1 // -*- mode: cpp; mode: fold -*-
3 // $Id: strutl.cc,v 1.48 2003/07/18 14:15:11 mdz Exp $
4 /* ######################################################################
6 String Util - Some useful string functions.
8 These have been collected from here and there to do all sorts of useful
9 things to strings. They are useful in file parsers, URI handlers and
10 especially in APT methods.
12 This source is placed in the Public Domain, do with it what you will
13 It was originally written by Jason Gunthorpe <jgg@gpu.srv.ualberta.ca>
15 ##################################################################### */
18 #include <apt-pkg/strutl.h>
19 #include <apt-pkg/fileutl.h>
20 #include <apt-pkg/error.h>
39 // UTF8ToCodeset - Convert some UTF-8 string for some codeset /*{{{*/
40 // ---------------------------------------------------------------------
41 /* This is handy to use before display some information for enduser */
42 bool UTF8ToCodeset(const char *codeset
, const string
&orig
, string
*dest
)
47 size_t insize
, bufsize
;
50 cd
= iconv_open(codeset
, "UTF-8");
51 if (cd
== (iconv_t
)(-1)) {
52 // Something went wrong
54 _error
->Error("conversion from 'UTF-8' to '%s' not available",
62 insize
= bufsize
= orig
.size();
64 inptr
= (char *)inbuf
;
65 outbuf
= new char[bufsize
];
66 size_t lastError
= -1;
70 char *outptr
= outbuf
;
71 size_t outsize
= bufsize
;
72 size_t const err
= iconv(cd
, &inptr
, &insize
, &outptr
, &outsize
);
73 dest
->append(outbuf
, outptr
- outbuf
);
74 if (err
== (size_t)(-1))
81 // replace a series of unknown multibytes with a single "?"
82 if (lastError
!= insize
) {
83 lastError
= insize
- 1;
95 outbuf
= new char[bufsize
];
109 // strstrip - Remove white space from the front and back of a string /*{{{*/
110 // ---------------------------------------------------------------------
111 /* This is handy to use when parsing a file. It also removes \n's left
112 over from fgets and company */
113 char *_strstrip(char *String
)
115 for (;*String
!= 0 && (*String
== ' ' || *String
== '\t'); String
++);
120 char *End
= String
+ strlen(String
) - 1;
121 for (;End
!= String
- 1 && (*End
== ' ' || *End
== '\t' || *End
== '\n' ||
122 *End
== '\r'); End
--);
128 // strtabexpand - Converts tabs into 8 spaces /*{{{*/
129 // ---------------------------------------------------------------------
131 char *_strtabexpand(char *String
,size_t Len
)
133 for (char *I
= String
; I
!= I
+ Len
&& *I
!= 0; I
++)
137 if (I
+ 8 > String
+ Len
)
143 /* Assume the start of the string is 0 and find the next 8 char
149 Len
= 8 - ((String
- I
) % 8);
157 memmove(I
+ Len
,I
+ 1,strlen(I
) + 1);
158 for (char *J
= I
; J
+ Len
!= I
; *I
= ' ', I
++);
163 // ParseQuoteWord - Parse a single word out of a string /*{{{*/
164 // ---------------------------------------------------------------------
165 /* This grabs a single word, converts any % escaped characters to their
166 proper values and advances the pointer. Double quotes are understood
167 and striped out as well. This is for URI/URL parsing. It also can
168 understand [] brackets.*/
169 bool ParseQuoteWord(const char *&String
,string
&Res
)
171 // Skip leading whitespace
172 const char *C
= String
;
173 for (;*C
!= 0 && *C
== ' '; C
++);
177 // Jump to the next word
178 for (;*C
!= 0 && isspace(*C
) == 0; C
++)
182 for (C
++; *C
!= 0 && *C
!= '"'; C
++);
188 for (C
++; *C
!= 0 && *C
!= ']'; C
++);
194 // Now de-quote characters
197 const char *Start
= String
;
199 for (I
= Buffer
; I
< Buffer
+ sizeof(Buffer
) && Start
!= C
; I
++)
201 if (*Start
== '%' && Start
+ 2 < C
&&
202 isxdigit(Start
[1]) && isxdigit(Start
[2]))
207 *I
= (char)strtol(Tmp
,0,16);
220 // Skip ending white space
221 for (;*C
!= 0 && isspace(*C
) != 0; C
++);
226 // ParseCWord - Parses a string like a C "" expression /*{{{*/
227 // ---------------------------------------------------------------------
228 /* This expects a series of space separated strings enclosed in ""'s.
