3 * This file is part of LyX, the document processor.
4 * Licence details can be found in the file COPYING.
6 * \author Lars Gullik Bjønnes
7 * \author Jean-Marc Lasgouttes
9 * Full author contact details are available in file CREDITS.
14 #include "support/lstrings.h"
15 #include "support/lyxlib.h"
16 #include "support/convert.h"
20 #include <boost/tokenizer.hpp>
21 #include <boost/assert.hpp>
23 #ifndef I_AM_NOT_AFRAID_OF_HEADER_LIBRARIES
25 #include <boost/format.hpp>
35 #ifdef LIBC_WCTYPE_USES_UCS4
36 // We can use the libc ctype functions because we unset the LC_CTYPE
37 // category of the current locale in gettext.C
40 // Steal some code from somewhere else, e.g. glib (look at gunicode.h)
41 // The code that we currently use does not really work.
51 #ifndef CXX_GLOBAL_CSTD
61 int compare_no_case(string const & s, string const & s2)
63 string::const_iterator p = s.begin();
64 string::const_iterator p2 = s2.begin();
66 while (p != s.end() && p2 != s2.end()) {
67 int const lc1 = tolower(*p);
68 int const lc2 = tolower(*p2);
70 return (lc1 < lc2) ? -1 : 1;
75 if (s.size() == s2.size())
77 if (s.size() < s2.size())
83 int compare_no_case(docstring const & s, docstring const & s2)
85 docstring::const_iterator p = s.begin();
86 docstring::const_iterator p2 = s2.begin();
88 while (p != s.end() && p2 != s2.end()) {
89 char_type const lc1 = lowercase(*p);
90 char_type const lc2 = lowercase(*p2);
92 return (lc1 < lc2) ? -1 : 1;
97 if (s.size() == s2.size())
99 if (s.size() < s2.size())
107 template<typename Char>
108 Char ascii_tolower(Char c) {
109 if (c >= 'A' && c <= 'Z')
110 return c - 'A' + 'a';
117 int compare_ascii_no_case(string const & s, string const & s2)
119 string::const_iterator p = s.begin();
120 string::const_iterator p2 = s2.begin();
122 while (p != s.end() && p2 != s2.end()) {
123 int const lc1 = ascii_tolower(*p);
124 int const lc2 = ascii_tolower(*p2);
126 return (lc1 < lc2) ? -1 : 1;
131 if (s.size() == s2.size())
133 if (s.size() < s2.size())
139 int compare_ascii_no_case(docstring const & s, docstring const & s2)
141 docstring::const_iterator p = s.begin();
142 docstring::const_iterator p2 = s2.begin();
144 while (p != s.end() && p2 != s2.end()) {
145 char_type const lc1 = ascii_tolower(*p);
146 char_type const lc2 = ascii_tolower(*p2);
148 return (lc1 < lc2) ? -1 : 1;
153 if (s.size() == s2.size())
155 if (s.size() < s2.size())
161 int compare_no_case(string const & s, string const & s2, unsigned int len)
163 string::const_iterator p = s.begin();
164 string::const_iterator p2 = s2.begin();
166 while (i < len && p != s.end() && p2 != s2.end()) {
167 int const lc1 = tolower(*p);
168 int const lc2 = tolower(*p2);
170 return (lc1 < lc2) ? -1 : 1;
176 if (s.size() >= len && s2.size() >= len)
178 if (s.size() < s2.size())
184 bool isStrInt(string const & str)
186 if (str.empty()) return false;
188 // Remove leading and trailing white space chars.
189 string const tmpstr = trim(str);
190 if (tmpstr.empty()) return false;
192 string::const_iterator cit = tmpstr.begin();
193 if ((*cit) == '-') ++cit;
194 string::const_iterator end = tmpstr.end();
195 for (; cit != end; ++cit) {
196 if (!isdigit((*cit))) return false;
202 bool isStrUnsignedInt(string const & str)
204 if (str.empty()) return false;
206 // Remove leading and trailing white space chars.
207 string const tmpstr = trim(str);
208 if (tmpstr.empty()) return false;
210 string::const_iterator cit = tmpstr.begin();
211 string::const_iterator end = tmpstr.end();
212 for (; cit != end; ++cit) {
213 if (!isdigit((*cit))) return false;
219 bool isStrDbl(string const & str)
221 if (str.empty()) return false;
223 // Remove leading and trailing white space chars.
