]> git.saurik.com Git - apt.git/blame_incremental - apt-pkg/contrib/strutl.cc
apt-key: change to / before find to satisfy its CWD needs
[apt.git] / apt-pkg / contrib / strutl.cc
... / ...
CommitLineData
1// -*- mode: cpp; mode: fold -*-
2// Description /*{{{*/
3// $Id: strutl.cc,v 1.48 2003/07/18 14:15:11 mdz Exp $
4/* ######################################################################
5
6 String Util - Some useful string functions.
7
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.
11
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>
14
15 ##################################################################### */
16 /*}}}*/
17// Includes /*{{{*/
18#include <config.h>
19
20#include <apt-pkg/strutl.h>
21#include <apt-pkg/fileutl.h>
22#include <apt-pkg/error.h>
23
24#include <algorithm>
25#include <iomanip>
26#include <locale>
27#include <sstream>
28#include <string>
29#include <vector>
30
31#include <stddef.h>
32#include <stdlib.h>
33#include <time.h>
34#include <ctype.h>
35#include <string.h>
36#include <stdio.h>
37#include <unistd.h>
38#include <regex.h>
39#include <errno.h>
40#include <stdarg.h>
41#include <iconv.h>
42
43#include <apti18n.h>
44 /*}}}*/
45using namespace std;
46
47// Strip - Remove white space from the front and back of a string /*{{{*/
48// ---------------------------------------------------------------------
49namespace APT {
50 namespace String {
51std::string Strip(const std::string &str)
52{
53 // ensure we have at least one character
54 if (str.empty() == true)
55 return str;
56
57 char const * const s = str.c_str();
58 size_t start = 0;
59 for (; isspace(s[start]) != 0; ++start)
60 ; // find the first not-space
61
62 // string contains only whitespaces
63 if (s[start] == '\0')
64 return "";
65
66 size_t end = str.length() - 1;
67 for (; isspace(s[end]) != 0; --end)
68 ; // find the last not-space
69
70 return str.substr(start, end - start + 1);
71}
72
73bool Endswith(const std::string &s, const std::string &end)
74{
75 if (end.size() > s.size())
76 return false;
77 return (s.compare(s.size() - end.size(), end.size(), end) == 0);
78}
79
80bool Startswith(const std::string &s, const std::string &start)
81{
82 if (start.size() > s.size())
83 return false;
84 return (s.compare(0, start.size(), start) == 0);
85}
86
87}
88}
89 /*}}}*/
90// UTF8ToCodeset - Convert some UTF-8 string for some codeset /*{{{*/
91// ---------------------------------------------------------------------
92/* This is handy to use before display some information for enduser */
93bool UTF8ToCodeset(const char *codeset, const string &orig, string *dest)
94{
95 iconv_t cd;
96 const char *inbuf;
97 char *inptr, *outbuf;
98 size_t insize, bufsize;
99 dest->clear();
100
101 cd = iconv_open(codeset, "UTF-8");
102 if (cd == (iconv_t)(-1)) {
103 // Something went wrong
104 if (errno == EINVAL)
105 _error->Error("conversion from 'UTF-8' to '%s' not available",
106 codeset);
107 else
108 perror("iconv_open");
109
110 return false;
111 }
112
113 insize = bufsize = orig.size();
114 inbuf = orig.data();
115 inptr = (char *)inbuf;
116 outbuf = new char[bufsize];
117 size_t lastError = -1;
118
119 while (insize != 0)
120 {
121 char *outptr = outbuf;
122 size_t outsize = bufsize;
123 size_t const err = iconv(cd, &inptr, &insize, &outptr, &outsize);
124 dest->append(outbuf, outptr - outbuf);
125 if (err == (size_t)(-1))
126 {
127 switch (errno)
128 {
129 case EILSEQ:
130 insize--;
131 inptr++;
132 // replace a series of unknown multibytes with a single "?"
133 if (lastError != insize) {
134 lastError = insize - 1;
135 dest->append("?");
136 }
137 break;
138 case EINVAL:
139 insize = 0;
140 break;
141 case E2BIG:
142 if (outptr == outbuf)
143 {
144 bufsize *= 2;
145 delete[] outbuf;
146 outbuf = new char[bufsize];
147 }
148 break;
149 }
150 }
151 }
152
153 delete[] outbuf;
154
155 iconv_close(cd);
156
157 return true;
158}
159 /*}}}*/
160// strstrip - Remove white space from the front and back of a string /*{{{*/
161// ---------------------------------------------------------------------
162/* This is handy to use when parsing a file. It also removes \n's left
163 over from fgets and company */
164char *_strstrip(char *String)
165{
166 for (;*String != 0 && (*String == ' ' || *String == '\t'); String++);
167
168 if (*String == 0)
169 return String;
170 return _strrstrip(String);
171}
172 /*}}}*/
173// strrstrip - Remove white space from the back of a string /*{{{*/
174// ---------------------------------------------------------------------
175char *_strrstrip(char *String)
176{
177 char *End = String + strlen(String) - 1;
178 for (;End != String - 1 && (*End == ' ' || *End == '\t' || *End == '\n' ||
179 *End == '\r'); End--);
180 End++;
181 *End = 0;
182 return String;
183}
184 /*}}}*/
185// strtabexpand - Converts tabs into 8 spaces /*{{{*/
186// ---------------------------------------------------------------------
187/* */
188char *_strtabexpand(char *String,size_t Len)
189{
190 for (char *I = String; I != I + Len && *I != 0; I++)
191 {
192 if (*I != '\t')
193 continue;
194 if (I + 8 > String + Len)
195 {
196 *I = 0;
197 return String;
198 }
199
200 /* Assume the start of the string is 0 and find the next 8 char
201 division */
202 int Len;
203 if (String == I)
204 Len = 1;
205 else
206 Len = 8 - ((String - I) % 8);
207 Len -= 2;
208 if (Len <= 0)
209 {
210 *I = ' ';
211 continue;
212 }
213
214 memmove(I + Len,I + 1,strlen(I) + 1);
215 for (char *J = I; J + Len != I; *I = ' ', I++);
216 }
217 return String;
218}
219 /*}}}*/
220// ParseQuoteWord - Parse a single word out of a string /*{{{*/
221// ---------------------------------------------------------------------
222/* This grabs a single word, converts any % escaped characters to their
223 proper values and advances the pointer. Double quotes are understood
224 and striped out as well. This is for URI/URL parsing. It also can
225 understand [] brackets.*/
226bool ParseQuoteWord(const char *&String,string &Res)
227{
228 // Skip leading whitespace
229 const char *C = String;
230 for (;*C != 0 && *C == ' '; C++);
231 if (*C == 0)
232 return false;
233
234 // Jump to the next word
235 for (;*C != 0 && isspace(*C) == 0; C++)
236 {
237 if (*C == '"')
238 {
239 C = strchr(C + 1, '"');
240 if (C == NULL)
241 return false;
242 }
243 if (*C == '[')
244 {
245 C = strchr(C + 1, ']');
246 if (C == NULL)
247 return false;
248 }
249 }
250
251 // Now de-quote characters
252 char Buffer[1024];
253 char Tmp[3];
254 const char *Start = String;
255 char *I;
256 for (I = Buffer; I < Buffer + sizeof(Buffer) && Start != C; I++)
257 {
258 if (*Start == '%' && Start + 2 < C &&
259 isxdigit(Start[1]) && isxdigit(Start[2]))
260 {
261 Tmp[0] = Start[1];
262 Tmp[1] = Start[2];
263 Tmp[2] = 0;
264 *I = (char)strtol(Tmp,0,16);
265 Start += 3;
266 continue;
267 }
268 if (*Start != '"')
269 *I = *Start;
270 else
271 I--;
272 Start++;
273 }
274 *I = 0;
275 Res = Buffer;
276
277 // Skip ending white space
278 for (;*C != 0 && isspace(*C) != 0; C++);
279 String = C;
280 return true;
281}
282 /*}}}*/
283// ParseCWord - Parses a string like a C "" expression /*{{{*/
284// ---------------------------------------------------------------------
285/* This expects a series of space separated strings enclosed in ""'s.
