2 The software in this package is distributed under the GNU General
3 Public License version 2 (with a special exception described below).
5 A copy of GNU General Public License (GPL) is included in this distribution,
6 in the file COPYING.GPL.
8 As a special exception, if other files instantiate templates or use macros
9 or inline functions from this file, or you compile this file and link it
10 with other works to produce a work based on this file, this file
11 does not by itself cause the resulting work to be covered
12 by the GNU General Public License.
14 However the source code for this file must still be made available
15 in accordance with section (3) of the GNU General Public License.
17 This exception does not invalidate any other reasons why a work based
18 on this file might be covered by the GNU General Public License.
22 * (c) Copyright 2007-2008 by Intra2net AG
30 #include <cmath> // for round()
38 #include <boost/numeric/conversion/cast.hpp>
39 #include <boost/foreach.hpp>
41 #include <boost/assert.hpp>
42 #include <boost/shared_ptr.hpp>
43 #include <openssl/bio.h>
44 #include <openssl/evp.h>
46 #include <stringfunc.hxx>
57 const std::string hexDigitsLower("0123456789abcdef");
58 const std::string hexDigitsUpper("0123456789ABCDEF");
63 char operator() (char c)
65 return std::toupper(c);
67 }; // eo struct UpperFunc
72 char operator() (char c)
74 return std::tolower(c);
76 }; // eo struct LowerFunc
79 } // eo namespace <anonymous>
84 * default list of Whitespaces (" \t\r\n");
86 const std::string Whitespaces = " \t\r\n";
89 * default list of lineendings ("\r\n");
91 const std::string LineEndings= "\r\n";
96 * @brief checks if a string begins with a given prefix.
97 * @param[in,out] str the string which is tested
98 * @param prefix the prefix which should be tested for.
99 * @return @a true iff the prefix is not empty and the string begins with that prefix.
101 bool has_prefix(const std::string& str, const std::string& prefix)
103 if (prefix.empty() || str.empty() || str.size() < prefix.size() )
107 return str.compare(0, prefix.size(), prefix) == 0;
108 } // eo has_prefix(const std::string&,const std::string&)
112 * @brief checks if a string ends with a given suffix.
113 * @param[in,out] str the string which is tested
114 * @param suffix the suffix which should be tested for.
115 * @return @a true iff the suffix is not empty and the string ends with that suffix.
117 bool has_suffix(const std::string& str, const std::string& suffix)
119 if (suffix.empty() || str.empty() || str.size() < suffix.size() )
123 return str.compare(str.size() - suffix.size(), suffix.size(), suffix) == 0;
124 } // eo has_suffix(const std::string&,const std::string&)
128 * cut off characters from a given list from front and end of a string.
129 * @param[in,out] str the string which should be trimmed.
130 * @param charlist the list of characters to remove from beginning and end of string
131 * @return the result string.
133 std::string trim_mod(std::string& str, const std::string& charlist)
135 // first: trim the beginning:
136 std::string::size_type pos= str.find_first_not_of (charlist);
137 if (pos == std::string::npos)
139 // whole string consists of charlist (or is already empty)
145 // str starts with charlist
148 // now let's look at the tail:
149 pos= str.find_last_not_of(charlist) +1; // note: we already know there is at least one other char!
150 if ( pos < str.size() )
152 str.erase(pos, str.size()-pos);
155 } // eo trim_mod(std::string&,const std::string&)
160 * removes last character from a string when it is in a list of chars to be removed.
161 * @param[in,out] str the string.
162 * @param what the list of chars which will be tested for.
163 * @return the resulting string with last char removed (if applicable)
165 std::string chomp_mod(std::string& str, const std::string& what)
167 if (str.empty() || what.empty() )
171 if (what.find(str.at (str.size()-1) ) != std::string::npos)
173 str.erase(str.size() - 1);
176 } // eo chomp_mod(std::string&,const std::string&)
180 * @brief converts a string to lower case.
181 * @param[in,out] str the string to modify.
184 std::string to_lower_mod(std::string& str)
186 std::transform(str.begin(), str.end(), str.begin(), LowerFunc() );
188 } // eo to_lower_mod(std::string&)
192 * @brief converts a string to upper case.
193 * @param[in,out] str the string to modify.
196 std::string to_upper_mod(std::string& str)
198 std::transform( str.begin(), str.end(), str.begin(), UpperFunc() );
200 } // eo to_upper_mod(std::string&)
205 * cut off characters from a given list from front and end of a string.
206 * @param str the string which should be trimmed.
207 * @param charlist the list of characters to remove from beginning and end of string
208 * @return the result string.
