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&)
433 std::string join_string (
434 const char *const parts[], /* assumed NULL-terminated */
435 const std::string& delimiter
442 const char *const *cur = parts;
445 result = std::string (*cur);
447 while (*++cur != NULL) {
449 result += std::string (*cur);
465 * @brief returns a hex string from a binary string.
466 * @param str the (binary) string
467 * @param upper_case_digits determine whether to use upper case characters for digits A-F.
468 * @return the string in hex notation.
470 std::string convert_binary_to_hex(
471 const std::string& str,
472 bool upper_case_digits
476 std::string hexDigits(upper_case_digits ? hexDigitsUpper : hexDigitsLower);
477 for ( std::string::const_iterator it= str.begin();
481 result.push_back( hexDigits[ ( (*it) >> 4) & 0x0f ] );
482 result.push_back( hexDigits[ (*it) & 0x0f ] );
485 } // eo convert_binary_to_hex(const std::string&,bool)
489 * @brief converts a hex digit string to binary string.
490 * @param str hex digit string
491 * @return the binary string.
493 * The hex digit string may contains white spaces or colons which are treated
494 * as delimiters between hex digit groups.
496 * @todo rework the handling of half nibbles (consistency)!
498 std::string convert_hex_to_binary(
499 const std::string& str
501 throw (std::runtime_error)
505 bool hasNibble= false;
506 bool lastWasWS= true;
507 for ( std::string::const_iterator it= str.begin();
511 std::string::size_type p = hexDigitsLower.find( *it );
512 if (p== std::string::npos)
514 p= hexDigitsUpper.find( *it );
516 if (p == std::string::npos)
518 if ( ( Whitespaces.find( *it ) != std::string::npos) // is it a whitespace?
519 or ( *it == ':') // or a colon?
522 // we treat that as a valid delimiter:
525 // 1 nibble before WS is treate as lower part:
534 if (p == std::string::npos )
536 throw runtime_error("illegal character in hex digit string: " + str);
550 //we already had a nibble, so a char is complete now:
551 result.push_back( c );
556 // this is the first nibble of a new char:
562 //well, there is one nibble left
563 // let's do some heuristics:
566 // if the preceeding character was a white space (or a colon)
567 // we treat the nibble as lower part:
568 //( this is consistent with shortened hex notations where leading zeros are not noted)
569 result.push_back( c );
573 // if it was part of a hex digit chain, we treat it as UPPER part (!!)
574 result.push_back( c << 4 );
578 } // eo convert_hex_to_binary(const std::string&)
581 static list<string>& alloc_template_starts()
583 static list<string> result;
586 result.push_back("std::list");
587 result.push_back("std::vector");
592 string shorten_stl_types(const string &input)
594 string output = input;
596 // first: replace fixed string for std::string
597 replace_all(output, "std::basic_string<char, std::char_traits<char>, std::allocator<char> >",
600 // loop over list/vector/... that have an allocator, e.g.
601 // std::list< some_type_here, std::allocator<some_type_here> >
602 string::size_type start, comma, end, len, start_text_len;
604 string allocator_text;
605 BOOST_FOREACH(const string &start_text, alloc_template_starts())
610 start_text_len = start_text.length();
611 while( (start=output.find(start_text+"<", start)) != string::npos )
613 len = output.length();
614 start += start_text_len+1; // start next iter and tests here after opening bracket
616 // now comes the tricky part: find matching ',' and the closing '>' even if "subtype" is template again
618 n_open_brackets = 1; // the bracket right after start_text counts as first
619 while (comma < len && n_open_brackets > 0)
621 if (output[comma] == ',' && n_open_brackets == 1)
623 else if (output[comma] == '<')
625 else if (output[comma] == '>')
630 while (end < len && n_open_brackets > 0)
632 if (output[end] == '<')
634 else if (output[end] == '>')
637 if (n_open_brackets == 0)
638 break; // do not increment end
643 // check that start < comma < end < len && n_open_brackets == 0
644 if (start >= comma || comma >= end || end >= len || n_open_brackets != 0)
645 continue; // input seems to be of unexpected form
647 // check that type in allocator is same as until comma
648 string type = output.substr(start, comma-start);
649 if (type[type.length()-1] == '>')
650 allocator_text = string("std::allocator<") + type + " > ";
652 allocator_text = string("std::allocator<") + type + "> ";
653 if (output.substr(comma+2, end-comma-2) == allocator_text)
654 output.replace(comma+2, end-comma-2, "_alloc_");
661 typedef boost::shared_ptr<BIO> BIO_Ptr;
664 * @brief Converts openssl generic input/output to std::string
666 * Code adapted from keymakerd.
