1 // -*- mode: cpp; mode: fold -*-
3 // $Id: strutl.cc,v 1.30 1999/10/17 07:30:23 jgg Exp $
4 /* ######################################################################
6 String Util - Some usefull string functions.
8 These have been collected from here and there to do all sorts of usefull
9 things to strings. They are usefull 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 ##################################################################### */
19 #pragma implementation "apt-pkg/strutl.h"
22 #include <apt-pkg/strutl.h>
23 #include <apt-pkg/fileutl.h>
32 // strstrip - Remove white space from the front and back of a string /*{{{*/
33 // ---------------------------------------------------------------------
34 /* This is handy to use when parsing a file. It also removes \n's left
35 over from fgets and company */
36 char *_strstrip(char *String
)
38 for (;*String
!= 0 && (*String
== ' ' || *String
== '\t'); String
++);
43 char *End
= String
+ strlen(String
) - 1;
44 for (;End
!= String
- 1 && (*End
== ' ' || *End
== '\t' || *End
== '\n' ||
45 *End
== '\r'); End
--);
51 // strtabexpand - Converts tabs into 8 spaces /*{{{*/
52 // ---------------------------------------------------------------------
54 char *_strtabexpand(char *String
,size_t Len
)
56 for (char *I
= String
; I
!= I
+ Len
&& *I
!= 0; I
++)
60 if (I
+ 8 > String
+ Len
)
66 /* Assume the start of the string is 0 and find the next 8 char
72 Len
= 8 - ((String
- I
) % 8);
80 memmove(I
+ Len
,I
+ 1,strlen(I
) + 1);
81 for (char *J
= I
; J
+ Len
!= I
; *I
= ' ', I
++);
86 // ParseQuoteWord - Parse a single word out of a string /*{{{*/
87 // ---------------------------------------------------------------------
88 /* This grabs a single word, converts any % escaped characters to their
89 proper values and advances the pointer. Double quotes are understood
90 and striped out as well. This is for URI/URL parsing. */
91 bool ParseQuoteWord(const char *&String
,string
&Res
)
93 // Skip leading whitespace
94 const char *C
= String
;
95 for (;*C
!= 0 && *C
== ' '; C
++);
99 // Jump to the next word
100 for (;*C
!= 0 && isspace(*C
) == 0; C
++)
104 for (C
++;*C
!= 0 && *C
!= '"'; C
++);
110 // Now de-quote characters
113 const char *Start
= String
;
115 for (I
= Buffer
; I
< Buffer
+ sizeof(Buffer
) && Start
!= C
; I
++)
117 if (*Start
== '%' && Start
+ 2 < C
)
122 *I
= (char)strtol(Tmp
,0,16);
135 // Skip ending white space
136 for (;*C
!= 0 && isspace(*C
) != 0; C
++);
141 // ParseCWord - Parses a string like a C "" expression /*{{{*/
142 // ---------------------------------------------------------------------
143 /* This expects a series of space seperated strings enclosed in ""'s.
144 It concatenates the ""'s into a single string. */
145 bool ParseCWord(const char *String
,string
&Res
)
147 // Skip leading whitespace
148 const char *C
= String
;
149 for (;*C
!= 0 && *C
== ' '; C
++);
155 if (strlen(String
) >= sizeof(Buffer
))
162 for (C
++; *C
!= 0 && *C
!= '"'; C
++)
171 if (C
!= String
&& isspace(*C
) != 0 && isspace(C
[-1]) != 0)
173 if (isspace(*C
) == 0)
182 // QuoteString - Convert a string into quoted from /*{{{*/
183 // ---------------------------------------------------------------------
185 string
QuoteString(string Str
,const char *Bad
)
188 for (string::iterator I
= Str
.begin(); I
!= Str
.end(); I
++)
190 if (strchr(Bad
,*I
) != 0 || isprint(*I
) == 0 ||
191 *I
<= 0x20 || *I
>= 0x7F)
194 sprintf(Buf
,"%%%02x",(int)*I
);
203 // DeQuoteString - Convert a string from quoted from /*{{{*/
204 // ---------------------------------------------------------------------
205 /* This undoes QuoteString */
206 string
DeQuoteString(string Str
)
209 for (string::iterator I
= Str
.begin(); I
!= Str
.end(); I
++)
211 if (*I
== '%' && I
+ 2 < Str
.end())
217 Res
+= (char)strtol(Tmp
,0,16);
228 // SizeToStr - Convert a long into a human readable size /*{{{*/
229 // ---------------------------------------------------------------------
230 /* A max of 4 digits are shown before conversion to the next highest unit.
