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 <apt-pkg/strutl.h>
  19 #include <apt-pkg/fileutl.h>
  20 #include <apt-pkg/error.h>
  21
  22 #include <apti18n.h>
  23
  24 #include <ctype.h>
  25 #include <string.h>
  26 #include <stdio.h>
  27 #include <algorithm>
  28 #include <unistd.h>
  29 #include <regex.h>
  30 #include <errno.h>
  31 #include <stdarg.h>
  32 #include <iconv.h>
  33
  34 #include "config.h"
  35
  36 using namespace std;
  37                                                                         /*}}}*/
  38
  39 // UTF8ToCodeset - Convert some UTF-8 string for some codeset           /*{{{*/
  40 // ---------------------------------------------------------------------
  41 /* This is handy to use before display some information for enduser  */
  42 bool UTF8ToCodeset(const char *codeset, const string &orig, string *dest)
  43 {
  44   iconv_t cd;
  45   const char *inbuf;
  46   char *inptr, *outbuf;
  47   size_t insize, bufsize;
  48   dest->clear();
  49
  50   cd = iconv_open(codeset, "UTF-8");
  51   if (cd == (iconv_t)(-1)) {
  52      // Something went wrong
  53      if (errno == EINVAL)
  54         _error->Error("conversion from 'UTF-8' to '%s' not available",
  55                codeset);
  56      else
  57         perror("iconv_open");
  58
  59      return false;
  60   }
  61
  62   insize = bufsize = orig.size();
  63   inbuf = orig.data();
  64   inptr = (char *)inbuf;
  65   outbuf = new char[bufsize];
  66   size_t lastError = -1;
  67
  68   while (insize != 0)
  69   {
  70      char *outptr = outbuf;
  71      size_t outsize = bufsize;
  72      size_t const err = iconv(cd, &inptr, &insize, &outptr, &outsize);
  73      dest->append(outbuf, outptr - outbuf);
  74      if (err == (size_t)(-1))
  75      {
  76         switch (errno)
  77         {
  78         case EILSEQ:
  79            insize--;
  80            inptr++;
  81            // replace a series of unknown multibytes with a single "?"
  82            if (lastError != insize) {
  83               lastError = insize - 1;
  84               dest->append("?");
  85            }
  86            break;
  87         case EINVAL:
  88            insize = 0;
  89            break;
  90         case E2BIG:
  91            if (outptr == outbuf)
  92            {
  93               bufsize *= 2;
  94               delete[] outbuf;
  95               outbuf = new char[bufsize];
  96            }
  97            break;
  98         }
  99      }
 100   }
 101
 102   delete[] outbuf;
 103
 104   iconv_close(cd);
 105
 106   return true;
 107 }
 108                                                                         /*}}}*/
 109 // strstrip - Remove white space from the front and back of a string    /*{{{*/
 110 // ---------------------------------------------------------------------
 111 /* This is handy to use when parsing a file. It also removes \n's left
 112    over from fgets and company */
 113 char *_strstrip(char *String)
 114 {
 115    for (;*String != 0 && (*String == ' ' || *String == '\t'); String++);
 116
 117    if (*String == 0)
 118       return String;
 119
 120    char *End = String + strlen(String) - 1;
 121    for (;End != String - 1 && (*End == ' ' || *End == '\t' || *End == '\n' ||
 122                                *End == '\r'); End--);
 123    End++;
 124    *End = 0;
 125    return String;
 126 };
 127                                                                         /*}}}*/
 128 // strtabexpand - Converts tabs into 8 spaces                           /*{{{*/
 129 // ---------------------------------------------------------------------
 130 /* */
 131 char *_strtabexpand(char *String,size_t Len)
 132 {
 133    for (char *I = String; I != I + Len && *I != 0; I++)
 134    {
 135       if (*I != '\t')
 136          continue;
 137       if (I + 8 > String + Len)
 138       {
 139          *I = 0;
 140          return String;
 141       }
 142
 143       /* Assume the start of the string is 0 and find the next 8 char
 144          division */
 145       int Len;
 146       if (String == I)
 147          Len = 1;
 148       else
 149          Len = 8 - ((String - I) % 8);
 150       Len -= 2;
 151       if (Len <= 0)
 152       {
 153          *I = ' ';
 154          continue;
 155       }
 156
 157       memmove(I + Len,I + 1,strlen(I) + 1);
 158       for (char *J = I; J + Len != I; *I = ' ', I++);
 159    }
 160    return String;
 161 }
 162                                                                         /*}}}*/
 163 // ParseQuoteWord - Parse a single word out of a string                 /*{{{*/
 164 // ---------------------------------------------------------------------
 165 /* This grabs a single word, converts any % escaped characters to their
 166    proper values and advances the pointer. Double quotes are understood
 167    and striped out as well. This is for URI/URL parsing. It also can
 168    understand [] brackets.*/
 169 bool ParseQuoteWord(const char *&String,string &Res)
 170 {
 171    // Skip leading whitespace
 172    const char *C = String;
 173    for (;*C != 0 && *C == ' '; C++);
 174    if (*C == 0)
 175       return false;
 176
 177    // Jump to the next word
 178    for (;*C != 0 && isspace(*C) == 0; C++)
 179    {
 180       if (*C == '"')
 181       {
 182          for (C++; *C != 0 && *C != '"'; C++);
 183          if (*C == 0)
 184             return false;
 185       }
 186       if (*C == '[')
 187       {
 188          for (C++; *C != 0 && *C != ']'; C++);
 189          if (*C == 0)
 190             return false;
 191       }
 192    }
 193
 194    // Now de-quote characters
 195    char Buffer[1024];
 196    char Tmp[3];
 197    const char *Start = String;
 198    char *I;
 199    for (I = Buffer; I < Buffer + sizeof(Buffer) && Start != C; I++)
 200    {
 201       if (*Start == '%' && Start + 2 < C &&
 202           isxdigit(Start[1]) && isxdigit(Start[2]))
 203       {
 204          Tmp[0] = Start[1];
 205          Tmp[1] = Start[2];
 206          Tmp[2] = 0;
 207          *I = (char)strtol(Tmp,0,16);
 208          Start += 3;
 209          continue;
 210       }
 211       if (*Start != '"')
 212          *I = *Start;
 213       else
 214          I--;
 215       Start++;
 216    }
 217    *I = 0;
 218    Res = Buffer;
 219
 220    // Skip ending white space
 221    for (;*C != 0 && isspace(*C) != 0; C++);
 222    String = C;
 223    return true;
 224 }
 225                                                                         /*}}}*/
 226 // ParseCWord - Parses a string like a C "" expression                  /*{{{*/
 227 // ---------------------------------------------------------------------
 228 /* This expects a series of space separated strings enclosed in ""'s.
