take only its length to decide which datetime format we
 953    // encounter here. The date isn't stored.
 954    std::transform(weekday.begin(), weekday.end(), weekday.begin(), ::tolower);
 955    std::array<char const * const, 7> c_weekdays = {{ "sun", "mon", "tue", "wed", "thu", "fri", "sat" }};
 956    if (std::find(c_weekdays.begin(), c_weekdays.end(), weekday.substr(0,3)) == c_weekdays.end())
 957       return false;
 958
 959    switch (weekday.length())
 960    {
 961    case 4:
 962       // Sun, 06 Nov 1994 08:49:37 GMT ; RFC 822, updated by RFC 1123
 963       if (weekday[3] != ',')
 964          return false;
 965       ss >> day >> month >> year >> timespec >> zone;
 966       break;
 967    case 3:
 968       // Sun Nov  6 08:49:37 1994 ; ANSI C's asctime() format
 969       ss >> month >> day >> timespec >> year;
 970       zone = "UTC";
 971       break;
 972    case 0:
 973    case 1:
 974    case 2:
 975       return false;
 976    default:
 977       // Sunday, 06-Nov-94 08:49:37 GMT ; RFC 850, obsoleted by RFC 1036
 978       if (weekday[weekday.length() - 1] != ',')
 979          return false;
 980       ss >> datespec >> timespec >> zone;
 981       auto const expldate = VectorizeString(datespec, '-');
 982       if (expldate.size() != 3)
 983          return false;
 984       try {
 985          size_t pos;
 986          day = std::stoi(expldate[0], &pos);
 987          if (pos != expldate[0].length())
 988             return false;
 989          year = 1900 + std::stoi(expldate[2], &pos);
 990          if (pos != expldate[2].length())
 991             return false;
 992          strprintf(datespec, "%.4d-%.2d-%.2d", year, MonthConv(expldate[1].c_str()) + 1, day);
 993       } catch (...) {
 994          return false;
 995       }
 996       break;
 997    }
 998
 999    if (ss.fail() || ss.bad() || !ss.eof())
1000       return false;
1001
1002    if (zone != "GMT" && zone != "UTC" && zone != "Z") // RFC 822
1003    {
1004       // numeric timezones as a should of RFC 1123 and generally preferred
1005       try {
1006          size_t pos;
1007          auto const z = std::stoi(zone, &pos);
1008          if (z != 0 || pos != zone.length())
1009             return false;
1010       } catch (...) {
1011          return false;
1012       }
1013    }
1014
1015    if (datespec.empty())
1016    {
1017       if (month.empty())
1018          return false;
1019       strprintf(datespec, "%.4d-%.2d-%.2d", year, MonthConv(month.c_str()) + 1, day);
1020    }
1021
1022    std::string const datetime = datespec + ' ' + timespec;
1023    struct tm Tm;
1024    if (strptime(datetime.c_str(), "%Y-%m-%d %H:%M:%S", &Tm) == nullptr)
1025       return false;
1026    time = timegm(&Tm);
1027    return true;
1028 }
1029                                                                         /*}}}*/
1030 // FTPMDTMStrToTime - Converts a ftp modification date into a time_t    /*{{{*/
1031 // ---------------------------------------------------------------------
1032 /* */
1033 bool FTPMDTMStrToTime(const char* const str,time_t &time)
1034 {
1035    struct tm Tm;
1036    // MDTM includes no whitespaces but recommend and ignored by strptime
1037    if (strptime(str, "%Y %m %d %H %M %S", &Tm) == NULL)
1038       return false;
1039
1040    time = timegm(&Tm);
1041    return true;
1042 }
1043                                                                         /*}}}*/
1044 // StrToTime - Converts a string into a time_t                          /*{{{*/
1045 // ---------------------------------------------------------------------
1046 /* This handles all 3 popular time formats including RFC 1123, RFC 1036
1047    and the C library asctime format. It requires the GNU library function
1048    'timegm' to convert a struct tm in UTC to a time_t. For some bizzar
1049    reason the C library does not provide any such function :< This also
1050    handles the weird, but unambiguous FTP time format*/
1051 bool StrToTime(const string &Val,time_t &Result)
1052 {
1053    struct tm Tm;
1054    char Month[10];
1055
1056    // Skip the day of the week
1057    const char *I = strchr(Val.c_str(), ' ');
1058
1059    // Handle RFC 1123 time
1060    Month[0] = 0;
1061    if (sscanf(I," %2d %3s %4d %2d:%2d:%2d GMT",&Tm.tm_mday,Month,&Tm.tm_year,
