X-Git-Url: https://git.saurik.com/apt.git/blobdiff_plain/468126e3030ac3db47e4c7d361badf99c338fdf4..308b793694774eece8765d172b8e989d8ed29925:/apt-pkg/contrib/strutl.cc diff --git a/apt-pkg/contrib/strutl.cc b/apt-pkg/contrib/strutl.cc index 14ead61e1..160450366 100644 --- a/apt-pkg/contrib/strutl.cc +++ b/apt-pkg/contrib/strutl.cc @@ -15,10 +15,6 @@ ##################################################################### */ /*}}}*/ // Includes /*{{{*/ -#ifdef __GNUG__ -#pragma implementation "apt-pkg/strutl.h" -#endif - #include #include #include @@ -47,9 +43,10 @@ bool UTF8ToCodeset(const char *codeset, const string &orig, string *dest) { iconv_t cd; const char *inbuf; - char *inptr, *outbuf, *outptr; - size_t insize, outsize; - + char *inptr, *outbuf; + size_t insize, bufsize; + dest->clear(); + cd = iconv_open(codeset, "UTF-8"); if (cd == (iconv_t)(-1)) { // Something went wrong @@ -59,22 +56,49 @@ bool UTF8ToCodeset(const char *codeset, const string &orig, string *dest) else perror("iconv_open"); - // Clean the destination string - *dest = ""; - return false; } - insize = outsize = orig.size(); + insize = bufsize = orig.size(); inbuf = orig.data(); inptr = (char *)inbuf; - outbuf = new char[insize+1]; - outptr = outbuf; + outbuf = new char[bufsize]; + size_t lastError = -1; - iconv(cd, &inptr, &insize, &outptr, &outsize); - *outptr = '\0'; + while (insize != 0) + { + char *outptr = outbuf; + size_t outsize = bufsize; + size_t const err = iconv(cd, &inptr, &insize, &outptr, &outsize); + dest->append(outbuf, outptr - outbuf); + if (err == (size_t)(-1)) + { + switch (errno) + { + case EILSEQ: + insize--; + inptr++; + // replace a series of unknown multibytes with a single "?" + if (lastError != insize) { + lastError = insize - 1; + dest->append("?"); + } + break; + case EINVAL: + insize = 0; + break; + case E2BIG: + if (outptr == outbuf) + { + bufsize *= 2; + delete[] outbuf; + outbuf = new char[bufsize]; + } + break; + } + } + } - *dest = outbuf; delete[] outbuf; iconv_close(cd); @@ -174,7 +198,8 @@ bool ParseQuoteWord(const char *&String,string &Res) char *I; for (I = Buffer; I < Buffer + sizeof(Buffer) && Start != C; I++) { - if (*Start == '%' && Start + 2 < C) + if (*Start == '%' && Start + 2 < C && + isxdigit(Start[1]) && isxdigit(Start[2])) { Tmp[0] = Start[1]; Tmp[1] = Start[2]; @@ -249,7 +274,8 @@ string QuoteString(const string &Str, const char *Bad) for (string::const_iterator I = Str.begin(); I != Str.end(); I++) { if (strchr(Bad,*I) != 0 || isprint(*I) == 0 || - *I <= 0x20 || *I >= 0x7F) + *I == 0x25 || // percent '%' char + *I <= 0x20 || *I >= 0x7F) // control chars { char Buf[10]; sprintf(Buf,"%%%02x",(int)*I); @@ -265,11 +291,17 @@ string QuoteString(const string &Str, const char *Bad) // --------------------------------------------------------------------- /* This undoes QuoteString */ string DeQuoteString(const string &Str) +{ + return DeQuoteString(Str.begin(),Str.end()); +} +string DeQuoteString(string::const_iterator const &begin, + string::const_iterator const &end) { string Res; - for (string::const_iterator I = Str.begin(); I != Str.end(); I++) + for (string::const_iterator I = begin; I != end; I++) { - if (*I == '%' && I + 2 < Str.end()) + if (*I == '%' && I + 2 < end && + isxdigit(I[1]) && isxdigit(I[2])) { char Tmp[3]; Tmp[0] = I[1]; @@ -308,13 +340,13 @@ string SizeToStr(double Size) { if (ASize < 100 && I != 0) { - sprintf(S,"%.1f%c",ASize,Ext[I]); + sprintf(S,"%'.1f%c",ASize,Ext[I]); break; } if (ASize < 