 229    It concatenates the ""'s into a single string. */
 230 bool ParseCWord(const char *&String,string &Res)
 231 {
 232    // Skip leading whitespace
 233    const char *C = String;
 234    for (;*C != 0 && *C == ' '; C++);
 235    if (*C == 0)
 236       return false;
 237
 238    char Buffer[1024];
 239    char *Buf = Buffer;
 240    if (strlen(String) >= sizeof(Buffer))
 241        return false;
 242
 243    for (; *C != 0; C++)
 244    {
 245       if (*C == '"')
 246       {
 247          for (C++; *C != 0 && *C != '"'; C++)
 248             *Buf++ = *C;
 249
 250          if (*C == 0)
 251             return false;
 252
 253          continue;
 254       }
 255
 256       if (C != String && isspace(*C) != 0 && isspace(C[-1]) != 0)
 257          continue;
 258       if (isspace(*C) == 0)
 259          return false;
 260       *Buf++ = ' ';
 261    }
 262    *Buf = 0;
 263    Res = Buffer;
 264    String = C;
 265    return true;
 266 }
 267                                                                         /*}}}*/
 268 // QuoteString - Convert a string into quoted from                      /*{{{*/
 269 // ---------------------------------------------------------------------
 270 /* */
 271 string QuoteString(const string &Str, const char *Bad)
 272 {
 273    string Res;
 274    for (string::const_iterator I = Str.begin(); I != Str.end(); I++)
 275    {
 276       if (strchr(Bad,*I) != 0 || isprint(*I) == 0 ||
 277           *I == 0x25 || // percent '%' char
 278           *I <= 0x20 || *I >= 0x7F) // control chars
 279       {
 280          char Buf[10];
 281          sprintf(Buf,"%%%02x",(int)*I);
 282          Res += Buf;
 283       }
 284       else
 285          Res += *I;
 286    }
 287    return Res;
 288 }
 289                                                                         /*}}}*/
 290 // DeQuoteString - Convert a string from quoted from                    /*{{{*/
 291 // ---------------------------------------------------------------------
 292 /* This undoes QuoteString */
 293 string DeQuoteString(const string &Str)
 294 {
 295    return DeQuoteString(Str.begin(),Str.end());
 296 }
 297 string DeQuoteString(string::const_iterator const &begin,
 298                         string::const_iterator const &end)
 299 {
 300    string Res;
 301    for (string::const_iterator I = begin; I != end; I++)
 302    {
 303       if (*I == '%' && I + 2 < end &&
 304           isxdigit(I[1]) && isxdigit(I[2]))
 305       {
 306          char Tmp[3];
 307          Tmp[0] = I[1];
 308          Tmp[1] = I[2];
 309          Tmp[2] = 0;
 310          Res += (char)strtol(Tmp,0,16);
 311          I += 2;
 312          continue;
 313       }
 314       else
 315          Res += *I;
 316    }
 317    return Res;
 318 }
 319
 320                                                                         /*}}}*/
 321 // SizeToStr - Convert a long into a human readable size                /*{{{*/
 322 // ---------------------------------------------------------------------
 323 /* A max of 4 digits are shown before conversion to the next highest unit.
 324    The max length of the string will be 5 chars unless the size is > 10
 325    YottaBytes (E24) */
 326 string SizeToStr(double Size)
 327 {
 328    char S[300];
 329    double ASize;
 330    if (Size >= 0)
 331       ASize = Size;
 332    else
 333       ASize = -1*Size;
 334
 335    /* bytes, KiloBytes, MegaBytes, GigaBytes, TeraBytes, PetaBytes,
 336       ExaBytes, ZettaBytes, YottaBytes */
 337    char Ext[] = {'\0','k','M','G','T','P','E','Z','Y'};
 338    int I = 0;
 339    while (I <= 8)
 340    {
 341       if (ASize < 100 && I != 0)
 342       {
 343          sprintf(S,"%'.1f%c",ASize,Ext[I]);
 344          break;
 345       }
 346
 347       if (ASize < 10000)
 348       {
 349          sprintf(S,"%'.0f%c",ASize,Ext[I]);
 350          break;
 351       }
 352       ASize /= 1000.0;
 353       I++;
 354    }
 355
 356    return S;
 357 }
 358                                                                         /*}}}*/
 359 // TimeToStr - Convert the time into a string                           /*{{{*/
 360 // ---------------------------------------------------------------------
 361 /* Converts a number of seconds to a hms format */
 362 string TimeToStr(unsigned long Sec)
 363 {
 364    char S[300];
 365
 366    while (1)
 367    {
 368       if (Sec > 60*60*24)
 369       {
 370          //d means days, h means hours, min means minutes, s means seconds
 371          sprintf(S,_("%lid %lih %limin %lis"),Sec/60/60/24,(Sec/60/60) % 24,(Sec/60) % 60,Sec % 60);
 372          break;
 373       }
 374
 375       if (Sec > 60*60)
 376       {
 377          //h means hours, min means minutes, s means seconds
 378          sprintf(S,_("%lih %limin %lis"),Sec/60/60,(Sec/60) % 60,Sec % 60);
 379          break;
 380       }
 381
 382       if (Sec > 60)
 383       {
 384          //min means minutes, s means seconds
 385          sprintf(S,_("%limin %lis"),Sec/60,Sec % 60);
 386          break;
 387       }
 388
 389       //s means seconds
 390       sprintf(S,_("%lis"),Sec);
 391       break;
 392    }
 393
 394    return S;
 395 }
 396                                                                         /*}}}*/
 397 // SubstVar - Substitute a string for another string                    /*{{{*/
 398 // ---------------------------------------------------------------------
 399 /* This replaces all occurances of Subst with Contents in Str. */
 400 string SubstVar(const string &Str,const string &Subst,const string &Contents)
 401 {
 402    string::size_type Pos = 0;
 403    string::size_type OldPos = 0;
 404    string Temp;
 405