1062               &Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) != 6)
1063    {
1064       // Handle RFC 1036 time
1065       if (sscanf(I," %2d-%3s-%3d %2d:%2d:%2d GMT",&Tm.tm_mday,Month,
1066                  &Tm.tm_year,&Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) == 6)
1067          Tm.tm_year += 1900;
1068       else
1069       {
1070          // asctime format
1071          if (sscanf(I," %3s %2d %2d:%2d:%2d %4d",Month,&Tm.tm_mday,
1072                     &Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec,&Tm.tm_year) != 6)
1073          {
1074             // 'ftp' time
1075             if (sscanf(Val.c_str(),"%4d%2d%2d%2d%2d%2d",&Tm.tm_year,&Tm.tm_mon,
1076                        &Tm.tm_mday,&Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) != 6)
1077                return false;
1078             Tm.tm_mon--;
1079          }
1080       }
1081    }
1082
1083    Tm.tm_isdst = 0;
1084    if (Month[0] != 0)
1085       Tm.tm_mon = MonthConv(Month);
1086    else
1087       Tm.tm_mon = 0; // we don't have a month, so pick something
1088    Tm.tm_year -= 1900;
1089
1090    // Convert to local time and then to GMT
1091    Result = timegm(&Tm);
1092    return true;
1093 }
1094                                                                         /*}}}*/
1095 // StrToNum - Convert a fixed length string to a number                 /*{{{*/
1096 // ---------------------------------------------------------------------
1097 /* This is used in decoding the crazy fixed length string headers in
1098    tar and ar files. */
1099 bool StrToNum(const char *Str,unsigned long &Res,unsigned Len,unsigned Base)
1100 {
1101    char S[30];
1102    if (Len >= sizeof(S))
1103       return false;
1104    memcpy(S,Str,Len);
1105    S[Len] = 0;
1106
1107    // All spaces is a zero
1108    Res = 0;
1109    unsigned I;
1110    for (I = 0; S[I] == ' '; I++);
1111    if (S[I] == 0)
1112       return true;
1113
1114    char *End;
1115    Res = strtoul(S,&End,Base);
1116    if (End == S)
1117       return false;
1118
1119    return true;
1120 }
1121                                                                         /*}}}*/
1122 // StrToNum - Convert a fixed length string to a number                 /*{{{*/
1123 // ---------------------------------------------------------------------
1124 /* This is used in decoding the crazy fixed length string headers in
1125    tar and ar files. */
1126 bool StrToNum(const char *Str,unsigned long long &Res,unsigned Len,unsigned Base)
1127 {
1128    char S[30];
1129    if (Len >= sizeof(S))
1130       return false;
1131    memcpy(S,Str,Len);
1132    S[Len] = 0;
1133
1134    // All spaces is a zero
1135    Res = 0;
1136    unsigned I;
1137    for (I = 0; S[I] == ' '; I++);
1138    if (S[I] == 0)
1139       return true;
1140
1141    char *End;
1142    Res = strtoull(S,&End,Base);
1143    if (End == S)
1144       return false;
1145
1146    return true;
1147 }
1148                                                                         /*}}}*/
1149
1150 // Base256ToNum - Convert a fixed length binary to a number             /*{{{*/
1151 // ---------------------------------------------------------------------
1152 /* This is used in decoding the 256bit encoded fixed length fields in
1153    tar files */
1154 bool Base256ToNum(const char *Str,unsigned long long &Res,unsigned int Len)
1155 {
1156    if ((Str[0] & 0x80) == 0)
1157       return false;
1158    else
1159    {
1160       Res = Str[0] & 0x7F;
1161       for(unsigned int i = 1; i < Len; ++i)
1162          Res = (Res<<8) + Str[i];
1163       return true;
1164    }
1165 }
1166                                                                         /*}}}*/
1167 // Base256ToNum - Convert a fixed length binary to a number             /*{{{*/
1168 // ---------------------------------------------------------------------
1169 /* This is used in decoding the 256bit encoded fixed length fields in
1170    tar files */
1171 bool Base256ToNum(const char *Str,unsigned long &Res,unsigned int Len)
1172 {
1173    unsigned long long Num;
1174    bool rc;
1175
1176    rc = Base256ToNum(Str, Num, Len);
1177    Res = Num;
1178    if (Res != Num)
1179       return false;
1180
1181    return rc;
1182 }
1183                                                                         /*}}}*/
