// -*- mode: cpp; mode: fold -*-
// Description /*{{{*/
-// $Id: strutl.cc,v 1.31 1999/12/10 07:21:52 jgg Exp $
+// $Id: strutl.cc,v 1.48 2003/07/18 14:15:11 mdz Exp $
/* ######################################################################
- String Util - Some usefull string functions.
+ String Util - Some useful string functions.
- These have been collected from here and there to do all sorts of usefull
- things to strings. They are usefull in file parsers, URI handlers and
+ These have been collected from here and there to do all sorts of useful
+ things to strings. They are useful in file parsers, URI handlers and
especially in APT methods.
This source is placed in the Public Domain, do with it what you will
##################################################################### */
/*}}}*/
// Includes /*{{{*/
-#ifdef __GNUG__
-#pragma implementation "apt-pkg/strutl.h"
-#endif
-
#include <apt-pkg/strutl.h>
#include <apt-pkg/fileutl.h>
+#include <apt-pkg/error.h>
+#include <apti18n.h>
+
#include <ctype.h>
#include <string.h>
#include <stdio.h>
+#include <algorithm>
#include <unistd.h>
+#include <regex.h>
#include <errno.h>
+#include <stdarg.h>
+#include <iconv.h>
+
+#include "config.h"
+
+using namespace std;
/*}}}*/
+// UTF8ToCodeset - Convert some UTF-8 string for some codeset /*{{{*/
+// ---------------------------------------------------------------------
+/* This is handy to use before display some information for enduser */
+bool UTF8ToCodeset(const char *codeset, const string &orig, string *dest)
+{
+ iconv_t cd;
+ const char *inbuf;
+ char *inptr, *outbuf;
+ size_t insize, bufsize;
+ dest->clear();
+
+ cd = iconv_open(codeset, "UTF-8");
+ if (cd == (iconv_t)(-1)) {
+ // Something went wrong
+ if (errno == EINVAL)
+ _error->Error("conversion from 'UTF-8' to '%s' not available",
+ codeset);
+ else
+ perror("iconv_open");
+
+ return false;
+ }
+
+ insize = bufsize = orig.size();
+ inbuf = orig.data();
+ inptr = (char *)inbuf;
+ outbuf = new char[bufsize];
+ size_t lastError = -1;
+
+ 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;
+ }
+ }
+ }
+
+ delete[] outbuf;
+
+ iconv_close(cd);
+
+ return true;
+}
+ /*}}}*/
// strstrip - Remove white space from the front and back of a string /*{{{*/
// ---------------------------------------------------------------------
/* This is handy to use when parsing a file. It also removes \n's left
// ---------------------------------------------------------------------
/* This grabs a single word, converts any % escaped characters to their
proper values and advances the pointer. Double quotes are understood
- and striped out as well. This is for URI/URL parsing. */
+ and striped out as well. This is for URI/URL parsing. It also can
+ understand [] brackets.*/
bool ParseQuoteWord(const char *&String,string &Res)
{
// Skip leading whitespace
{
if (*C == '"')
{
- for (C++;*C != 0 && *C != '"'; C++);
- if (*C == 0)
+ C = strchr(C + 1, '"');
+ if (C == NULL)
+ return false;
+ }
+ if (*C == '[')
+ {
+ C = strchr(C + 1, ']');
+ if (C == NULL)
return false;
}
}
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];
/*}}}*/
// ParseCWord - Parses a string like a C "" expression /*{{{*/
// ---------------------------------------------------------------------
-/* This expects a series of space seperated strings enclosed in ""'s.
+/* This expects a series of space separated strings enclosed in ""'s.