10000) { - sprintf(S,"%.0f%c",ASize,Ext[I]); + sprintf(S,"%'.0f%c",ASize,Ext[I]); break; } ASize /= 1000.0; @@ -335,23 +367,27 @@ string TimeToStr(unsigned long Sec) { if (Sec > 60*60*24) { - sprintf(S,"%lid %lih%lim%lis",Sec/60/60/24,(Sec/60/60) % 24,(Sec/60) % 60,Sec % 60); + //d means days, h means hours, min means minutes, s means seconds + sprintf(S,_("%lid %lih %limin %lis"),Sec/60/60/24,(Sec/60/60) % 24,(Sec/60) % 60,Sec % 60); break; } if (Sec > 60*60) { - sprintf(S,"%lih%lim%lis",Sec/60/60,(Sec/60) % 60,Sec % 60); + //h means hours, min means minutes, s means seconds + sprintf(S,_("%lih %limin %lis"),Sec/60/60,(Sec/60) % 60,Sec % 60); break; } if (Sec > 60) { - sprintf(S,"%lim%lis",Sec/60,Sec % 60); + //min means minutes, s means seconds + sprintf(S,_("%limin %lis"),Sec/60,Sec % 60); break; } - - sprintf(S,"%lis",Sec); + + //s means seconds + sprintf(S,_("%lis"),Sec); break; } @@ -387,6 +423,17 @@ string SubstVar(string Str,const struct SubstVar *Vars) return Str; } /*}}}*/ +// OutputInDepth - return a string with separator multiplied with depth /*{{{*/ +// --------------------------------------------------------------------- +/* Returns a string with the supplied separator depth + 1 times in it */ +std::string OutputInDepth(const unsigned long Depth, const char* Separator) +{ + std::string output = ""; + for(unsigned long d=Depth+1; d > 0; d--) + output.append(Separator); + return output; +} + /*}}}*/ // URItoFileName - Convert the uri into a unique file name /*{{{*/ // --------------------------------------------------------------------- /* This converts a URI into a safe filename. It quotes all unsafe characters @@ -463,9 +510,9 @@ string Base64Encode(const string &S) return Final; } /*}}}*/ -// stringcmp - Arbitary string compare /*{{{*/ +// stringcmp - Arbitrary string compare /*{{{*/ // --------------------------------------------------------------------- -/* This safely compares two non-null terminated strings of arbitary +/* This safely compares two non-null terminated strings of arbitrary length */ int stringcmp(const char *A,const char *AEnd,const char *B,const char *BEnd) { @@ -521,13 +568,13 @@ int stringcmp(string::const_iterator A,string::const_iterator AEnd, } #endif /*}}}*/ -// stringcasecmp - Arbitary case insensitive string compare /*{{{*/ +// stringcasecmp - Arbitrary case insensitive string compare /*{{{*/ // --------------------------------------------------------------------- /* */ int stringcasecmp(const char *A,const char *AEnd,const char *B,const char *BEnd) { for (; A != AEnd && B != BEnd; A++, B++) - if (toupper(*A) != toupper(*B)) + if (tolower_ascii(*A) != tolower_ascii(*B)) break; if (A == AEnd && B == BEnd) @@ -536,7 +583,7 @@ int stringcasecmp(const char *A,const char *AEnd,const char *B,const char *BEnd) return 1; if (B == BEnd) return -1; - if (toupper(*A) < toupper(*B)) + if (tolower_ascii(*A) < tolower_ascii(*B)) return -1; return 1; } @@ -545,7 +592,7 @@ int stringcasecmp(string::const_iterator A,string::const_iterator AEnd, const char *B,const char *BEnd) { for (; A != AEnd && B != BEnd; A++, B++) - if (toupper(*A) != toupper(*B)) + if (tolower_ascii(*A) != tolower_ascii(*B)) break; if (A == AEnd && B == BEnd) @@ -554,7 +601,7 @@ int stringcasecmp(string::const_iterator A,string::const_iterator AEnd, return 1; if (B == BEnd) return -1; - if (toupper(*A) < toupper(*B)) + if (tolower_ascii(*A) < tolower_ascii(*B)) return -1; return 1; } @@ -562,7 +609,7 @@ int stringcasecmp(string::const_iterator