1184 // HexDigit - Convert a hex character into an integer                   /*{{{*/
1185 // ---------------------------------------------------------------------
1186 /* Helper for Hex2Num */
1187 static int HexDigit(int c)
1188 {
1189    if (c >= '0' && c <= '9')
1190       return c - '0';
1191    if (c >= 'a' && c <= 'f')
1192       return c - 'a' + 10;
1193    if (c >= 'A' && c <= 'F')
1194       return c - 'A' + 10;
1195    return -1;
1196 }
1197                                                                         /*}}}*/
1198 // Hex2Num - Convert a long hex number into a buffer                    /*{{{*/
1199 // ---------------------------------------------------------------------
1200 /* The length of the buffer must be exactly 1/2 the length of the string. */
1201 bool Hex2Num(const string &Str,unsigned char *Num,unsigned int Length)
1202 {
1203    return Hex2Num(APT::StringView(Str), Num, Length);
1204 }
1205
1206 bool Hex2Num(const APT::StringView Str,unsigned char *Num,unsigned int Length)
1207 {
1208    if (Str.length() != Length*2)
1209       return false;
1210
1211    // Convert each digit. We store it in the same order as the string
1212    int J = 0;
1213    for (auto I = Str.begin(); I != Str.end();J++, I += 2)
1214    {
1215       int first_half = HexDigit(I[0]);
1216       int second_half;
1217       if (first_half < 0)
1218          return false;
1219
1220       second_half = HexDigit(I[1]);
1221       if (second_half < 0)
1222          return false;
1223       Num[J] = first_half << 4;
1224       Num[J] += second_half;
1225    }
1226
1227    return true;
1228 }
1229                                                                         /*}}}*/
1230 // TokSplitString - Split a string up by a given token                  /*{{{*/
1231 // ---------------------------------------------------------------------
1232 /* This is intended to be a faster splitter, it does not use dynamic
1233    memories. Input is changed to insert nulls at each token location. */
1234 bool TokSplitString(char Tok,char *Input,char **List,
1235                     unsigned long ListMax)
1236 {
1237    // Strip any leading spaces
1238    char *Start = Input;
1239    char *Stop = Start + strlen(Start);
1240    for (; *Start != 0 && isspace(*Start) != 0; Start++);
1241
1242    unsigned long Count = 0;
1243    char *Pos = Start;
1244    while (Pos != Stop)
1245    {
1246       // Skip to the next Token
1247       for (; Pos != Stop && *Pos != Tok; Pos++);
1248
1249       // Back remove spaces
1250       char *End = Pos;
1251       for (; End > Start && (End[-1] == Tok || isspace(End[-1]) != 0); End--);
1252       *End = 0;
1253
1254       List[Count++] = Start;
1255       if (Count >= ListMax)
1256       {
1257          List[Count-1] = 0;
1258          return false;
1259       }
1260
1261       // Advance pos
1262       for (; Pos != Stop && (*Pos == Tok || isspace(*Pos) != 0 || *Pos == 0); Pos++);
1263       Start = Pos;
1264    }
1265
1266    List[Count] = 0;
1267    return true;
1268 }
1269                                                                         /*}}}*/
1270 // VectorizeString - Split a string up into a vector of strings         /*{{{*/
1271 // ---------------------------------------------------------------------
1272 /* This can be used to split a given string up into a vector, so the
1273    propose is the same as in the method above and this one is a bit slower
1274    also, but the advantage is that we have an iteratable vector */
1275 vector<string> VectorizeString(string const &haystack, char const &split)
1276 {
1277    vector<string> exploded;
1278    if (haystack.empty() == true)
1279       return exploded;
1280    string::const_iterator start = haystack.begin();
1281    string::const_iterator end = start;
1282    do {
1283       for (; end != haystack.end() && *end != split; ++end);
1284       exploded.push_back(string(start, end));
1285       start = end + 1;
1286    } while (end != haystack.end() && (++end) != haystack.end());
1287    return exploded;
1288 }
1289                                                                         /*}}}*/
1290 // StringSplit - split a string into a string vector by token           /*{{{*/
1291 // ---------------------------------------------------------------------
1292 /* See header for details.