It concatenates the ""'s into a single string. */
-bool ParseCWord(const char *String,string &Res)
+bool ParseCWord(const char *&String,string &Res)
{
// Skip leading whitespace
const char *C = String;
if (isspace(*C) == 0)
return false;
*Buf++ = ' ';
- }
+ }
*Buf = 0;
Res = Buffer;
+ String = C;
return true;
}
/*}}}*/
// QuoteString - Convert a string into quoted from /*{{{*/
// ---------------------------------------------------------------------
/* */
-string QuoteString(string Str,const char *Bad)
+string QuoteString(const string &Str, const char *Bad)
{
string Res;
- for (string::iterator I = Str.begin(); I != Str.end(); I++)
+ 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);
// DeQuoteString - Convert a string from quoted from /*{{{*/
// ---------------------------------------------------------------------
/* This undoes QuoteString */
-string DeQuoteString(string Str)
+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::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];
{
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;
{
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;
}
// SubstVar - Substitute a string for another string /*{{{*/
// ---------------------------------------------------------------------
/* This replaces all occurances of Subst with Contents in Str. */
-string SubstVar(string Str,string Subst,string Contents)
+string SubstVar(const string &Str,const string &Subst,const string &Contents)
{
string::size_type Pos = 0;
string::size_type OldPos = 0;
return Temp + string(Str,OldPos);
}
+
+string SubstVar(string Str,const struct SubstVar *Vars)
+{
+ for (; Vars->Subst != 0; Vars++)
+ Str = SubstVar(Str,Vars->Subst,*Vars->Contents);
+ 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
and converts / to _ and removes the scheme identifier. The resulting
file name should be unique and never occur again for a different file */
-string URItoFileName(string URI)
+string URItoFileName(const string &URI)
{
// Nuke 'sensitive' items
::URI U(URI);
- U.User = string();
- U.Password = string();
- U.Access = "";
+ U.User.clear();
+ U.Password.clear();
+ U.Access.clear();
// "\x00-\x20{}|\\\\^\\[\\]<>\"\x7F-\xFF";
- URI = QuoteString(U,"\\|{}[]<>\"^~_=!@#$%^&*");
- string::iterator J = URI.begin();
- for (; J != URI.end(); J++)
- if (*J == '/')
- *J = '_';
- return URI;
+ string NewURI = QuoteString(U,"\\|{}[]<>\"^~_=!@#$%^&*");
+ replace(NewURI.begin(),NewURI.end(),'/','_');
+ return NewURI;
}
/*}}}*/
// Base64Encode - Base64 Encoding routine for short strings /*{{{*/
from wget and then patched and bug fixed.
This spec can be found in rfc2045 */
-string Base64Encode(string S)
+string Base64Encode(const string &S)
{
// Conversion table.
static char tbl[64] = {'A','B','C','D','E','F','G','H',
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)
{
return -1;
return 1;
}
+
+#if __GNUC__ >= 3
+int stringcmp(string::const_iterator A,string::const_iterator AEnd,
+ const char *B,const char *BEnd)
+{
+ for (; A != AEnd && B != BEnd; A++, B++)
+ if (*A != *B)
+ break;
+
+ if (A == AEnd && B == BEnd)
+ return 0;
+ if (A == AEnd)
+ return 1;
+ if (B == BEnd)
+ return -1;
+ if (*A < *B)
+ return -1;
+ return 1;
+}
+int stringcmp(string::const_iterator A,string::const_iterator AEnd,
+ string::const_iterator B,string::const_iterator BEnd)
+{
+ for (; A != AEnd && B != BEnd; A++, B++)
+ if (*A != *B)
+ break;
+
+ if (A == AEnd && B == BEnd)
+ return 0;
+ if (A == AEnd)
+ return 1;
+ if (B == BEnd)
+ return -1;
+ if (*A < *B)
+ return -1;
+ return 1;
+}
+#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)
return 1;
if (B == BEnd)
return -1;
- if (toupper(*A) < toupper(*B))
+ if (tolower_ascii(*A) < tolower_ascii(*B))
return -1;
return 1;
}
+#if __GNUC__ >= 3
+int stringcasecmp(string::const_iterator A,string::const_iterator AEnd,
+ const char *B,const char *BEnd)
+{
+ for (; A != AEnd && B != BEnd; A++, B++)
+ if (tolower_ascii(*A) != tolower_ascii(*B))
+ break;
+
+ if (A == AEnd && B == BEnd)
+ return 0;
+ if (A == AEnd)
+ return 1;
+ if (B == BEnd)
+ return -1;
+ if (tolower_ascii(*A) < tolower_ascii(*B))
+ return -1;
+ return 1;
+}
+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 (tolower_ascii(*A) != tolower_ascii(*B))
+ break;
+
+ if (A == AEnd && B == BEnd)
+ return 0;
+ if (A == AEnd)
+ return 1;
+ if (B == BEnd)
+ return -1;
+ if (tolower_ascii(*A) < tolower_ascii(*B))
+ return -1;
+ return 1;
+}
+#endif
/*}}}*/
// LookupTag - Lookup the value of a tag in a taged string /*{{{*/
// ---------------------------------------------------------------------
/* The format is like those used in package files and the method
communication system */
-string LookupTag(string Message,const char *Tag,const char *Default)
+string LookupTag(const string &Message,const char *Tag,const char *Default)
{
// Look for a matching tag.