 406    while (OldPos < Str.length() &&
 407           (Pos = Str.find(Subst,OldPos)) != string::npos)
 408    {
 409       Temp += string(Str,OldPos,Pos) + Contents;
 410       OldPos = Pos + Subst.length();
 411    }
 412
 413    if (OldPos == 0)
 414       return Str;
 415
 416    return Temp + string(Str,OldPos);
 417 }
 418
 419 string SubstVar(string Str,const struct SubstVar *Vars)
 420 {
 421    for (; Vars->Subst != 0; Vars++)
 422       Str = SubstVar(Str,Vars->Subst,*Vars->Contents);
 423    return Str;
 424 }
 425                                                                         /*}}}*/
 426 // OutputInDepth - return a string with separator multiplied with depth /*{{{*/
 427 // ---------------------------------------------------------------------
 428 /* Returns a string with the supplied separator depth + 1 times in it */
 429 std::string OutputInDepth(const unsigned long Depth, const char* Separator)
 430 {
 431    std::string output = "";
 432    for(unsigned long d=Depth+1; d > 0; d--)
 433       output.append(Separator);
 434    return output;
 435 }
 436                                                                         /*}}}*/
 437 // URItoFileName - Convert the uri into a unique file name              /*{{{*/
 438 // ---------------------------------------------------------------------
 439 /* This converts a URI into a safe filename. It quotes all unsafe characters
 440    and converts / to _ and removes the scheme identifier. The resulting
 441    file name should be unique and never occur again for a different file */
 442 string URItoFileName(const string &URI)
 443 {
 444    // Nuke 'sensitive' items
 445    ::URI U(URI);
 446    U.User.clear();
 447    U.Password.clear();
 448    U.Access.clear();
 449
 450    // "\x00-\x20{}|\\\\^\\[\\]<>\"\x7F-\xFF";
 451    string NewURI = QuoteString(U,"\\|{}[]<>\"^~_=!@#$%^&*");
 452    replace(NewURI.begin(),NewURI.end(),'/','_');
 453    return NewURI;
 454 }
 455                                                                         /*}}}*/
 456 // Base64Encode - Base64 Encoding routine for short strings             /*{{{*/
 457 // ---------------------------------------------------------------------
 458 /* This routine performs a base64 transformation on a string. It was ripped
 459    from wget and then patched and bug fixed.
 460
 461    This spec can be found in rfc2045 */
 462 string Base64Encode(const string &S)
 463 {
 464    // Conversion table.
 465    static char tbl[64] = {'A','B','C','D','E','F','G','H',
 466                           'I','J','K','L','M','N','O','P',
 467                           'Q','R','S','T','U','V','W','X',
 468                           'Y','Z','a','b','c','d','e','f',
 469                           'g','h','i','j','k','l','m','n',
 470                           'o','p','q','r','s','t','u','v',
 471                           'w','x','y','z','0','1','2','3',
 472                           '4','5','6','7','8','9','+','/'};
 473
 474    // Pre-allocate some space
 475    string Final;
 476    Final.reserve((4*S.length() + 2)/3 + 2);
 477
 478    /* Transform the 3x8 bits to 4x6 bits, as required by
 479       base64.  */
 480    for (string::const_iterator I = S.begin(); I < S.end(); I += 3)
 481    {
 482       char Bits[3] = {0,0,0};
 483       Bits[0] = I[0];
 484       if (I + 1 < S.end())
 485          Bits[1] = I[1];
 486       if (I + 2 < S.end())
 487          Bits[2] = I[2];
 488
 489       Final += tbl[Bits[0] >> 2];
 490       Final += tbl[((Bits[0] & 3) << 4) + (Bits[1] >> 4)];
 491
 492       if (I + 1 >= S.end())
 493          break;
 494
 495       Final += tbl[((Bits[1] & 0xf) << 2) + (Bits[2] >> 6)];
 496
 497       if (I + 2 >= S.end())
 498          break;
 499
 500       Final += tbl[Bits[2] & 0x3f];
 501    }
 502
 503    /* Apply the padding elements, this tells how many bytes the remote
 504       end should discard */
 505    if (S.length() % 3 == 2)
 506       Final += '=';
 507    if (S.length() % 3 == 1)
 508       Final += "==";
 509
 510    return Final;
 511 }
 512                                                                         /*}}}*/
 513 // stringcmp - Arbitrary string compare                                 /*{{{*/
 514 // ---------------------------------------------------------------------
 515 /* This safely compares two non-null terminated strings of arbitrary
 516    length */
 517 int stringcmp(const char *A,const char *AEnd,const char *B,const char *BEnd)
 518 {
 519    for (; A != AEnd && B != BEnd; A++, B++)
 520       if (*A != *B)
 521          break;
 522
 523    if (A == AEnd && B == BEnd)
 524       return 0;
 525    if (A == AEnd)
 526       return 1;
 527    if (B == BEnd)
 528       return -1;
 529    if (*A < *B)
 530       return -1;
 531    return 1;
 532 }
 533
 534 #if __GNUC__ >= 3
 535 int stringcmp(string::const_iterator A,string::const_iterator AEnd,
 536               const char *B,const char *BEnd)
 537 {
 538    for (; A != AEnd && B != BEnd; A++, B++)
 539       if (*A != *B)
 540          break;
 541
 542    if (A == AEnd && B == BEnd)
 543       return 0;
 544    if (A == AEnd)
 545       return 1;
 546    if (B == BEnd)
 547       return -1;
 548    if (*A < *B)
 549       return -1;
 550    return 1;
 551 }
 552 int stringcmp(string::const_iterator A,string::const_iterator AEnd,
 553               string::const_iterator B,string::const_iterator BEnd)
 554 {
 555    for (; A != AEnd && B != BEnd; A++, B++)
 556       if (*A != *B)
 557          break;
 558
 559    if (A == AEnd && B == BEnd)
 560       return 0;
 561    if (A == AEnd)
 562       return 1;
 563    if (B == BEnd)
 564       return -1;
 565    if (*A < *B)
