// -*- mode: cpp; mode: fold -*-
// Description /*{{{*/
-// $Id: strutl.cc,v 1.39 2001/05/27 05:19:30 jgg Exp $
+// $Id: strutl.cc,v 1.48 2003/07/18 14:15:11 mdz Exp $
/* ######################################################################
String Util - Some useful string functions.
##################################################################### */
/*}}}*/
// 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 <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
// 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 (const char *I = Str.c_str(); *I != 0; I++)
+ for (string::const_iterator I = Str.begin(); I != Str.end(); I++)
{
if (strchr(Bad,*I) != 0 || isprint(*I) == 0 ||
*I <= 0x20 || *I >= 0x7F)
// DeQuoteString - Convert a string from quoted from /*{{{*/
// ---------------------------------------------------------------------
/* This undoes QuoteString */
-string DeQuoteString(string Str)
+string DeQuoteString(const string &Str)
{
string Res;
- for (const char *I = Str.c_str(); *I != 0; I++)
+ for (string::const_iterator I = Str.begin(); I != Str.end(); I++)
{
- if (*I == '%' && I + 2 < Str.c_str() + Str.length())
+ if (*I == '%' && I + 2 < Str.end())
{
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 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,"\\|{}[]<>\"^~_=!@#$%^&*");
- char *J = const_cast<char *>(URI.c_str());
- for (; *J != 0; J++)
- if (*J == '/')
- *J = '_';
- return URI;
+ string NewURI = QuoteString(U,"\\|{}[]<>\"^~_=!@#$%^&*");
+ 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',
/* Transform the 3x8 bits to 4x6 bits, as required by
base64. */
- for (const char *I = S.c_str(); I < (S.c_str() + S.length()); I += 3)
+ for (string::const_iterator I = S.begin(); I < S.end(); I += 3)
{
char Bits[3] = {0,0,0};
Bits[0] = I[0];
- if (I + 1 < S.c_str() + S.length())
+ if (I + 1 < S.end())
Bits[1] = I[1];
- if (I + 2 < S.c_str() + S.length())
+ if (I + 2 < S.end())
Bits[2] = I[2];
Final += tbl[Bits[0] >> 2];
Final += tbl[((Bits[0] & 3) << 4) + (Bits[1] >> 4)];
- if (I + 1 >= S.c_str() + S.length())
+ if (I + 1 >= S.end())
break;
Final += tbl[((Bits[1] & 0xf) << 2) + (Bits[2] >> 6)];
- if (I + 2 >= S.c_str() + S.length())
+ if (I + 2 >= S.end())
break;
Final += tbl[Bits[2] & 0x3f];
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)
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 (toupper(*A) != toupper(*B))
+ break;
+
+ if (A == AEnd && B == BEnd)
+ return 0;
+ if (A == AEnd)
+ return 1;
+ if (B == BEnd)
+ return -1;
+ if (toupper(*A) < toupper(*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 (toupper(*A) != toupper(*B))
+ break;
+
+ if (A == AEnd && B == BEnd)
+ return 0;
+ if (A == AEnd)
+ return 1;
+ if (B == BEnd)
+ return -1;
+ if (toupper(*A) < toupper(*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 (const char *I = Message.c_str(); I + Length < Message.c_str() + Message.length(); 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
- const char *J;
+ string::const_iterator J;
I += Length + 1;
- for (; isspace(*I) != 0 && *I != 0; I++);
- for (J = I; *J != '\n' && *J != 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 != 0; 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)
+int StringToBool(const string &Text,int Default)
{
char *End;
int Res = strtol(Text.c_str(),&End,0);
// ---------------------------------------------------------------------
/* 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)
{
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;
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
+
+/* 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!
static time_t timegm(struct tm *t)
{
time_t tl, tb;
'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];
// 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(string Str,unsigned char *Num,unsigned int Length)
+bool Hex2Num(const string &Str,unsigned char *Num,unsigned int Length)
{
if (Str.length() != Length*2)
return false;
return true;
}
/*}}}*/
+// ExplodeString - Split a string up into a vector /*{{{*/
+// ---------------------------------------------------------------------
+/* 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 an iteratable vector */
+vector<string> ExplodeString(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 /*{{{*/
// ---------------------------------------------------------------------
/* */
/*}}}*/
// ioprintf - C format string outputter to C++ iostreams /*{{{*/
// ---------------------------------------------------------------------
-/* This is used to make the internationalization strinc easier to translate
- and to allow reordering of parameters */
+/* 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[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);
- // std::ostream::operator <<(out, (const char *)S);
+ 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;
+ unsigned long 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;
+}
+ /*}}}*/
+
+// tolower_ascii - tolower() function that ignores the locale /*{{{*/
+// ---------------------------------------------------------------------
+/* */
+int tolower_ascii(int c)
+{
+ if (c >= 'A' and c <= 'Z')
+ return c + 32;
+ return c;
}
/*}}}*/
// ---------------------------------------------------------------------
/* The domain list is a comma seperate list of domains that are suffix
matched against the argument */
-bool CheckDomainList(string Host,string List)
+bool CheckDomainList(const string &Host,const string &List)
{
- const char *Start = List.c_str();
- for (const char *Cur = List.c_str(); *Cur != 0; Cur++)
+ string::const_iterator Start = List.begin();
+ for (string::const_iterator Cur = List.begin(); Cur <= List.end(); Cur++)
{
- if (*Cur != ',')
+ if (Cur < List.end() && *Cur != ',')
continue;
// Match the end of the string..