286 It concatenates the ""'s into a single string. */
287bool ParseCWord(const char *&String,string &Res)
288{
289 // Skip leading whitespace
290 const char *C = String;
291 for (;*C != 0 && *C == ' '; C++);
292 if (*C == 0)
293 return false;
294
295 char Buffer[1024];
296 char *Buf = Buffer;
297 if (strlen(String) >= sizeof(Buffer))
298 return false;
299
300 for (; *C != 0; C++)
301 {
302 if (*C == '"')
303 {
304 for (C++; *C != 0 && *C != '"'; C++)
305 *Buf++ = *C;
306
307 if (*C == 0)
308 return false;
309
310 continue;
311 }
312
313 if (C != String && isspace(*C) != 0 && isspace(C[-1]) != 0)
314 continue;
315 if (isspace(*C) == 0)
316 return false;
317 *Buf++ = ' ';
318 }
319 *Buf = 0;
320 Res = Buffer;
321 String = C;
322 return true;
323}
324 /*}}}*/
325// QuoteString - Convert a string into quoted from /*{{{*/
326// ---------------------------------------------------------------------
327/* */
328string QuoteString(const string &Str, const char *Bad)
329{
330 std::stringstream Res;
331 for (string::const_iterator I = Str.begin(); I != Str.end(); ++I)
332 {
333 if (strchr(Bad,*I) != 0 || isprint(*I) == 0 ||
334 *I == 0x25 || // percent '%' char
335 *I <= 0x20 || *I >= 0x7F) // control chars
336 {
337 ioprintf(Res, "%%%02hhx", *I);
338 }
339 else
340 Res << *I;
341 }
342 return Res.str();
343}
344 /*}}}*/
345// DeQuoteString - Convert a string from quoted from /*{{{*/
346// ---------------------------------------------------------------------
347/* This undoes QuoteString */
348string DeQuoteString(const string &Str)
349{
350 return DeQuoteString(Str.begin(),Str.end());
351}
352string DeQuoteString(string::const_iterator const &begin,
353 string::const_iterator const &end)
354{
355 string Res;
356 for (string::const_iterator I = begin; I != end; ++I)
357 {
358 if (*I == '%' && I + 2 < end &&
359 isxdigit(I[1]) && isxdigit(I[2]))
360 {
361 char Tmp[3];
362 Tmp[0] = I[1];
363 Tmp[1] = I[2];
364 Tmp[2] = 0;
365 Res += (char)strtol(Tmp,0,16);
366 I += 2;
367 continue;
368 }
369 else
370 Res += *I;
371 }
372 return Res;
373}
374
375 /*}}}*/
376// SizeToStr - Convert a long into a human readable size /*{{{*/
377// ---------------------------------------------------------------------
378/* A max of 4 digits are shown before conversion to the next highest unit.
379 The max length of the string will be 5 chars unless the size is > 10
380 YottaBytes (E24) */
381string SizeToStr(double Size)
382{
383 double ASize;
384 if (Size >= 0)
385 ASize = Size;
386 else
387 ASize = -1*Size;
388
389 /* bytes, KiloBytes, MegaBytes, GigaBytes, TeraBytes, PetaBytes,
390 ExaBytes, ZettaBytes, YottaBytes */
391 char Ext[] = {'\0','k','M','G','T','P','E','Z','Y'};
392 int I = 0;
393 while (I <= 8)
394 {
395 if (ASize < 100 && I != 0)
396 {
397 std::string S;
398 strprintf(S, "%'.1f %c", ASize, Ext[I]);
399 return S;
400 }
401
402 if (ASize < 10000)
403 {
404 std::string S;
405 strprintf(S, "%'.0f %c", ASize, Ext[I]);
406 return S;
407 }
408 ASize /= 1000.0;
409 I++;
410 }
411 return "";
412}
413 /*}}}*/
414// TimeToStr - Convert the time into a string /*{{{*/
415// ---------------------------------------------------------------------
416/* Converts a number of seconds to a hms format */
417string TimeToStr(unsigned long Sec)
418{
419 std::string S;
420 if (Sec > 60*60*24)
421 {
422 //TRANSLATOR: d means days, h means hours, min means minutes, s means seconds
423 strprintf(S,_("%lid %lih %limin %lis"),Sec/60/60/24,(Sec/60/60) % 24,(Sec/60) % 60,Sec % 60);
424 }
425 else if (Sec > 60*60)
426 {
427 //TRANSLATOR: h means hours, min means minutes, s means seconds
428 strprintf(S,_("%lih %limin %lis"),Sec/60/60,(Sec/60) % 60,Sec % 60);
429 }
430 else if (Sec > 60)
431 {
432 //TRANSLATOR: min means minutes, s means seconds
433 strprintf(S,_("%limin %lis"),Sec/60,Sec % 60);
434 }
435 else
436 {
437 //TRANSLATOR: s means seconds
438 strprintf(S,_("%lis"),Sec);
439 }
440 return S;
441}
442 /*}}}*/
443// SubstVar - Substitute a string for another string /*{{{*/
444// ---------------------------------------------------------------------
445/* This replaces all occurrences of Subst with Contents in Str. */
446string SubstVar(const string &Str,const string &Subst,const string &Contents)
447{
448 if (Subst.empty() == true)
449 return Str;
450
451 string::size_type Pos = 0;
452 string::size_type OldPos = 0;
453 string Temp;
454
455 while (OldPos < Str.length() &&
456 (Pos = Str.find(Subst,OldPos)) != string::npos)
457 {
458 if (OldPos != Pos)
459 Temp.append(Str, OldPos, Pos - OldPos);
460 if (Contents.empty() == false)
461 Temp.append(Contents);
462 OldPos = Pos + Subst.length();
463 }
464
465 if (OldPos == 0)
466 return Str;
467
468 if (OldPos >= Str.length())
469 return Temp;
470
471 Temp.append(Str, OldPos, string::npos);
472 return Temp;
473}
474string SubstVar(string Str,const struct SubstVar *Vars)
475{
476 for (; Vars->Subst != 0; Vars++)
477 Str = SubstVar(Str,Vars->Subst,*Vars->Contents);
478 return Str;
479}
480 /*}}}*/
481// OutputInDepth - return a string with separator multiplied with depth /*{{{*/
482// ---------------------------------------------------------------------
483/* Returns a string with the supplied separator depth + 1 times in it */
484std::string OutputInDepth(const unsigned long Depth, const char* Separator)
485{
486 std::string output = "";
487 for(unsigned long d=Depth+1; d > 0; d--)
488 output.append(Separator);
489 return output;
490}
491 /*}}}*/
492// URItoFileName - Convert the uri into a unique file name /*{{{*/
493// ---------------------------------------------------------------------
494/* This converts a URI into a safe filename. It quotes all unsafe characters
495 and converts / to _ and removes the scheme identifier. The resulting
496 file name should be unique and never occur again for a different file */
497string URItoFileName(const string &URI)
498{
499 // Nuke 'sensitive' items
500 ::URI U(URI);
501 U.User.clear();
502 U.Password.clear();
503 U.Access.clear();
504
505 // "\x00-\x20{}|\\\\^\\[\\]<>\"\x7F-\xFF";
506 string NewURI = QuoteString(U,"\\|{}[]<>\"^~_=!@#$%^&*");
507 replace(NewURI.begin(),NewURI.end(),'/','_');
508 return NewURI;
509}
510 /*}}}*/
511// Base64Encode - Base64 Encoding routine for short strings /*{{{*/
512// ---------------------------------------------------------------------
513/* This routine performs a base64 transformation on a string. It was ripped
514 from wget and then patched and bug fixed.