1293  */
1294 vector<string> StringSplit(std::string const &s, std::string const &sep,
1295                            unsigned int maxsplit)
1296 {
1297    vector<string> split;
1298    size_t start, pos;
1299
1300    // no separator given, this is bogus
1301    if(sep.size() == 0)
1302       return split;
1303
1304    start = pos = 0;
1305    while (pos != string::npos)
1306    {
1307       pos = s.find(sep, start);
1308       split.push_back(s.substr(start, pos-start));
1309
1310       // if maxsplit is reached, the remaining string is the last item
1311       if(split.size() >= maxsplit)
1312       {
1313          split[split.size()-1] = s.substr(start);
1314          break;
1315       }
1316       start = pos+sep.size();
1317    }
1318    return split;
1319 }
1320                                                                         /*}}}*/
1321 // RegexChoice - Simple regex list/list matcher                         /*{{{*/
1322 // ---------------------------------------------------------------------
1323 /* */
1324 unsigned long RegexChoice(RxChoiceList *Rxs,const char **ListBegin,
1325                       const char **ListEnd)
1326 {
1327    for (RxChoiceList *R = Rxs; R->Str != 0; R++)
1328       R->Hit = false;
1329
1330    unsigned long Hits = 0;
1331    for (; ListBegin < ListEnd; ++ListBegin)
1332    {
1333       // Check if the name is a regex
1334       const char *I;
1335       bool Regex = true;
1336       for (I = *ListBegin; *I != 0; I++)
1337          if (*I == '.' || *I == '?' || *I == '*' || *I == '|')
1338             break;
1339       if (*I == 0)
1340          Regex = false;
1341
1342       // Compile the regex pattern
1343       regex_t Pattern;
1344       if (Regex == true)
1345          if (regcomp(&Pattern,*ListBegin,REG_EXTENDED | REG_ICASE |
1346                      REG_NOSUB) != 0)
1347             Regex = false;
1348
1349       // Search the list
1350       bool Done = false;
1351       for (RxChoiceList *R = Rxs; R->Str != 0; R++)
1352       {
1353          if (R->Str[0] == 0)
1354             continue;
1355
1356          if (strcasecmp(R->Str,*ListBegin) != 0)
1357          {
1358             if (Regex == false)
1359                continue;
1360             if (regexec(&Pattern,R->Str,0,0,0) != 0)
1361                continue;
1362          }
1363          Done = true;
1364
1365          if (R->Hit == false)
1366             Hits++;
1367
1368          R->Hit = true;
1369       }
1370
1371       if (Regex == true)
1372          regfree(&Pattern);
1373
1374       if (Done == false)
1375          _error->Warning(_("Selection %s not found"),*ListBegin);
1376    }
1377
1378    return Hits;
1379 }
1380                                                                         /*}}}*/
1381 // {str,io}printf - C format string outputter to C++ strings/iostreams  /*{{{*/
1382 // ---------------------------------------------------------------------
1383 /* This is used to make the internationalization strings easier to translate
1384    and to allow reordering of parameters */
1385 static bool iovprintf(ostream &out, const char *format,
1386                       va_list &args, ssize_t &size) {
1387    char *S = (char*)malloc(size);
1388    ssize_t const n = vsnprintf(S, size, format, args);
1389    if (n > -1 && n < size) {
1390       out << S;
1391       free(S);
1392       return true;
1393    } else {
1394       if (n > -1)
1395          size = n + 1;
1396       else
1397          size *= 2;
1398    }
1399    free(S);
1400    return false;
1401 }
1402 void ioprintf(ostream &out,const char *format,...)
1403 {
1404    va_list args;
1405    ssize_t size = 400;
1406    while (true) {
1407       bool ret;
1408       va_start(args,format);
1409       ret = iovprintf(out, format, args, size);
1410       va_end(args);
1411       if (ret == true)
1412          return;
1413    }
1414 }
1415 void strprintf(string &out,const char *format,...)