668 * @param bio Openssl's generic input/output
669 * @return :string STL string
671 static std::string _convert_BIO_to_string(BIO *input)
676 long written = BIO_get_mem_data(input, &output);
677 if (written <= 0 || output == NULL)
680 rtn.assign(output, written); //lint !e534 !e732
685 * @brief base64 encode a string using OpenSSL base64 functions
687 * Data size limit is 2GB on 32 bit (LONG_MAX)
689 * @param input String to encode
690 * @param one_line Encode all data as one line, no wrapping with line feeds
691 * @return base64 encoded string
693 std::string base64_encode(const std::string &input, bool one_line)
695 // check for empty buffer
699 // safety check to ensure our check afer BIO_write() works
700 if (input.size() >= LONG_MAX)
701 throw runtime_error("base64 encode: Too much data");
703 // setup encoder. Note: BIO_free_all frees both BIOs.
704 BIO_Ptr base64_encoder(BIO_new(BIO_f_base64()), BIO_free_all);
705 BIO *encoder_bio = base64_encoder.get();
707 BIO_set_flags(encoder_bio, BIO_FLAGS_BASE64_NO_NL);
709 // chain output buffer and encoder together
710 BIO *encoded_result = BIO_new(BIO_s_mem());
711 BIO_push(encoder_bio, encoded_result);
714 long written = BIO_write(encoder_bio, input.c_str(), input.size());
715 if ((unsigned)written != input.size())
718 out << "base64 encoding failed: input size: "
719 << input.size() << " vs. output size: " << written;
720 throw runtime_error(out.str());
722 if (BIO_flush(encoder_bio) != 1)
723 throw runtime_error("base64 encode: BIO_flush() failed");
725 return _convert_BIO_to_string(encoded_result);
729 * @brief base64 decode a string using OpenSSL base64 functions
731 * @param input String to decode
732 * @param one_line Expect all base64 data in one line. Input with line feeds will fail.
733 * @return base64 decoded string
735 std::string base64_decode(const std::string &input, bool one_line)
737 // check for empty buffer
741 // safety check for BIO_new_mem_buf()
742 if (input.size() >= INT_MAX)
743 throw runtime_error("base64 decode: Too much data");
745 // setup encoder. Note: BIO_free_all frees both BIOs.
746 BIO_Ptr base64_decoder(BIO_new(BIO_f_base64()), BIO_free_all);
747 BIO *bio_base64 = base64_decoder.get();
749 BIO_set_flags(bio_base64, BIO_FLAGS_BASE64_NO_NL);
751 // chain input buffer and decoder together
752 BIO *bio_input = BIO_new_mem_buf((void*)input.c_str(), input.size());
753 bio_input = BIO_push(bio_base64, bio_input);
755 BIO_Ptr decoded_result(BIO_new(BIO_s_mem()), BIO_free_all);
756 BIO *bio_decoded = decoded_result.get();
757 const int convbuf_size = 512;
758 char convbuf[convbuf_size];
761 while((read_bytes = BIO_read(bio_input, convbuf, convbuf_size)) > 0)
763 BOOST_ASSERT(read_bytes <= convbuf_size);
764 long written_bytes = BIO_write(bio_decoded, convbuf, read_bytes);
765 if (written_bytes != read_bytes)
768 out << "base64 decoding failed: read_bytes: "
769 << read_bytes << " vs. written_bytes: " << written_bytes;
770 throw runtime_error(out.str());
773 if (read_bytes == -2 || read_bytes == -1)
774 throw runtime_error("base64 decode: Error during decoding");
776 return _convert_BIO_to_string(bio_decoded);
779 } // eo namespace I2n
784 std::string iso_to_utf8(const std::string& isostring)
788 iconv_t i2utf8 = iconv_open("UTF-8", "ISO-8859-1");
790 if (iso_to_utf8 == (iconv_t)-1)
791 throw runtime_error("iconv can't convert from ISO-8859-1 to UTF-8");
793 size_t in_size=isostring.size();
