// -*- mode: cpp; mode: fold -*-
// Description /*{{{*/
-// $Id: strutl.cc,v 1.28 1999/08/02 03:07:48 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
#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 <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, *outptr;
+ size_t insize, outsize;
+
+ 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");
+
+ // Clean the destination string
+ *dest = "";
+
+ return false;
+ }
+
+ insize = outsize = orig.size();
+ inbuf = orig.data();
+ inptr = (char *)inbuf;
+ outbuf = new char[insize+1];
+ outptr = outbuf;
+
+ iconv(cd, &inptr, &insize, &outptr, &outsize);
+ *outptr = '\0';
+
+ *dest = outbuf;
+ 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++);
+ for (C++; *C != 0 && *C != '"'; C++);
+ if (*C == 0)
+ return false;
+ }
+ if (*C == '[')
+ {
+ for (C++; *C != 0 && *C != ']'; C++);
if (*C == 0)
return false;
}
/*}}}*/
// 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)
// DeQuoteString - Convert a string from quoted from /*{{{*/
// ---------------------------------------------------------------------
/* This undoes QuoteString */
-string DeQuoteString(string Str)
+string DeQuoteString(const string &Str)
{
string Res;
- for (string::iterator I = Str.begin(); I != Str.end(); I++)
+ for (string::const_iterator I = Str.begin(); I != Str.end(); I++)
{
if (*I == '%' && I + 2 < Str.end())
{
// 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;
+}
/*}}}*/
// 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)
{
- string::const_iterator I = URI.begin() + URI.find(':') + 1;
- for (; I < URI.end() && *I == '/'; I++);
-
+ // Nuke 'sensitive' items
+ ::URI U(URI);
+ U.User.clear();
+ U.Password.clear();
+ U.Access.clear();
+
// "\x00-\x20{}|\\\\^\\[\\]<>\"\x7F-\xFF";
- URI = QuoteString(string(I,URI.end() - I),"\\|{}[]<>\"^~_=!@#$%^&*");
- string::iterator J = URI.begin();
- for (; J != URI.end(); J++)
- if (*J == '/')
- *J = '_';
- return URI;
+ string NewURI = QuoteString(U,"\\|{}[]<>\"^~_=!@#$%^&*");
+ 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 -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 /*{{{*/
// ---------------------------------------------------------------------
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 (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--);
- return string(I,J-I);
+ return string(I,J);
}
for (; *I != '\n' && I < Message.end(); I++);
// ---------------------------------------------------------------------
/* 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);
fancy buffering is used. */
bool ReadMessages(int Fd, vector<string> &List)
{
- char Buffer[4000];
+ char Buffer[64000];
char *End = Buffer;
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);
// Fix up the buffer
for (; I < End && *I == '\n'; I++);
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;
/* This handles all 3 populare 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 :<*/
-bool StrToTime(string Val,time_t &Result)
+ reason the C library does not provide any such function :< This also
+ handles the weird, but unambiguous FTP time format*/
+bool StrToTime(const string &Val,time_t &Result)
{
struct tm Tm;
char Month[10];
for (;*I != 0 && *I != ' '; I++);
// 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) != 6)
{
// asctime format
if (sscanf(I," %3s %d %d:%d:%d %d",Month,&Tm.tm_mday,
&Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec,&Tm.tm_year) != 6)
- return false;
+ {
+ // 'ftp' time
+ 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--;
+ }
}
}
Tm.tm_isdst = 0;
- Tm.tm_mon = MonthConv(Month);
+ if (Month[0] != 0)
+ Tm.tm_mon = MonthConv(Month);
Tm.tm_year -= 1900;
// Convert to local time and then to GMT
// 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 != 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;
+}
+ /*}}}*/
+// 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[400];
+ vsnprintf(S,sizeof(S),format,args);
+ out << 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;
+}
+ /*}}}*/
+
+// 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;
+}
+ /*}}}*/
// 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
+ // Locate the first colon, this separates the scheme
for (; I < U.end() && *I != ':' ; I++);
string::const_iterator FirstColon = I;
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;
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);
+ User.assign(FirstColon,SecondColon);
if (SecondColon < At)
- Password = string(U,SecondColon - U.begin() + 1,At - SecondColon - 1);
+ Password.assign(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 /*{{{*/
/* */
URI::operator string()
{
- string Res = Access + ':';
+ string Res;
+
+ if (Access.empty() == false)
+ Res = Access + ':';
+
if (Host.empty() == false)
- {
- Res += "//";
+ {
+ if (Access.empty() == false)
+ Res += "//";
+
if (User.empty() == false)
{
- Res += "//" + User;
+ Res += User;
if (Password.empty() == false)
Res += ":" + Password;
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;
+}
+ /*}}}*/
projects / apt.git / blobdiff