515
516 This spec can be found in rfc2045 */
517string Base64Encode(const string &S)
518{
519 // Conversion table.
520 static char tbl[64] = {'A','B','C','D','E','F','G','H',
521 'I','J','K','L','M','N','O','P',
522 'Q','R','S','T','U','V','W','X',
523 'Y','Z','a','b','c','d','e','f',
524 'g','h','i','j','k','l','m','n',
525 'o','p','q','r','s','t','u','v',
526 'w','x','y','z','0','1','2','3',
527 '4','5','6','7','8','9','+','/'};
528
529 // Pre-allocate some space
530 string Final;
531 Final.reserve((4*S.length() + 2)/3 + 2);
532
533 /* Transform the 3x8 bits to 4x6 bits, as required by
534 base64. */
535 for (string::const_iterator I = S.begin(); I < S.end(); I += 3)
536 {
537 char Bits[3] = {0,0,0};
538 Bits[0] = I[0];
539 if (I + 1 < S.end())
540 Bits[1] = I[1];
541 if (I + 2 < S.end())
542 Bits[2] = I[2];
543
544 Final += tbl[Bits[0] >> 2];
545 Final += tbl[((Bits[0] & 3) << 4) + (Bits[1] >> 4)];
546
547 if (I + 1 >= S.end())
548 break;
549
550 Final += tbl[((Bits[1] & 0xf) << 2) + (Bits[2] >> 6)];
551
552 if (I + 2 >= S.end())
553 break;
554
555 Final += tbl[Bits[2] & 0x3f];
556 }
557
558 /* Apply the padding elements, this tells how many bytes the remote
559 end should discard */
560 if (S.length() % 3 == 2)
561 Final += '=';
562 if (S.length() % 3 == 1)
563 Final += "==";
564
565 return Final;
566}
567 /*}}}*/
568// stringcmp - Arbitrary string compare /*{{{*/
569// ---------------------------------------------------------------------
570/* This safely compares two non-null terminated strings of arbitrary
571 length */
572int stringcmp(const char *A,const char *AEnd,const char *B,const char *BEnd)
573{
574 for (; A != AEnd && B != BEnd; A++, B++)
575 if (*A != *B)
576 break;
577
578 if (A == AEnd && B == BEnd)
579 return 0;
580 if (A == AEnd)
581 return 1;
582 if (B == BEnd)
583 return -1;
584 if (*A < *B)
585 return -1;
586 return 1;
587}
588
589#if __GNUC__ >= 3
590int stringcmp(string::const_iterator A,string::const_iterator AEnd,
591 const char *B,const char *BEnd)
592{
593 for (; A != AEnd && B != BEnd; A++, B++)
594 if (*A != *B)
595 break;
596
597 if (A == AEnd && B == BEnd)
598 return 0;
599 if (A == AEnd)
600 return 1;
601 if (B == BEnd)
602 return -1;
603 if (*A < *B)
604 return -1;
605 return 1;
606}
607int stringcmp(string::const_iterator A,string::const_iterator AEnd,
608 string::const_iterator B,string::const_iterator BEnd)
609{
610 for (; A != AEnd && B != BEnd; A++, B++)
611 if (*A != *B)
612 break;
613
614 if (A == AEnd && B == BEnd)
615 return 0;
616 if (A == AEnd)
617 return 1;
618 if (B == BEnd)
619 return -1;
620 if (*A < *B)
621 return -1;
622 return 1;
623}
624#endif
625 /*}}}*/
626// stringcasecmp - Arbitrary case insensitive string compare /*{{{*/
627// ---------------------------------------------------------------------
628/* */
629int stringcasecmp(const char *A,const char *AEnd,const char *B,const char *BEnd)
630{
631 for (; A != AEnd && B != BEnd; A++, B++)
632 if (tolower_ascii(*A) != tolower_ascii(*B))
633 break;
634
635 if (A == AEnd && B == BEnd)
636 return 0;
637 if (A == AEnd)
638 return 1;
639 if (B == BEnd)
640 return -1;
641 if (tolower_ascii(*A) < tolower_ascii(*B))
642 return -1;
643 return 1;
644}
645#if __GNUC__ >= 3
646int stringcasecmp(string::const_iterator A,string::const_iterator AEnd,
647 const char *B,const char *BEnd)
648{
649 for (; A != AEnd && B != BEnd; A++, B++)
650 if (tolower_ascii(*A) != tolower_ascii(*B))
651 break;
652
653 if (A == AEnd && B == BEnd)
654 return 0;
655 if (A == AEnd)
656 return 1;
657 if (B == BEnd)
658 return -1;
659 if (tolower_ascii(*A) < tolower_ascii(*B))
660 return -1;
661 return 1;
662}
663int stringcasecmp(string::const_iterator A,string::const_iterator AEnd,
664 string::const_iterator B,string::const_iterator BEnd)
665{
666 for (; A != AEnd && B != BEnd; A++, B++)
667 if (tolower_ascii(*A) != tolower_ascii(*B))
668 break;
669
670 if (A == AEnd && B == BEnd)
671 return 0;
672 if (A == AEnd)
673 return 1;
674 if (B == BEnd)
675 return -1;
676 if (tolower_ascii(*A) < tolower_ascii(*B))
677 return -1;
678 return 1;
679}
680#endif
681 /*}}}*/
682// LookupTag - Lookup the value of a tag in a taged string /*{{{*/
683// ---------------------------------------------------------------------
684/* The format is like those used in package files and the method
685 communication system */
686string LookupTag(const string &Message,const char *Tag,const char *Default)
687{
688 // Look for a matching tag.