794 size_t out_size=in_size*4;
796 char *buf = (char *)malloc(out_size+1);
798 throw runtime_error("out of memory for iconv buffer");
800 char *in = (char *)isostring.c_str();
802 iconv(i2utf8, &in, &in_size, &out, &out_size);
804 buf[isostring.size()*4-out_size]=0;
814 std::string utf8_to_iso(const std::string& utf8string)
818 iconv_t utf82iso = iconv_open("ISO-8859-1","UTF-8");
820 if (utf82iso == (iconv_t)-1)
821 throw runtime_error("iconv can't convert from UTF-8 to ISO-8859-1");
823 size_t in_size=utf8string.size();
824 size_t out_size=in_size;
826 char *buf = (char *)malloc(out_size+1);
828 throw runtime_error("out of memory for iconv buffer");
830 char *in = (char *)utf8string.c_str();
832 iconv(utf82iso, &in, &in_size, &out, &out_size);
834 buf[utf8string.size()-out_size]=0;
839 iconv_close(utf82iso);
844 wchar_t* utf8_to_wbuf(const std::string& utf8string)
846 iconv_t utf82wstr = iconv_open("UCS-4LE","UTF-8");
848 if (utf82wstr == (iconv_t)-1)
849 throw runtime_error("iconv can't convert from UTF-8 to UCS-4");
851 size_t in_size=utf8string.size();
852 size_t out_size= (in_size+1)*sizeof(wchar_t);
854 wchar_t *buf = (wchar_t *)malloc(out_size);
856 throw runtime_error("out of memory for iconv buffer");
858 char *in = (char *)utf8string.c_str();
859 char *out = (char*) buf;
860 if (iconv(utf82wstr, &in, &in_size, &out, &out_size) == (size_t)-1)
861 throw runtime_error("error converting char encodings");
863 buf[ ( (utf8string.size()+1)*sizeof(wchar_t)-out_size) /sizeof(wchar_t) ]=0;
865 iconv_close(utf82wstr);
870 std::string utf7imap_to_utf8(const std::string& utf7imapstring)
874 iconv_t utf7imap2utf8 = iconv_open("UTF-8","UTF-7-IMAP");
876 if (utf7imap2utf8 == (iconv_t)-1)
877 throw runtime_error("iconv can't convert from UTF-7-IMAP to UTF-8");
879 size_t in_size=utf7imapstring.size();
880 size_t out_size=in_size*4;
882 char *buf = (char *)malloc(out_size+1);
884 throw runtime_error("out of memory for iconv buffer");
886 char *in = (char *)utf7imapstring.c_str();
888 iconv(utf7imap2utf8, &in, &in_size, &out, &out_size);
890 buf[utf7imapstring.size()*4-out_size]=0;
895 iconv_close(utf7imap2utf8);
900 std::string utf8_to_utf7imap(const std::string& utf8string)
904 iconv_t utf82utf7imap = iconv_open("UTF-7-IMAP", "UTF-8");
906 if (utf82utf7imap == (iconv_t)-1)
907 throw runtime_error("iconv can't convert from UTF-7-IMAP to UTF-8");
909 // UTF-7 is base64 encoded, a buffer 10x as large
910 // as the utf-8 buffer should be enough. If not the string will be truncated.
911 size_t in_size=utf8string.size();
912 size_t out_size=in_size*10;
914 char *buf = (char *)malloc(out_size+1);
916 throw runtime_error("out of memory for iconv buffer");
918 char *in = (char *)utf8string.c_str();
920 iconv(utf82utf7imap, &in, &in_size, &out, &out_size);
922 buf[utf8string.size()*10-out_size]= 0;
927 iconv_close(utf82utf7imap);
932 // Tokenize string by (html) tags
933 void tokenize_by_tag(vector<pair<string,bool> > &tokenized, const std::string &input)
935 string::size_type pos, len = input.size();
936 bool inside_tag = false;
939 for (pos = 0; pos < len; pos++)
941 if (input[pos] == '<')
945 if (!current.empty() )
947 tokenized.push_back( make_pair(current, false) );
951 current += input[pos];
953 else if (input[pos] == '>' && inside_tag)
955 current += input[pos];
957 if (!current.empty() )
959 tokenized.push_back( make_pair(current, true) );
964 current += input[pos];
967 // String left over in buffer?