229 It concatenates the ""'s into a single string. */
230 bool ParseCWord(const char *&String
,string
&Res
)
232 // Skip leading whitespace
233 const char *C
= String
;
234 for (;*C
!= 0 && *C
== ' '; C
++);
240 if (strlen(String
) >= sizeof(Buffer
))
247 for (C
++; *C
!= 0 && *C
!= '"'; C
++)
256 if (C
!= String
&& isspace(*C
) != 0 && isspace(C
[-1]) != 0)
258 if (isspace(*C
) == 0)
268 // QuoteString - Convert a string into quoted from /*{{{*/
269 // ---------------------------------------------------------------------
271 string
QuoteString(const string
&Str
, const char *Bad
)
274 for (string::const_iterator I
= Str
.begin(); I
!= Str
.end(); I
++)
276 if (strchr(Bad
,*I
) != 0 || isprint(*I
) == 0 ||
277 *I
== 0x25 || // percent '%' char
278 *I
<= 0x20 || *I
>= 0x7F) // control chars
281 sprintf(Buf
,"%%%02x",(int)*I
);
290 // DeQuoteString - Convert a string from quoted from /*{{{*/
291 // ---------------------------------------------------------------------
292 /* This undoes QuoteString */
293 string
DeQuoteString(const string
&Str
)
295 return DeQuoteString(Str
.begin(),Str
.end());
297 string
DeQuoteString(string::const_iterator
const &begin
,
298 string::const_iterator
const &end
)
301 for (string::const_iterator I
= begin
; I
!= end
; I
++)
303 if (*I
== '%' && I
+ 2 < end
&&
304 isxdigit(I
[1]) && isxdigit(I
[2]))
310 Res
+= (char)strtol(Tmp
,0,16);
321 // SizeToStr - Convert a long into a human readable size /*{{{*/
322 // ---------------------------------------------------------------------
323 /* A max of 4 digits are shown before conversion to the next highest unit.
324 The max length of the string will be 5 chars unless the size is > 10
326 string
SizeToStr(double Size
)
335 /* bytes, KiloBytes, MegaBytes, GigaBytes, TeraBytes, PetaBytes,
336 ExaBytes, ZettaBytes, YottaBytes */
337 char Ext
[] = {'\0','k','M','G','T','P','E','Z','Y'};
341 if (ASize
< 100 && I
!= 0)
343 sprintf(S
,"%'.1f %c",ASize
,Ext
[I
]);
349 sprintf(S
,"%'.0f %c",ASize
,Ext
[I
]);
359 // TimeToStr - Convert the time into a string /*{{{*/
360 // ---------------------------------------------------------------------
361 /* Converts a number of seconds to a hms format */
362 string
TimeToStr(unsigned long Sec
)
370 //d means days, h means hours, min means minutes, s means seconds
371 sprintf(S
,_("%lid %lih %limin %lis"),Sec
/60/60/24,(Sec
/60/60) % 24,(Sec
/60) % 60,Sec
% 60);
377 //h means hours, min means minutes, s means seconds
378 sprintf(S
,_("%lih %limin %lis"),Sec
/60/60,(Sec
/60) % 60,Sec
% 60);
384 //min means minutes, s means seconds
385 sprintf(S
,_("%limin %lis"),Sec
/60,Sec
% 60);
390 sprintf(S
,_("%lis"),Sec
);
397 // SubstVar - Substitute a string for another string /*{{{*/
398 // ---------------------------------------------------------------------
399 /* This replaces all occurances of Subst with Contents in Str. */
400 string
SubstVar(const string
&Str
,const string
&Subst
,const string
&Contents
)
402 string::size_type Pos
= 0;
403 string::size_type OldPos
= 0;
406 while (OldPos
< Str
.length() &&
407 (Pos
= Str
.find(Subst
,OldPos
)) != string::npos
)
409 Temp
+= string(Str
,OldPos
,Pos
) + Contents
;
410 OldPos
= Pos
+ Subst
.length();
416 return Temp
+ string(Str
,OldPos
);
419 string
SubstVar(string Str
,const struct SubstVar
*Vars
)
421 for (; Vars
->Subst
!= 0; Vars
++)
422 Str
= SubstVar(Str
,Vars
->Subst
,*Vars
->Contents
);
426 // OutputInDepth - return a string with separator multiplied with depth /*{{{*/
427 // ---------------------------------------------------------------------
428 /* Returns a string with the supplied separator depth + 1 times in it */
429 std::string
OutputInDepth(const unsigned long Depth
, const char* Separator
)
431 std::string output
= "";
432 for(unsigned long d
=Depth
+1; d
> 0; d
--)
433 output
.append(Separator
);
437 // URItoFileName - Convert the uri into a unique file name /*{{{*/
438 // ---------------------------------------------------------------------
439 /* This converts a URI into a safe filename. It quotes all unsafe characters
440 and converts / to _ and removes the scheme identifier. The resulting
441 file name should be unique and never occur again for a different file */
442 string
URItoFileName(const string
&URI
)
444 // Nuke 'sensitive' items
450 // "\x00-\x20{}|\\\\^\\[\\]<>\"\x7F-\xFF";
451 string NewURI
= QuoteString(U
,"\\|{}[]<>\"^~_=!@#$%^&*");
452 replace(NewURI
.begin(),NewURI
.end(),'/','_');
456 // Base64Encode - Base64 Encoding routine for short strings /*{{{*/
457 // ---------------------------------------------------------------------
458 /* This routine performs a base64 transformation on a string. It was ripped
459 from wget and then patched and bug fixed.
