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 struct local_lowercase {
365 char operator()(char c) const {
370 struct local_uppercase {
371 char operator()(char c) const {
376 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());
391 string const uppercase(string const & a)
394 transform(tmp.begin(), tmp.end(), tmp.begin(), local_uppercase());
399 string const ascii_lowercase(string const & a)
402 transform(tmp.begin(), tmp.end(), tmp.begin(),
403 local_ascii_lowercase());
408 bool prefixIs(string const & a, string const & pre)
410 string::size_type const prelen = pre.length();
411 string::size_type const alen = a.length();
413 if (prelen > alen || a.empty())
416 #if defined(STD_STRING_IS_GOOD)
417 return a.compare(0, prelen, pre) == 0;
419 return ::strncmp(a.c_str(), pre.c_str(), prelen) == 0;
425 bool prefixIs(docstring const & a, docstring const & pre)
427 docstring::size_type const prelen = pre.length();
428 docstring::size_type const alen = a.length();
430 if (prelen > alen || a.empty())
433 return a.compare(0, prelen, pre) == 0;
437 bool suffixIs(string const & a, char c)
439 if (a.empty()) return false;
440 return a[a.length() - 1] == c;
444 bool suffixIs(string const & a, string const & suf)
446 string::size_type const suflen = suf.length();
447 string::size_type const alen = a.length();
452 #if !defined(USE_INCLUDED_STRING) && !defined(STD_STRING_IS_GOOD)
453 string tmp(a, alen - suflen);
454 return ::strncmp(tmp.c_str(), suf.c_str(), suflen) == 0;
456 return a.compare(alen - suflen, suflen, suf) == 0;
462 bool containsOnly(string const & s, string const & cset)
464 return s.find_first_not_of(cset) == string::npos;
468 // ale970405+lasgoutt-970425
469 // rewritten to use new string (Lgb)
470 string const token(string const & a, char delim, int n)
472 if (a.empty()) return string();
474 string::size_type k = 0;
475 string::size_type i = 0;
477 // Find delimiter or end of string
479 if ((i = a.find(delim, i)) == string::npos)
483 // i is now the n'th delim (or string::npos)
484 if (i == string::npos) return string();
485 k = a.find(delim, i);
486 // k is now the n'th + 1 delim (or string::npos)
488 return a.substr(i, k - i);
492 docstring const token(docstring const & a, char_type delim, int n)
494 if (a.empty()) return docstring();
496 string::size_type k = 0;
497 string::size_type i = 0;
499 // Find delimiter or end of string
501 if ((i = a.find(delim, i)) == docstring::npos)
505 // i is now the n'th delim (or string::npos)
506 if (i == docstring::npos) return docstring();
507 k = a.find(delim, i);
508 // k is now the n'th + 1 delim (or string::npos)
510 return a.substr(i, k - i);
514 // this could probably be faster and/or cleaner, but it seems to work (JMarc)
515 // rewritten to use new string (Lgb)
516 int tokenPos(string const & a, char delim, string const & tok)
522 while (!str.empty()) {
523 str = split(str, tmptok, delim);
534 /// Substitute all \a oldchar with \a newchar
535 template<typename Ch> inline
536 std::basic_string<Ch> const subst_char(std::basic_string<Ch> const & a,
537 Ch oldchar, Ch newchar)
539 typedef std::basic_string<Ch> String;
541 typename String::iterator lit = tmp.begin();
542 typename String::iterator end = tmp.end();
543 for (; lit != end; ++lit)
544 if ((*lit) == oldchar)
550 /// substitutes all instances of \a oldstr with \a newstr
551 template<typename String> inline
552 String const subst_string(String const & a,
553 String const & oldstr, String const & newstr)
555 BOOST_ASSERT(!oldstr.empty());
557 typename String::size_type i = 0;
558 typename String::size_type const olen = oldstr.length();
559 while ((i = lstr.find(oldstr, i)) != string::npos) {
560 lstr.replace(i, olen, newstr);
561 i += newstr.length(); // We need to be sure that we dont
562 // use the same i over and over again.