210 std::string trim (const std::string& str, const std::string& charlist)
212 // first: trim the beginning:
213 std::string::size_type pos0= str.find_first_not_of(charlist);
214 if (pos0 == std::string::npos)
216 // whole string consists of charlist (or is already empty)
217 return std::string();
219 // now let's look at the end:
220 std::string::size_type pos1= str.find_last_not_of(charlist);
221 return str.substr(pos0, pos1 - pos0 + 1);
222 } // eo trim(const std:.string&,const std::string&)
226 * removes last character from a string when it is in a list of chars to be removed.
227 * @param str the string.
228 * @param what the list of chars which will be tested for.
229 * @return the resulting string with last char removed (if applicable)
231 std::string chomp (const std::string& str, const std::string& what)
233 if (str.empty() || what.empty() )
237 if (what.find(str.at (str.size()-1) ) != std::string::npos)
239 return str.substr(0, str.size()-1);
242 } // eo chomp(const std:.string&,const std::string&)
246 * @brief returns a lower case version of a given string.
247 * @param str the string
248 * @return the lower case version of the string
250 std::string to_lower (const std::string& str)
252 std::string result(str);
253 return to_lower_mod(result);
254 } // eo to_lower(const std::string&)
258 * @brief returns a upper case version of a given string.
259 * @param str the string
260 * @return the upper case version of the string
262 std::string to_upper(const std::string& str)
264 std::string result(str);
265 return to_upper_mod(result);
266 } // eo to_upper(const std::string&)
271 * @brief removes a given suffix from a string.
272 * @param str the string.
273 * @param suffix the suffix which should be removed if the string ends with it.
274 * @return the string without the suffix.
276 * If the string ends with the suffix, it is removed. If the the string doesn't end
277 * with the suffix the original string is returned.
279 std::string remove_suffix(const std::string& str, const std::string& suffix)
281 if (has_suffix(str,suffix) )
283 return str.substr(0, str.size()-suffix.size() );
286 } // eo remove_suffix(const std::string&,const std::string&)
291 * @brief removes a given prefix from a string.
292 * @param str the string.
293 * @param prefix the prefix which should be removed if the string begins with it.
294 * @return the string without the prefix.
296 * If the string begins with the prefix, it is removed. If the the string doesn't begin
297 * with the prefix the original string is returned.
299 std::string remove_prefix(const std::string& str, const std::string& prefix)
301 if (has_prefix(str,prefix) )
303 return str.substr( prefix.size() );
306 } // eo remove_prefix(const std::string&,const std::string&)
310 * split a string to key and value delimited by a given delimiter.
311 * The resulting key and value strings are trimmed (Whitespaces removed at beginning and end).
312 * @param str the string which should be splitted.
313 * @param[out] key the resulting key
314 * @param[out] value the resulting value
315 * @param delimiter the delimiter between key and value; default is '='.
316 * @return @a true if the split was successful.
319 const std::string& str,
324 std::string::size_type pos = str.find (delimiter);
325 if (pos == std::string::npos) return false;
326 key= str.substr(0,pos);
327 value= str.substr(pos+1);
331 } // eo pair_split(const std::string&,std::string&,std::string&,char)
335 * splits a string by given delimiter
337 * @param[in] str the string which should be splitted.
338 * @param[out] result the list resulting from splitting @a str.
339 * @param[in] delimiter the delimiter (word/phrase) at which @a str should be splitted.
340 * @param[in] omit_empty should empty parts not be stored?
341 * @param[in] trim_list list of characters the parts should be trimmed by.
342 * (empty string results in no trim)
345 const std::string& str,
346 std::list<std::string>& result,
347 const std::string& delimiter,
349 const std::string& trim_list
352 std::string::size_type pos, last_pos=0;
353 bool delimiter_found= false;
354 while ( last_pos < str.size() && last_pos != std::string::npos)
356 pos= str.find(delimiter, last_pos);
358 if (pos == std::string::npos)
360 part= str.substr(last_pos);
361 delimiter_found= false;
365 part= str.substr(last_pos, pos-last_pos);
366 delimiter_found=true;
368 if (pos != std::string::npos)
370 last_pos= pos+ delimiter.size();
374 last_pos= std::string::npos;
376 if (!trim_list.empty() ) trim_mod (part, trim_list);
377 if (omit_empty && part.empty() ) continue;
378 result.push_back( part );
380 // if the string ends with a delimiter we need to append an empty string if no omit_empty
382 // (this way we keep the split result consistent to a join operation)
383 if (delimiter_found && !omit_empty)
385 result.push_back("");
387 } // eo split_string(const std::string&,std::list< std::string >&,const std::string&,bool,const std::string&)
390 /** call split_string with list<string>, converts result to vector; vector is clear()-ed first
392 * Note: Uses 3 O(n)-operations: list.size, vector.resize and std::swap_ranges;
393 * not sure whether there is a better way to do this
396 const std::string& str,
397 std::vector<std::string>& result,
398 const std::string& delimiter,
400 const std::string& trim_list
403 std::list<std::string> tmp;
404 split_string(str, tmp, delimiter, omit_empty, trim_list);
405 std::size_t size = tmp.size(); // this is O(n)
407 result.resize(size); // also O(n)
408 std::swap_ranges(tmp.begin(), tmp.end(), result.begin()); // also O(n)
412 * splits a string by a given delimiter
413 * @param str the string which should be splitted.