1416 {
1417    va_list args;
1418    ssize_t size = 400;
1419    std::ostringstream outstr;
1420    while (true) {
1421       bool ret;
1422       va_start(args,format);
1423       ret = iovprintf(outstr, format, args, size);
1424       va_end(args);
1425       if (ret == true)
1426          break;
1427    }
1428    out = outstr.str();
1429 }
1430                                                                         /*}}}*/
1431 // safe_snprintf - Safer snprintf                                       /*{{{*/
1432 // ---------------------------------------------------------------------
1433 /* This is a snprintf that will never (ever) go past 'End' and returns a
1434    pointer to the end of the new string. The returned string is always null
1435    terminated unless Buffer == end. This is a better alterantive to using
1436    consecutive snprintfs. */
1437 char *safe_snprintf(char *Buffer,char *End,const char *Format,...)
1438 {
1439    va_list args;
1440    int Did;
1441
1442    if (End <= Buffer)
1443       return End;
1444    va_start(args,Format);
1445    Did = vsnprintf(Buffer,End - Buffer,Format,args);
1446    va_end(args);
1447
1448    if (Did < 0 || Buffer + Did > End)
1449       return End;
1450    return Buffer + Did;
1451 }
1452                                                                         /*}}}*/
1453 // StripEpoch - Remove the version "epoch" from a version string        /*{{{*/
1454 // ---------------------------------------------------------------------
1455 string StripEpoch(const string &VerStr)
1456 {
1457    size_t i = VerStr.find(":");
1458    if (i == string::npos)
1459       return VerStr;
1460    return VerStr.substr(i+1);
1461 }
1462                                                                         /*}}}*/
1463
1464 // tolower_ascii - tolower() function that ignores the locale           /*{{{*/
1465 // ---------------------------------------------------------------------
1466 /* This little function is the most called method we have and tries
1467    therefore to do the absolut minimum - and is notable faster than
1468    standard tolower/toupper and as a bonus avoids problems with different
1469    locales - we only operate on ascii chars anyway. */
1470 #undef tolower_ascii
1471 int tolower_ascii(int const c) APT_CONST APT_COLD;
1472 int tolower_ascii(int const c)
1473 {
1474    return tolower_ascii_inline(c);
1475 }
1476                                                                         /*}}}*/
1477
1478 // isspace_ascii - isspace() function that ignores the locale           /*{{{*/
1479 // ---------------------------------------------------------------------
1480 /* This little function is one of the most called methods we have and tries
1481    therefore to do the absolut minimum - and is notable faster than
1482    standard isspace() and as a bonus avoids problems with different
1483    locales - we only operate on ascii chars anyway. */
1484 #undef isspace_ascii
1485 int isspace_ascii(int const c) APT_CONST APT_COLD;
1486 int isspace_ascii(int const c)
1487 {
1488    return isspace_ascii_inline(c);
1489 }
1490                                                                         /*}}}*/
1491
1492 // CheckDomainList - See if Host is in a , separate list                /*{{{*/
1493 // ---------------------------------------------------------------------
1494 /* The domain list is a comma separate list of domains that are suffix
1495    matched against the argument */
1496 bool CheckDomainList(const string &Host,const string &List)
1497 {
1498    string::const_iterator Start = List.begin();
1499    for (string::const_iterator Cur = List.begin(); Cur <= List.end(); ++Cur)
1500    {
1501       if (Cur < List.end() && *Cur != ',')
1502          continue;
1503
1504       // Match the end of the string..