570 string const subst(string const & a, char oldchar, char newchar)
572 return subst_char(a, oldchar, newchar);
576 docstring const subst(docstring const & a,
577 char_type oldchar, char_type newchar)
579 return subst_char(a, oldchar, newchar);
583 string const subst(string const & a,
584 string const & oldstr, string const & newstr)
586 return subst_string(a, oldstr, newstr);
590 docstring const subst(docstring const & a,
591 docstring const & oldstr, docstring const & newstr)
593 return subst_string(a, oldstr, newstr);
597 docstring const trim(docstring const & a, char const * p)
601 if (a.empty() || !*p)
604 docstring s = lyx::from_ascii(p);
605 docstring::size_type r = a.find_last_not_of(s);
606 docstring::size_type l = a.find_first_not_of(s);
608 // Is this the minimal test? (lgb)
609 if (r == docstring::npos && l == docstring::npos)
612 return a.substr(l, r - l + 1);
616 string const trim(string const & a, char const * p)
620 if (a.empty() || !*p)
623 string::size_type r = a.find_last_not_of(p);
624 string::size_type l = a.find_first_not_of(p);
626 // Is this the minimal test? (lgb)
627 if (r == string::npos && l == string::npos)
630 return a.substr(l, r - l + 1);
634 string const rtrim(string const & a, char const * p)
638 if (a.empty() || !*p)
641 string::size_type r = a.find_last_not_of(p);
643 // Is this test really needed? (Lgb)
644 if (r == string::npos)
647 return a.substr(0, r + 1);
651 string const ltrim(string const & a, char const * p)
654 if (a.empty() || !*p)
656 string::size_type l = a.find_first_not_of(p);
657 if (l == string::npos)
659 return a.substr(l, string::npos);
663 docstring const ltrim(docstring const & a, char const * p)
666 if (a.empty() || !*p)
668 size_t l = a.find_first_not_of(from_ascii(p));
669 if (l == docstring::npos)
671 return a.substr(l, docstring::npos);
676 template<typename String, typename Char> inline
677 String const doSplit(String const & a, String & piece, Char delim)
680 typename String::size_type i = a.find(delim);
681 if (i == a.length() - 1) {
682 piece = a.substr(0, i);
683 } else if (i != String::npos) {
684 piece = a.substr(0, i);
685 tmp = a.substr(i + 1);
688 tmp = a.substr(i + 1);
698 string const split(string const & a, string & piece, char delim)
700 return doSplit(a, piece, delim);
704 docstring const split(docstring const & a, docstring & piece, char_type delim)
706 return doSplit(a, piece, delim);
710 string const split(string const & a, char delim)
713 string::size_type i = a.find(delim);
714 if (i != string::npos) // found delim
715 tmp = a.substr(i + 1);
721 string const rsplit(string const & a, string & piece, char delim)
724 string::size_type i = a.rfind(delim);
725 if (i != string::npos) { // delimiter was found
726 piece = a.substr(0, i);
727 tmp = a.substr(i + 1);
728 } else { // delimiter was not found
735 docstring const escape(docstring const & lab)
737 char_type hexdigit[16] = { '0', '1', '2', '3', '4', '5', '6', '7',
738 '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' };
740 for (docstring::size_type i = 0; i < lab.length(); ++i) {
741 char_type c = lab[i];
742 if (c >= 128 || c == '=' || c == '%') {
743 // Although char_type is a 32 bit type we know that
744 // UCS4 occupies only 21 bits, so we don't need to
745 // encode bigger values. Test for 2^24 because we
746 // can encode that with the 6 hex digits that are
747 // needed for 21 bits anyway.
748 BOOST_ASSERT(c < (1 << 24));
750 enc += hexdigit[(c>>20) & 15];
751 enc += hexdigit[(c>>16) & 15];
752 enc += hexdigit[(c>>12) & 15];
753 enc += hexdigit[(c>> 8) & 15];
754 enc += hexdigit[(c>> 4) & 15];
755 enc += hexdigit[ c & 15];
764 /// gives a vector of stringparts which have the delimiter delim
765 vector<string> const getVectorFromString(string const & str,
766 string const & delim)
768 // Lars would like this code to go, but for now his replacement (below)
769 // doesn't fullfil the same function. I have, therefore, reactivated the
770 // old code for now. Angus 11 Nov 2002.