461 This spec can be found in rfc2045 */
462 string
Base64Encode(const string
&S
)
465 static char tbl
[64] = {'A','B','C','D','E','F','G','H',
466 'I','J','K','L','M','N','O','P',
467 'Q','R','S','T','U','V','W','X',
468 'Y','Z','a','b','c','d','e','f',
469 'g','h','i','j','k','l','m','n',
470 'o','p','q','r','s','t','u','v',
471 'w','x','y','z','0','1','2','3',
472 '4','5','6','7','8','9','+','/'};
474 // Pre-allocate some space
476 Final
.reserve((4*S
.length() + 2)/3 + 2);
478 /* Transform the 3x8 bits to 4x6 bits, as required by
480 for (string::const_iterator I
= S
.begin(); I
< S
.end(); I
+= 3)
482 char Bits
[3] = {0,0,0};
489 Final
+= tbl
[Bits
[0] >> 2];
490 Final
+= tbl
[((Bits
[0] & 3) << 4) + (Bits
[1] >> 4)];
492 if (I
+ 1 >= S
.end())
495 Final
+= tbl
[((Bits
[1] & 0xf) << 2) + (Bits
[2] >> 6)];
497 if (I
+ 2 >= S
.end())
500 Final
+= tbl
[Bits
[2] & 0x3f];
503 /* Apply the padding elements, this tells how many bytes the remote
504 end should discard */
505 if (S
.length() % 3 == 2)
507 if (S
.length() % 3 == 1)
513 // stringcmp - Arbitrary string compare /*{{{*/
514 // ---------------------------------------------------------------------
515 /* This safely compares two non-null terminated strings of arbitrary
517 int stringcmp(const char *A
,const char *AEnd
,const char *B
,const char *BEnd
)
519 for (; A
!= AEnd
&& B
!= BEnd
; A
++, B
++)
523 if (A
== AEnd
&& B
== BEnd
)
535 int stringcmp(string::const_iterator A
,string::const_iterator AEnd
,
536 const char *B
,const char *BEnd
)
538 for (; A
!= AEnd
&& B
!= BEnd
; A
++, B
++)
542 if (A
== AEnd
&& B
== BEnd
)
552 int stringcmp(string::const_iterator A
,string::const_iterator AEnd
,
553 string::const_iterator B
,string::const_iterator BEnd
)
555 for (; A
!= AEnd
&& B
!= BEnd
; A
++, B
++)
559 if (A
== AEnd
&& B
== BEnd
)
571 // stringcasecmp - Arbitrary case insensitive string compare /*{{{*/
572 // ---------------------------------------------------------------------
574 int stringcasecmp(const char *A
,const char *AEnd
,const char *B
,const char *BEnd
)
576 for (; A
!= AEnd
&& B
!= BEnd
; A
++, B
++)
577 if (tolower_ascii(*A
) != tolower_ascii(*B
))
580 if (A
== AEnd
&& B
== BEnd
)
586 if (tolower_ascii(*A
) < tolower_ascii(*B
))
591 int stringcasecmp(string::const_iterator A
,string::const_iterator AEnd
,
592 const char *B
,const char *BEnd
)
594 for (; A
!= AEnd
&& B
!= BEnd
; A
++, B
++)
595 if (tolower_ascii(*A
) != tolower_ascii(*B
))
598 if (A
== AEnd
&& B
== BEnd
)
604 if (tolower_ascii(*A
) < tolower_ascii(*B
))
608 int stringcasecmp(string::const_iterator A
,string::const_iterator AEnd
,
609 string::const_iterator B
,string::const_iterator BEnd
)
611 for (; A
!= AEnd
&& B
!= BEnd
; A
++, B
++)
612 if (tolower_ascii(*A
) != tolower_ascii(*B
))
615 if (A
== AEnd
&& B
== BEnd
)
621 if (tolower_ascii(*A
) < tolower_ascii(*B
))
627 // LookupTag - Lookup the value of a tag in a taged string /*{{{*/
628 // ---------------------------------------------------------------------
629 /* The format is like those used in package files and the method
630 communication system */
631 string
LookupTag(const string
&Message
,const char *Tag
,const char *Default
)
633 // Look for a matching tag.