 566       return -1;
 567    return 1;
 568 }
 569 #endif
 570                                                                         /*}}}*/
 571 // stringcasecmp - Arbitrary case insensitive string compare            /*{{{*/
 572 // ---------------------------------------------------------------------
 573 /* */
 574 int stringcasecmp(const char *A,const char *AEnd,const char *B,const char *BEnd)
 575 {
 576    for (; A != AEnd && B != BEnd; A++, B++)
 577       if (tolower_ascii(*A) != tolower_ascii(*B))
 578          break;
 579
 580    if (A == AEnd && B == BEnd)
 581       return 0;
 582    if (A == AEnd)
 583       return 1;
 584    if (B == BEnd)
 585       return -1;
 586    if (tolower_ascii(*A) < tolower_ascii(*B))
 587       return -1;
 588    return 1;
 589 }
 590 #if __GNUC__ >= 3
 591 int stringcasecmp(string::const_iterator A,string::const_iterator AEnd,
 592                   const char *B,const char *BEnd)
 593 {
 594    for (; A != AEnd && B != BEnd; A++, B++)
 595       if (tolower_ascii(*A) != tolower_ascii(*B))
 596          break;
 597
 598    if (A == AEnd && B == BEnd)
 599       return 0;
 600    if (A == AEnd)
 601       return 1;
 602    if (B == BEnd)
 603       return -1;
 604    if (tolower_ascii(*A) < tolower_ascii(*B))
 605       return -1;
 606    return 1;
 607 }
 608 int stringcasecmp(string::const_iterator A,string::const_iterator AEnd,
 609                   string::const_iterator B,string::const_iterator BEnd)
 610 {
 611    for (; A != AEnd && B != BEnd; A++, B++)
 612       if (tolower_ascii(*A) != tolower_ascii(*B))
 613          break;
 614
 615    if (A == AEnd && B == BEnd)
 616       return 0;
 617    if (A == AEnd)
 618       return 1;
 619    if (B == BEnd)
 620       return -1;
 621    if (tolower_ascii(*A) < tolower_ascii(*B))
 622       return -1;
 623    return 1;
 624 }
 625 #endif
 626                                                                         /*}}}*/
 627 // LookupTag - Lookup the value of a tag in a taged string              /*{{{*/
 628 // ---------------------------------------------------------------------
 629 /* The format is like those used in package files and the method
 630    communication system */
 631 string LookupTag(const string &Message,const char *Tag,const char *Default)
 632 {
 633    // Look for a matching tag.
 634    int Length = strlen(Tag);
 635    for (string::const_iterator I = Message.begin(); I + Length < Message.end(); I++)
 636    {
 637       // Found the tag
 638       if (I[Length] == ':' && stringcasecmp(I,I+Length,Tag) == 0)
 639       {
 640          // Find the end of line and strip the leading/trailing spaces
 641          string::const_iterator J;
 642          I += Length + 1;
 643          for (; isspace(*I) != 0 && I < Message.end(); I++);
 644          for (J = I; *J != '\n' && J < Message.end(); J++);
 645          for (; J > I && isspace(J[-1]) != 0; J--);
 646
 647          return string(I,J);
 648       }
 649
 650       for (; *I != '\n' && I < Message.end(); I++);
 651    }
 652
 653    // Failed to find a match
 654    if (Default == 0)
 655       return string();
 656    return Default;
 657 }
 658                                                                         /*}}}*/
 659 // StringToBool - Converts a string into a boolean                      /*{{{*/
 660 // ---------------------------------------------------------------------
 661 /* This inspects the string to see if it is true or if it is false and
 662    then returns the result. Several varients on true/false are checked. */
 663 int StringToBool(const string &Text,int Default)
 664 {
 665    char *End;
 666    int Res = strtol(Text.c_str(),&End,0);
 667    if (End != Text.c_str() && Res >= 0 && Res <= 1)
 668       return Res;
 669
 670    // Check for positives
 671    if (strcasecmp(Text.c_str(),"no") == 0 ||
 672        strcasecmp(Text.c_str(),"false") == 0 ||
 673        strcasecmp(Text.c_str(),"without") == 0 ||
 674        strcasecmp(Text.c_str(),"off") == 0 ||
 675        strcasecmp(Text.c_str(),"disable") == 0)
 676       return 0;
 677
 678    // Check for negatives
 679    if (strcasecmp(Text.c_str(),"yes") == 0 ||
 680        strcasecmp(Text.c_str(),"true") == 0 ||
 681        strcasecmp(Text.c_str(),"with") == 0 ||
 682        strcasecmp(Text.c_str(),"on") == 0 ||
 683        strcasecmp(Text.c_str(),"enable") == 0)
 684       return 1;
 685
 686    return Default;
 687 }
 688                                                                         /*}}}*/
 689 // TimeRFC1123 - Convert a time_t into RFC1123 format                   /*{{{*/
 690 // ---------------------------------------------------------------------
 691 /* This converts a time_t into a string time representation that is
 692    year 2000 complient and timezone neutral */
 693 string TimeRFC1123(time_t Date)
 694 {
 695    struct tm Conv = *gmtime(&Date);
 696    char Buf[300];
 697
 698    const char *Day[] = {"Sun","Mon","Tue","Wed","Thu","Fri","Sat"};
 699    const char *Month[] = {"Jan","Feb","Mar","Apr","May","Jun","Jul",
 700                           "Aug","Sep","Oct","Nov","Dec"};
 701
 702    sprintf(Buf,"%s, %02i %s %i %02i:%02i:%02i GMT",Day[Conv.tm_wday],
 703            Conv.tm_mday,Month[Conv.tm_mon],Conv.tm_year+1900,Conv.tm_hour,
 704            Conv.tm_min,Conv.tm_sec);
 705    return Buf;
 706 }
 707                                                                         /*}}}*/
 708 // ReadMessages - Read messages from the FD                             /*{{{*/
 709 // ---------------------------------------------------------------------
 710 /* This pulls full messages from the input FD into the message buffer.