231 The max length of the string will be 5 chars unless the size is > 10
233 string
SizeToStr(double Size
)
242 /* bytes, KiloBytes, MegaBytes, GigaBytes, TeraBytes, PetaBytes,
243 ExaBytes, ZettaBytes, YottaBytes */
244 char Ext
[] = {'\0','k','M','G','T','P','E','Z','Y'};
248 if (ASize
< 100 && I
!= 0)
250 sprintf(S
,"%.1f%c",ASize
,Ext
[I
]);
256 sprintf(S
,"%.0f%c",ASize
,Ext
[I
]);
266 // TimeToStr - Convert the time into a string /*{{{*/
267 // ---------------------------------------------------------------------
268 /* Converts a number of seconds to a hms format */
269 string
TimeToStr(unsigned long Sec
)
277 sprintf(S
,"%lid %lih%lim%lis",Sec
/60/60/24,(Sec
/60/60) % 24,(Sec
/60) % 60,Sec
% 60);
283 sprintf(S
,"%lih%lim%lis",Sec
/60/60,(Sec
/60) % 60,Sec
% 60);
289 sprintf(S
,"%lim%lis",Sec
/60,Sec
% 60);
293 sprintf(S
,"%lis",Sec
);
300 // SubstVar - Substitute a string for another string /*{{{*/
301 // ---------------------------------------------------------------------
302 /* This replaces all occurances of Subst with Contents in Str. */
303 string
SubstVar(string Str
,string Subst
,string Contents
)
305 string::size_type Pos
= 0;
306 string::size_type OldPos
= 0;
309 while (OldPos
< Str
.length() &&
310 (Pos
= Str
.find(Subst
,OldPos
)) != string::npos
)
312 Temp
+= string(Str
,OldPos
,Pos
) + Contents
;
313 OldPos
= Pos
+ Subst
.length();
319 return Temp
+ string(Str
,OldPos
);
322 // URItoFileName - Convert the uri into a unique file name /*{{{*/
323 // ---------------------------------------------------------------------
324 /* This converts a URI into a safe filename. It quotes all unsafe characters
325 and converts / to _ and removes the scheme identifier. The resulting
326 file name should be unique and never occur again for a different file */
327 string
URItoFileName(string URI
)
329 // Nuke 'sensitive' items
332 U
.Password
= string();
335 // "\x00-\x20{}|\\\\^\\[\\]<>\"\x7F-\xFF";
336 URI
= QuoteString(U
,"\\|{}[]<>\"^~_=!@#$%^&*");
337 string::iterator J
= URI
.begin();
338 for (; J
!= URI
.end(); J
++)
344 // Base64Encode - Base64 Encoding routine for short strings /*{{{*/
345 // ---------------------------------------------------------------------
346 /* This routine performs a base64 transformation on a string. It was ripped
347 from wget and then patched and bug fixed.