689 int Length = strlen(Tag);
690 for (string::const_iterator I = Message.begin(); I + Length < Message.end(); ++I)
691 {
692 // Found the tag
693 if (I[Length] == ':' && stringcasecmp(I,I+Length,Tag) == 0)
694 {
695 // Find the end of line and strip the leading/trailing spaces
696 string::const_iterator J;
697 I += Length + 1;
698 for (; isspace_ascii(*I) != 0 && I < Message.end(); ++I);
699 for (J = I; *J != '\n' && J < Message.end(); ++J);
700 for (; J > I && isspace_ascii(J[-1]) != 0; --J);
701
702 return string(I,J);
703 }
704
705 for (; *I != '\n' && I < Message.end(); ++I);
706 }
707
708 // Failed to find a match
709 if (Default == 0)
710 return string();
711 return Default;
712}
713 /*}}}*/
714// StringToBool - Converts a string into a boolean /*{{{*/
715// ---------------------------------------------------------------------
716/* This inspects the string to see if it is true or if it is false and
717 then returns the result. Several varients on true/false are checked. */
718int StringToBool(const string &Text,int Default)
719{
720 char *ParseEnd;
721 int Res = strtol(Text.c_str(),&ParseEnd,0);
722 // ensure that the entire string was converted by strtol to avoid
723 // failures on "apt-cache show -a 0ad" where the "0" is converted
724 const char *TextEnd = Text.c_str()+Text.size();
725 if (ParseEnd == TextEnd && Res >= 0 && Res <= 1)
726 return Res;
727
728 // Check for positives
729 if (strcasecmp(Text.c_str(),"no") == 0 ||
730 strcasecmp(Text.c_str(),"false") == 0 ||
731 strcasecmp(Text.c_str(),"without") == 0 ||
732 strcasecmp(Text.c_str(),"off") == 0 ||
733 strcasecmp(Text.c_str(),"disable") == 0)
734 return 0;
735
736 // Check for negatives
737 if (strcasecmp(Text.c_str(),"yes") == 0 ||
738 strcasecmp(Text.c_str(),"true") == 0 ||
739 strcasecmp(Text.c_str(),"with") == 0 ||
740 strcasecmp(Text.c_str(),"on") == 0 ||
741 strcasecmp(Text.c_str(),"enable") == 0)
742 return 1;
743
744 return Default;
745}
746 /*}}}*/
747// TimeRFC1123 - Convert a time_t into RFC1123 format /*{{{*/
748// ---------------------------------------------------------------------
749/* This converts a time_t into a string time representation that is
750 year 2000 complient and timezone neutral */
751string TimeRFC1123(time_t Date)
752{
753 struct tm Conv;
754 if (gmtime_r(&Date, &Conv) == NULL)
755 return "";
756
757 auto const posix = std::locale("C.UTF-8");
758 std::ostringstream datestr;
759 datestr.imbue(posix);
760 datestr << std::put_time(&Conv, "%a, %d %b %Y %H:%M:%S GMT");
761 return datestr.str();
762}
763 /*}}}*/
764// ReadMessages - Read messages from the FD /*{{{*/
765// ---------------------------------------------------------------------
766/* This pulls full messages from the input FD into the message buffer.
767 It assumes that messages will not pause during transit so no
768 fancy buffering is used.
769
770 In particular: this reads blocks from the input until it believes
771 that it's run out of input text. Each block is terminated by a
772 double newline ('\n' followed by '\n').
773 */
774bool ReadMessages(int Fd, vector<string> &List)
775{
776 char Buffer[64000];
777 // Represents any left-over from the previous iteration of the
778 // parse loop. (i.e., if a message is split across the end
779 // of the buffer, it goes here)
780 string PartialMessage;
781
782 do {
783 int const Res = read(Fd, Buffer, sizeof(Buffer));
784 if (Res < 0 && errno == EINTR)
785 continue;
786
787 // process we read from has died
788 if (Res == 0)
789 return false;
790
791 // No data
792#if EAGAIN != EWOULDBLOCK
793 if (Res < 0 && (errno == EAGAIN || errno == EWOULDBLOCK))
794#else
795 if (Res < 0 && errno == EAGAIN)
796#endif
797 return true;
798 if (Res < 0)
799 return false;
800
801 // extract the message(s) from the buffer
802 char const *Start = Buffer;
803 char const * const End = Buffer + Res;
804
805 char const * NL = (char const *) memchr(Start, '\n', End - Start);
806 if (NL == NULL)
807 {
808 // end of buffer: store what we have so far and read new data in
809 PartialMessage.append(Start, End - Start);
810 Start = End;
811 }
812 else
813 ++NL;
814
815 if (PartialMessage.empty() == false && Start < End)
816 {
817 // if we start with a new line, see if the partial message we have ended with one
818 // so that we properly detect records ending between two read() runs
819 // cases are: \n|\n , \r\n|\r\n and \r\n\r|\n
820 // the case \r|\n\r\n is handled by the usual double-newline handling
821 if ((NL - Start) == 1 || ((NL - Start) == 2 && *Start == '\r'))
822 {
823 if (APT::String::Endswith(PartialMessage, "\n") || APT::String::Endswith(PartialMessage, "\r\n\r"))
824 {
825 PartialMessage.erase(PartialMessage.find_last_not_of("\r\n") + 1);
826 List.push_back(PartialMessage);
827 PartialMessage.clear();
828 while (NL < End && (*NL == '\n' || *NL == '\r')) ++NL;
829 Start = NL;
830 }
831 }
832 }
833
834 while (Start < End) {
835 char const * NL2 = (char const *) memchr(NL, '\n', End - NL);
836 if (NL2 == NULL)
837 {
838 // end of buffer: store what we have so far and read new data in
839 PartialMessage.append(Start, End - Start);
840 break;
841 }
842 ++NL2;
843
844 // did we find a double newline?
845 if ((NL2 - NL) == 1 || ((NL2 - NL) == 2 && *NL == '\r'))
846 {
847 PartialMessage.append(Start, NL2 - Start);
848 PartialMessage.erase(PartialMessage.find_last_not_of("\r\n") + 1);
849 List.push_back(PartialMessage);
850 PartialMessage.clear();
851 while (NL2 < End && (*NL2 == '\n' || *NL2 == '\r')) ++NL2;
852 Start = NL2;
853 }
854 NL = NL2;
855 }
856
857 // we have read at least one complete message and nothing left
858 if (PartialMessage.empty() == true)
859 return true;
860
861 if (WaitFd(Fd) == false)
862 return false;
863 } while (true);
864}
865 /*}}}*/
866// MonthConv - Converts a month string into a number /*{{{*/
867// ---------------------------------------------------------------------
868/* This was lifted from the boa webserver which lifted it from 'wn-v1.07'
869 Made it a bit more robust with a few tolower_ascii though. */
870static int MonthConv(char *Month)
871{
872 switch (tolower_ascii(*Month))
873 {
874 case 'a':
875 return tolower_ascii(Month[1]) == 'p'?3:7;
876 case 'd':
877 return 11;
878 case 'f':
879 return 1;
880 case 'j':
881 if (tolower_ascii(Month[1]) == 'a')
882 return 0;
883 return tolower_ascii(Month[2]) == 'n'?5:6;
884 case 'm':
885 return tolower_ascii(Month[2]) == 'r'?2:4;
886 case 'n':
887 return 10;
888 case 'o':
889 return 9;
890 case 's':
891 return 8;
892
893 // Pretend it is January..