1505       if ((Host.size() >= (unsigned)(Cur - Start)) &&
1506           Cur - Start != 0 &&
1507           stringcasecmp(Host.end() - (Cur - Start),Host.end(),Start,Cur) == 0)
1508          return true;
1509
1510       Start = Cur + 1;
1511    }
1512    return false;
1513 }
1514                                                                         /*}}}*/
1515 // strv_length - Return the length of a NULL-terminated string array    /*{{{*/
1516 // ---------------------------------------------------------------------
1517 /* */
1518 size_t strv_length(const char **str_array)
1519 {
1520    size_t i;
1521    for (i=0; str_array[i] != NULL; i++)
1522       /* nothing */
1523       ;
1524    return i;
1525 }
1526                                                                         /*}}}*/
1527 // DeEscapeString - unescape (\0XX and \xXX) from a string              /*{{{*/
1528 // ---------------------------------------------------------------------
1529 /* */
1530 string DeEscapeString(const string &input)
1531 {
1532    char tmp[3];
1533    string::const_iterator it;
1534    string output;
1535    for (it = input.begin(); it != input.end(); ++it)
1536    {
1537       // just copy non-escape chars
1538       if (*it != '\\')
1539       {
1540          output += *it;
1541          continue;
1542       }
1543
1544       // deal with double escape
1545       if (*it == '\\' &&
1546           (it + 1 < input.end()) &&  it[1] == '\\')
1547       {
1548          // copy
1549          output += *it;
1550          // advance iterator one step further
1551          ++it;
1552          continue;
1553       }
1554
1555       // ensure we have a char to read
1556       if (it + 1 == input.end())
1557          continue;
1558
1559       // read it
1560       ++it;
1561       switch (*it)
1562       {
1563          case '0':
1564             if (it + 2 <= input.end()) {
1565                tmp[0] = it[1];
1566                tmp[1] = it[2];
1567                tmp[2] = 0;
1568                output += (char)strtol(tmp, 0, 8);
1569                it += 2;
1570             }
1571             break;
1572          case 'x':
1573             if (it + 2 <= input.end()) {
1574                tmp[0] = it[1];
1575                tmp[1] = it[2];
1576                tmp[2] = 0;
1577                output += (char)strtol(tmp, 0, 16);
1578                it += 2;
1579             }
1580             break;
1581          default:
1582             // FIXME: raise exception here?
1583             break;
1584       }
1585    }
1586    return output;
1587 }
1588                                                                         /*}}}*/
1589 // URI::CopyFrom - Copy from an object                                  /*{{{*/
1590 // ---------------------------------------------------------------------
1591 /* This parses the URI into all of its components */
1592 void URI::CopyFrom(const string &U)
1593 {
1594    string::const_iterator I = U.begin();
1595
1596    // Locate the first colon, this separates the scheme
1597    for (; I < U.end() && *I != ':' ; ++I);
1598    string::const_iterator FirstColon = I;
1599
1600    /* Determine if this is a host type URI with a leading double //
1601       and then search for the first single / */
1602    string::const_iterator SingleSlash = I;
1603    if (I + 3 < U.end() && I[1] == '/' && I[2] == '/')
1604       SingleSlash += 3;
1605
1606    /* Find the / indicating the end of the hostname, ignoring /'s in the
1607       square brackets */
1608    bool InBracket = false;
1609    for (; SingleSlash < U.end() && (*SingleSlash != '/' || InBracket == true); ++SingleSlash)
1610    {
1611       if (*SingleSlash == '[')
1612          InBracket = true;
1613       if (InBracket == true && *SingleSlash == ']')
1614          InBracket = false;
1615    }
1616
1617    if (SingleSlash > U.end())
1618       SingleSlash = U.end();
1619
1620    // We can now write the access and path specifiers
1621    Access.assign(U.begin(),FirstColon);
1622    if (SingleSlash != U.end())
1623       Path.assign(SingleSlash,U.end());
1624    if (Path.empty() == true)
1625       Path = "/";
1626
1627    // Now we attempt to locate a user:pass@host fragment
1628    if (FirstColon + 2 <= U.end() && FirstColon[1] == '/' && FirstColon[2] == '/')
1629       FirstColon += 3;
1630    else
1631       FirstColon += 1;
1632    if (FirstColon >= U.end())
1633       return;
1634
1635    if (FirstColon > SingleSlash)
1636       FirstColon = SingleSlash;
1637
1638    // Find the colon...