349 This spec can be found in rfc2045 */
350 string
Base64Encode(string S
)
353 static char tbl
[64] = {'A','B','C','D','E','F','G','H',
354 'I','J','K','L','M','N','O','P',
355 'Q','R','S','T','U','V','W','X',
356 'Y','Z','a','b','c','d','e','f',
357 'g','h','i','j','k','l','m','n',
358 'o','p','q','r','s','t','u','v',
359 'w','x','y','z','0','1','2','3',
360 '4','5','6','7','8','9','+','/'};
362 // Pre-allocate some space
364 Final
.reserve((4*S
.length() + 2)/3 + 2);
366 /* Transform the 3x8 bits to 4x6 bits, as required by
368 for (string::const_iterator I
= S
.begin(); I
< S
.end(); I
+= 3)
370 char Bits
[3] = {0,0,0};
377 Final
+= tbl
[Bits
[0] >> 2];
378 Final
+= tbl
[((Bits
[0] & 3) << 4) + (Bits
[1] >> 4)];
380 if (I
+ 1 >= S
.end())
383 Final
+= tbl
[((Bits
[1] & 0xf) << 2) + (Bits
[2] >> 6)];
385 if (I
+ 2 >= S
.end())
388 Final
+= tbl
[Bits
[2] & 0x3f];
391 /* Apply the padding elements, this tells how many bytes the remote
392 end should discard */
393 if (S
.length() % 3 == 2)
395 if (S
.length() % 3 == 1)
401 // stringcmp - Arbitary string compare /*{{{*/
402 // ---------------------------------------------------------------------
403 /* This safely compares two non-null terminated strings of arbitary
405 int stringcmp(const char *A
,const char *AEnd
,const char *B
,const char *BEnd
)
407 for (; A
!= AEnd
&& B
!= BEnd
; A
++, B
++)
411 if (A
== AEnd
&& B
== BEnd
)
422 // stringcasecmp - Arbitary case insensitive string compare /*{{{*/
423 // ---------------------------------------------------------------------
425 int stringcasecmp(const char *A
,const char *AEnd
,const char *B
,const char *BEnd
)
427 for (; A
!= AEnd
&& B
!= BEnd
; A
++, B
++)
428 if (toupper(*A
) != toupper(*B
))
431 if (A
== AEnd
&& B
== BEnd
)
437 if (toupper(*A
) < toupper(*B
))
442 // LookupTag - Lookup the value of a tag in a taged string /*{{{*/
443 // ---------------------------------------------------------------------
444 /* The format is like those used in package files and the method
445 communication system */
446 string
LookupTag(string Message
,const char *Tag
,const char *Default
)
448 // Look for a matching tag.
449 int Length
= strlen(Tag
);
450 for (string::iterator I
= Message
.begin(); I
+ Length
< Message
.end(); I
++)
453 if (I
[Length
] == ':' && stringcasecmp(I
,I
+Length
,Tag
) == 0)
455 // Find the end of line and strip the leading/trailing spaces
458 for (; isspace(*I
) != 0 && I
< Message
.end(); I
++);
459 for (J
= I
; *J
!= '\n' && J
< Message
.end(); J
++);
460 for (; J
> I
&& isspace(J
[-1]) != 0; J
--);
462 return string(I
,J
-I
);
465 for (; *I
!= '\n' && I
< Message
.end(); I
++);
468 // Failed to find a match
474 // StringToBool - Converts a string into a boolean /*{{{*/
475 // ---------------------------------------------------------------------
476 /* This inspects the string to see if it is true or if it is false and
477 then returns the result. Several varients on true/false are checked. */
478 int StringToBool(string Text
,int Default
= -1)
481 int Res
= strtol(Text
.c_str(),&End
,0);
482 if (End
!= Text
.c_str() && Res
>= 0 && Res
<= 1)
485 // Check for positives
486 if (strcasecmp(Text
.c_str(),"no") == 0 ||
487 strcasecmp(Text
.c_str(),"false") == 0 ||
488 strcasecmp(Text
.c_str(),"without") == 0 ||
489 strcasecmp(Text
.c_str(),"off") == 0 ||
490 strcasecmp(Text
.c_str(),"disable") == 0)
493 // Check for negatives
494 if (strcasecmp(Text
.c_str(),"yes") == 0 ||
495 strcasecmp(Text
.c_str(),"true") == 0 ||
496 strcasecmp(Text
.c_str(),"with") == 0 ||
497 strcasecmp(Text
.c_str(),"on") == 0 ||
498 strcasecmp(Text
.c_str(),"enable") == 0)
504 // TimeRFC1123 - Convert a time_t into RFC1123 format /*{{{*/
505 // ---------------------------------------------------------------------
506 /* This converts a time_t into a string time representation that is
507 year 2000 complient and timezone neutral */
508 string
TimeRFC1123(time_t Date
)
510 struct tm Conv
= *gmtime(&Date
);
513 const char *Day
[] = {"Sun","Mon","Tue","Wed","Thu","Fri","Sat"};
514 const char *Month
[] = {"Jan","Feb","Mar","Apr","May","Jun","Jul",
515 "Aug","Sep","Oct","Nov","Dec"};
517 sprintf(Buf
,"%s, %02i %s %i %02i:%02i:%02i GMT",Day
[Conv
.tm_wday
],
518 Conv
.tm_mday
,Month
[Conv
.tm_mon
],Conv
.tm_year
+1900,Conv
.tm_hour
,
519 Conv
.tm_min
,Conv
.tm_sec
);
523 // ReadMessages - Read messages from the FD /*{{{*/
524 // ---------------------------------------------------------------------
525 /* This pulls full messages from the input FD into the message buffer.