894 default:
895 return 0;
896 }
897}
898 /*}}}*/
899// timegm - Internal timegm if the gnu version is not available /*{{{*/
900// ---------------------------------------------------------------------
901/* Converts struct tm to time_t, assuming the data in tm is UTC rather
902 than local timezone (mktime assumes the latter).
903
904 This function is a nonstandard GNU extension that is also present on
905 the BSDs and maybe other systems. For others we follow the advice of
906 the manpage of timegm and use his portable replacement. */
907#ifndef HAVE_TIMEGM
908static time_t timegm(struct tm *t)
909{
910 char *tz = getenv("TZ");
911 setenv("TZ", "", 1);
912 tzset();
913 time_t ret = mktime(t);
914 if (tz)
915 setenv("TZ", tz, 1);
916 else
917 unsetenv("TZ");
918 tzset();
919 return ret;
920}
921#endif
922 /*}}}*/
923// RFC1123StrToTime - Converts a HTTP1.1 full date strings into a time_t /*{{{*/
924// ---------------------------------------------------------------------
925/* tries to parses a full date as specified in RFC7231 §7.1.1.1
926 with one exception: HTTP/1.1 valid dates need to have GMT as timezone.
927 As we encounter dates from UTC or with a numeric timezone in other places,
928 we allow them here to to be able to reuse the method. Either way, a date
929 must be in UTC or parsing will fail. Previous implementations of this
930 method used to ignore the timezone and assume always UTC. */
931bool RFC1123StrToTime(const char* const str,time_t &time)
932{
933 struct tm t;
934 auto const &posix = std::locale("C.UTF-8");
935 auto const parse_time = [&](char const * const s, bool const has_timezone) {
936 std::istringstream ss(str);
937 ss.imbue(posix);
938 ss >> std::get_time(&t, s);
939 if (has_timezone && ss.fail() == false)
940 {
941 std::string timezone;
942 ss >> timezone;
943 if (timezone.empty())
944 return false;
945 if (timezone != "GMT" && timezone != "UTC" && timezone != "Z") // RFC 822
946 {
947 // numeric timezones as a should of RFC 1123 and generally preferred
948 try {
949 size_t pos;
950 auto const zone = std::stoi(timezone, &pos);
951 if (zone != 0 || pos != timezone.length())
952 return false;
953 } catch (...) {
954 return false;
955 }
956 }
957 }
958 t.tm_isdst = 0;
959 return ss.fail() == false;
960 };
961
962 bool const good =
963 // Sun, 06 Nov 1994 08:49:37 GMT ; RFC 822, updated by RFC 1123
964 parse_time("%a, %d %b %Y %H:%M:%S", true) ||
965 // Sunday, 06-Nov-94 08:49:37 GMT ; RFC 850, obsoleted by RFC 1036
966 parse_time("%A, %d-%b-%y %H:%M:%S", true) ||
967 // Sun Nov 6 08:49:37 1994 ; ANSI C's asctime() format
968 parse_time("%c", false); // "%a %b %d %H:%M:%S %Y"
969 if (good == false)
970 return false;
971
972 time = timegm(&t);
973 return true;
974}
975 /*}}}*/
976// FTPMDTMStrToTime - Converts a ftp modification date into a time_t /*{{{*/
977// ---------------------------------------------------------------------
978/* */
979bool FTPMDTMStrToTime(const char* const str,time_t &time)
980{
981 struct tm Tm;
982 // MDTM includes no whitespaces but recommend and ignored by strptime
983 if (strptime(str, "%Y %m %d %H %M %S", &Tm) == NULL)
984 return false;
985
986 time = timegm(&Tm);
987 return true;
988}
989 /*}}}*/
990// StrToTime - Converts a string into a time_t /*{{{*/
991// ---------------------------------------------------------------------
992/* This handles all 3 popular time formats including RFC 1123, RFC 1036
993 and the C library asctime format. It requires the GNU library function
994 'timegm' to convert a struct tm in UTC to a time_t. For some bizzar
995 reason the C library does not provide any such function :< This also
996 handles the weird, but unambiguous FTP time format*/
997bool StrToTime(const string &Val,time_t &Result)
998{
999 struct tm Tm;
1000 char Month[10];
1001
1002 // Skip the day of the week
1003 const char *I = strchr(Val.c_str(), ' ');
1004
1005 // Handle RFC 1123 time
1006 Month[0] = 0;
1007 if (sscanf(I," %2d %3s %4d %2d:%2d:%2d GMT",&Tm.tm_mday,Month,&Tm.tm_year,
1008 &Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) != 6)
1009 {
1010 // Handle RFC 1036 time
1011 if (sscanf(I," %2d-%3s-%3d %2d:%2d:%2d GMT",&Tm.tm_mday,Month,
1012 &Tm.tm_year,&Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) == 6)
1013 Tm.tm_year += 1900;
1014 else
1015 {
1016 // asctime format
1017 if (sscanf(I," %3s %2d %2d:%2d:%2d %4d",Month,&Tm.tm_mday,
1018 &Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec,&Tm.tm_year) != 6)
1019 {
1020 // 'ftp' time
1021 if (sscanf(Val.c_str(),"%4d%2d%2d%2d%2d%2d",&Tm.tm_year,&Tm.tm_mon,
1022 &Tm.tm_mday,&Tm.tm_hour,&Tm.tm_min,&Tm.tm_sec) != 6)
1023 return false;
1024 Tm.tm_mon--;
1025 }
1026 }
1027 }
1028
1029 Tm.tm_isdst = 0;
1030 if (Month[0] != 0)
1031 Tm.tm_mon = MonthConv(Month);
1032 else
1033 Tm.tm_mon = 0; // we don't have a month, so pick something
1034 Tm.tm_year -= 1900;
1035
1036 // Convert to local time and then to GMT
1037 Result = timegm(&Tm);
1038 return true;
1039}
1040 /*}}}*/
1041// StrToNum - Convert a fixed length string to a number /*{{{*/
1042// ---------------------------------------------------------------------
1043/* This is used in decoding the crazy fixed length string headers in
1044 tar and ar files. */
1045bool StrToNum(const char *Str,unsigned long &Res,unsigned Len,unsigned Base)
1046{
1047 char S[30];
1048 if (Len >= sizeof(S))
1049 return false;
1050 memcpy(S,Str,Len);
1051 S[Len] = 0;
1052
1053 // All spaces is a zero
1054 Res = 0;
1055 unsigned I;
1056 for (I = 0; S[I] == ' '; I++);
1057 if (S[I] == 0)
