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&)
490 * @brief returns a hex string from a binary string.
491 * @param str the (binary) string
492 * @param upper_case_digits determine whether to use upper case characters for digits A-F.
493 * @return the string in hex notation.
495 std::string convert_binary_to_hex(
496 const std::string& str,
497 bool upper_case_digits
501 std::string hexDigits(upper_case_digits ? hexDigitsUpper : hexDigitsLower);
502 for ( std::string::const_iterator it= str.begin();
506 result.push_back( hexDigits[ ( (*it) >> 4) & 0x0f ] );
507 result.push_back( hexDigits[ (*it) & 0x0f ] );
510 } // eo convert_binary_to_hex(const std::string&,bool)
514 * @brief converts a hex digit string to binary string.
515 * @param str hex digit string
516 * @return the binary string.
518 * The hex digit string may contains white spaces or colons which are treated
519 * as delimiters between hex digit groups.
521 * @todo rework the handling of half nibbles (consistency)!
523 std::string convert_hex_to_binary(
524 const std::string& str
526 throw (std::runtime_error)
530 bool hasNibble= false;
531 bool lastWasWS= true;
532 for ( std::string::const_iterator it= str.begin();
536 std::string::size_type p = hexDigitsLower.find( *it );
537 if (p== std::string::npos)
539 p= hexDigitsUpper.find( *it );
541 if (p == std::string::npos)
543 if ( ( Whitespaces.find( *it ) != std::string::npos) // is it a whitespace?
544 or ( *it == ':') // or a colon?
547 // we treat that as a valid delimiter:
550 // 1 nibble before WS is treate as lower part:
559 if (p == std::string::npos )
561 throw runtime_error("illegal character in hex digit string: " + str);
575 //we already had a nibble, so a char is complete now:
576 result.push_back( c );
581 // this is the first nibble of a new char:
587 //well, there is one nibble left
588 // let's do some heuristics:
591 // if the preceeding character was a white space (or a colon)
592 // we treat the nibble as lower part:
593 //( this is consistent with shortened hex notations where leading zeros are not noted)
594 result.push_back( c );
598 // if it was part of a hex digit chain, we treat it as UPPER part (!!)
599 result.push_back( c << 4 );
603 } // eo convert_hex_to_binary(const std::string&)
606 static list<string>& alloc_template_starts()
608 static list<string> result;
611 result.push_back("std::list");
612 result.push_back("std::vector");
617 string shorten_stl_types(const string &input)
619 string output = input;
621 // first: replace fixed string for std::string
622 replace_all(output, "std::basic_string<char, std::char_traits<char>, std::allocator<char> >",
625 // loop over list/vector/... that have an allocator, e.g.
626 // std::list< some_type_here, std::allocator<some_type_here> >
627 string::size_type start, comma, end, len, start_text_len;
629 string allocator_text;
630 BOOST_FOREACH(const string &start_text, alloc_template_starts())
635 start_text_len = start_text.length();
636 while( (start=output.find(start_text+"<", start)) != string::npos )
638 len = output.length();
639 start += start_text_len+1; // start next iter and tests here after opening bracket
641 // now comes the tricky part: find matching ',' and the closing '>' even if "subtype" is template again
643 n_open_brackets = 1; // the bracket right after start_text counts as first
644 while (comma < len && n_open_brackets > 0)
646 if (output[comma] == ',' && n_open_brackets == 1)
648 else if (output[comma] == '<')
650 else if (output[comma] == '>')
655 while (end < len && n_open_brackets > 0)
657 if (output[end] == '<')
659 else if (output[end] == '>')
662 if (n_open_brackets == 0)
663 break; // do not increment end
668 // check that start < comma < end < len && n_open_brackets == 0
669 if (start >= comma || comma >= end || end >= len || n_open_brackets != 0)
670 continue; // input seems to be of unexpected form
672 // check that type in allocator is same as until comma
673 string type = output.substr(start, comma-start);
674 if (type[type.length()-1] == '>')
675 allocator_text = string("std::allocator<") + type + " > ";
677 allocator_text = string("std::allocator<") + type + "> ";
678 if (output.substr(comma+2, end-comma-2) == allocator_text)
679 output.replace(comma+2, end-comma-2, "_alloc_");
686 typedef boost::shared_ptr<BIO> BIO_Ptr;
689 * @brief Converts openssl generic input/output to std::string
691 * Code adapted from keymakerd.