 711    It assumes that messages will not pause during transit so no
 712    fancy buffering is used.
 713
 714    In particular: this reads blocks from the input until it believes
 715    that it's run out of input text.  Each block is terminated by a
 716    double newline ('\n' followed by '\n').  As noted below, there is a
 717    bug in this code: it assumes that all the blocks have been read if
 718    it doesn't see additional text in the buffer after the last one is
 719    parsed, which will cause it to lose blocks if the last block
 720    coincides with the end of the buffer.
 721  */
 722 bool ReadMessages(int Fd, vector<string> &List)
 723 {
 724    char Buffer[64000];
 725    char *End = Buffer;
 726    // Represents any left-over from the previous iteration of the
 727    // parse loop.  (i.e., if a message is split across the end
 728    // of the buffer, it goes here)
 729    string PartialMessage;
 730
 731    while (1)
 732    {
 733       int Res = read(Fd,End,sizeof(Buffer) - (End-Buffer));
 734       if (Res < 0 && errno == EINTR)
 735          continue;
 736
 737       // Process is dead, this is kind of bad..
 738       if (Res == 0)
 739          return false;
 740
 741       // No data
 742       if (Res < 0 && errno == EAGAIN)
 743          return true;
 744       if (Res < 0)
 745          return false;
 746
 747       End += Res;
 748
 749       // Look for the end of the message
 750       for (char *I = Buffer; I + 1 < End; I++)
 751       {
 752          if (I[0] != '\n' || I[1] != '\n')
 753             continue;
 754
 755          // Pull the message out
 756          string Message(Buffer,I-Buffer);
 757          PartialMessage += Message;
 758
 759          // Fix up the buffer
 760          for (; I < End && *I == '\n'; I++);
 761          End -= I-Buffer;
 762          memmove(Buffer,I,End-Buffer);
 763          I = Buffer;
 764
 765          List.push_back(PartialMessage);
 766          PartialMessage.clear();
 767       }
 768       if (End != Buffer)
 769         {
 770           // If there's text left in the buffer, store it
 771           // in PartialMessage and throw the rest of the buffer
 772           // away.  This allows us to handle messages that
 773           // are longer than the static buffer size.
 774           PartialMessage += string(Buffer, End);
 775           End = Buffer;
 776         }
 777       else
 778         {
 779           // BUG ALERT: if a message block happens to end at a
 780           // multiple of 64000 characters, this will cause it to
 781           // terminate early, leading to a badly formed block and
 782           // probably crashing the method.  However, this is the only
 783           // way we have to find the end of the message block.  I have
 784           // an idea of how to fix this, but it will require changes
 785           // to the protocol (essentially to mark the beginning and
 786           // end of the block).
 787           //
 788           //  -- dburrows 2008-04-02
 789           return true;
 790         }
 791
 792       if (WaitFd(Fd) == false)
 793          return false;
 794    }
 795 }
 796                                                                         /*}}}*/
 797 // MonthConv - Converts a month string into a number                    /*{{{*/
 798 // ---------------------------------------------------------------------
 799 /* This was lifted from the boa webserver which lifted it from 'wn-v1.07'
 800    Made it a bit more robust with a few tolower_ascii though. */
 801 static int MonthConv(char *Month)
 802 {
 803    switch (tolower_ascii(*Month))
 804    {
 805       case 'a':
 806       return tolower_ascii(Month[1]) == 'p'?3:7;
 807       case 'd':
 808       return 11;
 809       case 'f':
 810       return 1;
 811       case 'j':
 812       if (tolower_ascii(Month[1]) == 'a')
 813          return 0;
 814       return tolower_ascii(Month[2]) == 'n'?5:6;
 815       case 'm':
 816       return tolower_ascii(Month[2]) == 'r'?2:4;
 817       case 'n':
 818       return 10;
 819       case 'o':
 820       return 9;
 821       case 's':
 822       return 8;
 823
 824       // Pretend it is January..
 825       default:
 826       return 0;
 827    }
 828 }
 829                                                                         /*}}}*/
 830 // timegm - Internal timegm function if gnu is not available            /*{{{*/
 831 // ---------------------------------------------------------------------
 832 /* Ripped this evil little function from wget - I prefer the use of
 833    GNU timegm if possible as this technique will have interesting problems
 834    with leap seconds, timezones and other.
 835
 836    Converts struct tm to time_t, assuming the data in tm is UTC rather
 837    than local timezone (mktime assumes the latter).
 838
 839    Contributed by Roger Beeman <beeman@cisco.com>, with the help of
 840    Mark Baushke <mdb@cisco.com> and the rest of the Gurus at CISCO. */
 841
 842 /* Turned it into an autoconf check, because GNU is not the only thing which
 843    can provide timegm. -- 2002-09-22, Joel Baker */
 844
 845 #ifndef HAVE_TIMEGM // Now with autoconf!