1058 return true;
1059
1060 char *End;
1061 Res = strtoul(S,&End,Base);
1062 if (End == S)
1063 return false;
1064
1065 return true;
1066}
1067 /*}}}*/
1068// StrToNum - Convert a fixed length string to a number /*{{{*/
1069// ---------------------------------------------------------------------
1070/* This is used in decoding the crazy fixed length string headers in
1071 tar and ar files. */
1072bool StrToNum(const char *Str,unsigned long long &Res,unsigned Len,unsigned Base)
1073{
1074 char S[30];
1075 if (Len >= sizeof(S))
1076 return false;
1077 memcpy(S,Str,Len);
1078 S[Len] = 0;
1079
1080 // All spaces is a zero
1081 Res = 0;
1082 unsigned I;
1083 for (I = 0; S[I] == ' '; I++);
1084 if (S[I] == 0)
1085 return true;
1086
1087 char *End;
1088 Res = strtoull(S,&End,Base);
1089 if (End == S)
1090 return false;
1091
1092 return true;
1093}
1094 /*}}}*/
1095
1096// Base256ToNum - Convert a fixed length binary to a number /*{{{*/
1097// ---------------------------------------------------------------------
1098/* This is used in decoding the 256bit encoded fixed length fields in
1099 tar files */
1100bool Base256ToNum(const char *Str,unsigned long long &Res,unsigned int Len)
1101{
1102 if ((Str[0] & 0x80) == 0)
1103 return false;
1104 else
1105 {
1106 Res = Str[0] & 0x7F;
1107 for(unsigned int i = 1; i < Len; ++i)
1108 Res = (Res<<8) + Str[i];
1109 return true;
1110 }
1111}
1112 /*}}}*/
1113// Base256ToNum - Convert a fixed length binary to a number /*{{{*/
1114// ---------------------------------------------------------------------
1115/* This is used in decoding the 256bit encoded fixed length fields in
1116 tar files */
1117bool Base256ToNum(const char *Str,unsigned long &Res,unsigned int Len)
1118{
1119 unsigned long long Num;
1120 bool rc;
1121
1122 rc = Base256ToNum(Str, Num, Len);
1123 Res = Num;
1124 if (Res != Num)
1125 return false;
1126
1127 return rc;
1128}
1129 /*}}}*/
1130// HexDigit - Convert a hex character into an integer /*{{{*/
1131// ---------------------------------------------------------------------
1132/* Helper for Hex2Num */
1133static int HexDigit(int c)
1134{
1135 if (c >= '0' && c <= '9')
1136 return c - '0';
1137 if (c >= 'a' && c <= 'f')
1138 return c - 'a' + 10;
1139 if (c >= 'A' && c <= 'F')
1140 return c - 'A' + 10;
1141 return -1;
1142}
1143 /*}}}*/
1144// Hex2Num - Convert a long hex number into a buffer /*{{{*/
1145// ---------------------------------------------------------------------
1146/* The length of the buffer must be exactly 1/2 the length of the string. */
1147bool Hex2Num(const string &Str,unsigned char *Num,unsigned int Length)
1148{
1149 return Hex2Num(APT::StringView(Str), Num, Length);
1150}
1151
1152bool Hex2Num(const APT::StringView Str,unsigned char *Num,unsigned int Length)
1153{
1154 if (Str.length() != Length*2)
1155 return false;
1156
1157 // Convert each digit. We store it in the same order as the string
1158 int J = 0;
1159 for (auto I = Str.begin(); I != Str.end();J++, I += 2)
1160 {
1161 int first_half = HexDigit(I[0]);
1162 int second_half;
1163 if (first_half < 0)
1164 return false;
1165
1166 second_half = HexDigit(I[1]);
1167 if (second_half < 0)
1168 return false;
1169 Num[J] = first_half << 4;
1170 Num[J] += second_half;
1171 }
1172
1173 return true;
1174}
1175 /*}}}*/
1176// TokSplitString - Split a string up by a given token /*{{{*/
1177// ---------------------------------------------------------------------
1178/* This is intended to be a faster splitter, it does not use dynamic
1179 memories. Input is changed to insert nulls at each token location. */
1180bool TokSplitString(char Tok,char *Input,char **List,
1181 unsigned long ListMax)
1182{
1183 // Strip any leading spaces
1184 char *Start = Input;
1185 char *Stop = Start + strlen(Start);
1186 for (; *Start != 0 && isspace(*Start) != 0; Start++);
1187
1188 unsigned long Count = 0;
1189 char *Pos = Start;
1190 while (Pos != Stop)
1191 {
1192 // Skip to the next Token
1193 for (; Pos != Stop && *Pos != Tok; Pos++);
1194
1195 // Back remove spaces
1196 char *End = Pos;
1197 for (; End > Start && (End[-1] == Tok || isspace(End[-1]) != 0); End--);
1198 *End = 0;
1199
1200 List[Count++] = Start;
1201 if (Count >= ListMax)
1202 {
1203 List[Count-1] = 0;
1204 return false;
1205 }
1206
1207 // Advance pos
1208 for (; Pos != Stop && (*Pos == Tok || isspace(*Pos) != 0 || *Pos == 0); Pos++);
1209 Start = Pos;
1210 }
1211
1212 List[Count] = 0;
1213 return true;
1214}
1215 /*}}}*/
1216// VectorizeString - Split a string up into a vector of strings /*{{{*/
1217// ---------------------------------------------------------------------
1218/* This can be used to split a given string up into a vector, so the
1219 propose is the same as in the method above and this one is a bit slower
1220 also, but the advantage is that we have an iteratable vector */
1221vector<string> VectorizeString(string const &haystack, char const &split)
1222{
1223 vector<string> exploded;
1224 if (haystack.empty() == true)
1225 return exploded;
1226 string::const_iterator start = haystack.begin();
1227 string::const_iterator end = start;
1228 do {
1229 for (; end != haystack.end() && *end != split; ++end);
1230 exploded.push_back(string(start, end));
1231 start = end + 1;
1232 } while (end != haystack.end() && (++end) != haystack.end());
1233 return exploded;
1234}
1235 /*}}}*/
1236// StringSplit - split a string into a string vector by token /*{{{*/
1237// ---------------------------------------------------------------------
1238/* See header for details.