968 if (!current.empty() )
969 tokenized.push_back( make_pair(current, false) );
970 } // eo tokenize_by_tag
973 std::string strip_html_tags(const std::string &input)
975 // Pair first: string, second: isTag
976 vector<pair<string,bool> > tokenized;
977 tokenize_by_tag (tokenized, input);
980 vector<pair<string,bool> >::const_iterator token, tokens_end = tokenized.end();
981 for (token = tokenized.begin(); token != tokens_end; ++token)
983 output += token->first;
986 } // eo strip_html_tags
989 // Smart-encode HTML en
990 string smart_html_entities(const std::string &input)
992 // Pair first: string, second: isTag
993 vector<pair<string,bool> > tokenized;
994 tokenize_by_tag (tokenized, input);
997 vector<pair<string,bool> >::const_iterator token, tokens_end = tokenized.end();
998 for (token = tokenized.begin(); token != tokens_end; ++token)
1000 // keep HTML tags as they are
1002 output += token->first;
1004 output += html_entities(token->first);
1011 string::size_type find_8bit(const std::string &str)
1013 string::size_type l=str.size();
1014 for (string::size_type p=0; p < l; p++)
1015 if (static_cast<unsigned char>(str[p]) > 127)
1018 return string::npos;
1021 // encoded UTF-8 chars into HTML entities
1022 string html_entities(std::string str)
1025 replace_all (str, "&", "&");
1026 replace_all (str, "<", "<");
1027 replace_all (str, ">", ">");
1028 replace_all (str, "\"", """);
1029 replace_all (str, "'", "'");
1030 replace_all (str, "/", "/");
1033 replace_all (str, "\xC3\xA4", "ä");
1034 replace_all (str, "\xC3\xB6", "ö");
1035 replace_all (str, "\xC3\xBC", "ü");
1036 replace_all (str, "\xC3\x84", "Ä");
1037 replace_all (str, "\xC3\x96", "Ö");
1038 replace_all (str, "\xC3\x9C", "Ü");
1041 replace_all (str, "\xC3\x9F", "ß");
1043 // conversion of remaining non-ASCII chars needed?
1044 // just do if needed because of performance
1045 if (find_8bit(str) != string::npos)
1047 // convert to fixed-size encoding UTF-32
1048 wchar_t* wbuf=utf8_to_wbuf(str);
1049 ostringstream target;
1051 // replace all non-ASCII chars with HTML representation
1052 for (int p=0; wbuf[p] != 0; p++)
1054 unsigned int c=wbuf[p];
1057 target << static_cast<unsigned char>(c);
1059 target << "&#" << c << ';';
1068 } // eo html_entities(std::string)
1070 // convert HTML entities to something that can be viewed on a basic text console (restricted to ASCII-7)
1071 string html_entities_to_console(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, "ä", "ae");
1083 replace_all (str, "ö", "oe");
1084 replace_all (str, "ü", "ue");
1085 replace_all (str, "Ä", "Ae");
1086 replace_all (str, "Ö", "Oe");
1087 replace_all (str, "Ü", "Ue");
1090 replace_all (str, "ß", "ss");
1095 // find_html_comments + remove_html_comments(str, comments)
1096 void remove_html_comments(string &str)
1098 vector<CommentZone> comments = find_html_comments(str);
1099 remove_html_comments(str, comments);
1102 // find all html comments, behaving correctly if they are nested; ignores comment tags ("<!--FOO .... BAR-->")
1103 // If there are invalid comments ("-->" before "<!--" or different number of closing and opening tags),
1104 // then the unknown index of corresponding start/end tag will be represented by a string::npos
1105 // Indices are from start of start tag until first index after closing tag
1106 vector<CommentZone> find_html_comments(const std::string &str)
1108 static const string START = "<!--";
1109 static const string CLOSE = "-->";
1110 static const string::size_type START_LEN = START.length();
1111 static const string::size_type CLOSE_LEN = CLOSE.length();
1113 vector<CommentZone> comments;
1115 // in order to find nested comments, need either recursion or a stack
1116 vector<string::size_type> starts; // stack of start tags
1118 string::size_type pos = 0;
1119 string::size_type len = str.length();
1120 string::size_type next_start, next_close;
1122 while (pos < len) // not really needed but just in case
1124 next_start = str.find(START, pos);
1125 next_close = str.find(CLOSE, pos);
1127 if ( (next_start == string::npos) && (next_close == string::npos) )
1128 break; // we are done
1130 else if ( (next_start == string::npos) || (next_close < next_start) ) // close one comment (pop)
1132 if (starts.empty()) // closing tag without a start
1133 comments.push_back(CommentZone(string::npos, next_close+CLOSE_LEN));
1136 comments.push_back(CommentZone(starts.back(), next_close+CLOSE_LEN));
1139 pos = next_close + CLOSE_LEN;
1142 else if ( (next_close == string::npos) || (next_start < next_close) ) // start a new comment (push)
1144 starts.push_back(next_start);
1145 pos = next_start + START_LEN;
1149 // add comments that have no closing tag from back to front (important for remove_html_comments!)