224 string const tmpstr = trim(str);
225 if (tmpstr.empty()) return false;
226 // if (1 < tmpstr.count('.')) return false;
228 string::const_iterator cit = tmpstr.begin();
229 bool found_dot(false);
230 if ((*cit) == '-') ++cit;
231 string::const_iterator end = tmpstr.end();
232 for (; cit != end; ++cit) {
252 bool isHexChar(char_type c)
264 c == 'a' || c == 'A' ||
265 c == 'b' || c == 'B' ||
266 c == 'c' || c == 'C' ||
267 c == 'd' || c == 'D' ||
268 c == 'e' || c == 'E' ||
269 c == 'f' || c == 'F';
275 bool isHex(docstring const & str)
279 if (str.length() > 2 && str[0] == '0' &&
280 (str[1] == 'x' || str[1] == 'X'))
283 int const len = str.length();
285 for (; index < len; ++index) {
286 if (!isHexChar(str[index]))
293 int hexToInt(docstring const & str)
295 string s = to_ascii(str);
297 sscanf(s.c_str(), "%x", &h);
302 bool isAscii(docstring const & str)
304 int const len = str.length();
305 for (int i = 0; i < len; ++i)
312 char lowercase(char c)
314 return char(tolower(c));
318 char uppercase(char c)
320 return char(toupper(c));
325 // for lowercase() and uppercase() function below when wchar_t is not used:
326 // 1) std::tolower() and std::toupper() are templates that
327 // compile fine with char_type. With the test (c >= 256) we
328 // do not trust these function to do the right thing with
330 // 2) these functions use the current locale, which is wrong
331 // if it is not latin1 based (latin1 is a subset of UCS4).
333 char_type lowercase(char_type c)
335 #ifdef LIBC_WCTYPE_USES_UCS4
346 char_type uppercase(char_type c)
348 #ifdef LIBC_WCTYPE_USES_UCS4
361 // since we cannot use std::tolower and std::toupper directly in the
362 // calls to std::transform yet, we use these helper clases. (Lgb)
364 template<typename Char> struct local_lowercase {
365 Char operator()(Char c) const {
370 struct local_uppercase {
371 char operator()(char c) const {
376 template<typename Char> struct local_ascii_lowercase {
377 Char operator()(Char c) const {
378 return ascii_tolower(c);
382 } // end of anon namespace
384 string const lowercase(string const & a)
387 transform(tmp.begin(), tmp.end(), tmp.begin(), local_lowercase<char>());
392 docstring const lowercase(docstring const & a)
395 transform(tmp.begin(), tmp.end(), tmp.begin(), local_lowercase<char_type>());
400 string const uppercase(string const & a)
403 transform(tmp.begin(), tmp.end(), tmp.begin(), local_uppercase());
408 string const ascii_lowercase(string const & a)
411 transform(tmp.begin(), tmp.end(), tmp.begin(),
412 local_ascii_lowercase<char>());
417 docstring const ascii_lowercase(docstring const & a)
420 transform(tmp.begin(), tmp.end(), tmp.begin(),
421 local_ascii_lowercase<char_type>());
426 bool prefixIs(docstring const & a, char_type c)
434 bool prefixIs(string const & a, string const & pre)
436 string::size_type const prelen = pre.length();
437 string::size_type const alen = a.length();
439 if (prelen > alen || a.empty())
442 #if defined(STD_STRING_IS_GOOD)
443 return a.compare(0, prelen, pre) == 0;
445 return ::strncmp(a.c_str(), pre.c_str(), prelen) == 0;
451 bool prefixIs(docstring const & a, docstring const & pre)
453 docstring::size_type const prelen = pre.length();
454 docstring::size_type const alen = a.length();
456 if (prelen > alen || a.empty())
459 return a.compare(0, prelen, pre) == 0;
463 bool suffixIs(string const & a, char c)
465 if (a.empty()) return false;
466 return a[a.length() - 1] == c;