A,string::const_iterator AEnd, string::const_iterator B,string::const_iterator BEnd) { for (; A != AEnd && B != BEnd; A++, B++) - if (toupper(*A) != toupper(*B)) + if (tolower_ascii(*A) != tolower_ascii(*B)) break; if (A == AEnd && B == BEnd) @@ -571,7 +618,7 @@ int stringcasecmp(string::const_iterator A,string::const_iterator AEnd, return 1; if (B == BEnd) return -1; - if (toupper(*A) < toupper(*B)) + if (tolower_ascii(*A) < tolower_ascii(*B)) return -1; return 1; } @@ -662,11 +709,24 @@ string TimeRFC1123(time_t Date) // --------------------------------------------------------------------- /* This pulls full messages from the input FD into the message buffer. It assumes that messages will not pause during transit so no - fancy buffering is used. */ + fancy buffering is used. + + In particular: this reads blocks from the input until it believes + that it's run out of input text. Each block is terminated by a + double newline ('\n' followed by '\n'). As noted below, there is a + bug in this code: it assumes that all the blocks have been read if + it doesn't see additional text in the buffer after the last one is + parsed, which will cause it to lose blocks if the last block + coincides with the end of the buffer. + */ bool ReadMessages(int Fd, vector &List) { char Buffer[64000]; char *End = Buffer; + // Represents any left-over from the previous iteration of the + // parse loop. (i.e., if a message is split across the end + // of the buffer, it goes here) + string PartialMessage; while (1) { @@ -694,6 +754,7 @@ bool ReadMessages(int Fd, vector &List) // Pull the message out string Message(Buffer,I-Buffer); + PartialMessage += Message; // Fix up the buffer for (; I < End && *I == '\n'; I++); @@ -701,10 +762,32 @@ bool ReadMessages(int Fd, vector &List) memmove(Buffer,I,End-Buffer); I = Buffer; - List.push_back(Message); + List.push_back(PartialMessage); + PartialMessage.clear(); } - if (End == Buffer) - return true; + if (End != Buffer) + { + // If there's text left in the buffer, store it + // in PartialMessage and throw the rest of the buffer + // away. This allows us to handle messages that + // are longer than the static buffer size. + PartialMessage += string(Buffer, End); + End = Buffer; + } + else + { + // BUG ALERT: if a message block happens to end at a + // multiple of 64000 characters, this will cause it to + // terminate early, leading to a badly formed block and + // probably crashing the method. However, this is the only + // way we have to find the end of the message block. I have + // an idea of how to fix this, but it will require changes + // to the protocol (essentially to mark the beginning and + // end of the block). + // + // -- dburrows 2008-04-02 + return true; + } if (WaitFd(Fd) == false) return false; @@ -714,28 +797,28 @@ bool ReadMessages(int Fd, vector &List) // MonthConv - Converts a month string into a number /*{{{*/ // --------------------------------------------------------------------- /* This was lifted from the boa webserver which lifted it from 'wn-v1.07' - Made it a bit more robust with a few touppers though. */ + Made it a bit more robust with a few tolower_ascii though. */ static int MonthConv(char *Month) { - switch (toupper(*Month)) + switch (tolower_ascii(*Month)) { - case 'A': - return toupper(Month[1]) == 'P'?3:7; - case 'D': + case 'a': + return tolower_ascii(Month[1]) == 'p'?3:7; + case 'd': return 11; - case 'F': + case 'f': return 1; - case 'J': - if (toupper(Month[1]) == 'A') + case 'j': + if (tolower_ascii(Month[1]) == 'a') return 0; - return toupper(Month[2]) == 'N'?5:6; - case 'M': - return toupper(Month[2]) == 'R'?2:4; - case 'N': + return tolower_ascii(Month[2]) == 'n'?5:6; + case 'm': + return tolower_ascii(Month[2]) == 'r'?2:4; + case 'n': return 10; - case 'O': + case 'o': return 9; - case 'S': + case 's': return 8; // Pretend it is January.. @@ -744,34 +827,66 @@ static int MonthConv(char *Month) } } /*}}}*/ -// timegm - Internal timegm function if gnu is not available /*{{{*/ +// timegm - Internal timegm if the gnu version is not available /*{{{*/ // --------------------------------------------------------------------- -/* Ripped this evil little function from wget - I prefer the use of - GNU timegm if possible as this technique will have interesting problems - with leap seconds, timezones and other. - - Converts struct tm to time_t, assuming the data in tm is UTC rather +/* Converts struct tm to time_t, assuming the data in tm is UTC rather than local timezone (mktime assumes the latter). - - Contributed by Roger Beeman , with the help of - Mark Baushke and the rest of the Gurus at CISCO. */ - -/* Turned it into an autoconf check, because GNU is not the only thing which - can provide timegm. -- 2002-09-22, Joel Baker */ -#ifndef HAVE_TIMEGM // Now with autoconf! + This function is a nonstandard GNU extension that is also present on + the BSDs and maybe other systems. For others we follow the advice of + the manpage of timegm and use his portable replacement. */ +#ifndef HAVE_TIMEGM static time_t timegm(struct tm *t) { - time_t tl, tb; - - tl = mktime (t); - if (tl == -1) - return -1; - tb = mktime (gmtime (&tl)); - return (tl <= tb ? (tl + (tl - tb)) : (tl - (tb - tl))); + char *tz = getenv("TZ"); + setenv("TZ", "", 1); + tzset(); + time_t ret = mktime(t); + if (tz) + setenv("TZ", tz, 1); + else + unsetenv("TZ"); + tzset(); + return ret; } #endif /*}}}*/ +// FullDateToTime - Converts a HTTP1.1 full date strings into a time_t /*{{{*/ +// --------------------------------------------------------------------- +/* tries to parses a full date as specified in RFC2616 Section 3.3.1 + with one exception: All timezones (%Z) are accepted but the protocol + says that it MUST be GMT, but this one is equal to UTC which we will + encounter from time to time (e.g. in Release files) so we accept all + here and just assume it is GMT (or UTC) later on */ +bool RFC1123StrToTime(const char* const str,time_t &time) +{ + struct tm Tm; + // Sun, 06 Nov 1994 08:49:37 GMT ; RFC 822, updated by RFC 1123 + if (strptime(str, "%a, %d %b %Y %H:%M:%S %Z", &Tm) == NULL && + // Sunday, 06-Nov-94 08:49:37 GMT ; RFC 850, obsoleted by RFC 1036 + strptime(str, "%A, %d-%b-%y %H:%M:%S %Z", &Tm) == NULL && + // Sun Nov 6 08:49:37 1994 ; ANSI C's asctime() format + strptime(str, "%a %b %d %H:%M:%S %Y", &Tm) == NULL) + return false; + + time = timegm(&Tm); + return true; +} + /*}}}*/ +// FTPMDTMStrToTime - Converts a ftp modification date into a time_t /*{{{*/ +// --------------------------------------------------------------------- +/* */ +bool FTPMDTMStrToTime(const char* const str,time_t &time) +{ + struct tm Tm; + // MDTM includes no whitespaces but recommend and ignored by strptime + if (strptime(str, "%Y %m %d %H %M %S", &Tm) == NULL) + return false; + + time = timegm(&Tm); + return true; +} + /*}}}*/ // StrToTime - Converts a string into a time_t /*{{{*/ // --------------------------------------------------------------------- /* This handles all 3 populare time formats