X-Git-Url: https://git.saurik.com/apt.git/blobdiff_plain/cec8c646c79e40d928464b08a6f2e3ea62bb1611..d27daedb6a0bf672508072100f20233d08ccf0e0:/apt-pkg/contrib/strutl.cc?ds=inline diff --git a/apt-pkg/contrib/strutl.cc b/apt-pkg/contrib/strutl.cc index 072dda3ac..396cb7898 100644 --- a/apt-pkg/contrib/strutl.cc +++ b/apt-pkg/contrib/strutl.cc @@ -15,14 +15,20 @@ ##################################################################### */ /*}}}*/ // Includes /*{{{*/ +#include + #include #include #include -#include - +#include +#include +#include +#include +#include #include #include +#include #include #include #include @@ -31,11 +37,53 @@ #include #include -#include "config.h" - -using namespace std; +#include /*}}}*/ +using namespace std; + +// Strip - Remove white space from the front and back of a string /*{{{*/ +// --------------------------------------------------------------------- +namespace APT { + namespace String { +std::string Strip(const std::string &str) +{ + // ensure we have at least one character + if (str.empty() == true) + return str; + + char const * const s = str.c_str(); + size_t start = 0; + for (; isspace(s[start]) != 0; ++start) + ; // find the first not-space + + // string contains only whitespaces + if (s[start] == '\0') + return ""; + + size_t end = str.length() - 1; + for (; isspace(s[end]) != 0; --end) + ; // find the last not-space + + return str.substr(start, end - start + 1); +} +bool Endswith(const std::string &s, const std::string &end) +{ + if (end.size() > s.size()) + return false; + return (s.substr(s.size() - end.size(), s.size()) == end); +} + +bool Startswith(const std::string &s, const std::string &start) +{ + if (start.size() > s.size()) + return false; + return (s.substr(0, start.size()) == start); +} + +} +} + /*}}}*/ // UTF8ToCodeset - Convert some UTF-8 string for some codeset /*{{{*/ // --------------------------------------------------------------------- /* This is handy to use before display some information for enduser */ @@ -116,14 +164,20 @@ char *_strstrip(char *String) if (*String == 0) return String; - + return _strrstrip(String); +} + /*}}}*/ +// strrstrip - Remove white space from the back of a string /*{{{*/ +// --------------------------------------------------------------------- +char *_strrstrip(char *String) +{ char *End = String + strlen(String) - 1; for (;End != String - 1 && (*End == ' ' || *End == '\t' || *End == '\n' || *End == '\r'); End--); End++; *End = 0; return String; -}; +} /*}}}*/ // strtabexpand - Converts tabs into 8 spaces /*{{{*/ // --------------------------------------------------------------------- @@ -179,14 +233,14 @@ bool ParseQuoteWord(const char *&String,string &Res) { if (*C == '"') { - for (C++; *C != 0 && *C != '"'; C++); - if (*C == 0) + C = strchr(C + 1, '"'); + if (C == NULL) return false; } if (*C == '[') { - for (C++; *C != 0 && *C != ']'; C++); - if (*C == 0) + C = strchr(C + 1, ']'); + if (C == NULL) return false; } } @@ -270,21 +324,19 @@ bool ParseCWord(const char *&String,string &Res) /* */ string QuoteString(const string &Str, const char *Bad) { - string Res; - for (string::const_iterator I = Str.begin(); I != Str.end(); I++) + std::stringstream Res; + for (string::const_iterator I = Str.begin(); I != Str.end(); ++I) { - if (strchr(Bad,*I) != 0 || isprint(*I) == 0 || + if (strchr(Bad,*I) != 0 || isprint(*I) == 0 || *I == 0x25 || // percent '%' char *I <= 0x20 || *I >= 0x7F) // control chars { - char Buf[10]; - sprintf(Buf,"%%%02x",(int)*I); - Res += Buf; + ioprintf(Res, "%%%02hhx", *I); } else - Res += *I; + Res << *I; } - return Res; + return Res.str(); } /*}}}*/ // DeQuoteString - Convert a string from quoted from /*{{{*/ @@ -298,7 +350,7 @@ string DeQuoteString(string::const_iterator const &begin, string::const_iterator const &end) { string Res; - for (string::const_iterator I = begin; I != end; I++) + for (string::const_iterator I = begin; I != end; ++I) { if (*I == '%' && I + 2 < end && isxdigit(I[1]) && isxdigit(I[2])) @@ -325,13 +377,12 @@ string DeQuoteString(string::const_iterator const &begin, YottaBytes (E24) */ string SizeToStr(double Size) { - char S[300]; double ASize; if (Size >= 0) ASize = Size; else ASize = -1*Size; - + /* bytes, KiloBytes, MegaBytes, GigaBytes, TeraBytes, PetaBytes, ExaBytes, ZettaBytes, YottaBytes */ char Ext[] = {'\0','k','M','G','T','P','E','Z','Y'}; @@ -340,20 +391,21 @@ string SizeToStr(double Size) { if (ASize < 100 && I != 0) { - sprintf(S,"%'.1f %c",ASize,Ext[I]); - break; + std::string S; + strprintf(S, "%'.1f %c", ASize, Ext[I]); + return S; } - + if (ASize < 10000) { - sprintf(S,"%'.0f %c",ASize,Ext[I]); - break; + std::string S; + strprintf(S, "%'.0f %c", ASize, Ext[I]); + return S; } ASize /= 1000.0; I++; } - - return S; + return ""; } /*}}}*/ // TimeToStr - Convert the time into a string /*{{{*/ @@ -361,61 +413,59 @@ string SizeToStr(double Size) /* Converts a number of seconds to a hms format */ string TimeToStr(unsigned long Sec) { - char S[300]; - - while (1) + std::string S; + if (Sec > 60*60*24) { - if (Sec > 60*60*24) - { - //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) - { - //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) - { - //min means minutes, s means seconds - sprintf(S,_("%limin %lis"),Sec/60,Sec % 60); - break; - } - - //s means seconds - sprintf(S,_("%lis"),Sec); - break; + //TRANSLATOR: d means days, h means hours, min means minutes, s means seconds + strprintf(S,_("%lid %lih %limin %lis"),Sec/60/60/24,(Sec/60/60) % 24,(Sec/60) % 60,Sec % 60); + } + else if (Sec > 60*60) + { + //TRANSLATOR: h means hours, min means minutes, s means seconds + strprintf(S,_("%lih %limin %lis"),Sec/60/60,(Sec/60) % 60,Sec % 60); + } + else if (Sec > 60) + { + //TRANSLATOR: min means minutes, s means seconds + strprintf(S,_("%limin %lis"),Sec/60,Sec % 60); + } + else + { + //TRANSLATOR: s means seconds + strprintf(S,_("%lis"),Sec); } - return S; } /*}}}*/ // SubstVar - Substitute a string for another string /*{{{*/ // --------------------------------------------------------------------- -/* This replaces all occurances of Subst with Contents in Str. */ +/* This replaces all occurrences of Subst with Contents in Str. */ string SubstVar(const string &Str,const string &Subst,const string &Contents) { + if (Subst.empty() == true) + return Str; + string::size_type Pos = 0; string::size_type OldPos = 0; string Temp; - - while (OldPos < Str.length() && + + while (OldPos < Str.length() && (Pos = Str.find(Subst,OldPos)) != string::npos) { - Temp += string(Str,OldPos,Pos) + Contents; - OldPos = Pos + Subst.length(); + if (OldPos != Pos) + Temp.append(Str, OldPos, Pos - OldPos); + if (Contents.empty() == false) + Temp.append(Contents); + OldPos = Pos + Subst.length(); } - + if (OldPos == 0) return Str; - + + if (OldPos >= Str.length()) + return Temp; return Temp + string(Str,OldPos); } - string SubstVar(string Str,const struct SubstVar *Vars) { for (; Vars->Subst != 0; Vars++) @@ -632,7 +682,7 @@ string LookupTag(const string &Message,const char *Tag,const char *Default) { // Look for a matching tag. int