int Length = strlen(Tag);
- for (string::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)
{
// Find the end of line and strip the leading/trailing spaces
- string::iterator J;
+ 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-I);
+ return string(I,J);
}
- for (; *I != '\n' && I < Message.end(); I++);
+ for (; *I != '\n' && I < Message.end(); ++I);
}
// Failed to find a match
// ---------------------------------------------------------------------
/* This inspects the string to see if it is true or if it is false and
then returns the result. Several varients on true/false are checked. */
-int StringToBool(string Text,int Default = -1)
+int StringToBool(const string &Text,int Default)
{
char *End;
int Res = strtol(Text.c_str(),&End,0);
year 2000 complient and timezone neutral */
string TimeRFC1123(time_t Date)
{
- struct tm Conv = *gmtime(&Date);
- char Buf[300];
+ struct tm Conv;
+ if (gmtime_r(&Date, &Conv) == NULL)
+ return "";
+ char Buf[300];
const char *Day[] = {"Sun","Mon","Tue","Wed","Thu","Fri","Sat"};
const char *Month[] = {"Jan","Feb","Mar","Apr","May","Jun","Jul",
"Aug","Sep","Oct","Nov","Dec"};
- sprintf(Buf,"%s, %02i %s %i %02i:%02i:%02i GMT",Day[Conv.tm_wday],
+ snprintf(Buf, sizeof(Buf), "%s, %02i %s %i %02i:%02i:%02i GMT",Day[Conv.tm_wday],
Conv.tm_mday,Month[Conv.tm_mon],Conv.tm_year+1900,Conv.tm_hour,
Conv.tm_min,Conv.tm_sec);
return Buf;
// ---------------------------------------------------------------------
/* 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<string> &List)
{
- char Buffer[4000];
+ 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)
{
return false;
// No data
- if (Res <= 0)
+ if (Res < 0 && errno == EAGAIN)
return true;
-
+ if (Res < 0)
+ return false;
+
End += Res;
// Look for the end of the message
continue;
// Pull the message out
- string Message(Buffer,0,I-Buffer);
+ string Message(Buffer,I-Buffer);
+ PartialMessage += Message;
// Fix up the buffer
for (; I < End && *I == '\n'; I++);
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;
// 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..
}
}
/*}}}*/
-// 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 <beeman@cisco.com>, with the help of
- Mark Baushke <mdb@cisco.com> and the rest of the Gurus at CISCO. */
-#ifndef __USE_MISC // glib sets this
+
+ 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;
+ setlocale (LC_ALL,"C");
+ bool const invalid =
+ // Sun, 06 Nov 1994 08:49:37 GMT ; RFC 822, updated by RFC 1123
+ (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);
+ setlocale (LC_ALL,"");
+ if (invalid == true)
+ 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
'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
handles the weird, but unambiguous FTP time format*/
-bool StrToTime(string Val,time_t &Result)
+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,
&Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec,&Tm.tm_year) != 6)
{
// 'ftp' time
- if (sscanf(I,"%4d%2d%2d%2d%2d%2d",&Tm.tm_year,&Tm.tm_mon,
+ if (sscanf(Val.c_str(),"%4d%2d%2d%2d%2d%2d",&Tm.tm_year,&Tm.tm_mon,
&Tm.tm_mday,&Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) != 6)
return false;
Tm.tm_mon--;
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)
+{
+ if ((Str[0] & 0x80) == 0)
+ return false;
+ else
+ {
+ Res = Str[0] & 0x7F;
+ for(unsigned int i = 1; i < Len; ++i)
+ Res = (Res<<8) + Str[i];
+ return true;
+ }
+}
+ /*}}}*/
// HexDigit - Convert a hex character into an integer /*{{{*/
// ---------------------------------------------------------------------
/* Helper for Hex2Num */
// Hex2Num - Convert a long hex number into a buffer /*{{{*/
// ---------------------------------------------------------------------
/* The length of the buffer must be exactly 1/2 the length of the string. */
-bool Hex2Num(const char *Start,const char *End,unsigned char *Num,
- unsigned int Length)
+bool Hex2Num(const string &Str,unsigned char *Num,unsigned int Length)
{
- if (End - Start != (signed)(Length*2))
+ if (Str.length() != Length*2)
return false;