1150 while (!starts.empty())
1152 comments.push_back(CommentZone(starts.back(), string::npos));
1159 // remove all html comments foundby find_html_comments
1160 void remove_html_comments(std::string &str, const vector<CommentZone> &comments)
1162 // remember position where last removal started
1163 string::size_type last_removal_start = str.length();
1165 // Go from back to front to not mess up indices.
1166 // This requires that bigger comments, that contain smaller comments, come AFTER
1167 // the small contained comments in the comments vector (i.e. comments are ordered by
1168 // their closing tag, not their opening tag). This is true for results from find_html_comments
1169 BOOST_REVERSE_FOREACH(const CommentZone &comment, comments)
1171 if (comment.first == string::npos)
1173 str = str.replace(0, comment.second, ""); // comment starts "before" str --> delete from start
1174 break; // there can be no more
1176 else if (comment.first >= last_removal_start)
1178 continue; // this comment is inside another comment that we have removed already
1180 else if (comment.second == string::npos) // comment ends "after" str --> delete until end
1182 str = str.replace(comment.first, string::npos, "");
1183 last_removal_start = comment.first;
1187 str = str.replace(comment.first, comment.second-comment.first, "");
1188 last_removal_start = comment.first;
1193 bool replace_all(string &base, const char *ist, const char *soll)
1197 return replace_all(base,&i,&s);
1200 bool replace_all(string &base, const string &ist, const char *soll)
1203 return replace_all(base,&ist,&s);
1206 bool replace_all(string &base, const string *ist, const string *soll)
1208 return replace_all(base,*ist,*soll);
1211 bool replace_all(string &base, const char *ist, const string *soll)
1214 return replace_all(base,&i,soll);
1217 bool replace_all(string &base, const string &ist, const string &soll)
1219 bool found_ist = false;
1220 string::size_type a=0;
1223 throw runtime_error ("replace_all called with empty search string");
1225 while ( (a=base.find(ist,a) ) != string::npos)
1227 base.replace(a,ist.size(),soll);
1236 * @brief replaces all characters that could be problematic or impose a security risk when being logged
1237 * @param str the original string
1238 * @param replace_with the character to replace the unsafe chars with
1239 * @return a string that is safe to send to syslog or other logfiles
1241 * All chars between 0x20 (space) and 0x7E (~) (including) are considered safe for logging.
1242 * See e.g. RFC 5424, section 8.2 or the posix character class "printable".
1243 * This eliminates all possible problems with NUL, control characters, 8 bit chars, UTF8.