470 bool suffixIs(docstring const & a, char_type c)
474 return a[a.length() - 1] == c;
478 bool suffixIs(string const & a, string const & suf)
480 string::size_type const suflen = suf.length();
481 string::size_type const alen = a.length();
486 #if !defined(USE_INCLUDED_STRING) && !defined(STD_STRING_IS_GOOD)
487 string tmp(a, alen - suflen);
488 return ::strncmp(tmp.c_str(), suf.c_str(), suflen) == 0;
490 return a.compare(alen - suflen, suflen, suf) == 0;
496 bool containsOnly(string const & s, string const & cset)
498 return s.find_first_not_of(cset) == string::npos;
502 // ale970405+lasgoutt-970425
503 // rewritten to use new string (Lgb)
504 string const token(string const & a, char delim, int n)
506 if (a.empty()) return string();
508 string::size_type k = 0;
509 string::size_type i = 0;
511 // Find delimiter or end of string
513 if ((i = a.find(delim, i)) == string::npos)
517 // i is now the n'th delim (or string::npos)
518 if (i == string::npos) return string();
519 k = a.find(delim, i);
520 // k is now the n'th + 1 delim (or string::npos)
522 return a.substr(i, k - i);
526 docstring const token(docstring const & a, char_type delim, int n)
528 if (a.empty()) return docstring();
530 string::size_type k = 0;
531 string::size_type i = 0;
533 // Find delimiter or end of string
535 if ((i = a.find(delim, i)) == docstring::npos)
539 // i is now the n'th delim (or string::npos)
540 if (i == docstring::npos) return docstring();
541 k = a.find(delim, i);
542 // k is now the n'th + 1 delim (or string::npos)
544 return a.substr(i, k - i);
548 // this could probably be faster and/or cleaner, but it seems to work (JMarc)
549 // rewritten to use new string (Lgb)
550 int tokenPos(string const & a, char delim, string const & tok)
556 while (!str.empty()) {
557 str = split(str, tmptok, delim);
568 /// Substitute all \a oldchar with \a newchar
569 template<typename Ch> inline
570 std::basic_string<Ch> const subst_char(std::basic_string<Ch> const & a,
571 Ch oldchar, Ch newchar)
573 typedef std::basic_string<Ch> String;
575 typename String::iterator lit = tmp.begin();
576 typename String::iterator end = tmp.end();
577 for (; lit != end; ++lit)
578 if ((*lit) == oldchar)
584 /// substitutes all instances of \a oldstr with \a newstr
585 template<typename String> inline
586 String const subst_string(String const & a,
587 String const & oldstr, String const & newstr)
589 BOOST_ASSERT(!oldstr.empty());
591 typename String::size_type i = 0;
592 typename String::size_type const olen = oldstr.length();
593 while ((i = lstr.find(oldstr, i)) != string::npos) {
594 lstr.replace(i, olen, newstr);
595 i += newstr.length(); // We need to be sure that we dont
596 // use the same i over and over again.
604 string const subst(string const & a, char oldchar, char newchar)
606 return subst_char(a, oldchar, newchar);
610 docstring const subst(docstring const & a,
611 char_type oldchar, char_type newchar)
613 return subst_char(a, oldchar, newchar);
617 string const subst(string const & a,
618 string const & oldstr, string const & newstr)
620 return subst_string(a, oldstr, newstr);
624 docstring const subst(docstring const & a,
625 docstring const & oldstr, docstring const & newstr)
627 return subst_string(a, oldstr, newstr);
631 docstring const trim(docstring const & a, char const * p)
635 if (a.empty() || !*p)
638 docstring s = lyx::from_ascii(p);
639 docstring::size_type r = a.find_last_not_of(s);
640 docstring::size_type l = a.find_first_not_of(s);