414 * @param delimiter delimiter the delimiter (word/phrase) at which @a str should be splitted.
415 * @param[in] omit_empty should empty parts not be stored?
416 * @param[in] trim_list list of characters the parts should be trimmed by.
417 * (empty string results in no trim)
418 * @return the list resulting from splitting @a str.
420 std::list<std::string> split_string(
421 const std::string& str,
422 const std::string& delimiter,
424 const std::string& trim_list
427 std::list<std::string> result;
428 split_string(str, result, delimiter, omit_empty, trim_list);
430 } // eo split_string(const std::string&,const std::string&,bool,const std::string&)
434 * @brief joins a list of strings into a single string.
436 * This funtion is (basically) the reverse operation of @a split_string.
438 * @param parts the list of strings.
439 * @param delimiter the delimiter which is inserted between the strings.
440 * @return the joined string.
442 std::string join_string(
443 const std::list< std::string >& parts,
444 const std::string& delimiter
448 if (! parts.empty() )
450 std::list< std::string >::const_iterator it= parts.begin();
452 while ( ++it != parts.end() )
459 } // eo join_string(const std::list< std::string >&,const std::string&)
462 /** @brief same as join_string for list, except uses a vector */
463 std::string join_string(
464 const std::vector< std::string >& parts,
465 const std::string& delimiter
469 if (! parts.empty() )
471 std::vector< std::string >::const_iterator it= parts.begin();
473 while ( ++it != parts.end() )
480 } // eo join_string(const std::vector< std::string >&,const std::string&)
482 std::string join_string (
483 const char *const parts[], /* assumed NULL-terminated */
484 const std::string& delimiter
491 const char *const *cur = parts;
494 result = std::string (*cur);
496 while (*++cur != NULL) {
498 result += std::string (*cur);
514 * @brief returns a hex string from a binary string.
515 * @param str the (binary) string
516 * @param upper_case_digits determine whether to use upper case characters for digits A-F.
517 * @return the string in hex notation.
519 std::string convert_binary_to_hex(
520 const std::string& str,
521 bool upper_case_digits
525 std::string hexDigits(upper_case_digits ? hexDigitsUpper : hexDigitsLower);
526 for ( std::string::const_iterator it= str.begin();
530 result.push_back( hexDigits[ ( (*it) >> 4) & 0x0f ] );
531 result.push_back( hexDigits[ (*it) & 0x0f ] );
534 } // eo convert_binary_to_hex(const std::string&,bool)
538 * @brief converts a hex digit string to binary string.
539 * @param str hex digit string
540 * @return the binary string.
542 * The hex digit string may contains white spaces or colons which are treated
543 * as delimiters between hex digit groups.
545 * @todo rework the handling of half nibbles (consistency)!
547 std::string convert_hex_to_binary(
548 const std::string& str
550 throw (std::runtime_error)
554 bool hasNibble= false;
555 bool lastWasWS= true;
556 for ( std::string::const_iterator it= str.begin();
560 std::string::size_type p = hexDigitsLower.find( *it );
561 if (p== std::string::npos)
563 p= hexDigitsUpper.find( *it );
565 if (p == std::string::npos)
567 if ( ( Whitespaces.find( *it ) != std::string::npos) // is it a whitespace?
568 or ( *it == ':') // or a colon?
571 // we treat that as a valid delimiter:
574 // 1 nibble before WS is treate as lower part:
583 if (p == std::string::npos )
585 throw runtime_error("illegal character in hex digit string: " + str);
599 //we already had a nibble, so a char is complete now:
600 result.push_back( c );
605 // this is the first nibble of a new char:
611 //well, there is one nibble left
612 // let's do some heuristics:
615 // if the preceeding character was a white space (or a colon)
616 // we treat the nibble as lower part:
617 //( this is consistent with shortened hex notations where leading zeros are not noted)
618 result.push_back( c );
622 // if it was part of a hex digit chain, we treat it as UPPER part (!!)