Length = strlen(Tag); - for (string::const_iterator I = Message.begin(); I + Length < Message.end(); I++) + for (string::const_iterator I = Message.begin(); I + Length < Message.end(); ++I) { // Found the tag if (I[Length] == ':' && stringcasecmp(I,I+Length,Tag) == 0) @@ -640,14 +690,14 @@ string LookupTag(const string &Message,const char *Tag,const char *Default) // Find the end of line and strip the leading/trailing spaces string::const_iterator J; I += Length + 1; - for (; isspace(*I) != 0 && I < Message.end(); I++); - for (J = I; *J != '\n' && J < Message.end(); J++); - for (; J > I && isspace(J[-1]) != 0; J--); + for (; isspace(*I) != 0 && I < Message.end(); ++I); + for (J = I; *J != '\n' && J < Message.end(); ++J); + for (; J > I && isspace(J[-1]) != 0; --J); return string(I,J); } - for (; *I != '\n' && I < Message.end(); I++); + for (; *I != '\n' && I < Message.end(); ++I); } // Failed to find a match @@ -662,9 +712,12 @@ string LookupTag(const string &Message,const char *Tag,const char *Default) then returns the result. Several varients on true/false are checked. */ int StringToBool(const string &Text,int Default) { - char *End; - int Res = strtol(Text.c_str(),&End,0); - if (End != Text.c_str() && Res >= 0 && Res <= 1) + char *ParseEnd; + int Res = strtol(Text.c_str(),&ParseEnd,0); + // ensure that the entire string was converted by strtol to avoid + // failures on "apt-cache show -a 0ad" where the "0" is converted + const char *TextEnd = Text.c_str()+Text.size(); + if (ParseEnd == TextEnd && Res >= 0 && Res <= 1) return Res; // Check for positives @@ -715,85 +768,94 @@ string TimeRFC1123(time_t Date) 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. + double newline ('\n' followed by '\n'). */ 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) - { - int Res = read(Fd,End,sizeof(Buffer) - (End-Buffer)); + + do { + int const Res = read(Fd, Buffer, sizeof(Buffer)); if (Res < 0 && errno == EINTR) continue; - - // Process is dead, this is kind of bad.. + + // process we read from has died if (Res == 0) return false; - + // No data - if (Res < 0 && errno == EAGAIN) + if (Res < 0 && (errno == EAGAIN || errno == EWOULDBLOCK)) return true; if (Res < 0) return false; - - End += Res; - - // Look for the end of the message - for (char *I = Buffer; I + 1 < End; I++) + + // extract the message(s) from the buffer + char const *Start = Buffer; + char const * const End = Buffer + Res; + + char const * NL = (char const *) memchr(Start, '\n', End - Start); + if (NL == NULL) { - if (I[0] != '\n' || I[1] != '\n') - continue; - - // Pull the message out - string Message(Buffer,I-Buffer); - PartialMessage += Message; - - // Fix up the buffer - for (; I < End && *I == '\n'; I++); - End -= I-Buffer; - memmove(Buffer,I,End-Buffer); - I = Buffer; - - List.push_back(PartialMessage); - PartialMessage.clear(); + // end of buffer: store what we have so far and read new data in + PartialMessage.append(Start, End - Start); + Start = End; } - 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; - } + ++NL; + + if (PartialMessage.empty() == false && Start < End) + { + // if we start with a new line, see if the partial message we have ended with one + // so that we properly detect records ending between two read() runs + // cases are: \n|\n , \r\n|\r\n and \r\n\r|\n + // the case \r|\n\r\n is handled by the usual double-newline handling + if ((NL - Start) == 1 || ((NL - Start) == 2 && *Start == '\r')) + { + if (APT::String::Endswith(PartialMessage, "\n") || APT::String::Endswith(PartialMessage, "\r\n\r")) + { + PartialMessage.erase(PartialMessage.find_last_not_of("\r\n") + 