693 * @param bio Openssl's generic input/output
694 * @return :string STL string
696 static std::string _convert_BIO_to_string(BIO *input)
701 long written = BIO_get_mem_data(input, &output);
702 if (written <= 0 || output == NULL)
705 rtn.assign(output, written); //lint !e534 !e732
710 * @brief base64 encode a string using OpenSSL base64 functions
712 * Data size limit is 2GB on 32 bit (LONG_MAX)
714 * @param input String to encode
715 * @param one_line Encode all data as one line, no wrapping with line feeds
716 * @return base64 encoded string
718 std::string base64_encode(const std::string &input, bool one_line)
720 // check for empty buffer
724 // safety check to ensure our check afer BIO_write() works
725 if (input.size() >= LONG_MAX)
726 throw runtime_error("base64 encode: Too much data");
728 // setup encoder. Note: BIO_free_all frees both BIOs.
729 BIO_Ptr base64_encoder(BIO_new(BIO_f_base64()), BIO_free_all);
730 BIO *encoder_bio = base64_encoder.get();
732 BIO_set_flags(encoder_bio, BIO_FLAGS_BASE64_NO_NL);
734 // chain output buffer and encoder together
735 BIO *encoded_result = BIO_new(BIO_s_mem());
736 BIO_push(encoder_bio, encoded_result);
739 long written = BIO_write(encoder_bio, input.c_str(), input.size());
740 if ((unsigned)written != input.size())
743 out << "base64 encoding failed: input size: "
744 << input.size() << " vs. output size: " << written;
745 throw runtime_error(out.str());
747 if (BIO_flush(encoder_bio) != 1)
748 throw runtime_error("base64 encode: BIO_flush() failed");
750 return _convert_BIO_to_string(encoded_result);
754 * @brief base64 decode a string using OpenSSL base64 functions
756 * @param input String to decode
757 * @param one_line Expect all base64 data in one line. Input with line feeds will fail.
758 * @return base64 decoded string
760 std::string base64_decode(const std::string &input, bool one_line)
762 // check for empty buffer
766 // safety check for BIO_new_mem_buf()
767 if (input.size() >= INT_MAX)
768 throw runtime_error("base64 decode: Too much data");
770 // setup encoder. Note: BIO_free_all frees both BIOs.
771 BIO_Ptr base64_decoder(BIO_new(BIO_f_base64()), BIO_free_all);
772 BIO *bio_base64 = base64_decoder.get();
774 BIO_set_flags(bio_base64, BIO_FLAGS_BASE64_NO_NL);
776 // chain input buffer and decoder together
777 BIO *bio_input = BIO_new_mem_buf((void*)input.c_str(), input.size());
778 bio_input = BIO_push(bio_base64, bio_input);
780 BIO_Ptr decoded_result(BIO_new(BIO_s_mem()), BIO_free_all);
781 BIO *bio_decoded = decoded_result.get();
782 const int convbuf_size = 512;
783 char convbuf[convbuf_size];
786 while((read_bytes = BIO_read(bio_input, convbuf, convbuf_size)) > 0)
788 BOOST_ASSERT(read_bytes <= convbuf_size);
789 long written_bytes = BIO_write(bio_decoded, convbuf, read_bytes);
790 if (written_bytes != read_bytes)
793 out << "base64 decoding failed: read_bytes: "
794 << read_bytes << " vs. written_bytes: " << written_bytes;
795 throw runtime_error(out.str());
798 if (read_bytes == -2 || read_bytes == -1)
799 throw runtime_error("base64 decode: Error during decoding");
801 return _convert_BIO_to_string(bio_decoded);
804 } // eo namespace I2n
809 std::string iso_to_utf8(const std::string& isostring)
813 iconv_t i2utf8 = iconv_open("UTF-8", "ISO-8859-1");
815 if (iso_to_utf8 == (iconv_t)-1)
816 throw runtime_error("iconv can't convert from ISO-8859-1 to UTF-8");
818 size_t in_size=isostring.size();