623 result.push_back( c << 4 );
627 } // eo convert_hex_to_binary(const std::string&)
630 static list<string>& alloc_template_starts()
632 static list<string> result;
635 result.push_back("std::list");
636 result.push_back("std::vector");
641 string shorten_stl_types(const string &input)
643 string output = input;
645 // first: replace fixed string for std::string
646 replace_all(output, "std::basic_string<char, std::char_traits<char>, std::allocator<char> >",
649 // loop over list/vector/... that have an allocator, e.g.
650 // std::list< some_type_here, std::allocator<some_type_here> >
651 string::size_type start, comma, end, len, start_text_len;
653 string allocator_text;
654 BOOST_FOREACH(const string &start_text, alloc_template_starts())
659 start_text_len = start_text.length();
660 while( (start=output.find(start_text+"<", start)) != string::npos )
662 len = output.length();
663 start += start_text_len+1; // start next iter and tests here after opening bracket
665 // now comes the tricky part: find matching ',' and the closing '>' even if "subtype" is template again
667 n_open_brackets = 1; // the bracket right after start_text counts as first
668 while (comma < len && n_open_brackets > 0)
670 if (output[comma] == ',' && n_open_brackets == 1)
672 else if (output[comma] == '<')
674 else if (output[comma] == '>')
679 while (end < len && n_open_brackets > 0)
681 if (output[end] == '<')
683 else if (output[end] == '>')
686 if (n_open_brackets == 0)
687 break; // do not increment end
692 // check that start < comma < end < len && n_open_brackets == 0
693 if (start >= comma || comma >= end || end >= len || n_open_brackets != 0)
694 continue; // input seems to be of unexpected form
696 // check that type in allocator is same as until comma
697 string type = output.substr(start, comma-start);
698 if (type[type.length()-1] == '>')
699 allocator_text = string("std::allocator<") + type + " > ";
701 allocator_text = string("std::allocator<") + type + "> ";
702 if (output.substr(comma+2, end-comma-2) == allocator_text)
703 output.replace(comma+2, end-comma-2, "_alloc_");
710 typedef boost::shared_ptr<BIO> BIO_Ptr;
713 * @brief Converts openssl generic input/output to std::string
715 * Code adapted from keymakerd.
717 * @param bio Openssl's generic input/output
718 * @return :string STL string
720 static std::string _convert_BIO_to_string(BIO *input)
725 long written = BIO_get_mem_data(input, &output);
726 if (written <= 0 || output == NULL)
729 rtn.assign(output, written); //lint !e534 !e732
734 * @brief base64 encode a string using OpenSSL base64 functions
736 * Data size limit is 2GB on 32 bit (LONG_MAX)
738 * @param input String to encode
739 * @param one_line Encode all data as one line, no wrapping with line feeds
740 * @return base64 encoded string
742 std::string base64_encode(const std::string &input, bool one_line)
744 // check for empty buffer
748 // safety check to ensure our check afer BIO_write() works
749 if (input.size() >= LONG_MAX)
750 throw runtime_error("base64 encode: Too much data");
752 // setup encoder. Note: BIO_free_all frees both BIOs.
753 BIO_Ptr base64_encoder(BIO_new(BIO_f_base64()), BIO_free_all);
754 BIO *encoder_bio = base64_encoder.get();
756 BIO_set_flags(encoder_bio, BIO_FLAGS_BASE64_NO_NL);
758 // chain output buffer and encoder together
759 BIO *encoded_result = BIO_new(BIO_s_mem());
760 BIO_push(encoder_bio, encoded_result);
763 long written = BIO_write(encoder_bio, input.c_str(), input.size());
764 if ((unsigned)written != input.size())
767 out << "base64 encoding failed: input size: "
768 << input.size() << " vs. output size: " << written;
769 throw runtime_error(out.str());
771 if (BIO_flush(encoder_bio) != 1)
772 throw runtime_error("base64 encode: BIO_flush() failed");
774 return _convert_BIO_to_string(encoded_result);
778 * @brief base64 decode a string using OpenSSL base64 functions
780 * @param input String to decode
781 * @param one_line Expect all base64 data in one line. Input with line feeds will fail.
782 * @return base64 decoded string
784 std::string base64_decode(const std::string &input, bool one_line)
786 // check for empty buffer
790 // safety check for BIO_new_mem_buf()
791 if (input.size() >= INT_MAX)
792 throw runtime_error("base64 decode: Too much data");
794 // setup encoder. Note: BIO_free_all frees both BIOs.