1639    I = FirstColon + 1;
1640    if (I > SingleSlash)
1641       I = SingleSlash;
1642    for (; I < SingleSlash && *I != ':'; ++I);
1643    string::const_iterator SecondColon = I;
1644
1645    // Search for the @ after the colon
1646    for (; I < SingleSlash && *I != '@'; ++I);
1647    string::const_iterator At = I;
1648
1649    // Now write the host and user/pass
1650    if (At == SingleSlash)
1651    {
1652       if (FirstColon < SingleSlash)
1653          Host.assign(FirstColon,SingleSlash);
1654    }
1655    else
1656    {
1657       Host.assign(At+1,SingleSlash);
1658       // username and password must be encoded (RFC 3986)
1659       User.assign(DeQuoteString(FirstColon,SecondColon));
1660       if (SecondColon < At)
1661          Password.assign(DeQuoteString(SecondColon+1,At));
1662    }
1663
1664    // Now we parse the RFC 2732 [] hostnames.
1665    unsigned long PortEnd = 0;
1666    InBracket = false;
1667    for (unsigned I = 0; I != Host.length();)
1668    {
1669       if (Host[I] == '[')
1670       {
1671          InBracket = true;
1672          Host.erase(I,1);
1673          continue;
1674       }
1675
1676       if (InBracket == true && Host[I] == ']')
1677       {
1678          InBracket = false;
1679          Host.erase(I,1);
1680          PortEnd = I;
1681          continue;
1682       }
1683       I++;
1684    }
1685
1686    // Tsk, weird.
1687    if (InBracket == true)
1688    {
1689       Host.clear();
1690       return;
1691    }
1692
1693    // Now we parse off a port number from the hostname
1694    Port = 0;
1695    string::size_type Pos = Host.rfind(':');
1696    if (Pos == string::npos || Pos < PortEnd)
1697       return;
1698
1699    Port = atoi(string(Host,Pos+1).c_str());
1700    Host.assign(Host,0,Pos);
1701 }
1702                                                                         /*}}}*/
1703 // URI::operator string - Convert the URI to a string                   /*{{{*/
1704 // ---------------------------------------------------------------------
1705 /* */
1706 URI::operator string()
1707 {
1708    std::stringstream Res;
1709
1710    if (Access.empty() == false)
1711       Res << Access << ':';
1712
1713    if (Host.empty() == false)
1714    {
1715       if (Access.empty() == false)
1716          Res << "//";
1717
1718       if (User.empty() == false)
1719       {
1720          // FIXME: Technically userinfo is permitted even less
1721          // characters than these, but this is not conveniently
1722          // expressed with a blacklist.
1723          Res << QuoteString(User, ":/?#[]@");
1724          if (Password.empty() == false)
1725             Res << ":" << QuoteString(Password, ":/?#[]@");
1726          Res << "@";
1727       }
1728
1729       // Add RFC 2732 escaping characters
1730       if (Access.empty() == false && Host.find_first_of("/:") != string::npos)
1731          Res << '[' << Host << ']';
1732       else
1733          Res << Host;
1734
1735       if (Port != 0)
1736          Res << ':' << std::to_string(Port);
1737    }
1738
1739    if (Path.empty() == false)
1740    {
1741       if (Path[0] != '/')
1742          Res << "/" << Path;
1743       else
1744          Res << Path;
1745    }
1746
1747    return Res.str();
1748 }
1749                                                                         /*}}}*/
1750 // URI::SiteOnly - Return the schema and site for the URI               /*{{{*/
1751 string URI::SiteOnly(const string &URI)
1752 {
1753    ::URI U(URI);
1754    U.User.clear();
1755    U.Password.clear();
1756    U.Path.clear();
1757    return U;
1758 }
1759                                                                         /*}}}*/
1760 // URI::ArchiveOnly - Return the schema, site and cleaned path for the URI /*{{{*/
1761 string URI::ArchiveOnly(const string &URI)
1762 {
1763    ::URI U(URI);
1764    U.User.clear();
1765    U.Password.clear();
1766    if (U.Path.empty() == false && U.Path[U.Path.length() - 1] == '/')
1767       U.Path.erase(U.Path.length() - 1);
1768    return U;
1769 }
1770                                                                         /*}}}*/
1771 // URI::NoUserPassword - Return the schema, site and path for the URI   /*{{{*/
1772 string URI::NoUserPassword(const string &URI)
1773 {
1774    ::URI U(URI);
1775    U.User.clear();
1776    U.Password.clear();
1777    return U;
1778 }
1779                                                                         /*}}}*/