819 size_t out_size=in_size*4;
821 char *buf = (char *)malloc(out_size+1);
823 throw runtime_error("out of memory for iconv buffer");
825 char *in = (char *)isostring.c_str();
827 iconv(i2utf8, &in, &in_size, &out, &out_size);
829 buf[isostring.size()*4-out_size]=0;
839 std::string utf8_to_iso(const std::string& utf8string)
843 iconv_t utf82iso = iconv_open("ISO-8859-1","UTF-8");
845 if (utf82iso == (iconv_t)-1)
846 throw runtime_error("iconv can't convert from UTF-8 to ISO-8859-1");
848 size_t in_size=utf8string.size();
849 size_t out_size=in_size;
851 char *buf = (char *)malloc(out_size+1);
853 throw runtime_error("out of memory for iconv buffer");
855 char *in = (char *)utf8string.c_str();
857 iconv(utf82iso, &in, &in_size, &out, &out_size);
859 buf[utf8string.size()-out_size]=0;
864 iconv_close(utf82iso);
869 wchar_t* utf8_to_wbuf(const std::string& utf8string)
871 iconv_t utf82wstr = iconv_open("UCS-4LE","UTF-8");
873 if (utf82wstr == (iconv_t)-1)
874 throw runtime_error("iconv can't convert from UTF-8 to UCS-4");
876 size_t in_size=utf8string.size();
877 size_t out_size= (in_size+1)*sizeof(wchar_t);
879 wchar_t *buf = (wchar_t *)malloc(out_size);
881 throw runtime_error("out of memory for iconv buffer");
883 char *in = (char *)utf8string.c_str();
884 char *out = (char*) buf;
885 if (iconv(utf82wstr, &in, &in_size, &out, &out_size) == (size_t)-1)
886 throw runtime_error("error converting char encodings");
888 buf[ ( (utf8string.size()+1)*sizeof(wchar_t)-out_size) /sizeof(wchar_t) ]=0;
890 iconv_close(utf82wstr);
895 std::string utf7imap_to_utf8(const std::string& utf7imapstring)
899 iconv_t utf7imap2utf8 = iconv_open("UTF-8","UTF-7-IMAP");
901 if (utf7imap2utf8 == (iconv_t)-1)
902 throw runtime_error("iconv can't convert from UTF-7-IMAP to UTF-8");
904 size_t in_size=utf7imapstring.size();
905 size_t out_size=in_size*4;
907 char *buf = (char *)malloc(out_size+1);
909 throw runtime_error("out of memory for iconv buffer");
911 char *in = (char *)utf7imapstring.c_str();
913 iconv(utf7imap2utf8, &in, &in_size, &out, &out_size);
915 buf[utf7imapstring.size()*4-out_size]=0;
920 iconv_close(utf7imap2utf8);
925 std::string utf8_to_utf7imap(const std::string& utf8string)
929 iconv_t utf82utf7imap = iconv_open("UTF-7-IMAP", "UTF-8");
931 if (utf82utf7imap == (iconv_t)-1)
932 throw runtime_error("iconv can't convert from UTF-7-IMAP to UTF-8");
934 // UTF-7 is base64 encoded, a buffer 10x as large
935 // as the utf-8 buffer should be enough. If not the string will be truncated.
936 size_t in_size=utf8string.size();
937 size_t out_size=in_size*10;
939 char *buf = (char *)malloc(out_size+1);
941 throw runtime_error("out of memory for iconv buffer");
943 char *in = (char *)utf8string.c_str();
945 iconv(utf82utf7imap, &in, &in_size, &out, &out_size);
947 buf[utf8string.size()*10-out_size]= 0;
952 iconv_close(utf82utf7imap);
957 // Tokenize string by (html) tags
958 void tokenize_by_tag(vector<pair<string,bool> > &tokenized, const std::string &input)
960 string::size_type pos, len = input.size();
961 bool inside_tag = false;
964 for (pos = 0; pos < len; pos++)
966 if (input[pos] == '<')
970 if (!current.empty() )
972 tokenized.push_back( make_pair(current, false) );
976 current += input[pos];
978 else if (input[pos] == '>' && inside_tag)
980 current += input[pos];
982 if (!current.empty() )
984 tokenized.push_back( make_pair(current, true) );
989 current += input[pos];
992 // String left over in buffer?