795 BIO_Ptr base64_decoder(BIO_new(BIO_f_base64()), BIO_free_all);
796 BIO *bio_base64 = base64_decoder.get();
798 BIO_set_flags(bio_base64, BIO_FLAGS_BASE64_NO_NL);
800 // chain input buffer and decoder together
801 BIO *bio_input = BIO_new_mem_buf((void*)input.c_str(), input.size());
802 bio_input = BIO_push(bio_base64, bio_input);
804 BIO_Ptr decoded_result(BIO_new(BIO_s_mem()), BIO_free_all);
805 BIO *bio_decoded = decoded_result.get();
806 const int convbuf_size = 512;
807 char convbuf[convbuf_size];
810 while((read_bytes = BIO_read(bio_input, convbuf, convbuf_size)) > 0)
812 BOOST_ASSERT(read_bytes <= convbuf_size);
813 long written_bytes = BIO_write(bio_decoded, convbuf, read_bytes);
814 if (written_bytes != read_bytes)
817 out << "base64 decoding failed: read_bytes: "
818 << read_bytes << " vs. written_bytes: " << written_bytes;
819 throw runtime_error(out.str());
822 if (read_bytes == -2 || read_bytes == -1)
823 throw runtime_error("base64 decode: Error during decoding");
825 return _convert_BIO_to_string(bio_decoded);
828 } // eo namespace I2n
833 std::string iso_to_utf8(const std::string& isostring)
837 iconv_t i2utf8 = iconv_open("UTF-8", "ISO-8859-1");
839 if (iso_to_utf8 == (iconv_t)-1)
840 throw runtime_error("iconv can't convert from ISO-8859-1 to UTF-8");
842 size_t in_size=isostring.size();
843 size_t out_size=in_size*4;
845 char *buf = (char *)malloc(out_size+1);
847 throw runtime_error("out of memory for iconv buffer");
849 char *in = (char *)isostring.c_str();
851 iconv(i2utf8, &in, &in_size, &out, &out_size);
853 buf[isostring.size()*4-out_size]=0;
863 std::string utf8_to_iso(const std::string& utf8string)
867 iconv_t utf82iso = iconv_open("ISO-8859-1","UTF-8");
869 if (utf82iso == (iconv_t)-1)
870 throw runtime_error("iconv can't convert from UTF-8 to ISO-8859-1");
872 size_t in_size=utf8string.size();
873 size_t out_size=in_size;
875 char *buf = (char *)malloc(out_size+1);
877 throw runtime_error("out of memory for iconv buffer");
879 char *in = (char *)utf8string.c_str();
881 iconv(utf82iso, &in, &in_size, &out, &out_size);
883 buf[utf8string.size()-out_size]=0;
888 iconv_close(utf82iso);
893 wchar_t* utf8_to_wbuf(const std::string& utf8string)
895 iconv_t utf82wstr = iconv_open("UCS-4LE","UTF-8");
897 if (utf82wstr == (iconv_t)-1)
898 throw runtime_error("iconv can't convert from UTF-8 to UCS-4");
900 size_t in_size=utf8string.size();
901 size_t out_size= (in_size+1)*sizeof(wchar_t);
903 wchar_t *buf = (wchar_t *)malloc(out_size);
905 throw runtime_error("out of memory for iconv buffer");
907 char *in = (char *)utf8string.c_str();
908 char *out = (char*) buf;
909 if (iconv(utf82wstr, &in, &in_size, &out, &out_size) == (size_t)-1)
910 throw runtime_error("error converting char encodings");
912 buf[ ( (utf8string.size()+1)*sizeof(wchar_t)-out_size) /sizeof(wchar_t) ]=0;
914 iconv_close(utf82wstr);
919 std::string utf7imap_to_utf8(const std::string& utf7imapstring)
923 iconv_t utf7imap2utf8 = iconv_open("UTF-8","UTF-7-IMAP");
925 if (utf7imap2utf8 == (iconv_t)-1)
926 throw runtime_error("iconv can't convert from UTF-7-IMAP to UTF-8");
928 size_t in_size=utf7imapstring.size();
929 size_t out_size=in_size*4;
931 char *buf = (char *)malloc(out_size+1);
933 throw runtime_error("out of memory for iconv buffer");
935 char *in = (char *)utf7imapstring.c_str();
937 iconv(utf7imap2utf8, &in, &in_size, &out, &out_size);
939 buf[utf7imapstring.size()*4-out_size]=0;
944 iconv_close(utf7imap2utf8);
949 std::string utf8_to_utf7imap(const std::string& utf8string)
953 iconv_t utf82utf7imap = iconv_open("UTF-7-IMAP", "UTF-8");
955 if (utf82utf7imap == (iconv_t)-1)
956 throw runtime_error("iconv can't convert from UTF-7-IMAP to UTF-8");
958 // UTF-7 is base64 encoded, a buffer 10x as large
959 // as the utf-8 buffer should be enough. If not the string will be truncated.