including RFC 1123, RFC 1036 @@ -925,6 +1040,24 @@ bool TokSplitString(char Tok,char *Input,char **List, return true; } /*}}}*/ +// VectorizeString - Split a string up into a vector of strings /*{{{*/ +// --------------------------------------------------------------------- +/* This can be used to split a given string up into a vector, so the + propose is the same as in the method above and this one is a bit slower + also, but the advantage is that we have an iteratable vector */ +vector VectorizeString(string const &haystack, char const &split) +{ + string::const_iterator start = haystack.begin(); + string::const_iterator end = start; + vector exploded; + do { + for (; end != haystack.end() && *end != split; ++end); + exploded.push_back(string(start, end)); + start = end + 1; + } while (end != haystack.end() && (++end) != haystack.end()); + return exploded; +} + /*}}}*/ // RegexChoice - Simple regex list/list matcher /*{{{*/ // --------------------------------------------------------------------- /* */ @@ -995,11 +1128,26 @@ void ioprintf(ostream &out,const char *format,...) va_start(args,format); // sprintf the description - char S[400]; + char S[4096]; vsnprintf(S,sizeof(S),format,args); out << S; } /*}}}*/ +// strprintf - C format string outputter to C++ strings /*{{{*/ +// --------------------------------------------------------------------- +/* This is used to make the internationalization strings easier to translate + and to allow reordering of parameters */ +void strprintf(string &out,const char *format,...) +{ + va_list args; + va_start(args,format); + + // sprintf the description + char S[4096]; + vsnprintf(S,sizeof(S),format,args); + out = string(S); +} + /*}}}*/ // safe_snprintf - Safer snprintf /*{{{*/ // --------------------------------------------------------------------- /* This is a snprintf that will never (ever) go past 'End' and returns a @@ -1023,6 +1171,20 @@ char *safe_snprintf(char *Buffer,char *End,const char *Format,...) } /*}}}*/ +// tolower_ascii - tolower() function that ignores the locale /*{{{*/ +// --------------------------------------------------------------------- +/* This little function is the most called method we have and tries + therefore to do the absolut minimum - and is noteable faster than + standard tolower/toupper and as a bonus avoids problems with different + locales - we only operate on ascii chars anyway. */ +int tolower_ascii(int const c) +{ + if (c >= 'A' && c <= 'Z') + return c + 32; + return c; +} + /*}}}*/ + // CheckDomainList - See if Host is in a , seperate list /*{{{*/ // --------------------------------------------------------------------- /* The domain list is a comma seperate list of domains that are suffix @@ -1116,9 +1278,10 @@ void URI::CopyFrom(const string &U) else { Host.assign(At+1,SingleSlash); - User.assign(FirstColon,SecondColon); + // username and password must be encoded (RFC 3986) + User.assign(DeQuoteString(FirstColon,SecondColon)); if (SecondColon < At) - Password.assign(SecondColon+1,At); + Password.assign(DeQuoteString(SecondColon+1,At)); } // Now we parse the RFC 2732 [] hostnames. @@ -1222,3 +1385,15 @@ string URI::SiteOnly(const string &URI) return U; } /*}}}*/ +// URI::NoUserPassword - Return the schema, site and path for the URI /*{{{*/ +// --------------------------------------------------------------------- +/* */ +string URI::NoUserPassword(const string &URI) +{ + ::URI U(URI); + U.User.clear(); + U.Password.clear(); + U.Port = 0; + return U; +} + /*}}}*/