993 if (!current.empty() )
994 tokenized.push_back( make_pair(current, false) );
995 } // eo tokenize_by_tag
998 std::string strip_html_tags(const std::string &input)
1000 // Pair first: string, second: isTag
1001 vector<pair<string,bool> > tokenized;
1002 tokenize_by_tag (tokenized, input);
1005 vector<pair<string,bool> >::const_iterator token, tokens_end = tokenized.end();
1006 for (token = tokenized.begin(); token != tokens_end; ++token)
1008 output += token->first;
1011 } // eo strip_html_tags
1014 // Smart-encode HTML en
1015 string smart_html_entities(const std::string &input)
1017 // Pair first: string, second: isTag
1018 vector<pair<string,bool> > tokenized;
1019 tokenize_by_tag (tokenized, input);
1022 vector<pair<string,bool> >::const_iterator token, tokens_end = tokenized.end();
1023 for (token = tokenized.begin(); token != tokens_end; ++token)
1025 // keep HTML tags as they are
1027 output += token->first;
1029 output += html_entities(token->first);
1036 string::size_type find_8bit(const std::string &str)
1038 string::size_type l=str.size();
1039 for (string::size_type p=0; p < l; p++)
1040 if (static_cast<unsigned char>(str[p]) > 127)
1043 return string::npos;
1046 // encoded UTF-8 chars into HTML entities
1047 string html_entities(std::string str)
1050 replace_all (str, "&", "&");
1051 replace_all (str, "<", "<");
1052 replace_all (str, ">", ">");
1053 replace_all (str, "\"", """);
1054 replace_all (str, "'", "'");
1055 replace_all (str, "/", "/");
1058 replace_all (str, "\xC3\xA4", "ä");
1059 replace_all (str, "\xC3\xB6", "ö");
1060 replace_all (str, "\xC3\xBC", "ü");
1061 replace_all (str, "\xC3\x84", "Ä");
1062 replace_all (str, "\xC3\x96", "Ö");
1063 replace_all (str, "\xC3\x9C", "Ü");
1066 replace_all (str, "\xC3\x9F", "ß");
1068 // conversion of remaining non-ASCII chars needed?
1069 // just do if needed because of performance
1070 if (find_8bit(str) != string::npos)
1072 // convert to fixed-size encoding UTF-32
1073 wchar_t* wbuf=utf8_to_wbuf(str);
1074 ostringstream target;
1076 // replace all non-ASCII chars with HTML representation
1077 for (int p=0; wbuf[p] != 0; p++)
1079 unsigned int c=wbuf[p];
1082 target << static_cast<unsigned char>(c);
1084 target << "&#" << c << ';';
1093 } // eo html_entities(std::string)
1095 // convert HTML entities to something that can be viewed on a basic text console (restricted to ASCII-7)
1096 string html_entities_to_console(std::string str)
1099 replace_all (str, "&", "&");
1100 replace_all (str, "<", "<");
1101 replace_all (str, ">", ">");
1102 replace_all (str, """, "\"");
1103 replace_all (str, "'", "'");
1104 replace_all (str, "/", "/");
1107 replace_all (str, "ä", "ae");
1108 replace_all (str, "ö", "oe");
1109 replace_all (str, "ü", "ue");
1110 replace_all (str, "Ä", "Ae");
1111 replace_all (str, "Ö", "Oe");
1112 replace_all (str, "Ü", "Ue");
1115 replace_all (str, "ß", "ss");
1120 // find_html_comments + remove_html_comments(str, comments)
1121 void remove_html_comments(string &str)
1123 vector<CommentZone> comments = find_html_comments(str);
1124 remove_html_comments(str, comments);
1127 // find all html comments, behaving correctly if they are nested; ignores comment tags ("<!--FOO .... BAR-->")
1128 // If there are invalid comments ("-->" before "<!--" or different number of closing and opening tags),
1129 // then the unknown index of corresponding start/end tag will be represented by a string::npos
1130 // Indices are from start of start tag until first index after closing tag
1131 vector<CommentZone> find_html_comments(const std::string &str)
1133 static const string START = "<!--";
1134 static const string CLOSE = "-->";
1135 static const string::size_type START_LEN = START.length();
1136 static const string::size_type CLOSE_LEN = CLOSE.length();
1138 vector<CommentZone> comments;
1140 // in order to find nested comments, need either recursion or a stack
1141 vector<string::size_type> starts; // stack of start tags
1143 string::size_type pos = 0;
1144 string::size_type len = str.length();
1145 string::size_type next_start, next_close;
1147 while (pos < len) // not really needed but just in case
1149 next_start = str.find(START, pos);
1150 next_close = str.find(CLOSE, pos);
1152 if ( (next_start == string::npos) && (next_close == string::npos) )
1153 break; // we are done
1155 else if ( (next_start == string::npos) || (next_close < next_start) ) // close one comment (pop)
1157 if (starts.empty()) // closing tag without a start
1158 comments.push_back(CommentZone(string::npos, next_close+CLOSE_LEN));
1161 comments.push_back(CommentZone(starts.back(), next_close+CLOSE_LEN));
1164 pos = next_close + CLOSE_LEN;
1167 else if ( (next_close == string::npos) || (next_start < next_close) ) // start a new comment (push)
1169 starts.push_back(next_start);
1170 pos = next_start + START_LEN;
1174 // add comments that have no closing tag from back to front (important for remove_html_comments!)