526 It assumes that messages will not pause during transit so no
527 fancy buffering is used. */
528 bool ReadMessages(int Fd
, vector
<string
> &List
)
535 int Res
= read(Fd
,End
,sizeof(Buffer
) - (End
-Buffer
));
536 if (Res
< 0 && errno
== EINTR
)
539 // Process is dead, this is kind of bad..
549 // Look for the end of the message
550 for (char *I
= Buffer
; I
+ 1 < End
; I
++)
552 if (I
[0] != '\n' || I
[1] != '\n')
555 // Pull the message out
556 string
Message(Buffer
,0,I
-Buffer
);
559 for (; I
< End
&& *I
== '\n'; I
++);
561 memmove(Buffer
,I
,End
-Buffer
);
564 List
.push_back(Message
);
569 if (WaitFd(Fd
) == false)
574 // MonthConv - Converts a month string into a number /*{{{*/
575 // ---------------------------------------------------------------------
576 /* This was lifted from the boa webserver which lifted it from 'wn-v1.07'
577 Made it a bit more robust with a few touppers though. */
578 static int MonthConv(char *Month
)
580 switch (toupper(*Month
))
583 return toupper(Month
[1]) == 'P'?3:7;
589 if (toupper(Month
[1]) == 'A')
591 return toupper(Month
[2]) == 'N'?5:6;
593 return toupper(Month
[2]) == 'R'?2:4;
601 // Pretend it is January..
607 // timegm - Internal timegm function if gnu is not available /*{{{*/
608 // ---------------------------------------------------------------------
609 /* Ripped this evil little function from wget - I prefer the use of
610 GNU timegm if possible as this technique will have interesting problems
611 with leap seconds, timezones and other.
613 Converts struct tm to time_t, assuming the data in tm is UTC rather
614 than local timezone (mktime assumes the latter).
616 Contributed by Roger Beeman <beeman@cisco.com>, with the help of
617 Mark Baushke <mdb@cisco.com> and the rest of the Gurus at CISCO. */
618 #ifndef __USE_MISC // glib sets this
619 static time_t timegm(struct tm
*t
)
626 tb
= mktime (gmtime (&tl
));
627 return (tl
<= tb
? (tl
+ (tl
- tb
)) : (tl
- (tb
- tl
)));
631 // StrToTime - Converts a string into a time_t /*{{{*/
632 // ---------------------------------------------------------------------
633 /* This handles all 3 populare time formats including RFC 1123, RFC 1036
634 and the C library asctime format. It requires the GNU library function
635 'timegm' to convert a struct tm in UTC to a time_t. For some bizzar
636 reason the C library does not provide any such function :<*/
637 bool StrToTime(string Val
,time_t &Result
)
641 const char *I
= Val
.c_str();
643 // Skip the day of the week
644 for (;*I
!= 0 && *I
!= ' '; I
++);
646 // Handle RFC 1123 time
647 if (sscanf(I
," %d %3s %d %d:%d:%d GMT",&Tm
.tm_mday
,Month
,&Tm
.tm_year
,
648 &Tm
.tm_hour
,&Tm
.tm_min
,&Tm
.tm_sec
) != 6)
650 // Handle RFC 1036 time
651 if (sscanf(I
," %d-%3s-%d %d:%d:%d GMT",&Tm
.tm_mday
,Month
,
652 &Tm
.tm_year
,&Tm
.tm_hour
,&Tm
.tm_min
,&Tm
.tm_sec
) == 6)
657 if (sscanf(I
," %3s %d %d:%d:%d %d",Month
,&Tm
.tm_mday
,
658 &Tm
.tm_hour
,&Tm
.tm_min
,&Tm
.tm_sec
,&Tm
.tm_year
) != 6)
664 Tm
.tm_mon
= MonthConv(Month
);
667 // Convert to local time and then to GMT
668 Result
= timegm(&Tm
);
672 // StrToNum - Convert a fixed length string to a number /*{{{*/
673 // ---------------------------------------------------------------------
674 /* This is used in decoding the crazy fixed length string headers in
676 bool StrToNum(const char *Str
,unsigned long &Res
,unsigned Len
,unsigned Base
)
679 if (Len
>= sizeof(S
))