634 int Length
= strlen(Tag
);
635 for (string::const_iterator I
= Message
.begin(); I
+ Length
< Message
.end(); I
++)
638 if (I
[Length
] == ':' && stringcasecmp(I
,I
+Length
,Tag
) == 0)
640 // Find the end of line and strip the leading/trailing spaces
641 string::const_iterator J
;
643 for (; isspace(*I
) != 0 && I
< Message
.end(); I
++);
644 for (J
= I
; *J
!= '\n' && J
< Message
.end(); J
++);
645 for (; J
> I
&& isspace(J
[-1]) != 0; J
--);
650 for (; *I
!= '\n' && I
< Message
.end(); I
++);
653 // Failed to find a match
659 // StringToBool - Converts a string into a boolean /*{{{*/
660 // ---------------------------------------------------------------------
661 /* This inspects the string to see if it is true or if it is false and
662 then returns the result. Several varients on true/false are checked. */
663 int StringToBool(const string
&Text
,int Default
)
666 int Res
= strtol(Text
.c_str(),&End
,0);
667 if (End
!= Text
.c_str() && Res
>= 0 && Res
<= 1)
670 // Check for positives
671 if (strcasecmp(Text
.c_str(),"no") == 0 ||
672 strcasecmp(Text
.c_str(),"false") == 0 ||
673 strcasecmp(Text
.c_str(),"without") == 0 ||
674 strcasecmp(Text
.c_str(),"off") == 0 ||
675 strcasecmp(Text
.c_str(),"disable") == 0)
678 // Check for negatives
679 if (strcasecmp(Text
.c_str(),"yes") == 0 ||
680 strcasecmp(Text
.c_str(),"true") == 0 ||
681 strcasecmp(Text
.c_str(),"with") == 0 ||
682 strcasecmp(Text
.c_str(),"on") == 0 ||
683 strcasecmp(Text
.c_str(),"enable") == 0)
689 // TimeRFC1123 - Convert a time_t into RFC1123 format /*{{{*/
690 // ---------------------------------------------------------------------
691 /* This converts a time_t into a string time representation that is
692 year 2000 complient and timezone neutral */
693 string
TimeRFC1123(time_t Date
)
695 struct tm Conv
= *gmtime(&Date
);
698 const char *Day
[] = {"Sun","Mon","Tue","Wed","Thu","Fri","Sat"};
699 const char *Month
[] = {"Jan","Feb","Mar","Apr","May","Jun","Jul",
700 "Aug","Sep","Oct","Nov","Dec"};
702 sprintf(Buf
,"%s, %02i %s %i %02i:%02i:%02i GMT",Day
[Conv
.tm_wday
],
703 Conv
.tm_mday
,Month
[Conv
.tm_mon
],Conv
.tm_year
+1900,Conv
.tm_hour
,
704 Conv
.tm_min
,Conv
.tm_sec
);
708 // ReadMessages - Read messages from the FD /*{{{*/
709 // ---------------------------------------------------------------------
710 /* This pulls full messages from the input FD into the message buffer.
711 It assumes that messages will not pause during transit so no
712 fancy buffering is used.
714 In particular: this reads blocks from the input until it believes
715 that it's run out of input text. Each block is terminated by a
716 double newline ('\n' followed by '\n'). As noted below, there is a
717 bug in this code: it assumes that all the blocks have been read if
718 it doesn't see additional text in the buffer after the last one is
719 parsed, which will cause it to lose blocks if the last block
720 coincides with the end of the buffer.
722 bool ReadMessages(int Fd
, vector
<string
> &List
)
726 // Represents any left-over from the previous iteration of the
727 // parse loop. (i.e., if a message is split across the end
728 // of the buffer, it goes here)
729 string PartialMessage
;
733 int Res
= read(Fd
,End
,sizeof(Buffer
) - (End
-Buffer
));
734 if (Res
< 0 && errno
== EINTR
)
737 // Process is dead, this is kind of bad..
