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(string const & a, string const & pre)
428 string::size_type const prelen = pre.length();
429 string::size_type const alen = a.length();
431 if (prelen > alen || a.empty())
434 #if defined(STD_STRING_IS_GOOD)
435 return a.compare(0, prelen, pre) == 0;
437 return ::strncmp(a.c_str(), pre.c_str(), prelen) == 0;
443 bool prefixIs(docstring const & a, docstring const & pre)
445 docstring::size_type const prelen = pre.length();
446 docstring::size_type const alen = a.length();
448 if (prelen > alen || a.empty())
451 return a.compare(0, prelen, pre) == 0;
455 bool suffixIs(string const & a, char c)
457 if (a.empty()) return false;
458 return a[a.length() - 1] == c;
462 bool suffixIs(string const & a, string const & suf)
464 string::size_type const suflen = suf.length();
465 string::size_type const alen = a.length();
470 #if !defined(USE_INCLUDED_STRING) && !defined(STD_STRING_IS_GOOD)
471 string tmp(a, alen - suflen);
472 return ::strncmp(tmp.c_str(), suf.c_str(), suflen) == 0;
474 return a.compare(alen - suflen, suflen, suf) == 0;
480 bool containsOnly(string const & s, string const & cset)
482 return s.find_first_not_of(cset) == string::npos;
486 // ale970405+lasgoutt-970425
487 // rewritten to use new string (Lgb)
488 string const token(string const & a, char delim, int n)
490 if (a.empty()) return string();
492 string::size_type k = 0;
493 string::size_type i = 0;
495 // Find delimiter or end of string
497 if ((i = a.find(delim, i)) == string::npos)
501 // i is now the n'th delim (or string::npos)
502 if (i == string::npos) return string();
503 k = a.find(delim, i);
504 // k is now the n'th + 1 delim (or string::npos)
506 return a.substr(i, k - i);
510 docstring const token(docstring const & a, char_type delim, int n)
512 if (a.empty()) return docstring();
514 string::size_type k = 0;
515 string::size_type i = 0;
517 // Find delimiter or end of string
519 if ((i = a.find(delim, i)) == docstring::npos)
523 // i is now the n'th delim (or string::npos)
524 if (i == docstring::npos) return docstring();
525 k = a.find(delim, i);
526 // k is now the n'th + 1 delim (or string::npos)
528 return a.substr(i, k - i);
532 // this could probably be faster and/or cleaner, but it seems to work (JMarc)
533 // rewritten to use new string (Lgb)
534 int tokenPos(string const & a, char delim, string const & tok)
540 while (!str.empty()) {
541 str = split(str, tmptok, delim);
552 /// Substitute all \a oldchar with \a newchar
553 template<typename Ch> inline
554 std::basic_string<Ch> const subst_char(std::basic_string<Ch> const & a,
555 Ch oldchar, Ch newchar)
557 typedef std::basic_string<Ch> String;
559 typename String::iterator lit = tmp.begin();
560 typename String::iterator end = tmp.end();
561 for (; lit != end; ++lit)
562 if ((*lit) == oldchar)
568 /// substitutes all instances of \a oldstr with \a newstr
569 template<typename String> inline
570 String const subst_string(String const & a,
571 String const & oldstr, String const & newstr)
573 BOOST_ASSERT(!oldstr.empty());
575 typename String::size_type i = 0;
576 typename String::size_type const olen = oldstr.length();
577 while ((i = lstr.find(oldstr, i)) != string::npos) {
578 lstr.replace(i, olen, newstr);
579 i += newstr.length(); // We need to be sure that we dont
580 // use the same i over and over again.
588 string const subst(string const & a, char oldchar, char newchar)
590 return subst_char(a, oldchar, newchar);
594 docstring const subst(docstring const & a,
595 char_type oldchar, char_type newchar)
597 return subst_char(a, oldchar, newchar);
601 string const subst(string const & a,
602 string const & oldstr, string const & newstr)
604 return subst_string(a, oldstr, newstr);
608 docstring const subst(docstring const & a,
609 docstring const & oldstr, docstring const & newstr)
611 return subst_string(a, oldstr, newstr);
615 docstring const trim(docstring const & a, char const * p)
619 if (a.empty() || !*p)
622 docstring s = lyx::from_ascii(p);
623 docstring::size_type r = a.find_last_not_of(s);
624 docstring::size_type l = a.find_first_not_of(s);
626 // Is this the minimal test? (lgb)
627 if (r == docstring::npos && l == docstring::npos)
630 return a.substr(l, r - l + 1);
634 string const trim(string const & a, char const * p)
638 if (a.empty() || !*p)
641 string::size_type r = a.find_last_not_of(p);
642 string::size_type l = a.find_first_not_of(p);
644 // Is this the minimal test? (lgb)
645 if (r == string::npos && l == string::npos)
648 return a.substr(l, r - l + 1);
652 string const rtrim(string const & a, char const * p)
656 if (a.empty() || !*p)
659 string::size_type r = a.find_last_not_of(p);
661 // Is this test really needed? (Lgb)
662 if (r == string::npos)
665 return a.substr(0, r + 1);
669 string const ltrim(string const & a, char const * p)
672 if (a.empty() || !*p)
674 string::size_type l = a.find_first_not_of(p);
675 if (l == string::npos)
677 return a.substr(l, string::npos);