684 // All spaces is a zero
687 for (I
= 0; S
[I
] == ' '; I
++);
692 Res
= strtoul(S
,&End
,Base
);
699 // HexDigit - Convert a hex character into an integer /*{{{*/
700 // ---------------------------------------------------------------------
701 /* Helper for Hex2Num */
702 static int HexDigit(int c
)
704 if (c
>= '0' && c
<= '9')
706 if (c
>= 'a' && c
<= 'f')
708 if (c
>= 'A' && c
<= 'F')
713 // Hex2Num - Convert a long hex number into a buffer /*{{{*/
714 // ---------------------------------------------------------------------
715 /* The length of the buffer must be exactly 1/2 the length of the string. */
716 bool Hex2Num(const char *Start
,const char *End
,unsigned char *Num
,
719 if (End
- Start
!= (signed)(Length
*2))
722 // Convert each digit. We store it in the same order as the string
724 for (const char *I
= Start
; I
< End
;J
++, I
+= 2)
726 if (isxdigit(*I
) == 0 || isxdigit(I
[1]) == 0)
729 Num
[J
] = HexDigit(I
[0]) << 4;
730 Num
[J
] += HexDigit(I
[1]);
737 // URI::CopyFrom - Copy from an object /*{{{*/
738 // ---------------------------------------------------------------------
739 /* This parses the URI into all of its components */
740 void URI::CopyFrom(string U
)
742 string::const_iterator I
= U
.begin();
744 // Locate the first colon, this seperates the scheme
745 for (; I
< U
.end() && *I
!= ':' ; I
++);
746 string::const_iterator FirstColon
= I
;
748 /* Determine if this is a host type URI with a leading double //
749 and then search for the first single / */
750 string::const_iterator SingleSlash
= I
;
751 if (I
+ 3 < U
.end() && I
[1] == '/' && I
[2] == '/')
753 for (; SingleSlash
< U
.end() && *SingleSlash
!= '/'; SingleSlash
++);
754 if (SingleSlash
> U
.end())
755 SingleSlash
= U
.end();
757 // We can now write the access and path specifiers
758 Access
= string(U
,0,FirstColon
- U
.begin());
759 if (SingleSlash
!= U
.end())
760 Path
= string(U
,SingleSlash
- U
.begin());
761 if (Path
.empty() == true)
764 // Now we attempt to locate a user:pass@host fragment
765 if (FirstColon
[1] == '/' && FirstColon
[2] == '/')
769 if (FirstColon
>= U
.end())
772 if (FirstColon
> SingleSlash
)
773 FirstColon
= SingleSlash
;
779 for (; I
< SingleSlash
&& *I
!= ':'; I
++);
780 string::const_iterator SecondColon
= I
;
782 // Search for the @ after the colon
783 for (; I
< SingleSlash
&& *I
!= '@'; I
++);
784 string::const_iterator At
= I
;
786 // Now write the host and user/pass
787 if (At
== SingleSlash
)
789 if (FirstColon
< SingleSlash
)
790 Host
= string(U
,FirstColon
- U
.begin(),SingleSlash
- FirstColon
);
794 Host
= string(U
,At
- U
.begin() + 1,SingleSlash
- At
- 1);
795 User
= string(U
,FirstColon
- U
.begin(),SecondColon
- FirstColon
);
796 if (SecondColon
< At
)
797 Password
= string(U
,SecondColon
- U
.begin() + 1,At
- SecondColon
- 1);
800 // Now we parse off a port number from the hostname
802 string::size_type Pos
= Host
.rfind(':');
803 if (Pos
== string::npos
)
806 Port
= atoi(string(Host
,Pos
+1).c_str());
807 Host
= string(Host
,0,Pos
);
810 // URI::operator string - Convert the URI to a string /*{{{*/
811 // ---------------------------------------------------------------------
813 URI::operator string()
817 if (Access
.empty() == false)
820 if (Host
.empty() == false)
822 if (Access
.empty() == false)
825 if (User
.empty() == false)
828 if (Password
.empty() == false)
829 Res
+= ":" + Password
;
837 sprintf(S
,":%u",Port
);
842 if (Path
.empty() == false)