742 if (Res
< 0 && errno
== EAGAIN
)
749 // Look for the end of the message
750 for (char *I
= Buffer
; I
+ 1 < End
; I
++)
752 if (I
[0] != '\n' || I
[1] != '\n')
755 // Pull the message out
756 string
Message(Buffer
,I
-Buffer
);
757 PartialMessage
+= Message
;
760 for (; I
< End
&& *I
== '\n'; I
++);
762 memmove(Buffer
,I
,End
-Buffer
);
765 List
.push_back(PartialMessage
);
766 PartialMessage
.clear();
770 // If there's text left in the buffer, store it
771 // in PartialMessage and throw the rest of the buffer
772 // away. This allows us to handle messages that
773 // are longer than the static buffer size.
774 PartialMessage
+= string(Buffer
, End
);
779 // BUG ALERT: if a message block happens to end at a
780 // multiple of 64000 characters, this will cause it to
781 // terminate early, leading to a badly formed block and
782 // probably crashing the method. However, this is the only
783 // way we have to find the end of the message block. I have
784 // an idea of how to fix this, but it will require changes
785 // to the protocol (essentially to mark the beginning and
786 // end of the block).
788 // -- dburrows 2008-04-02
792 if (WaitFd(Fd
) == false)
797 // MonthConv - Converts a month string into a number /*{{{*/
798 // ---------------------------------------------------------------------
799 /* This was lifted from the boa webserver which lifted it from 'wn-v1.07'
800 Made it a bit more robust with a few tolower_ascii though. */
801 static int MonthConv(char *Month
)
803 switch (tolower_ascii(*Month
))
806 return tolower_ascii(Month
[1]) == 'p'?3:7;
812 if (tolower_ascii(Month
[1]) == 'a')
814 return tolower_ascii(Month
[2]) == 'n'?5:6;
816 return tolower_ascii(Month
[2]) == 'r'?2:4;
824 // Pretend it is January..
830 // timegm - Internal timegm if the gnu version is not available /*{{{*/
831 // ---------------------------------------------------------------------
832 /* Converts struct tm to time_t, assuming the data in tm is UTC rather
833 than local timezone (mktime assumes the latter).
835 This function is a nonstandard GNU extension that is also present on
836 the BSDs and maybe other systems. For others we follow the advice of
837 the manpage of timegm and use his portable replacement. */
839 static time_t timegm(struct tm
*t
)
841 char *tz
= getenv("TZ");
844 time_t ret
= mktime(t
);
854 // FullDateToTime - Converts a HTTP1.1 full date strings into a time_t /*{{{*/
855 // ---------------------------------------------------------------------
856 /* tries to parses a full date as specified in RFC2616 Section 3.3.1
857 with one exception: All timezones (%Z) are accepted but the protocol
858 says that it MUST be GMT, but this one is equal to UTC which we will
859 encounter from time to time (e.g. in Release files) so we accept all
860 here and just assume it is GMT (or UTC) later on */
861 bool RFC1123StrToTime(const char* const str
,time_t &time
)
864 setlocale (LC_ALL
,"C");
866 // Sun, 06 Nov 1994 08:49:37 GMT ; RFC 822, updated by RFC 1123
867 (strptime(str
, "%a, %d %b %Y %H:%M:%S %Z", &Tm
) == NULL
&&
868 // Sunday, 06-Nov-94 08:49:37 GMT ; RFC 850, obsoleted by RFC 1036
869 strptime(str
, "%A, %d-%b-%y %H:%M:%S %Z", &Tm
) == NULL
&&
870 // Sun Nov 6 08:49:37 1994 ; ANSI C's asctime() format
871 strptime(str
, "%a %b %d %H:%M:%S %Y", &Tm
) == NULL
);
872 setlocale (LC_ALL
,"");
880 // FTPMDTMStrToTime - Converts a ftp modification date into a time_t /*{{{*/
881 // ---------------------------------------------------------------------
883 bool FTPMDTMStrToTime(const char* const str
,time_t &time
)
886 // MDTM includes no whitespaces but recommend and ignored by strptime
887 if (strptime(str
, "%Y %m %d %H %M %S", &Tm
) == NULL
)
894 // StrToTime - Converts a string into a time_t /*{{{*/
895 // ---------------------------------------------------------------------
896 /* This handles all 3 populare time formats including RFC 1123, RFC 1036