1239 */
1240vector<string> StringSplit(std::string const &s, std::string const &sep,
1241 unsigned int maxsplit)
1242{
1243 vector<string> split;
1244 size_t start, pos;
1245
1246 // no separator given, this is bogus
1247 if(sep.size() == 0)
1248 return split;
1249
1250 start = pos = 0;
1251 while (pos != string::npos)
1252 {
1253 pos = s.find(sep, start);
1254 split.push_back(s.substr(start, pos-start));
1255
1256 // if maxsplit is reached, the remaining string is the last item
1257 if(split.size() >= maxsplit)
1258 {
1259 split[split.size()-1] = s.substr(start);
1260 break;
1261 }
1262 start = pos+sep.size();
1263 }
1264 return split;
1265}
1266 /*}}}*/
1267// RegexChoice - Simple regex list/list matcher /*{{{*/
1268// ---------------------------------------------------------------------
1269/* */
1270unsigned long RegexChoice(RxChoiceList *Rxs,const char **ListBegin,
1271 const char **ListEnd)
1272{
1273 for (RxChoiceList *R = Rxs; R->Str != 0; R++)
1274 R->Hit = false;
1275
1276 unsigned long Hits = 0;
1277 for (; ListBegin < ListEnd; ++ListBegin)
1278 {
1279 // Check if the name is a regex
1280 const char *I;
1281 bool Regex = true;
1282 for (I = *ListBegin; *I != 0; I++)
1283 if (*I == '.' || *I == '?' || *I == '*' || *I == '|')
1284 break;
1285 if (*I == 0)
1286 Regex = false;
1287
1288 // Compile the regex pattern
1289 regex_t Pattern;
1290 if (Regex == true)
1291 if (regcomp(&Pattern,*ListBegin,REG_EXTENDED | REG_ICASE |
1292 REG_NOSUB) != 0)
1293 Regex = false;
1294
1295 // Search the list
1296 bool Done = false;
1297 for (RxChoiceList *R = Rxs; R->Str != 0; R++)
1298 {
1299 if (R->Str[0] == 0)
1300 continue;
1301
1302 if (strcasecmp(R->Str,*ListBegin) != 0)
1303 {
1304 if (Regex == false)
1305 continue;
1306 if (regexec(&Pattern,R->Str,0,0,0) != 0)
1307 continue;
1308 }
1309 Done = true;
1310
1311 if (R->Hit == false)
1312 Hits++;
1313
1314 R->Hit = true;
1315 }
1316
1317 if (Regex == true)
1318 regfree(&Pattern);
1319
1320 if (Done == false)
1321 _error->Warning(_("Selection %s not found"),*ListBegin);
1322 }
1323
1324 return Hits;
1325}
1326 /*}}}*/
1327// {str,io}printf - C format string outputter to C++ strings/iostreams /*{{{*/
1328// ---------------------------------------------------------------------
1329/* This is used to make the internationalization strings easier to translate
1330 and to allow reordering of parameters */
1331static bool iovprintf(ostream &out, const char *format,
1332 va_list &args, ssize_t &size) {
1333 char *S = (char*)malloc(size);
1334 ssize_t const n = vsnprintf(S, size, format, args);
1335 if (n > -1 && n < size) {
1336 out << S;
1337 free(S);
1338 return true;
1339 } else {
1340 if (n > -1)
1341 size = n + 1;
1342 else
1343 size *= 2;
1344 }
1345 free(S);
1346 return false;
1347}
1348void ioprintf(ostream &out,const char *format,...)
1349{
1350 va_list args;
1351 ssize_t size = 400;
1352 while (true) {
1353 bool ret;
1354 va_start(args,format);
1355 ret = iovprintf(out, format, args, size);
1356 va_end(args);
1357 if (ret == true)
1358 return;
1359 }
1360}
1361void strprintf(string &out,const char *format,...)
1362{
1363 va_list args;
1364 ssize_t size = 400;
1365 std::ostringstream outstr;
1366 while (true) {
1367 bool ret;
1368 va_start(args,format);
1369 ret = iovprintf(outstr, format, args, size);
1370 va_end(args);
1371 if (ret == true)
1372 break;
1373 }
1374 out = outstr.str();
1375}
1376 /*}}}*/
1377// safe_snprintf - Safer snprintf /*{{{*/
1378// ---------------------------------------------------------------------
1379/* This is a snprintf that will never (ever) go past 'End' and returns a
1380 pointer to the end of the new string. The returned string is always null
1381 terminated unless Buffer == end. This is a better alterantive to using
1382 consecutive snprintfs. */
1383char *safe_snprintf(char *Buffer,char *End,const char *Format,...)
1384{
1385 va_list args;
1386 int Did;
1387
1388 if (End <= Buffer)
1389 return End;
1390 va_start(args,Format);
1391 Did = vsnprintf(Buffer,End - Buffer,Format,args);
1392 va_end(args);
1393
1394 if (Did < 0 || Buffer + Did > End)
1395 return End;
1396 return Buffer + Did;
1397}
1398 /*}}}*/
1399// StripEpoch - Remove the version "epoch" from a version string /*{{{*/
1400// ---------------------------------------------------------------------
1401string StripEpoch(const string &VerStr)
1402{
1403 size_t i = VerStr.find(":");
1404 if (i == string::npos)
1405 return VerStr;
1406 return VerStr.substr(i+1);
1407}
1408 /*}}}*/
1409
1410// tolower_ascii - tolower() function that ignores the locale /*{{{*/
1411// ---------------------------------------------------------------------
1412/* This little function is the most called method we have and tries
1413 therefore to do the absolut minimum - and is notable faster than
1414 standard tolower/toupper and as a bonus avoids problems with different
1415 locales - we only operate on ascii chars anyway. */
1416#undef tolower_ascii
1417int tolower_ascii(int const c) APT_CONST APT_COLD;
1418int tolower_ascii(int const c)
1419{
1420 return tolower_ascii_inline(c);
1421}
1422 /*}}}*/
1423
1424// isspace_ascii - isspace() function that ignores the locale /*{{{*/
1425// ---------------------------------------------------------------------
1426/* This little function is one of the most called methods we have and tries
1427 therefore to do the absolut minimum - and is notable faster than
1428 standard isspace() and as a bonus avoids problems with different
1429 locales - we only operate on ascii chars anyway. */
1430#undef isspace_ascii
1431int isspace_ascii(int const c) APT_CONST APT_COLD;
1432int isspace_ascii(int const c)
1433{
1434 return isspace_ascii_inline(c);
1435}
1436 /*}}}*/
1437
1438// CheckDomainList - See if Host is in a , separate list /*{{{*/
1439// ---------------------------------------------------------------------
1440/* The domain list is a comma separate list of domains that are suffix
1441 matched against the argument */
1442bool CheckDomainList(const string &Host,const string &List)
1443{
1444 string::const_iterator Start = List.begin();
1445 for (string::const_iterator Cur = List.begin(); Cur <= List.end(); ++Cur)
1446 {
1447 if (Cur < List.end() && *Cur != ',')
1448 continue;
1449
1450 // Match the end of the string..
1451 if ((Host.size() >= (unsigned)(Cur - Start)) &&
1452 Cur - Start != 0 &&
1453 stringcasecmp(Host.end() - (Cur - Start),Host.end(),Start,Cur) == 0)
1454 return true;
1455
1456 Start = Cur + 1;
1457 }
1458 return false;
1459}
1460 /*}}}*/
1461// strv_length - Return the length of a NULL-terminated string array /*{{{*/
1462// ---------------------------------------------------------------------
1463/* */
1464size_t strv_length(const char **str_array)
1465{
1466 size_t i;
1467 for (i=0; str_array[i] != NULL; i++)
1468 /* nothing */
1469 ;
1470 return i;
1471}
1472 /*}}}*/
1473// DeEscapeString - unescape (\0XX and \xXX) from a string /*{{{*/
1474// ---------------------------------------------------------------------
1475/* */
1476string DeEscapeString(const string &input)
1477{
1478 char tmp[3];
1479 string::const_iterator it;
1480 string output;
1481 for (it = input.begin(); it != input.end(); ++it)
1482 {
1483 // just copy non-escape chars
1484 if (*it != '\\')
1485 {
1486 output += *it;
1487 continue;
1488 }
1489
1490 // deal with double escape
1491 if (*it == '\\' &&
1492 (it + 1 < input.end()) && it[1] == '\\')
1493 {
1494 // copy
1495 output += *it;
1496 // advance iterator one step further
1497 ++it;
1498 continue;
1499 }
1500
1501 // ensure we have a char to read
1502 if (it + 1 == input.end())
1503 continue;
1504
1505 // read it
1506 ++it;
1507 switch (*it)
1508 {
1509 case '0':
1510 if (it + 2 <= input.end()) {
1511 tmp[0] = it[1];
1512 tmp[1] = it[2];
1513 tmp[2] = 0;
1514 output += (char)strtol(tmp, 0, 8);
1515 it += 2;
1516 }
1517 break;
1518 case 'x':
1519 if (it + 2 <= input.end()) {
1520 tmp[0] = it[1];
1521 tmp[1] = it[2];
1522 tmp[2] = 0;
1523 output += (char)strtol(tmp, 0, 16);
1524 it += 2;
1525 }
1526 break;
1527 default:
1528 // FIXME: raise exception here?