642 // Is this the minimal test? (lgb)
643 if (r == docstring::npos && l == docstring::npos)
646 return a.substr(l, r - l + 1);
650 string const trim(string const & a, char const * p)
654 if (a.empty() || !*p)
657 string::size_type r = a.find_last_not_of(p);
658 string::size_type l = a.find_first_not_of(p);
660 // Is this the minimal test? (lgb)
661 if (r == string::npos && l == string::npos)
664 return a.substr(l, r - l + 1);
668 string const rtrim(string const & a, char const * p)
672 if (a.empty() || !*p)
675 string::size_type r = a.find_last_not_of(p);
677 // Is this test really needed? (Lgb)
678 if (r == string::npos)
681 return a.substr(0, r + 1);
685 docstring const rtrim(docstring const & a, char const * p)
689 if (a.empty() || !*p)
692 docstring::size_type r = a.find_last_not_of(from_ascii(p));
694 // Is this test really needed? (Lgb)
695 if (r == docstring::npos)
698 return a.substr(0, r + 1);
702 string const ltrim(string const & a, char const * p)
705 if (a.empty() || !*p)
707 string::size_type l = a.find_first_not_of(p);
708 if (l == string::npos)
710 return a.substr(l, string::npos);
714 docstring const ltrim(docstring const & a, char const * p)
717 if (a.empty() || !*p)
719 size_t l = a.find_first_not_of(from_ascii(p));
720 if (l == docstring::npos)
722 return a.substr(l, docstring::npos);
727 template<typename String, typename Char> inline
728 String const doSplit(String const & a, String & piece, Char delim)
731 typename String::size_type i = a.find(delim);
732 if (i == a.length() - 1) {
733 piece = a.substr(0, i);
734 } else if (i != String::npos) {
735 piece = a.substr(0, i);
736 tmp = a.substr(i + 1);
739 tmp = a.substr(i + 1);
749 string const split(string const & a, string & piece, char delim)
751 return doSplit(a, piece, delim);
755 docstring const split(docstring const & a, docstring & piece, char_type delim)
757 return doSplit(a, piece, delim);
761 string const split(string const & a, char delim)
764 string::size_type i = a.find(delim);
765 if (i != string::npos) // found delim
766 tmp = a.substr(i + 1);
772 string const rsplit(string const & a, string & piece, char delim)
775 string::size_type i = a.rfind(delim);
776 if (i != string::npos) { // delimiter was found
777 piece = a.substr(0, i);
778 tmp = a.substr(i + 1);
779 } else { // delimiter was not found
786 docstring const escape(docstring const & lab)
788 char_type hexdigit[16] = { '0', '1', '2', '3', '4', '5', '6', '7',
789 '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' };
791 for (docstring::size_type i = 0; i < lab.length(); ++i) {
792 char_type c = lab[i];
793 if (c >= 128 || c == '=' || c == '%') {
794 // Although char_type is a 32 bit type we know that
795 // UCS4 occupies only 21 bits, so we don't need to
796 // encode bigger values. Test for 2^24 because we
797 // can encode that with the 6 hex digits that are
798 // needed for 21 bits anyway.
799 BOOST_ASSERT(c < (1 << 24));
801 enc += hexdigit[(c>>20) & 15];
802 enc += hexdigit[(c>>16) & 15];
803 enc += hexdigit[(c>>12) & 15];
804 enc += hexdigit[(c>> 8) & 15];
805 enc += hexdigit[(c>> 4) & 15];
806 enc += hexdigit[ c & 15];
817 template<typename String> vector<String> const
818 getVectorFromStringT(String const & str, String const & delim)
820 // Lars would like this code to go, but for now his replacement (below)
821 // doesn't fullfil the same function. I have, therefore, reactivated the
822 // old code for now. Angus 11 Nov 2002.