// Convert each digit. We store it in the same order as the string
int J = 0;
- for (const char *I = Start; I < End;J++, I += 2)
+ for (string::const_iterator I = Str.begin(); I != Str.end();J++, I += 2)
{
if (isxdigit(*I) == 0 || isxdigit(I[1]) == 0)
return false;
return true;
}
/*}}}*/
+// TokSplitString - Split a string up by a given token /*{{{*/
+// ---------------------------------------------------------------------
+/* This is intended to be a faster splitter, it does not use dynamic
+ memories. Input is changed to insert nulls at each token location. */
+bool TokSplitString(char Tok,char *Input,char **List,
+ unsigned long ListMax)
+{
+ // Strip any leading spaces
+ char *Start = Input;
+ char *Stop = Start + strlen(Start);
+ for (; *Start != 0 && isspace(*Start) != 0; Start++);
+
+ unsigned long Count = 0;
+ char *Pos = Start;
+ while (Pos != Stop)
+ {
+ // Skip to the next Token
+ for (; Pos != Stop && *Pos != Tok; Pos++);
+
+ // Back remove spaces
+ char *End = Pos;
+ for (; End > Start && (End[-1] == Tok || isspace(End[-1]) != 0); End--);
+ *End = 0;
+
+ List[Count++] = Start;
+ if (Count >= ListMax)
+ {
+ List[Count-1] = 0;
+ return false;
+ }
+
+ // Advance pos
+ for (; Pos != Stop && (*Pos == Tok || isspace(*Pos) != 0 || *Pos == 0); Pos++);
+ Start = Pos;
+ }
+
+ List[Count] = 0;
+ 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<string> VectorizeString(string const &haystack, char const &split)
+{
+ string::const_iterator start = haystack.begin();
+ string::const_iterator end = start;
+ vector<string> 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 /*{{{*/
+// ---------------------------------------------------------------------
+/* */
+unsigned long RegexChoice(RxChoiceList *Rxs,const char **ListBegin,
+ const char **ListEnd)
+{
+ for (RxChoiceList *R = Rxs; R->Str != 0; R++)
+ R->Hit = false;
+
+ unsigned long Hits = 0;
+ for (; ListBegin != ListEnd; ListBegin++)
+ {
+ // Check if the name is a regex
+ const char *I;
+ bool Regex = true;
+ for (I = *ListBegin; *I != 0; I++)
+ if (*I == '.' || *I == '?' || *I == '*' || *I == '|')
+ break;
+ if (*I == 0)
+ Regex = false;
+
+ // Compile the regex pattern
+ regex_t Pattern;
+ if (Regex == true)
+ if (regcomp(&Pattern,*ListBegin,REG_EXTENDED | REG_ICASE |
+ REG_NOSUB) != 0)
+ Regex = false;
+
+ // Search the list
+ bool Done = false;
+ for (RxChoiceList *R = Rxs; R->Str != 0; R++)
+ {
+ if (R->Str[0] == 0)
+ continue;
+
+ if (strcasecmp(R->Str,*ListBegin) != 0)
+ {
+ if (Regex == false)
+ continue;
+ if (regexec(&Pattern,R->Str,0,0,0) != 0)
+ continue;
+ }
+ Done = true;
+
+ if (R->Hit == false)
+ Hits++;
+
+ R->Hit = true;
+ }
+
+ if (Regex == true)
+ regfree(&Pattern);
+
+ if (Done == false)
+ _error->Warning(_("Selection %s not found"),*ListBegin);
+ }
+
+ return Hits;
+}
+ /*}}}*/
+// ioprintf - C format string outputter to C++ 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,...)
+{
+ va_list args;
+ va_start(args,format);
+
+ // sprintf the description
+ 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
+ pointer to the end of the new string. The returned string is always null
+ terminated unless Buffer == end. This is a better alterantive to using
+ consecutive snprintfs. */
+char *safe_snprintf(char *Buffer,char *End,const char *Format,...)
+{
+ va_list args;
+ int Did;
+
+ va_start(args,Format);
+
+ if (End <= Buffer)
+ return End;
+
+ Did = vsnprintf(Buffer,End - Buffer,Format,args);
+ if (Did < 0 || Buffer + Did > End)
+ return End;
+ return Buffer + Did;
+}
+ /*}}}*/
+// StripEpoch - Remove the version "epoch" from a version string /*{{{*/
+// ---------------------------------------------------------------------
+string StripEpoch(const string &VerStr)
+{
+ size_t i = VerStr.find(":");
+ if (i == string::npos)
+ 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
+ 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
+ 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)
+ {
+ if (Cur < List.end() && *Cur != ',')
+ continue;
+
+ // Match the end of the string..