897 and the C library asctime format. It requires the GNU library function
898 'timegm' to convert a struct tm in UTC to a time_t. For some bizzar
899 reason the C library does not provide any such function :< This also
900 handles the weird, but unambiguous FTP time format*/
901 bool StrToTime(const string
&Val
,time_t &Result
)
905 const char *I
= Val
.c_str();
907 // Skip the day of the week
908 for (;*I
!= 0 && *I
!= ' '; I
++);
910 // Handle RFC 1123 time
912 if (sscanf(I
," %d %3s %d %d:%d:%d GMT",&Tm
.tm_mday
,Month
,&Tm
.tm_year
,
913 &Tm
.tm_hour
,&Tm
.tm_min
,&Tm
.tm_sec
) != 6)
915 // Handle RFC 1036 time
916 if (sscanf(I
," %d-%3s-%d %d:%d:%d GMT",&Tm
.tm_mday
,Month
,
917 &Tm
.tm_year
,&Tm
.tm_hour
,&Tm
.tm_min
,&Tm
.tm_sec
) == 6)
922 if (sscanf(I
," %3s %d %d:%d:%d %d",Month
,&Tm
.tm_mday
,
923 &Tm
.tm_hour
,&Tm
.tm_min
,&Tm
.tm_sec
,&Tm
.tm_year
) != 6)
926 if (sscanf(Val
.c_str(),"%4d%2d%2d%2d%2d%2d",&Tm
.tm_year
,&Tm
.tm_mon
,
927 &Tm
.tm_mday
,&Tm
.tm_hour
,&Tm
.tm_min
,&Tm
.tm_sec
) != 6)
936 Tm
.tm_mon
= MonthConv(Month
);
939 // Convert to local time and then to GMT
940 Result
= timegm(&Tm
);
944 // StrToNum - Convert a fixed length string to a number /*{{{*/
945 // ---------------------------------------------------------------------
946 /* This is used in decoding the crazy fixed length string headers in
948 bool StrToNum(const char *Str
,unsigned long &Res
,unsigned Len
,unsigned Base
)
951 if (Len
>= sizeof(S
))
956 // All spaces is a zero
959 for (I
= 0; S
[I
] == ' '; I
++);
964 Res
= strtoul(S
,&End
,Base
);
971 // Base256ToNum - Convert a fixed length binary to a number /*{{{*/
972 // ---------------------------------------------------------------------
973 /* This is used in decoding the 256bit encoded fixed length fields in
975 bool Base256ToNum(const char *Str
,unsigned long &Res
,unsigned Len
)
978 if ((Str
[0] & 0x80) == 0)
984 Res
= (Res
<<8) + Str
[i
];
989 // HexDigit - Convert a hex character into an integer /*{{{*/
990 // ---------------------------------------------------------------------
991 /* Helper for Hex2Num */
992 static int HexDigit(int c
)
994 if (c
>= '0' && c
<= '9')
996 if (c
>= 'a' && c
<= 'f')
998 if (c
>= 'A' && c
<= 'F')
1003 // Hex2Num - Convert a long hex number into a buffer /*{{{*/
1004 // ---------------------------------------------------------------------
1005 /* The length of the buffer must be exactly 1/2 the length of the string. */
1006 bool Hex2Num(const string
&Str
,unsigned char *Num
,unsigned int Length
)
1008 if (Str
.length() != Length
*2)
1011 // Convert each digit. We store it in the same order as the string
1013 for (string::const_iterator I
= Str
.begin(); I
!= Str
.end();J
++, I
+= 2)
1015 if (isxdigit(*I
) == 0 || isxdigit(I
[1]) == 0)
1018 Num
[J
] = HexDigit(I
[0]) << 4;
1019 Num
[J
] += HexDigit(I
[1]);
1025 // TokSplitString - Split a string up by a given token /*{{{*/
1026 // ---------------------------------------------------------------------
1027 /* This is intended to be a faster splitter, it does not use dynamic
1028 memories. Input is changed to insert nulls at each token location. */
1029 bool TokSplitString(char Tok
,char *Input
,char **List
,
1030 unsigned long ListMax
)
1032 // Strip any leading spaces
1033 char *Start
= Input
;
1034 char *Stop
= Start
+ strlen(Start
);
1035 for (; *Start
!= 0 && isspace(*Start
) != 0; Start
++);
1037 unsigned long Count
= 0;
1041 // Skip to the next Token
1042 for (; Pos
!= Stop
&& *Pos
!= Tok
; Pos
++);
1044 // Back remove spaces
1046 for (; End
> Start
&& (End
[-1] == Tok
|| isspace(End
[-1]) != 0); End
--);
1049 List
[Count
++] = Start
;
1050 if (Count
>= ListMax
)
1057 for (; Pos
!= Stop
&& (*Pos
== Tok
|| isspace(*Pos
) != 0 || *Pos
== 0); Pos
++);
1065 // VectorizeString - Split a string up into a vector of strings /*{{{*/
1066 // ---------------------------------------------------------------------
1067 /* This can be used to split a given string up into a vector, so the