827 String keys = rtrim(str);
829 typename String::size_type const idx = keys.find(delim);
830 if (idx == String::npos) {
831 vec.push_back(ltrim(keys));
834 String const key = trim(keys.substr(0, idx));
837 typename String::size_type const start = idx + delim.size();
838 keys = keys.substr(start);
842 typedef boost::char_separator<typename String::value_type> Separator;
843 typedef boost::tokenizer<Separator, typename String::const_iterator, String> Tokenizer;
844 Separator sep(delim.c_str());
845 Tokenizer tokens(str, sep);
846 return vector<String>(tokens.begin(), tokens.end());
853 vector<string> const getVectorFromString(string const & str,
854 string const & delim)
856 return getVectorFromStringT<string>(str, delim);
860 vector<docstring> const getVectorFromString(docstring const & str,
861 docstring const & delim)
863 return getVectorFromStringT<docstring>(str, delim);
867 // the same vice versa
868 string const getStringFromVector(vector<string> const & vec,
869 string const & delim)
873 for (vector<string>::const_iterator it = vec.begin();
874 it != vec.end(); ++it) {
875 string item = trim(*it);
886 int findToken(char const * const str[], string const & search_token)
890 while (str[i][0] && str[i] != search_token)
898 docstring const externalLineEnding(docstring const & str)
900 #if defined(__APPLE__)
901 // The MAC clipboard uses \r for lineendings, and we use \n
902 return subst(str, '\n', '\r');
903 #elif defined (_WIN32) || (defined (__CYGWIN__) && defined (X_DISPLAY_MISSING))
904 // Windows clipboard uses \r\n for lineendings, and we use \n
905 return subst(str, lyx::from_ascii("\n"), lyx::from_ascii("\r\n"));
912 docstring const internalLineEnding(docstring const & str)
914 docstring const s = subst(str,
915 lyx::from_ascii("\r\n"), lyx::from_ascii("\n"));
916 return subst(s, '\r', '\n');
920 #ifndef I_AM_NOT_AFRAID_OF_HEADER_LIBRARIES
924 docstring bformat(docstring const & fmt, int arg1)
926 return (boost::basic_format<char_type>(fmt) % arg1).str();
931 docstring bformat(docstring const & fmt, long arg1)
933 return (boost::basic_format<char_type>(fmt) % arg1).str();
938 docstring bformat(docstring const & fmt, unsigned int arg1)
940 return (boost::basic_format<char_type>(fmt) % arg1).str();
945 docstring bformat<docstring>(docstring const & fmt, docstring arg1)
947 return (boost::basic_format<char_type>(fmt) % arg1).str();
952 docstring bformat(docstring const & fmt, char * arg1)
954 return (boost::basic_format<char_type>(fmt) % arg1).str();
959 docstring bformat(docstring const & fmt, int arg1, int arg2)
961 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
966 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2)
968 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
973 docstring bformat(docstring const & fmt, char const * arg1, docstring arg2)
975 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
980 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2, docstring arg3)
982 return (boost::basic_format<char_type>(fmt) % arg1 % arg2 % arg3).str();
987 docstring bformat(docstring const & fmt,
988 docstring arg1, docstring arg2, docstring arg3, docstring arg4)
990 return (boost::basic_format<char_type>(fmt) % arg1 % arg2 % arg3 % arg4).str();
996 docstring bformat(docstring const & fmt, int arg1)
998 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
999 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
1000 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1005 docstring bformat(docstring const & fmt, long arg1)
1007 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
1008 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
1009 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1014 docstring bformat(docstring const & fmt, unsigned int arg1)
1016 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
1017 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
1018 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1023 docstring bformat(docstring const & fmt, docstring arg1)
1025 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1026 docstring const str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
1027 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1032 docstring bformat(docstring const & fmt, char * arg1)
1034 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1035 docstring const str = subst(fmt, lyx::from_ascii("%1$s"), lyx::from_ascii(arg1));
1036 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1041 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2)
1043 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1044 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1045 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
1046 str = subst(str, lyx::from_ascii("%2$s"), arg2);
1047 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1052 docstring bformat(docstring const & fmt, char const * arg1, docstring arg2)
1054 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1055 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1056 docstring str = subst(fmt, lyx::from_ascii("%1$s"), lyx::from_ascii(arg1));
1057 str = subst(fmt, lyx::from_ascii("%2$s"), arg2);
1058 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1063 docstring bformat(docstring const & fmt, int arg1, int arg2)
1065 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
1066 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$d")));
1067 docstring str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
1068 str = subst(str, lyx::from_ascii("%2$d"), convert<docstring>(arg2));
1069 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1074 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2, docstring arg3)
1076 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1077 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1078 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%3$s")));
1079 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
1080 str = subst(str, lyx::from_ascii("%2$s"), arg2);
1081 str = subst(str, lyx::from_ascii("%3$s"), arg3);
1082 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1087 docstring bformat(docstring const & fmt,
1088 docstring arg1, docstring arg2, docstring arg3, docstring arg4)
1090 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1091 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1092 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%3$s")));
1093 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%4$s")));
1094 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
1095 str = subst(str, lyx::from_ascii("%2$s"), arg2);
1096 str = subst(str, lyx::from_ascii("%3$s"), arg3);
1097 str = subst(str, lyx::from_ascii("%4$s"), arg4);
1098 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1104 } // namespace support