+ if ((Host.size() >= (unsigned)(Cur - Start)) &&
+ Cur - Start != 0 &&
+ stringcasecmp(Host.end() - (Cur - Start),Host.end(),Start,Cur) == 0)
+ return true;
+
+ Start = Cur + 1;
+ }
+ return false;
+}
+ /*}}}*/
+// DeEscapeString - unescape (\0XX and \xXX) from a string /*{{{*/
+// ---------------------------------------------------------------------
+/* */
+string DeEscapeString(const string &input)
+{
+ char tmp[3];
+ string::const_iterator it, escape_start;
+ string output, octal, hex;
+ 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 */
-void URI::CopyFrom(string U)
+void URI::CopyFrom(const string &U)
{
string::const_iterator I = U.begin();
- // Locate the first colon, this seperates the scheme
- for (; I < U.end() && *I != ':' ; I++);
+ // Locate the first colon, this separates the scheme
+ for (; I < U.end() && *I != ':' ; ++I);
string::const_iterator FirstColon = I;
/* Determine if this is a host type URI with a leading double //
string::const_iterator SingleSlash = I;
if (I + 3 < U.end() && I[1] == '/' && I[2] == '/')
SingleSlash += 3;
- for (; SingleSlash < U.end() && *SingleSlash != '/'; SingleSlash++);
+
+ /* 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)
+ {
+ if (*SingleSlash == '[')
+ InBracket = true;
+ if (InBracket == true && *SingleSlash == ']')
+ InBracket = false;
+ }
+
if (SingleSlash > U.end())
SingleSlash = U.end();
// We can now write the access and path specifiers
- Access = string(U,0,FirstColon - U.begin());
+ Access.assign(U.begin(),FirstColon);
if (SingleSlash != U.end())
- Path = string(U,SingleSlash - U.begin());
+ Path.assign(SingleSlash,U.end());
if (Path.empty() == true)
Path = "/";
// Now we attempt to locate a user:pass@host fragment
- if (FirstColon[1] == '/' && FirstColon[2] == '/')
+ if (FirstColon + 2 <= U.end() && FirstColon[1] == '/' && FirstColon[2] == '/')
FirstColon += 3;
else
FirstColon += 1;
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
if (At == SingleSlash)
{
if (FirstColon < SingleSlash)
- Host = string(U,FirstColon - U.begin(),SingleSlash - FirstColon);
+ Host.assign(FirstColon,SingleSlash);
}
else
{
- Host = string(U,At - U.begin() + 1,SingleSlash - At - 1);
- User = string(U,FirstColon - U.begin(),SecondColon - FirstColon);
+ Host.assign(At+1,SingleSlash);
+ // username and password must be encoded (RFC 3986)
+ User.assign(DeQuoteString(FirstColon,SecondColon));
if (SecondColon < At)
- Password = string(U,SecondColon - U.begin() + 1,At - SecondColon - 1);
+ Password.assign(DeQuoteString(SecondColon+1,At));
}
+ // Now we parse the RFC 2732 [] hostnames.
+ unsigned long PortEnd = 0;
+ InBracket = false;
+ for (unsigned I = 0; I != Host.length();)
+ {
+ if (Host[I] == '[')
+ {
+ InBracket = true;
+ Host.erase(I,1);
+ continue;
+ }
+
+ if (InBracket == true && Host[I] == ']')
+ {
+ InBracket = false;
+ Host.erase(I,1);
+ PortEnd = I;
+ continue;
+ }
+ I++;
+ }
+
+ // Tsk, weird.
+ if (InBracket == true)
+ {
+ Host.clear();
+ return;
+ }
+
// Now we parse off a port number from the hostname
Port = 0;
string::size_type Pos = Host.rfind(':');
- if (Pos == string::npos)
+ if (Pos == string::npos || Pos < PortEnd)
return;
Port = atoi(string(Host,Pos+1).c_str());
- Host = string(Host,0,Pos);
+ Host.assign(Host,0,Pos);
}
/*}}}*/
// URI::operator string - Convert the URI to a string /*{{{*/
Res = Access + ':';
if (Host.empty() == false)
- {
+ {
if (Access.empty() == false)
Res += "//";
-
+
if (User.empty() == false)
{
Res += User;
Res += "@";
}
- Res += Host;
+ // Add RFC 2732 escaping characters
+ if (Access.empty() == false &&
+ (Host.find('/') != string::npos || Host.find(':') != string::npos))
+ Res += '[' + Host + ']';
+ else
+ Res += Host;
+
if (Port != 0)
{
char S[30];
return Res;
}
/*}}}*/
+// 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::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;
+}
+ /*}}}*/