1068 propose is the same as in the method above and this one is a bit slower
1069 also, but the advantage is that we have an iteratable vector */
1070 vector
<string
> VectorizeString(string
const &haystack
, char const &split
)
1072 string::const_iterator start
= haystack
.begin();
1073 string::const_iterator end
= start
;
1074 vector
<string
> exploded
;
1076 for (; end
!= haystack
.end() && *end
!= split
; ++end
);
1077 exploded
.push_back(string(start
, end
));
1079 } while (end
!= haystack
.end() && (++end
) != haystack
.end());
1083 // RegexChoice - Simple regex list/list matcher /*{{{*/
1084 // ---------------------------------------------------------------------
1086 unsigned long RegexChoice(RxChoiceList
*Rxs
,const char **ListBegin
,
1087 const char **ListEnd
)
1089 for (RxChoiceList
*R
= Rxs
; R
->Str
!= 0; R
++)
1092 unsigned long Hits
= 0;
1093 for (; ListBegin
!= ListEnd
; ListBegin
++)
1095 // Check if the name is a regex
1098 for (I
= *ListBegin
; *I
!= 0; I
++)
1099 if (*I
== '.' || *I
== '?' || *I
== '*' || *I
== '|')
1104 // Compile the regex pattern
1107 if (regcomp(&Pattern
,*ListBegin
,REG_EXTENDED
| REG_ICASE
|
1113 for (RxChoiceList
*R
= Rxs
; R
->Str
!= 0; R
++)
1118 if (strcasecmp(R
->Str
,*ListBegin
) != 0)
1122 if (regexec(&Pattern
,R
->Str
,0,0,0) != 0)
1127 if (R
->Hit
== false)
1137 _error
->Warning(_("Selection %s not found"),*ListBegin
);
1143 // ioprintf - C format string outputter to C++ iostreams /*{{{*/
1144 // ---------------------------------------------------------------------
1145 /* This is used to make the internationalization strings easier to translate
1146 and to allow reordering of parameters */
1147 void ioprintf(ostream
&out
,const char *format
,...)
1150 va_start(args
,format
);
1152 // sprintf the description
1154 vsnprintf(S
,sizeof(S
),format
,args
);
1158 // strprintf - C format string outputter to C++ strings /*{{{*/
1159 // ---------------------------------------------------------------------
1160 /* This is used to make the internationalization strings easier to translate
1161 and to allow reordering of parameters */
1162 void strprintf(string
&out
,const char *format
,...)
1165 va_start(args
,format
);
1167 // sprintf the description
1169 vsnprintf(S
,sizeof(S
),format
,args
);
1173 // safe_snprintf - Safer snprintf /*{{{*/
1174 // ---------------------------------------------------------------------
1175 /* This is a snprintf that will never (ever) go past 'End' and returns a
1176 pointer to the end of the new string. The returned string is always null
1177 terminated unless Buffer == end. This is a better alterantive to using
1178 consecutive snprintfs. */
1179 char *safe_snprintf(char *Buffer
,char *End
,const char *Format
,...)
1184 va_start(args
,Format
);
1189 Did
= vsnprintf(Buffer
,End
- Buffer
,Format
,args
);
1190 if (Did
< 0 || Buffer
+ Did
> End
)
1192 return Buffer
+ Did
;
1196 // tolower_ascii - tolower() function that ignores the locale /*{{{*/
1197 // ---------------------------------------------------------------------
1198 /* This little function is the most called method we have and tries
1199 therefore to do the absolut minimum - and is noteable faster than
1200 standard tolower/toupper and as a bonus avoids problems with different
1201 locales - we only operate on ascii chars anyway. */
1202 int tolower_ascii(int const c
)
1204 if (c
>= 'A' && c
<= 'Z')
1210 // CheckDomainList - See if Host is in a , seperate list /*{{{*/
1211 // ---------------------------------------------------------------------
1212 /* The domain list is a comma seperate list of domains that are suffix
1213 matched against the argument */
1214 bool CheckDomainList(const string
&Host
,const string
&List
)
1216 string::const_iterator Start
= List
.begin();
1217 for (string::const_iterator Cur
= List
.begin(); Cur
<= List
.end(); Cur
++)
1219 if (Cur
< List
.end() && *Cur
!= ',')
1222 // Match the end of the string..