1); + List.push_back(PartialMessage); + PartialMessage.clear(); + while (NL < End && (*NL == '\n' || *NL == '\r')) ++NL; + Start = NL; + } + } + } + + while (Start < End) { + char const * NL2 = (char const *) memchr(NL, '\n', End - NL); + if (NL2 == NULL) + { + // end of buffer: store what we have so far and read new data in + PartialMessage.append(Start, End - Start); + break; + } + ++NL2; + + // did we find a double newline? + if ((NL2 - NL) == 1 || ((NL2 - NL) == 2 && *NL == '\r')) + { + PartialMessage.append(Start, NL2 - Start); + PartialMessage.erase(PartialMessage.find_last_not_of("\r\n") + 1); + List.push_back(PartialMessage); + PartialMessage.clear(); + while (NL2 < End && (*NL2 == '\n' || *NL2 == '\r')) ++NL2; + Start = NL2; + } + NL = NL2; + } + + // we have read at least one complete message and nothing left + if (PartialMessage.empty() == true) + return true; if (WaitFd(Fd) == false) return false; - } + } while (true); } /*}}}*/ // MonthConv - Converts a month string into a number /*{{{*/ @@ -895,7 +957,7 @@ bool FTPMDTMStrToTime(const char* const str,time_t &time) /*}}}*/ // StrToTime - Converts a string into a time_t /*{{{*/ // --------------------------------------------------------------------- -/* This handles all 3 populare time formats including RFC 1123, RFC 1036 +/* This handles all 3 popular time formats including RFC 1123, RFC 1036 and the C library asctime format. It requires the GNU library function 'timegm' to convert a struct tm in UTC to a time_t. For some bizzar reason the C library does not provide any such function :< This also @@ -904,24 +966,23 @@ bool StrToTime(const string &Val,time_t &Result) { struct tm Tm; char Month[10]; - const char *I = Val.c_str(); - + // Skip the day of the week - for (;*I != 0 && *I != ' '; I++); - + const char *I = strchr(Val.c_str(), ' '); + // Handle RFC 1123 time Month[0] = 0; - if (sscanf(I," %d %3s %d %d:%d:%d GMT",&Tm.tm_mday,Month,&Tm.tm_year, + if (sscanf(I," %2d %3s %4d %2d:%2d:%2d GMT",&Tm.tm_mday,Month,&Tm.tm_year, &Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) != 6) { // Handle RFC 1036 time - if (sscanf(I," %d-%3s-%d %d:%d:%d GMT",&Tm.tm_mday,Month, + if (sscanf(I," %2d-%3s-%3d %2d:%2d:%2d GMT",&Tm.tm_mday,Month, &Tm.tm_year,&Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) == 6) Tm.tm_year += 1900; else { // asctime format - if (sscanf(I," %3s %d %d:%d:%d %d",Month,&Tm.tm_mday, + if (sscanf(I," %3s %2d %2d:%2d:%2d %4d",Month,&Tm.tm_mday, &Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec,&Tm.tm_year) != 6) { // 'ftp' time @@ -936,6 +997,8 @@ bool StrToTime(const string &Val,time_t &Result) Tm.tm_isdst = 0; if (Month[0] != 0) Tm.tm_mon = MonthConv(Month); + else + Tm.tm_mon = 0; // we don't have a month, so pick something Tm.tm_year -= 1900; // Convert to local time and then to GMT @@ -970,11 +1033,39 @@ bool StrToNum(const char *Str,unsigned long &Res,unsigned Len,unsigned Base) return true; } /*}}}*/ +// StrToNum - Convert a fixed length string to a number /*{{{*/ +// --------------------------------------------------------------------- +/* This is used in decoding the crazy fixed length string headers in + tar and ar files. */ +bool StrToNum(const char *Str,unsigned long long &Res,unsigned Len,unsigned Base) +{ + char S[30]; + if (Len >= sizeof(S)) + return false; + memcpy(S,Str,Len); + S[Len] = 0; + + // All spaces is a zero + Res = 0; + unsigned I; + for (I = 0; S[I] == ' '; I++); + if (S[I] == 