 846 static time_t timegm(struct tm *t)
 847 {
 848    time_t tl, tb;
 849
 850    tl = mktime (t);
 851    if (tl == -1)
 852       return -1;
 853    tb = mktime (gmtime (&tl));
 854    return (tl <= tb ? (tl + (tl - tb)) : (tl - (tb - tl)));
 855 }
 856 #endif
 857                                                                         /*}}}*/
 858 // StrToTime - Converts a string into a time_t                          /*{{{*/
 859 // ---------------------------------------------------------------------
 860 /* This handles all 3 populare time formats including RFC 1123, RFC 1036
 861    and the C library asctime format. It requires the GNU library function
 862    'timegm' to convert a struct tm in UTC to a time_t. For some bizzar
 863    reason the C library does not provide any such function :< This also
 864    handles the weird, but unambiguous FTP time format*/
 865 bool StrToTime(const string &Val,time_t &Result)
 866 {
 867    struct tm Tm;
 868    char Month[10];
 869    const char *I = Val.c_str();
 870
 871    // Skip the day of the week
 872    for (;*I != 0  && *I != ' '; I++);
 873
 874    // Handle RFC 1123 time
 875    Month[0] = 0;
 876    if (sscanf(I," %d %3s %d %d:%d:%d GMT",&Tm.tm_mday,Month,&Tm.tm_year,
 877               &Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) != 6)
 878    {
 879       // Handle RFC 1036 time
 880       if (sscanf(I," %d-%3s-%d %d:%d:%d GMT",&Tm.tm_mday,Month,
 881                  &Tm.tm_year,&Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) == 6)
 882          Tm.tm_year += 1900;
 883       else
 884       {
 885          // asctime format
 886          if (sscanf(I," %3s %d %d:%d:%d %d",Month,&Tm.tm_mday,
 887                     &Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec,&Tm.tm_year) != 6)
 888          {
 889             // 'ftp' time
 890             if (sscanf(Val.c_str(),"%4d%2d%2d%2d%2d%2d",&Tm.tm_year,&Tm.tm_mon,
 891                        &Tm.tm_mday,&Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) != 6)
 892                return false;
 893             Tm.tm_mon--;
 894          }
 895       }
 896    }
 897
 898    Tm.tm_isdst = 0;
 899    if (Month[0] != 0)
 900       Tm.tm_mon = MonthConv(Month);
 901    Tm.tm_year -= 1900;
 902
 903    // Convert to local time and then to GMT
 904    Result = timegm(&Tm);
 905    return true;
 906 }
 907                                                                         /*}}}*/
 908 // StrToNum - Convert a fixed length string to a number                 /*{{{*/
 909 // ---------------------------------------------------------------------
 910 /* This is used in decoding the crazy fixed length string headers in
 911    tar and ar files. */
 912 bool StrToNum(const char *Str,unsigned long &Res,unsigned Len,unsigned Base)
 913 {
 914    char S[30];
 915    if (Len >= sizeof(S))
 916       return false;
 917    memcpy(S,Str,Len);
 918    S[Len] = 0;
 919
 920    // All spaces is a zero
 921    Res = 0;
 922    unsigned I;
 923    for (I = 0; S[I] == ' '; I++);
 924    if (S[I] == 0)
 925       return true;
 926
 927    char *End;
 928    Res = strtoul(S,&End,Base);
 929    if (End == S)
 930       return false;
 931
 932    return true;
 933 }
 934                                                                         /*}}}*/
 935 // HexDigit - Convert a hex character into an integer                   /*{{{*/
 936 // ---------------------------------------------------------------------
 937 /* Helper for Hex2Num */
 938 static int HexDigit(int c)
 939 {
 940    if (c >= '0' && c <= '9')
 941       return c - '0';
 942    if (c >= 'a' && c <= 'f')
 943       return c - 'a' + 10;
 944    if (c >= 'A' && c <= 'F')
 945       return c - 'A' + 10;
 946    return 0;
 947 }
 948                                                                         /*}}}*/
 949 // Hex2Num - Convert a long hex number into a buffer                    /*{{{*/
 950 // ---------------------------------------------------------------------
 951 /* The length of the buffer must be exactly 1/2 the length of the string. */
 952 bool Hex2Num(const string &Str,unsigned char *Num,unsigned int Length)
 953 {
 954    if (Str.length() != Length*2)
 955       return false;
 956
 957    // Convert each digit. We store it in the same order as the string
 958    int J = 0;
 959    for (string::const_iterator I = Str.begin(); I != Str.end();J++, I += 2)
 960    {
 961       if (isxdigit(*I) == 0 || isxdigit(I[1]) == 0)
 962          return false;
 963
 964       Num[J] = HexDigit(I[0]) << 4;
 965       Num[J] += HexDigit(I[1]);
 966    }
 967
 968    return true;
 969 }
 970                                                                         /*}}}*/
 971 // TokSplitString - Split a string up by a given token                  /*{{{*/
 972 // ---------------------------------------------------------------------
 973 /* This is intended to be a faster splitter, it does not use dynamic
 974    memories. Input is changed to insert nulls at each token location. */
 975 bool TokSplitString(char Tok,char *Input,char **List,
 976                     unsigned long ListMax)
 977 {
 978    // Strip any leading spaces
 979    char *Start = Input;
 980    char *Stop = Start + strlen(Start);
 981    for (; *Start != 0 && isspace(*Start) != 0; Start++);
 982
 983    unsigned long Count = 0;
 984    char *Pos = Start;
 985    while (Pos != Stop)
 986    {
 987       // Skip to the next Token
 988       for (; Pos != Stop && *Pos != Tok; Pos++);
 989
 990       // Back remove spaces
 991       char *End = Pos;
 992       for (; End > Start && (End[-1] == Tok || isspace(End[-1]) != 0); End--);
 993       *End = 0;
 994
 995       List[Count++] = Start;
 996       if (Count >= ListMax)
 997       {
 998          List[Count-1] = 0;
 999          return false;
1000       }
1001
1002       // Advance pos
1003       for (; Pos != Stop && (*Pos == Tok || isspace(*Pos) != 0 || *Pos == 0); Pos++);
1004       Start = Pos;
1005    }
1006
1007    List[Count] = 0;
1008    return true;
1009 }
1010                                                                         /*}}}*/
1011 // VectorizeString - Split a string up into a vector of strings         /*{{{*/