960 size_t in_size=utf8string.size();
961 size_t out_size=in_size*10;
963 char *buf = (char *)malloc(out_size+1);
965 throw runtime_error("out of memory for iconv buffer");
967 char *in = (char *)utf8string.c_str();
969 iconv(utf82utf7imap, &in, &in_size, &out, &out_size);
971 buf[utf8string.size()*10-out_size]= 0;
976 iconv_close(utf82utf7imap);
981 // Tokenize string by (html) tags
982 void tokenize_by_tag(vector<pair<string,bool> > &tokenized, const std::string &input)
984 string::size_type pos, len = input.size();
985 bool inside_tag = false;
988 for (pos = 0; pos < len; pos++)
990 if (input[pos] == '<')
994 if (!current.empty() )
996 tokenized.push_back( make_pair(current, false) );
1000 current += input[pos];
1002 else if (input[pos] == '>' && inside_tag)
1004 current += input[pos];
1006 if (!current.empty() )
1008 tokenized.push_back( make_pair(current, true) );
1013 current += input[pos];
1016 // String left over in buffer?
1017 if (!current.empty() )
1018 tokenized.push_back( make_pair(current, false) );
1019 } // eo tokenize_by_tag
1022 std::string strip_html_tags(const std::string &input)
1024 // Pair first: string, second: isTag
1025 vector<pair<string,bool> > tokenized;
1026 tokenize_by_tag (tokenized, input);
1029 vector<pair<string,bool> >::const_iterator token, tokens_end = tokenized.end();
1030 for (token = tokenized.begin(); token != tokens_end; ++token)
1032 output += token->first;
1035 } // eo strip_html_tags
1038 // Smart-encode HTML en
1039 string smart_html_entities(const std::string &input)
1041 // Pair first: string, second: isTag
1042 vector<pair<string,bool> > tokenized;
1043 tokenize_by_tag (tokenized, input);
1046 vector<pair<string,bool> >::const_iterator token, tokens_end = tokenized.end();
1047 for (token = tokenized.begin(); token != tokens_end; ++token)
1049 // keep HTML tags as they are
1051 output += token->first;
1053 output += html_entities(token->first);
1060 string::size_type find_8bit(const std::string &str)
1062 string::size_type l=str.size();
1063 for (string::size_type p=0; p < l; p++)
1064 if (static_cast<unsigned char>(str[p]) > 127)
1067 return string::npos;
1070 // encoded UTF-8 chars into HTML entities
1071 string html_entities(std::string str)
1074 replace_all (str, "&", "&");
1075 replace_all (str, "<", "<");
1076 replace_all (str, ">", ">");
1077 replace_all (str, "\"", """);
1078 replace_all (str, "'", "'");
1079 replace_all (str, "/", "/");
1082 replace_all (str, "\xC3\xA4", "ä");
1083 replace_all (str, "\xC3\xB6", "ö");
1084 replace_all (str, "\xC3\xBC", "ü");
1085 replace_all (str, "\xC3\x84", "Ä");
1086 replace_all (str, "\xC3\x96", "Ö");
1087 replace_all (str, "\xC3\x9C", "Ü");
1090 replace_all (str, "\xC3\x9F", "ß");
1092 // conversion of remaining non-ASCII chars needed?