0) + return true; + + char *End; + Res = strtoull(S,&End,Base); + if (End == S) + return false; + + return true; +} + /*}}}*/ + // Base256ToNum - Convert a fixed length binary to a number /*{{{*/ // --------------------------------------------------------------------- /* This is used in decoding the 256bit encoded fixed length fields in tar files */ -bool Base256ToNum(const char *Str,unsigned long &Res,unsigned int Len) +bool Base256ToNum(const char *Str,unsigned long long &Res,unsigned int Len) { if ((Str[0] & 0x80) == 0) return false; @@ -987,6 +1078,23 @@ bool Base256ToNum(const char *Str,unsigned long &Res,unsigned int Len) } } /*}}}*/ +// Base256ToNum - Convert a fixed length binary to a number /*{{{*/ +// --------------------------------------------------------------------- +/* This is used in decoding the 256bit encoded fixed length fields in + tar files */ +bool Base256ToNum(const char *Str,unsigned long &Res,unsigned int Len) +{ + unsigned long long Num; + bool rc; + + rc = Base256ToNum(Str, Num, Len); + Res = Num; + if (Res != Num) + return false; + + return rc; +} + /*}}}*/ // HexDigit - Convert a hex character into an integer /*{{{*/ // --------------------------------------------------------------------- /* Helper for Hex2Num */ @@ -1070,9 +1178,11 @@ bool TokSplitString(char Tok,char *Input,char **List, also, but the advantage is that we have an iteratable vector */ vector VectorizeString(string const &haystack, char const &split) { + vector exploded; + if (haystack.empty() == true) + return exploded; 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)); @@ -1081,6 +1191,37 @@ vector VectorizeString(string const &haystack, char const &split) return exploded; } /*}}}*/ +// StringSplit - split a string into a string vector by token /*{{{*/ +// --------------------------------------------------------------------- +/* See header for details. + */ +vector StringSplit(std::string const &s, std::string const &sep, + unsigned int maxsplit) +{ + vector split; + size_t start, pos; + + // no seperator given, this is bogus + if(sep.size() == 0) + return split; + + start = pos = 0; + while (pos != string::npos) + { + pos = s.find(sep, start); + split.push_back(s.substr(start, pos-start)); + + // if maxsplit is reached, the remaining string is the last item + if(split.size() >= maxsplit) + { + split[split.size()-1] = s.substr(start); + break; + } + start = pos+sep.size(); + } + return split; +} + /*}}}*/ // RegexChoice - Simple regex list/list matcher /*{{{*/ // --------------------------------------------------------------------- /* */ @@ -1091,7 +1232,7 @@ unsigned long RegexChoice(RxChoiceList *Rxs,const char **ListBegin, R->Hit = false; unsigned long Hits = 0; - for (; ListBegin != ListEnd; ListBegin++) + for (; ListBegin < ListEnd; ++ListBegin) { // Check if the name is a regex const char *I; @@ -1141,34 +1282,54 @@ unsigned long RegexChoice(RxChoiceList *Rxs,const char **ListBegin, return Hits; } /*}}}*/ -// ioprintf - C format string outputter to C++ iostreams /*{{{*/ +// {str,io}printf - C format string outputter to C++ strings/iostreams /*{{{*/ // --------------------------------------------------------------------- /* This is used to make the internationalization strings easier to translate and to allow reordering of parameters */ -void ioprintf(ostream &out,const char *format,...) +static bool iovprintf(ostream &out, const char *format, + va_list &args, ssize_t &size) { + char *S = (char*)malloc(size); + ssize_t const n = vsnprintf(S, size, format, args); + if (n > -1 && n < size) { + out << S; + free(S); + return true; + } else { + if (n > -1) + size = n + 1; + else + size *= 2; + } + free(S); + return false; +} +void ioprintf(ostream &out,const char *format,...) { va_list args; - va_start(args,format); - - // sprintf the description - char S[4096]; - vsnprintf(S,sizeof(S),format,args); - out << S; + ssize_t size = 400; + while (true) { + bool ret; + va_start(args,format); + ret = iovprintf(out, format, args, size); + va_end(args); + if (ret == true) + return; + } } - /*}}}*/ -// 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,...) +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); + ssize_t size = 400; + std::ostringstream outstr; + while (true) { + bool ret; + va_start(args,format); + ret = iovprintf(outstr, format, args, size); + va_end(args); + if (ret == true) + break; + } + out = outstr.str(); } /*}}}*/ // safe_snprintf - Safer snprintf /*{{{*/ @@ -1182,12 +1343,12 @@ char *safe_snprintf(char *Buffer,char *End,const char *Format,...) va_list args; int Did; - va_start(args,Format); - if (End <= Buffer) return End; - + va_start(args,Format); Did = vsnprintf(Buffer,End - Buffer,Format,args); + va_end(args); + if (Did < 0 || Buffer + Did > End) return End; return Buffer + Did; @@ -1202,11 +1363,11 @@ string StripEpoch(const string &VerStr) return VerStr; return VerStr.substr(i+1); } - + /*}}}*/ // 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 + therefore to do the absolut minimum - and is notable 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) @@ -1217,14 +1378,14 @@ int tolower_ascii(int const c) } /*}}}*/ -// CheckDomainList - See if Host is in a , seperate list /*{{{*/ +// CheckDomainList - See if Host is in a , separate list /*{{{*/ // --------------------------------------------------------------------- -/* The domain list is a comma seperate list of domains that are suffix +/* The domain list is a comma separate list of domains that are suffix matched against the argument */ bool CheckDomainList(const string &Host,const string &List) { string::const_iterator Start = List.begin(); - for (string::const_iterator Cur = List.begin(); Cur <= List.end(); Cur++) + for (string::const_iterator Cur = List.begin(); Cur <= List.end(); ++Cur) { if (Cur < List.end() && *Cur != ',') continue; @@ -1240,7 +1401,80 @@ bool CheckDomainList(const string &Host,const string &List) return false; } /*}}}*/ +// strv_length - Return the length of a NULL-terminated string array /*{{{*/ +// --------------------------------------------------------------------- +/* */ +size_t strv_length(const char **str_array) +{ + size_t i; + for (i=0; str_array[i] != NULL; i++) + /* nothing */ + ; + return i; +} + /*}}}*/ +// DeEscapeString - unescape (\0XX and \xXX) from a string /*{{{*/ +// --------------------------------------------------------------------- +/* */ +string DeEscapeString(const string &input) +{ + char tmp[3]; + string::const_iterator it; + string output; + for (it = input.begin(); it != input.end(); ++it) + { + // just copy non-escape chars + if (*it != '\\') + { + output += *it; + continue; + } + // deal with double escape + if (*it == '\\' && + (it + 1 < input.end()) && it[1] == '\\') + { + // copy + output += *it; + // advance iterator one step further + ++it; + continue; + } + + // ensure we have a char to read + if (it + 1 == input.end()) + continue; + + // read it + ++it; + switch (*it) + { + case '0': + if (it + 2 <= input.end()) { + tmp[0] = it[1]; + tmp[1] = it[2]; + tmp[2] = 0; + output += (char)strtol(tmp, 0, 8); + it += 2; + } + break; + case 'x': + if (it + 2 <= input.end()) { + tmp[0] = it[1]; + tmp[1] = it[2]; + tmp[2] = 0; + output += (char)strtol(tmp, 0, 16); + it += 2; + } + break; + default: + // FIXME: raise exception here? + break; + } + } + return output; +} + /*}}}*/ // URI::CopyFrom - Copy from an object /*{{{*/ // --------------------------------------------------------------------- /* This parses the URI into all of its components */ @@ -1249,7 +1483,7 @@ void URI::CopyFrom(const string &U) string::const_iterator I = U.begin(); // Locate the first colon, this separates the scheme - for (; I < U.end() && *I != ':' ; I++); + for (; I < U.end() && *I != ':' ; ++I); string::const_iterator FirstColon = I; /* Determine if this is a host type URI with a leading double // @@ -1261,7 +1495,7 @@ void URI::CopyFrom(const string &U) /* Find the / indicating the end of the hostname, ignoring /'s in the square brackets */ bool InBracket = false; - for (; SingleSlash < U.end() && (*SingleSlash != '/' || InBracket == true); SingleSlash++) + for (; SingleSlash < U.end() && (*SingleSlash != '/' || InBracket == true); ++SingleSlash) { if (*SingleSlash == '[') InBracket = true; @@ -1294,11 +1528,11 @@ void URI::CopyFrom(const string &U) I = FirstColon + 1; if (I > SingleSlash) I = SingleSlash; - for (; I < SingleSlash && *I != ':'; I++); + for (; I < SingleSlash && *I != ':'; ++I); string::const_iterator SecondColon = I; // Search for the @ after the colon - for (; I < SingleSlash && *I != '@'; I++); + for (; I < SingleSlash && *I != '@'; ++I); string::const_iterator At = I; // Now write the host and user/pass @@ -1360,72 +1594,75 @@ void URI::CopyFrom(const string &U) /* */ URI::operator string() { - string Res; - + std::stringstream Res; + if (Access.empty() == false) - Res = Access + ':'; - + Res << Access << ':'; + if (Host.empty() == false) - { + { if (Access.empty() == false) - Res += "//"; - + Res << "//"; + if (User.empty() == false) { - Res += User; + // FIXME: Technically userinfo is permitted even less + // characters than these, but this is not conveniently + // expressed with a blacklist. + Res << QuoteString(User, ":/?#[]@"); if (Password.empty() == false) - Res += ":" + Password; - Res += "@"; + Res << ":" << QuoteString(Password, ":/?#[]@"); + Res << "@"; } - + // Add RFC 2732 escaping characters - if (Access.empty() == false && - (Host.find('/') != string::npos || Host.find(':') != string::npos)) - Res += '[' + Host + ']'; + if (Access.empty() == false && Host.find_first_of("/:") != string::npos) + Res << '[' << Host << ']'; else - Res += Host; - + Res << Host; + if (Port != 0) - { - char S[30]; - sprintf(S,":%u",Port); - Res += S; - } + Res << ':' << Port; } - + if (Path.empty() == false) { if (Path[0] != '/') - Res += "/" + Path; + Res << "/" << Path; else - Res += Path; + Res << Path; } - - return Res; + + return Res.str(); } /*}}}*/ // URI::SiteOnly - Return the schema and site for the URI /*{{{*/ -// --------------------------------------------------------------------- -/* */ string URI::SiteOnly(const string &URI) { ::URI U(URI); U.User.clear(); U.Password.clear(); U.Path.clear(); - U.Port = 0; + return U; +} + /*}}}*/ +// URI::ArchiveOnly - Return the schema, site and cleaned path for the URI /*{{{*/ +string URI::ArchiveOnly(const string &URI) +{ + ::URI U(URI); + U.User.clear(); + U.Password.clear(); + if (U.Path.empty() == false && U.Path[U.Path.length() - 1] == '/') + U.Path.erase(U.Path.length() - 1); 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; } /*}}}*/