1246 std::string sanitize_for_logging(const std::string &str, const char replace_with)
1248 std::string output=str;
1250 const string::size_type len = output.size();
1251 for (std::string::size_type p=0; p < len; p++)
1252 if (output[p] < 0x20 || output[p] > 0x7E)
1253 output[p]=replace_with;
1259 string to_lower(const string &src)
1263 string::size_type pos, end = dst.size();
1264 for (pos = 0; pos < end; pos++)
1265 dst[pos] = tolower(dst[pos]);
1270 string to_upper(const string &src)
1274 string::size_type pos, end = dst.size();
1275 for (pos = 0; pos < end; pos++)
1276 dst[pos] = toupper(dst[pos]);
1282 const int MAX_UNIT_FORMAT_SYMBOLS = 6;
1284 const string shortUnitFormatSymbols[MAX_UNIT_FORMAT_SYMBOLS] = {
1293 const string longUnitFormatSymbols[MAX_UNIT_FORMAT_SYMBOLS] = {
1294 i18n_noop(" Bytes"),
1295 i18n_noop(" KBytes"),
1296 i18n_noop(" MBytes"),
1297 i18n_noop(" GBytes"),
1298 i18n_noop(" TBytes"),
1299 i18n_noop(" PBytes")
1303 static long double rounding_upwards(
1304 const long double number,
1305 const int rounding_multiplier
1308 long double rounded_number;
1309 rounded_number = number * rounding_multiplier;
1310 rounded_number += 0.5;
1311 rounded_number = (int64_t) (rounded_number);
1312 rounded_number = (long double) (rounded_number) / (long double) (rounding_multiplier);
1314 return rounded_number;
1318 string nice_unit_format(
1319 const int64_t input,
1320 const UnitFormat format,
1324 // select the system of units (decimal or binary)
1326 if (base == UnitBase1000)
1335 long double size = input;
1337 // check the size of the input number to fit in the appropriate symbol
1339 while (size > multiple)
1341 size = size / multiple;
1344 // rollback to the previous values and stop the loop when cannot
1345 // represent the number length.
1346 if (sizecount >= MAX_UNIT_FORMAT_SYMBOLS)
1348 size = size * multiple;
1354 // round the input number "half up" to multiples of 10
1355 const int rounding_multiplier = 10;
1356 size = rounding_upwards(size, rounding_multiplier);
1358 // format the input number, placing the appropriate symbol
1360 out.setf (ios::fixed);
1361 if (format == ShortUnitFormat)
1364 out << size << i18n( shortUnitFormatSymbols[sizecount].c_str() );
1369 out << size << i18n( longUnitFormatSymbols[sizecount].c_str() );
1373 } // eo nice_unit_format(int input)
1376 string nice_unit_format(
1378 const UnitFormat format,
1382 // round as double and cast to int64_t
1383 // cast raised overflow error near max val of int64_t (~9.2e18, see unittest)
1384 int64_t input_casted_and_rounded =
1385 boost::numeric_cast<int64_t>( round(input) );
1388 return nice_unit_format( input_casted_and_rounded, format, base );
1389 } // eo nice_unit_format(double input)
1392 string escape(const string &s)
1395 string::size_type p;
1398 while ( (p=out.find_first_of("\"\\",p) ) !=out.npos)
1400 out.insert (p,"\\");
1405 while ( (p=out.find_first_of("\r",p) ) !=out.npos)
1407 out.replace (p,1,"\\r");
1412 while ( (p=out.find_first_of("\n",p) ) !=out.npos)
1414 out.replace (p,1,"\\n");
1421 } // eo scape(const std::string&)
1424 string descape(const string &s, int startpos, int &endpos)
1428 if (s.at(startpos) != '"')
1429 throw out_of_range("value not type escaped string");
1431 out=s.substr(startpos+1);
1432 string::size_type p=0;
1434 // search for the end of the string
1435 while ( (p=out.find("\"",p) ) !=out.npos)
1440 // the " might be escaped with a backslash
1441 while (e>=0 && out.at (e) =='\\')
1443 if (escaped == false)
1457 // we now have the end of the string
1458 out=out.substr(0,p);
1460 // tell calling prog about the endposition
1461 endpos=startpos+p+1;
1463 // descape all \ stuff inside the string now
1465 while ( (p=out.find_first_of("\\",p) ) !=out.npos)
1467 switch (out.at(p+1) )
1470 out.replace(p,2,"\r");
1473 out.replace(p,2,"\n");
1482 } // eo descape(const std::string&,int,int&)
1485 string escape_shellarg(const string &input)
1487 string output = "'";
1488 string::const_iterator it, it_end = input.end();
1489 for (it = input.begin(); it != it_end; ++it)