1093 // just do if needed because of performance
1094 if (find_8bit(str) != string::npos)
1096 // convert to fixed-size encoding UTF-32
1097 wchar_t* wbuf=utf8_to_wbuf(str);
1098 ostringstream target;
1100 // replace all non-ASCII chars with HTML representation
1101 for (int p=0; wbuf[p] != 0; p++)
1103 unsigned int c=wbuf[p];
1106 target << static_cast<unsigned char>(c);
1108 target << "&#" << c << ';';
1117 } // eo html_entities(std::string)
1119 // convert HTML entities to something that can be viewed on a basic text console (restricted to ASCII-7)
1120 string html_entities_to_console(std::string str)
1123 replace_all (str, "&", "&");
1124 replace_all (str, "<", "<");
1125 replace_all (str, ">", ">");
1126 replace_all (str, """, "\"");
1127 replace_all (str, "'", "'");
1128 replace_all (str, "/", "/");
1131 replace_all (str, "ä", "ae");
1132 replace_all (str, "ö", "oe");
1133 replace_all (str, "ü", "ue");
1134 replace_all (str, "Ä", "Ae");
1135 replace_all (str, "Ö", "Oe");
1136 replace_all (str, "Ü", "Ue");
1139 replace_all (str, "ß", "ss");
1144 // find_html_comments + remove_html_comments(str, comments)
1145 void remove_html_comments(string &str)
1147 vector<CommentZone> comments = find_html_comments(str);
1148 remove_html_comments(str, comments);
1151 // find all html comments, behaving correctly if they are nested; ignores comment tags ("<!--FOO .... BAR-->")
1152 // If there are invalid comments ("-->" before "<!--" or different number of closing and opening tags),
1153 // then the unknown index of corresponding start/end tag will be represented by a string::npos
1154 // Indices are from start of start tag until first index after closing tag
1155 vector<CommentZone> find_html_comments(const std::string &str)
1157 static const string START = "<!--";
1158 static const string CLOSE = "-->";
1159 static const string::size_type START_LEN = START.length();
1160 static const string::size_type CLOSE_LEN = CLOSE.length();
1162 vector<CommentZone> comments;
1164 // in order to find nested comments, need either recursion or a stack
1165 vector<string::size_type> starts; // stack of start tags
1167 string::size_type pos = 0;
1168 string::size_type len = str.length();
1169 string::size_type next_start, next_close;
1171 while (pos < len) // not really needed but just in case
1173 next_start = str.find(START, pos);
1174 next_close = str.find(CLOSE, pos);
1176 if ( (next_start == string::npos) && (next_close == string::npos) )
1177 break; // we are done
1179 else if ( (next_start == string::npos) || (next_close < next_start) ) // close one comment (pop)
1181 if (starts.empty()) // closing tag without a start
1182 comments.push_back(CommentZone(string::npos, next_close+CLOSE_LEN));
1185 comments.push_back(CommentZone(starts.back(), next_close+CLOSE_LEN));
1188 pos = next_close + CLOSE_LEN;
1191 else if ( (next_close == string::npos) || (next_start < next_close) ) // start a new comment (push)
1193 starts.push_back(next_start);
1194 pos = next_start + START_LEN;
1198 // add comments that have no closing tag from back to front (important for remove_html_comments!)
1199 while (!starts.empty())
1201 comments.push_back(CommentZone(starts.back(), string::npos));
1208 // remove all html comments foundby find_html_comments
1209 void remove_html_comments(std::string &str, const vector<CommentZone> &comments)
1211 // remember position where last removal started
1212 string::size_type last_removal_start = str.length();
1214 // Go from back to front to not mess up indices.
1215 // This requires that bigger comments, that contain smaller comments, come AFTER
1216 // the small contained comments in the comments vector (i.e. comments are ordered by
1217 // their closing tag, not their opening tag). This is true for results from find_html_comments
1218 BOOST_REVERSE_FOREACH(const CommentZone &comment, comments)
1220 if (comment.first == string::npos)
1222 str = str.replace(0, comment.second, ""); // comment starts "before" str --> delete from start
1223 break; // there can be no more
1225 else if (comment.first >= last_removal_start)
1227 continue; // this comment is inside another comment that we have removed already
1229 else if (comment.second == string::npos) // comment ends "after" str --> delete until end
1231 str = str.replace(comment.first, string::npos, "");
1232 last_removal_start = comment.first;
1236 str = str.replace(comment.first, comment.second-comment.first, "");
1237 last_removal_start = comment.first;
1242 bool replace_all(string &base, const char *ist, const char *soll)
1246 return replace_all(base,&i,&s);
1249 bool replace_all(string &base, const string &ist, const char *soll)
1252 return replace_all(base,&ist,&s);
1255 bool replace_all(string &base, const string *ist, const string *soll)
1257 return replace_all(base,*ist,*soll);
1260 bool replace_all(string &base, const char *ist, const string *soll)
1263 return replace_all(base,&i,soll);
1266 bool replace_all(string &base, const string &ist, const string &soll)
1268 bool found_ist = false;
1269 string::size_type a=0;
1272 throw runtime_error ("replace_all called with empty search string");
1274 while ( (a=base.find(ist,a) ) != string::npos)
1276 base.replace(a,ist.size(),soll);
1285 * @brief replaces all characters that could be problematic or impose a security risk when being logged
1286 * @param str the original string
1287 * @param replace_with the character to replace the unsafe chars with
1288 * @return a string that is safe to send to syslog or other logfiles
1290 * All chars between 0x20 (space) and 0x7E (~) (including) are considered safe for logging.
1291 * See e.g. RFC 5424, section 8.2 or the posix character class "printable".
1292 * This eliminates all possible problems with NUL, control characters, 8 bit chars, UTF8.