1529 break;
1530 }
1531 }
1532 return output;
1533}
1534 /*}}}*/
1535// URI::CopyFrom - Copy from an object /*{{{*/
1536// ---------------------------------------------------------------------
1537/* This parses the URI into all of its components */
1538void URI::CopyFrom(const string &U)
1539{
1540 string::const_iterator I = U.begin();
1541
1542 // Locate the first colon, this separates the scheme
1543 for (; I < U.end() && *I != ':' ; ++I);
1544 string::const_iterator FirstColon = I;
1545
1546 /* Determine if this is a host type URI with a leading double //
1547 and then search for the first single / */
1548 string::const_iterator SingleSlash = I;
1549 if (I + 3 < U.end() && I[1] == '/' && I[2] == '/')
1550 SingleSlash += 3;
1551
1552 /* Find the / indicating the end of the hostname, ignoring /'s in the
1553 square brackets */
1554 bool InBracket = false;
1555 for (; SingleSlash < U.end() && (*SingleSlash != '/' || InBracket == true); ++SingleSlash)
1556 {
1557 if (*SingleSlash == '[')
1558 InBracket = true;
1559 if (InBracket == true && *SingleSlash == ']')
1560 InBracket = false;
1561 }
1562
1563 if (SingleSlash > U.end())
1564 SingleSlash = U.end();
1565
1566 // We can now write the access and path specifiers
1567 Access.assign(U.begin(),FirstColon);
1568 if (SingleSlash != U.end())
1569 Path.assign(SingleSlash,U.end());
1570 if (Path.empty() == true)
1571 Path = "/";
1572
1573 // Now we attempt to locate a user:pass@host fragment
1574 if (FirstColon + 2 <= U.end() && FirstColon[1] == '/' && FirstColon[2] == '/')
1575 FirstColon += 3;
1576 else
1577 FirstColon += 1;
1578 if (FirstColon >= U.end())
1579 return;
1580
1581 if (FirstColon > SingleSlash)
1582 FirstColon = SingleSlash;
1583
1584 // Find the colon...
1585 I = FirstColon + 1;
1586 if (I > SingleSlash)
1587 I = SingleSlash;
1588 for (; I < SingleSlash && *I != ':'; ++I);
1589 string::const_iterator SecondColon = I;
1590
1591 // Search for the @ after the colon
1592 for (; I < SingleSlash && *I != '@'; ++I);
1593 string::const_iterator At = I;
1594
1595 // Now write the host and user/pass
1596 if (At == SingleSlash)
1597 {
1598 if (FirstColon < SingleSlash)
1599 Host.assign(FirstColon,SingleSlash);
1600 }
1601 else
1602 {
1603 Host.assign(At+1,SingleSlash);
1604 // username and password must be encoded (RFC 3986)
1605 User.assign(DeQuoteString(FirstColon,SecondColon));
1606 if (SecondColon < At)
1607 Password.assign(DeQuoteString(SecondColon+1,At));
1608 }
1609
1610 // Now we parse the RFC 2732 [] hostnames.
1611 unsigned long PortEnd = 0;
1612 InBracket = false;
1613 for (unsigned I = 0; I != Host.length();)
1614 {
1615 if (Host[I] == '[')
1616 {
1617 InBracket = true;
1618 Host.erase(I,1);
1619 continue;
1620 }
1621
1622 if (InBracket == true && Host[I] == ']')
1623 {
1624 InBracket = false;
1625 Host.erase(I,1);
1626 PortEnd = I;
1627 continue;
1628 }
1629 I++;
1630 }
1631
1632 // Tsk, weird.
1633 if (InBracket == true)
1634 {
1635 Host.clear();
1636 return;
1637 }
1638
1639 // Now we parse off a port number from the hostname
1640 Port = 0;
1641 string::size_type Pos = Host.rfind(':');
1642 if (Pos == string::npos || Pos < PortEnd)
1643 return;
1644
1645 Port = atoi(string(Host,Pos+1).c_str());
1646 Host.assign(Host,0,Pos);
1647}
1648 /*}}}*/
1649// URI::operator string - Convert the URI to a string /*{{{*/
1650// ---------------------------------------------------------------------
1651/* */
1652URI::operator string()
1653{
1654 std::stringstream Res;
1655
1656 if (Access.empty() == false)
1657 Res << Access << ':';
1658
1659 if (Host.empty() == false)
1660 {
1661 if (Access.empty() == false)
1662 Res << "//";
1663
1664 if (User.empty() == false)
1665 {
1666 // FIXME: Technically userinfo is permitted even less
1667 // characters than these, but this is not conveniently
1668 // expressed with a blacklist.
1669 Res << QuoteString(User, ":/?#[]@");
1670 if (Password.empty() == false)
1671 Res << ":" << QuoteString(Password, ":/?#[]@");
1672 Res << "@";
1673 }
1674
1675 // Add RFC 2732 escaping characters
1676 if (Access.empty() == false && Host.find_first_of("/:") != string::npos)
1677 Res << '[' << Host << ']';
1678 else
1679 Res << Host;
1680
1681 if (Port != 0)
1682 Res << ':' << std::to_string(Port);
1683 }
1684
1685 if (Path.empty() == false)
1686 {
1687 if (Path[0] != '/')
1688 Res << "/" << Path;
1689 else
1690 Res << Path;
1691 }
1692
1693 return Res.str();
1694}
1695 /*}}}*/
1696// URI::SiteOnly - Return the schema and site for the URI /*{{{*/
1697string URI::SiteOnly(const string &URI)
1698{
1699 ::URI U(URI);
1700 U.User.clear();
1701 U.Password.clear();
1702 U.Path.clear();
1703 return U;
1704}
1705 /*}}}*/
1706// URI::ArchiveOnly - Return the schema, site and cleaned path for the URI /*{{{*/
1707string URI::ArchiveOnly(const string &URI)
1708{
1709 ::URI U(URI);
1710 U.User.clear();
1711 U.Password.clear();
1712 if (U.Path.empty() == false && U.Path[U.Path.length() - 1] == '/')
1713 U.Path.erase(U.Path.length() - 1);
1714 return U;
1715}
1716 /*}}}*/
1717// URI::NoUserPassword - Return the schema, site and path for the URI /*{{{*/
1718string URI::NoUserPassword(const string &URI)
1719{
1720 ::URI U(URI);
1721 U.User.clear();
1722 U.Password.clear();
1723 return U;
1724}
1725 /*}}}*/