775 string keys = rtrim(str);
777 string::size_type const idx = keys.find(delim);
778 if (idx == string::npos) {
779 vec.push_back(ltrim(keys));
782 string const key = trim(keys.substr(0, idx));
785 string::size_type const start = idx + delim.size();
786 keys = keys.substr(start);
790 boost::char_separator<char> sep(delim.c_str());
791 boost::tokenizer<boost::char_separator<char> > tokens(str, sep);
792 return vector<string>(tokens.begin(), tokens.end());
797 // the same vice versa
798 string const getStringFromVector(vector<string> const & vec,
799 string const & delim)
803 for (vector<string>::const_iterator it = vec.begin();
804 it != vec.end(); ++it) {
805 string item = trim(*it);
816 int findToken(char const * const str[], string const & search_token)
820 while (str[i][0] && str[i] != search_token)
828 docstring const externalLineEnding(docstring const & str)
830 #if defined(__APPLE__)
831 // The MAC clipboard uses \r for lineendings, and we use \n
832 return subst(str, '\n', '\r');
833 #elif defined (_WIN32) || (defined (__CYGWIN__) && defined (X_DISPLAY_MISSING))
834 // Windows clipboard uses \r\n for lineendings, and we use \n
835 return subst(str, lyx::from_ascii("\n"), lyx::from_ascii("\r\n"));
842 docstring const internalLineEnding(docstring const & str)
844 docstring const s = subst(str,
845 lyx::from_ascii("\r\n"), lyx::from_ascii("\n"));
846 return subst(s, '\r', '\n');
850 #ifndef I_AM_NOT_AFRAID_OF_HEADER_LIBRARIES
854 docstring bformat(docstring const & fmt, int arg1)
856 return (boost::basic_format<char_type>(fmt) % arg1).str();
861 docstring bformat(docstring const & fmt, long arg1)
863 return (boost::basic_format<char_type>(fmt) % arg1).str();
868 docstring bformat(docstring const & fmt, unsigned int arg1)
870 return (boost::basic_format<char_type>(fmt) % arg1).str();
875 docstring bformat<docstring>(docstring const & fmt, docstring arg1)
877 return (boost::basic_format<char_type>(fmt) % arg1).str();
882 docstring bformat(docstring const & fmt, char * arg1)
884 return (boost::basic_format<char_type>(fmt) % arg1).str();
889 docstring bformat(docstring const & fmt, int arg1, int arg2)
891 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
896 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2)
898 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
903 docstring bformat(docstring const & fmt, char const * arg1, docstring arg2)
905 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
910 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2, docstring arg3)
912 return (boost::basic_format<char_type>(fmt) % arg1 % arg2 % arg3).str();
917 docstring bformat(docstring const & fmt,
918 docstring arg1, docstring arg2, docstring arg3, docstring arg4)
920 return (boost::basic_format<char_type>(fmt) % arg1 % arg2 % arg3 % arg4).str();
926 docstring bformat(docstring const & fmt, int arg1)
928 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
929 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
930 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
935 docstring bformat(docstring const & fmt, long arg1)
937 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
938 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
939 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
944 docstring bformat(docstring const & fmt, unsigned int arg1)
946 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
947 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
948 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
953 docstring bformat(docstring const & fmt, docstring arg1)
955 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
956 docstring const str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
957 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
962 docstring bformat(docstring const & fmt, char * arg1)
964 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
965 docstring const str = subst(fmt, lyx::from_ascii("%1$s"), lyx::from_ascii(arg1));
966 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
971 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2)
973 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
974 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
975 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
976 str = subst(str, lyx::from_ascii("%2$s"), arg2);
977 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
982 docstring bformat(docstring const & fmt, char const * arg1, docstring arg2)
984 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
985 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
986 docstring str = subst(fmt, lyx::from_ascii("%1$s"), lyx::from_ascii(arg1));
987 str = subst(fmt, lyx::from_ascii("%2$s"), arg2);
988 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
993 docstring bformat(docstring const & fmt, int arg1, int arg2)
995 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
996 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$d")));
997 docstring str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
998 str = subst(str, lyx::from_ascii("%2$d"), convert<docstring>(arg2));
999 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1004 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2, docstring arg3)
1006 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1007 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1008 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%3$s")));
1009 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
1010 str = subst(str, lyx::from_ascii("%2$s"), arg2);
1011 str = subst(str, lyx::from_ascii("%3$s"), arg3);
1012 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1017 docstring bformat(docstring const & fmt,
1018 docstring arg1, docstring arg2, docstring arg3, docstring arg4)
1020 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1021 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1022 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%3$s")));
1023 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%4$s")));
1024 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
1025 str = subst(str, lyx::from_ascii("%2$s"), arg2);
1026 str = subst(str, lyx::from_ascii("%3$s"), arg3);
1027 str = subst(str, lyx::from_ascii("%4$s"), arg4);
1028 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1034 } // namespace support