- if ((Host.size() >= (unsigned)(Cur - List.c_str())) &&
+ if ((Host.size() >= (unsigned)(Cur - Start)) &&
Cur - Start != 0 &&
- stringcasecmp(Host.c_str() + Host.length() - (Cur - Start),Host.c_str()+Host.length(),Start,Cur) == 0)
+ stringcasecmp(Host.end() - (Cur - Start),Host.end(),Start,Cur) == 0)
return true;
Start = Cur + 1;
// 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)
{
- const char *I = U.c_str();
+ string::const_iterator I = U.begin();
// Locate the first colon, this separates the scheme
- for (; *I != 0 && *I != ':' ; I++);
- const char *FirstColon = I;
+ for (; I < U.end() && *I != ':' ; I++);
+ string::const_iterator FirstColon = I;
/* Determine if this is a host type URI with a leading double //
and then search for the first single / */
- const char *SingleSlash = I;
- if (I + 3 < U.c_str() + U.length() && I[1] == '/' && I[2] == '/')
+ string::const_iterator SingleSlash = I;
+ if (I + 3 < U.end() && I[1] == '/' && I[2] == '/')
SingleSlash += 3;
/* Find the / indicating the end of the hostname, ignoring /'s in the
square brackets */
bool InBracket = false;
- for (; SingleSlash < U.c_str() + U.length() && (*SingleSlash != '/' || InBracket == true); SingleSlash++)
+ for (; SingleSlash < U.end() && (*SingleSlash != '/' || InBracket == true); SingleSlash++)
{
if (*SingleSlash == '[')
InBracket = true;
InBracket = false;
}
- if (SingleSlash > U.c_str() + U.length())
- SingleSlash = U.c_str() + U.length();
+ if (SingleSlash > U.end())
+ SingleSlash = U.end();
// We can now write the access and path specifiers
- Access = string(U,0,FirstColon - U.c_str());
- if (*SingleSlash != 0)
- Path = string(U,SingleSlash - U.c_str());
+ Access.assign(U.begin(),FirstColon);
+ if (SingleSlash != U.end())
+ 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;
- if (FirstColon >= U.c_str() + U.length())
+ if (FirstColon >= U.end())
return;
if (FirstColon > SingleSlash)
if (I > SingleSlash)
I = SingleSlash;
for (; I < SingleSlash && *I != ':'; I++);
- const char *SecondColon = I;
+ string::const_iterator SecondColon = I;
// Search for the @ after the colon
for (; I < SingleSlash && *I != '@'; I++);
- const char *At = I;
+ string::const_iterator At = I;
// Now write the host and user/pass
if (At == SingleSlash)
{
if (FirstColon < SingleSlash)
- Host = string(U,FirstColon - U.c_str(),SingleSlash - FirstColon);
+ Host.assign(FirstColon,SingleSlash);
}
else
{
- Host = string(U,At - U.c_str() + 1,SingleSlash - At - 1);
- User = string(U,FirstColon - U.c_str(),SecondColon - FirstColon);
+ Host.assign(At+1,SingleSlash);
+ User.assign(FirstColon,SecondColon);
if (SecondColon < At)
- Password = string(U,SecondColon - U.c_str() + 1,At - SecondColon - 1);
+ Password.assign(SecondColon+1,At);
}
// Now we parse the RFC 2732 [] hostnames.
// Tsk, weird.
if (InBracket == true)
{
- Host = string();
+ Host.clear();
return;
}
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 /*{{{*/
// URI::SiteOnly - Return the schema and site for the URI /*{{{*/
// ---------------------------------------------------------------------
/* */
-string URI::SiteOnly(string 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 = string();
- U.Password = string();
- U.Path = string();
+ U.User.clear();
+ U.Password.clear();
U.Port = 0;
return U;
}
projects / apt.git / blobdiff