1175 while (!starts.empty())
1177 comments.push_back(CommentZone(starts.back(), string::npos));
1184 // remove all html comments foundby find_html_comments
1185 void remove_html_comments(std::string &str, const vector<CommentZone> &comments)
1187 // remember position where last removal started
1188 string::size_type last_removal_start = str.length();
1190 // Go from back to front to not mess up indices.
1191 // This requires that bigger comments, that contain smaller comments, come AFTER
1192 // the small contained comments in the comments vector (i.e. comments are ordered by
1193 // their closing tag, not their opening tag). This is true for results from find_html_comments
1194 BOOST_REVERSE_FOREACH(const CommentZone &comment, comments)
1196 if (comment.first == string::npos)
1198 str = str.replace(0, comment.second, ""); // comment starts "before" str --> delete from start
1199 break; // there can be no more
1201 else if (comment.first >= last_removal_start)
1203 continue; // this comment is inside another comment that we have removed already
1205 else if (comment.second == string::npos) // comment ends "after" str --> delete until end
1207 str = str.replace(comment.first, string::npos, "");
1208 last_removal_start = comment.first;
1212 str = str.replace(comment.first, comment.second-comment.first, "");
1213 last_removal_start = comment.first;
1218 bool replace_all(string &base, const char *ist, const char *soll)
1222 return replace_all(base,&i,&s);
1225 bool replace_all(string &base, const string &ist, const char *soll)
1228 return replace_all(base,&ist,&s);
1231 bool replace_all(string &base, const string *ist, const string *soll)
1233 return replace_all(base,*ist,*soll);
1236 bool replace_all(string &base, const char *ist, const string *soll)
1239 return replace_all(base,&i,soll);
1242 bool replace_all(string &base, const string &ist, const string &soll)
1244 bool found_ist = false;
1245 string::size_type a=0;
1248 throw runtime_error ("replace_all called with empty search string");
1250 while ( (a=base.find(ist,a) ) != string::npos)
1252 base.replace(a,ist.size(),soll);
1261 * @brief replaces all characters that could be problematic or impose a security risk when being logged
1262 * @param str the original string
1263 * @param replace_with the character to replace the unsafe chars with
1264 * @return a string that is safe to send to syslog or other logfiles
1266 * All chars between 0x20 (space) and 0x7E (~) (including) are considered safe for logging.
1267 * See e.g. RFC 5424, section 8.2 or the posix character class "printable".
1268 * This eliminates all possible problems with NUL, control characters, 8 bit chars, UTF8.