1223 if ((Host
.size() >= (unsigned)(Cur
- Start
)) &&
1225 stringcasecmp(Host
.end() - (Cur
- Start
),Host
.end(),Start
,Cur
) == 0)
1234 // URI::CopyFrom - Copy from an object /*{{{*/
1235 // ---------------------------------------------------------------------
1236 /* This parses the URI into all of its components */
1237 void URI::CopyFrom(const string
&U
)
1239 string::const_iterator I
= U
.begin();
1241 // Locate the first colon, this separates the scheme
1242 for (; I
< U
.end() && *I
!= ':' ; I
++);
1243 string::const_iterator FirstColon
= I
;
1245 /* Determine if this is a host type URI with a leading double //
1246 and then search for the first single / */
1247 string::const_iterator SingleSlash
= I
;
1248 if (I
+ 3 < U
.end() && I
[1] == '/' && I
[2] == '/')
1251 /* Find the / indicating the end of the hostname, ignoring /'s in the
1253 bool InBracket
= false;
1254 for (; SingleSlash
< U
.end() && (*SingleSlash
!= '/' || InBracket
== true); SingleSlash
++)
1256 if (*SingleSlash
== '[')
1258 if (InBracket
== true && *SingleSlash
== ']')
1262 if (SingleSlash
> U
.end())
1263 SingleSlash
= U
.end();
1265 // We can now write the access and path specifiers
1266 Access
.assign(U
.begin(),FirstColon
);
1267 if (SingleSlash
!= U
.end())
1268 Path
.assign(SingleSlash
,U
.end());
1269 if (Path
.empty() == true)
1272 // Now we attempt to locate a user:pass@host fragment
1273 if (FirstColon
+ 2 <= U
.end() && FirstColon
[1] == '/' && FirstColon
[2] == '/')
1277 if (FirstColon
>= U
.end())
1280 if (FirstColon
> SingleSlash
)
1281 FirstColon
= SingleSlash
;
1283 // Find the colon...
1285 if (I
> SingleSlash
)
1287 for (; I
< SingleSlash
&& *I
!= ':'; I
++);
1288 string::const_iterator SecondColon
= I
;
1290 // Search for the @ after the colon
1291 for (; I
< SingleSlash
&& *I
!= '@'; I
++);
1292 string::const_iterator At
= I
;
1294 // Now write the host and user/pass
1295 if (At
== SingleSlash
)
1297 if (FirstColon
< SingleSlash
)
1298 Host
.assign(FirstColon
,SingleSlash
);
1302 Host
.assign(At
+1,SingleSlash
);
1303 // username and password must be encoded (RFC 3986)
1304 User
.assign(DeQuoteString(FirstColon
,SecondColon
));
1305 if (SecondColon
< At
)
1306 Password
.assign(DeQuoteString(SecondColon
+1,At
));
1309 // Now we parse the RFC 2732 [] hostnames.
1310 unsigned long PortEnd
= 0;
1312 for (unsigned I
= 0; I
!= Host
.length();)
1321 if (InBracket
== true && Host
[I
] == ']')
1332 if (InBracket
== true)
1338 // Now we parse off a port number from the hostname
1340 string::size_type Pos
= Host
.rfind(':');
1341 if (Pos
== string::npos
|| Pos
< PortEnd
)
1344 Port
= atoi(string(Host
,Pos
+1).c_str());
1345 Host
.assign(Host
,0,Pos
);
1348 // URI::operator string - Convert the URI to a string /*{{{*/
1349 // ---------------------------------------------------------------------
1351 URI::operator string()
1355 if (Access
.empty() == false)
1358 if (Host
.empty() == false)
1360 if (Access
.empty() == false)
1363 if (User
.empty() == false)
1366 if (Password
.empty() == false)
1367 Res
+= ":" + Password
;
1371 // Add RFC 2732 escaping characters
1372 if (Access
.empty() == false &&
1373 (Host
.find('/') != string::npos
|| Host
.find(':') != string::npos
))
1374 Res
+= '[' + Host
+ ']';
1381 sprintf(S
,":%u",Port
);
1386 if (Path
.empty() == false)
1397 // URI::SiteOnly - Return the schema and site for the URI /*{{{*/
1398 // ---------------------------------------------------------------------
1400 string
URI::SiteOnly(const string
&URI
)
1410 // URI::NoUserPassword - Return the schema, site and path for the URI /*{{{*/
1411 // ---------------------------------------------------------------------
1413 string
URI::NoUserPassword(const string
&URI
)