1012 // ---------------------------------------------------------------------
1013 /* This can be used to split a given string up into a vector, so the
1014    propose is the same as in the method above and this one is a bit slower
1015    also, but the advantage is that we have an iteratable vector */
1016 vector<string> VectorizeString(string const &haystack, char const &split)
1017 {
1018    string::const_iterator start = haystack.begin();
1019    string::const_iterator end = start;
1020    vector<string> exploded;
1021    do {
1022       for (; end != haystack.end() && *end != split; ++end);
1023       exploded.push_back(string(start, end));
1024       start = end + 1;
1025    } while (end != haystack.end() && (++end) != haystack.end());
1026    return exploded;
1027 }
1028                                                                         /*}}}*/
1029 // RegexChoice - Simple regex list/list matcher                         /*{{{*/
1030 // ---------------------------------------------------------------------
1031 /* */
1032 unsigned long RegexChoice(RxChoiceList *Rxs,const char **ListBegin,
1033                       const char **ListEnd)
1034 {
1035    for (RxChoiceList *R = Rxs; R->Str != 0; R++)
1036       R->Hit = false;
1037
1038    unsigned long Hits = 0;
1039    for (; ListBegin != ListEnd; ListBegin++)
1040    {
1041       // Check if the name is a regex
1042       const char *I;
1043       bool Regex = true;
1044       for (I = *ListBegin; *I != 0; I++)
1045          if (*I == '.' || *I == '?' || *I == '*' || *I == '|')
1046             break;
1047       if (*I == 0)
1048          Regex = false;
1049
1050       // Compile the regex pattern
1051       regex_t Pattern;
1052       if (Regex == true)
1053          if (regcomp(&Pattern,*ListBegin,REG_EXTENDED | REG_ICASE |
1054                      REG_NOSUB) != 0)
1055             Regex = false;
1056
1057       // Search the list
1058       bool Done = false;
1059       for (RxChoiceList *R = Rxs; R->Str != 0; R++)
1060       {
1061          if (R->Str[0] == 0)
1062             continue;
1063
1064          if (strcasecmp(R->Str,*ListBegin) != 0)
1065          {
1066             if (Regex == false)
1067                continue;
1068             if (regexec(&Pattern,R->Str,0,0,0) != 0)
1069                continue;
1070          }
1071          Done = true;
1072
1073          if (R->Hit == false)
1074             Hits++;
1075
1076          R->Hit = true;
1077       }
1078
1079       if (Regex == true)
1080          regfree(&Pattern);
1081
1082       if (Done == false)
1083          _error->Warning(_("Selection %s not found"),*ListBegin);
1084    }
1085
1086    return Hits;
1087 }
1088                                                                         /*}}}*/
1089 // ioprintf - C format string outputter to C++ iostreams                /*{{{*/
1090 // ---------------------------------------------------------------------
1091 /* This is used to make the internationalization strings easier to translate
1092    and to allow reordering of parameters */
1093 void ioprintf(ostream &out,const char *format,...)
1094 {
1095    va_list args;
1096    va_start(args,format);
1097
1098    // sprintf the description
1099    char S[4096];
1100    vsnprintf(S,sizeof(S),format,args);
1101    out << S;
1102 }
1103                                                                         /*}}}*/
1104 // strprintf - C format string outputter to C++ strings                 /*{{{*/
1105 // ---------------------------------------------------------------------
1106 /* This is used to make the internationalization strings easier to translate
1107    and to allow reordering of parameters */
1108 void strprintf(string &out,const char *format,...)
1109 {
1110    va_list args;
1111    va_start(args,format);
1112
1113    // sprintf the description
1114    char S[4096];
1115    vsnprintf(S,sizeof(S),format,args);
1116    out = string(S);
1117 }
1118                                                                         /*}}}*/
1119 // safe_snprintf - Safer snprintf                                       /*{{{*/
1120 // ---------------------------------------------------------------------
1121 /* This is a snprintf that will never (ever) go past 'End' and returns a
1122    pointer to the end of the new string. The returned string is always null
1123    terminated unless Buffer == end. This is a better alterantive to using
1124    consecutive snprintfs. */
1125 char *safe_snprintf(char *Buffer,char *End,const char *Format,...)
1126 {
1127    va_list args;
1128    unsigned long Did;
1129
1130    va_start(args,Format);
1131
1132    if (End <= Buffer)
1133       return End;
1134
1135    Did = vsnprintf(Buffer,End - Buffer,Format,args);
1136    if (Did < 0 || Buffer + Did > End)
1137       return End;
1138    return Buffer + Did;
1139 }
1140                                                                         /*}}}*/
1141
1142 // tolower_ascii - tolower() function that ignores the locale           /*{{{*/
1143 // ---------------------------------------------------------------------
1144 /* This little function is the most called method we have and tries
1145    therefore to do the absolut minimum - and is noteable faster than
1146    standard tolower/toupper and as a bonus avoids problems with different
1147    locales - we only operate on ascii chars anyway. */
1148 int tolower_ascii(int const c)
1149 {
1150    if (c >= 'A' && c <= 'Z')
1151       return c + 32;
1152    return c;
1153 }
1154                                                                         /*}}}*/
1155
1156 // CheckDomainList - See if Host is in a , seperate list                /*{{{*/
1157 // ---------------------------------------------------------------------
1158 /* The domain list is a comma seperate list of domains that are suffix
1159    matched against the argument */
1160 bool CheckDomainList(const string &Host,const string &List)
1161 {
1162    string::const_iterator Start = List.begin();
1163    for (string::const_iterator Cur = List.begin(); Cur <= List.end(); Cur++)
1164    {
1165       if (Cur < List.end() && *Cur != ',')
1166          continue;
1167
1168       // Match the end of the string..