1271 std::string sanitize_for_logging(const std::string &str, const char replace_with)
1273 std::string output=str;
1275 const string::size_type len = output.size();
1276 for (std::string::size_type p=0; p < len; p++)
1277 if (output[p] < 0x20 || output[p] > 0x7E)
1278 output[p]=replace_with;
1284 string to_lower(const string &src)
1288 string::size_type pos, end = dst.size();
1289 for (pos = 0; pos < end; pos++)
1290 dst[pos] = tolower(dst[pos]);
1295 string to_upper(const string &src)
1299 string::size_type pos, end = dst.size();
1300 for (pos = 0; pos < end; pos++)
1301 dst[pos] = toupper(dst[pos]);
1307 const int MAX_UNIT_FORMAT_SYMBOLS = 6;
1309 const string shortUnitFormatSymbols[MAX_UNIT_FORMAT_SYMBOLS] = {
1318 const string longUnitFormatSymbols[MAX_UNIT_FORMAT_SYMBOLS] = {
1319 i18n_noop(" Bytes"),
1320 i18n_noop(" KBytes"),
1321 i18n_noop(" MBytes"),
1322 i18n_noop(" GBytes"),
1323 i18n_noop(" TBytes"),
1324 i18n_noop(" PBytes")
1328 static long double rounding_upwards(
1329 const long double number,
1330 const int rounding_multiplier
1333 long double rounded_number;
1334 rounded_number = number * rounding_multiplier;
1335 rounded_number += 0.5;
1336 rounded_number = (int64_t) (rounded_number);
1337 rounded_number = (long double) (rounded_number) / (long double) (rounding_multiplier);
1339 return rounded_number;
1343 string nice_unit_format(
1344 const int64_t input,
1345 const UnitFormat format,
1349 // select the system of units (decimal or binary)
1351 if (base == UnitBase1000)
1360 long double size = input;
1362 // check the size of the input number to fit in the appropriate symbol
1364 while (size > multiple)
1366 size = size / multiple;
1369 // rollback to the previous values and stop the loop when cannot
1370 // represent the number length.
1371 if (sizecount >= MAX_UNIT_FORMAT_SYMBOLS)
1373 size = size * multiple;
1379 // round the input number "half up" to multiples of 10
1380 const int rounding_multiplier = 10;
1381 size = rounding_upwards(size, rounding_multiplier);
1383 // format the input number, placing the appropriate symbol
1385 out.setf (ios::fixed);
1386 if (format == ShortUnitFormat)
1389 out << size << i18n( shortUnitFormatSymbols[sizecount].c_str() );
1394 out << size << i18n( longUnitFormatSymbols[sizecount].c_str() );
1398 } // eo nice_unit_format(int input)
1401 string nice_unit_format(
1403 const UnitFormat format,
1407 // round as double and cast to int64_t
1408 // cast raised overflow error near max val of int64_t (~9.2e18, see unittest)
1409 int64_t input_casted_and_rounded =
1410 boost::numeric_cast<int64_t>( round(input) );
1413 return nice_unit_format( input_casted_and_rounded, format, base );
1414 } // eo nice_unit_format(double input)
1417 string escape(const string &s)
1420 string::size_type p;
1423 while ( (p=out.find_first_of("\"\\",p) ) !=out.npos)
1425 out.insert (p,"\\");
1430 while ( (p=out.find_first_of("\r",p) ) !=out.npos)
1432 out.replace (p,1,"\\r");
1437 while ( (p=out.find_first_of("\n",p) ) !=out.npos)
1439 out.replace (p,1,"\\n");
1446 } // eo scape(const std::string&)
1449 string descape(const string &s, int startpos, int &endpos)
1453 if (s.at(startpos) != '"')
1454 throw out_of_range("value not type escaped string");
1456 out=s.substr(startpos+1);
1457 string::size_type p=0;
1459 // search for the end of the string
1460 while ( (p=out.find("\"",p) ) !=out.npos)
1465 // the " might be escaped with a backslash
1466 while (e>=0 && out.at (e) =='\\')
1468 if (escaped == false)
1482 // we now have the end of the string
1483 out=out.substr(0,p);
1485 // tell calling prog about the endposition
1486 endpos=startpos+p+1;
1488 // descape all \ stuff inside the string now
1490 while ( (p=out.find_first_of("\\",p) ) !=out.npos)
1492 switch (out.at(p+1) )
1495 out.replace(p,2,"\r");
1498 out.replace(p,2,"\n");
1507 } // eo descape(const std::string&,int,int&)
1510 string escape_shellarg(const string &input)
1512 string output = "'";
1513 string::const_iterator it, it_end = input.end();
1514 for (it = input.begin(); it != it_end; ++it)