1295 std::string sanitize_for_logging(const std::string &str, const char replace_with)
1297 std::string output=str;
1299 const string::size_type len = output.size();
1300 for (std::string::size_type p=0; p < len; p++)
1301 if (output[p] < 0x20 || output[p] > 0x7E)
1302 output[p]=replace_with;
1308 string to_lower(const string &src)
1312 string::size_type pos, end = dst.size();
1313 for (pos = 0; pos < end; pos++)
1314 dst[pos] = tolower(dst[pos]);
1319 string to_upper(const string &src)
1323 string::size_type pos, end = dst.size();
1324 for (pos = 0; pos < end; pos++)
1325 dst[pos] = toupper(dst[pos]);
1331 const int MAX_UNIT_FORMAT_SYMBOLS = 6;
1333 const string shortUnitFormatSymbols[MAX_UNIT_FORMAT_SYMBOLS] = {
1342 const string longUnitFormatSymbols[MAX_UNIT_FORMAT_SYMBOLS] = {
1343 i18n_noop(" Bytes"),
1344 i18n_noop(" KBytes"),
1345 i18n_noop(" MBytes"),
1346 i18n_noop(" GBytes"),
1347 i18n_noop(" TBytes"),
1348 i18n_noop(" PBytes")
1352 static long double rounding_upwards(
1353 const long double number,
1354 const int rounding_multiplier
1357 long double rounded_number;
1358 rounded_number = number * rounding_multiplier;
1359 rounded_number += 0.5;
1360 rounded_number = (int64_t) (rounded_number);
1361 rounded_number = (long double) (rounded_number) / (long double) (rounding_multiplier);
1363 return rounded_number;
1367 string nice_unit_format(
1368 const int64_t input,
1369 const UnitFormat format,
1373 // select the system of units (decimal or binary)
1375 if (base == UnitBase1000)
1384 long double size = input;
1386 // check the size of the input number to fit in the appropriate symbol
1388 while (size > multiple)
1390 size = size / multiple;
1393 // rollback to the previous values and stop the loop when cannot
1394 // represent the number length.
1395 if (sizecount >= MAX_UNIT_FORMAT_SYMBOLS)
1397 size = size * multiple;
1403 // round the input number "half up" to multiples of 10
1404 const int rounding_multiplier = 10;
1405 size = rounding_upwards(size, rounding_multiplier);
1407 // format the input number, placing the appropriate symbol
1409 out.setf (ios::fixed);
1410 if (format == ShortUnitFormat)
1413 out << size << i18n( shortUnitFormatSymbols[sizecount].c_str() );
1418 out << size << i18n( longUnitFormatSymbols[sizecount].c_str() );
1422 } // eo nice_unit_format(int input)
1425 string nice_unit_format(
1427 const UnitFormat format,
1431 // round as double and cast to int64_t
1432 // cast raised overflow error near max val of int64_t (~9.2e18, see unittest)
1433 int64_t input_casted_and_rounded =
1434 boost::numeric_cast<int64_t>( round(input) );
1437 return nice_unit_format( input_casted_and_rounded, format, base );
1438 } // eo nice_unit_format(double input)
1441 string escape(const string &s)
1444 string::size_type p;
1447 while ( (p=out.find_first_of("\"\\",p) ) !=out.npos)
1449 out.insert (p,"\\");
1454 while ( (p=out.find_first_of("\r",p) ) !=out.npos)
1456 out.replace (p,1,"\\r");
1461 while ( (p=out.find_first_of("\n",p) ) !=out.npos)
1463 out.replace (p,1,"\\n");
1470 } // eo scape(const std::string&)
1473 string descape(const string &s, int startpos, int &endpos)
1477 if (s.at(startpos) != '"')
1478 throw out_of_range("value not type escaped string");
1480 out=s.substr(startpos+1);
1481 string::size_type p=0;
1483 // search for the end of the string
1484 while ( (p=out.find("\"",p) ) !=out.npos)
1489 // the " might be escaped with a backslash
1490 while (e>=0 && out.at (e) =='\\')
1492 if (escaped == false)
1506 // we now have the end of the string
1507 out=out.substr(0,p);
1509 // tell calling prog about the endposition
1510 endpos=startpos+p+1;
1512 // descape all \ stuff inside the string now
1514 while ( (p=out.find_first_of("\\",p) ) !=out.npos)
1516 switch (out.at(p+1) )
1519 out.replace(p,2,"\r");
1522 out.replace(p,2,"\n");
1531 } // eo descape(const std::string&,int,int&)
1534 string escape_shellarg(const string &input)
1536 string output = "'";
1537 string::const_iterator it, it_end = input.end();
1538 for (it = input.begin(); it != it_end; ++it)