1169       if ((Host.size() >= (unsigned)(Cur - Start)) &&
1170           Cur - Start != 0 &&
1171           stringcasecmp(Host.end() - (Cur - Start),Host.end(),Start,Cur) == 0)
1172          return true;
1173
1174       Start = Cur + 1;
1175    }
1176    return false;
1177 }
1178                                                                         /*}}}*/
1179
1180 // URI::CopyFrom - Copy from an object                                  /*{{{*/
1181 // ---------------------------------------------------------------------
1182 /* This parses the URI into all of its components */
1183 void URI::CopyFrom(const string &U)
1184 {
1185    string::const_iterator I = U.begin();
1186
1187    // Locate the first colon, this separates the scheme
1188    for (; I < U.end() && *I != ':' ; I++);
1189    string::const_iterator FirstColon = I;
1190
1191    /* Determine if this is a host type URI with a leading double //
1192       and then search for the first single / */
1193    string::const_iterator SingleSlash = I;
1194    if (I + 3 < U.end() && I[1] == '/' && I[2] == '/')
1195       SingleSlash += 3;
1196
1197    /* Find the / indicating the end of the hostname, ignoring /'s in the
1198       square brackets */
1199    bool InBracket = false;
1200    for (; SingleSlash < U.end() && (*SingleSlash != '/' || InBracket == true); SingleSlash++)
1201    {
1202       if (*SingleSlash == '[')
1203          InBracket = true;
1204       if (InBracket == true && *SingleSlash == ']')
1205          InBracket = false;
1206    }
1207
1208    if (SingleSlash > U.end())
1209       SingleSlash = U.end();
1210
1211    // We can now write the access and path specifiers
1212    Access.assign(U.begin(),FirstColon);
1213    if (SingleSlash != U.end())
1214       Path.assign(SingleSlash,U.end());
1215    if (Path.empty() == true)
1216       Path = "/";
1217
1218    // Now we attempt to locate a user:pass@host fragment
1219    if (FirstColon + 2 <= U.end() && FirstColon[1] == '/' && FirstColon[2] == '/')
1220       FirstColon += 3;
1221    else
1222       FirstColon += 1;
1223    if (FirstColon >= U.end())
1224       return;
1225
1226    if (FirstColon > SingleSlash)
1227       FirstColon = SingleSlash;
1228
1229    // Find the colon...
1230    I = FirstColon + 1;
1231    if (I > SingleSlash)
1232       I = SingleSlash;
1233    for (; I < SingleSlash && *I != ':'; I++);
1234    string::const_iterator SecondColon = I;
1235
1236    // Search for the @ after the colon
1237    for (; I < SingleSlash && *I != '@'; I++);
1238    string::const_iterator At = I;
1239
1240    // Now write the host and user/pass
1241    if (At == SingleSlash)
1242    {
1243       if (FirstColon < SingleSlash)
1244          Host.assign(FirstColon,SingleSlash);
1245    }
1246    else
1247    {
1248       Host.assign(At+1,SingleSlash);
1249       // username and password must be encoded (RFC 3986)
1250       User.assign(DeQuoteString(FirstColon,SecondColon));
1251       if (SecondColon < At)
1252          Password.assign(DeQuoteString(SecondColon+1,At));
1253    }
1254
1255    // Now we parse the RFC 2732 [] hostnames.
1256    unsigned long PortEnd = 0;
1257    InBracket = false;
1258    for (unsigned I = 0; I != Host.length();)
1259    {
1260       if (Host[I] == '[')
1261       {
1262          InBracket = true;
1263          Host.erase(I,1);
1264          continue;
1265       }
1266
1267       if (InBracket == true && Host[I] == ']')
1268       {
1269          InBracket = false;
1270          Host.erase(I,1);
1271          PortEnd = I;
1272          continue;
1273       }
1274       I++;
1275    }
1276
1277    // Tsk, weird.
1278    if (InBracket == true)
1279    {
1280       Host.clear();
1281       return;
1282    }
1283
1284    // Now we parse off a port number from the hostname
1285    Port = 0;
1286    string::size_type Pos = Host.rfind(':');
1287    if (Pos == string::npos || Pos < PortEnd)
1288       return;
1289
1290    Port = atoi(string(Host,Pos+1).c_str());
1291    Host.assign(Host,0,Pos);
1292 }
1293                                                                         /*}}}*/
1294 // URI::operator string - Convert the URI to a string                   /*{{{*/
1295 // ---------------------------------------------------------------------
1296 /* */
1297 URI::operator string()
1298 {
1299    string Res;
1300
1301    if (Access.empty() == false)
1302       Res = Access + ':';
1303
1304    if (Host.empty() == false)
1305    {
1306       if (Access.empty() == false)
1307          Res += "//";
1308
1309       if (User.empty() == false)
1310       {
1311          Res +=  User;
1312          if (Password.empty() == false)
1313             Res += ":" + Password;
1314          Res += "@";
1315       }
1316
1317       // Add RFC 2732 escaping characters
1318       if (Access.empty() == false &&
1319           (Host.find('/') != string::npos || Host.find(':') != string::npos))
1320          Res += '[' + Host + ']';
1321       else
1322          Res += Host;
1323
1324       if (Port != 0)
1325       {
1326          char S[30];
1327          sprintf(S,":%u",Port);
1328          Res += S;
1329       }
1330    }
1331
1332    if (Path.empty() == false)
1333    {
1334       if (Path[0] != '/')
1335          Res += "/" + Path;
1336       else
1337          Res += Path;
1338    }
1339
1340    return Res;
1341 }
1342                                                                         /*}}}*/
1343 // URI::SiteOnly - Return the schema and site for the URI               /*{{{*/
1344 // ---------------------------------------------------------------------
1345 /* */
1346 string URI::SiteOnly(const string &URI)
1347 {
1348    ::URI U(URI);
1349    U.User.clear();
1350    U.Password.clear();
1351    U.Path.clear();
1352    U.Port = 0;
1353    return U;
1354 }
1355                                                                         /*}}}*/
1356 // URI::NoUserPassword - Return the schema, site and path for the URI   /*{{{*/
1357 // ---------------------------------------------------------------------
1358 /* */
1359 string URI::NoUserPassword(const string &URI)
1360 {
1361    ::URI U(URI);
1362    U.User.clear();
1363    U.Password.clear();
1364    U.Port = 0;
1365    return U;
1366 }
1367                                                                         /*}}}*/