681 docstring const ltrim(docstring const & a, char const * p)
684 if (a.empty() || !*p)
686 size_t l = a.find_first_not_of(from_ascii(p));
687 if (l == docstring::npos)
689 return a.substr(l, docstring::npos);
694 template<typename String, typename Char> inline
695 String const doSplit(String const & a, String & piece, Char delim)
698 typename String::size_type i = a.find(delim);
699 if (i == a.length() - 1) {
700 piece = a.substr(0, i);
701 } else if (i != String::npos) {
702 piece = a.substr(0, i);
703 tmp = a.substr(i + 1);
706 tmp = a.substr(i + 1);
716 string const split(string const & a, string & piece, char delim)
718 return doSplit(a, piece, delim);
722 docstring const split(docstring const & a, docstring & piece, char_type delim)
724 return doSplit(a, piece, delim);
728 string const split(string const & a, char delim)
731 string::size_type i = a.find(delim);
732 if (i != string::npos) // found delim
733 tmp = a.substr(i + 1);
739 string const rsplit(string const & a, string & piece, char delim)
742 string::size_type i = a.rfind(delim);
743 if (i != string::npos) { // delimiter was found
744 piece = a.substr(0, i);
745 tmp = a.substr(i + 1);
746 } else { // delimiter was not found
753 docstring const escape(docstring const & lab)
755 char_type hexdigit[16] = { '0', '1', '2', '3', '4', '5', '6', '7',
756 '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' };
758 for (docstring::size_type i = 0; i < lab.length(); ++i) {
759 char_type c = lab[i];
760 if (c >= 128 || c == '=' || c == '%') {
761 // Although char_type is a 32 bit type we know that
762 // UCS4 occupies only 21 bits, so we don't need to
763 // encode bigger values. Test for 2^24 because we
764 // can encode that with the 6 hex digits that are
765 // needed for 21 bits anyway.
766 BOOST_ASSERT(c < (1 << 24));
768 enc += hexdigit[(c>>20) & 15];
769 enc += hexdigit[(c>>16) & 15];
770 enc += hexdigit[(c>>12) & 15];
771 enc += hexdigit[(c>> 8) & 15];
772 enc += hexdigit[(c>> 4) & 15];
773 enc += hexdigit[ c & 15];
782 /// gives a vector of stringparts which have the delimiter delim
783 vector<string> const getVectorFromString(string const & str,
784 string const & delim)
786 // Lars would like this code to go, but for now his replacement (below)
787 // doesn't fullfil the same function. I have, therefore, reactivated the
788 // old code for now. Angus 11 Nov 2002.
793 string keys = rtrim(str);
795 string::size_type const idx = keys.find(delim);
796 if (idx == string::npos) {
797 vec.push_back(ltrim(keys));
800 string const key = trim(keys.substr(0, idx));
803 string::size_type const start = idx + delim.size();
804 keys = keys.substr(start);
808 boost::char_separator<char> sep(delim.c_str());
809 boost::tokenizer<boost::char_separator<char> > tokens(str, sep);
810 return vector<string>(tokens.begin(), tokens.end());
815 // the same vice versa
816 string const getStringFromVector(vector<string> const & vec,
817 string const & delim)
821 for (vector<string>::const_iterator it = vec.begin();
822 it != vec.end(); ++it) {
823 string item = trim(*it);
834 int findToken(char const * const str[], string const & search_token)
838 while (str[i][0] && str[i] != search_token)
846 docstring const externalLineEnding(docstring const & str)
848 #if defined(__APPLE__)
849 // The MAC clipboard uses \r for lineendings, and we use \n
850 return subst(str, '\n', '\r');
851 #elif defined (_WIN32) || (defined (__CYGWIN__) && defined (X_DISPLAY_MISSING))
852 // Windows clipboard uses \r\n for lineendings, and we use \n
853 return subst(str, lyx::from_ascii("\n"), lyx::from_ascii("\r\n"));
860 docstring const internalLineEnding(docstring const & str)
862 docstring const s = subst(str,
863 lyx::from_ascii("\r\n"), lyx::from_ascii("\n"));
864 return subst(s, '\r', '\n');
868 #ifndef I_AM_NOT_AFRAID_OF_HEADER_LIBRARIES
872 docstring bformat(docstring const & fmt, int arg1)
874 return (boost::basic_format<char_type>(fmt) % arg1).str();
879 docstring bformat(docstring const & fmt, long arg1)
881 return (boost::basic_format<char_type>(fmt) % arg1).str();
886 docstring bformat(docstring const & fmt, unsigned int arg1)
888 return (boost::basic_format<char_type>(fmt) % arg1).str();
893 docstring bformat<docstring>(docstring const & fmt, docstring arg1)
895 return (boost::basic_format<char_type>(fmt) % arg1).str();
900 docstring bformat(docstring const & fmt, char * arg1)
902 return (boost::basic_format<char_type>(fmt) % arg1).str();
907 docstring bformat(docstring const & fmt, int arg1, int arg2)
909 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
914 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2)
916 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
921 docstring bformat(docstring const & fmt, char const * arg1, docstring arg2)
923 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
928 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2, docstring arg3)
930 return (boost::basic_format<char_type>(fmt) % arg1 % arg2 % arg3).str();
935 docstring bformat(docstring const & fmt,
936 docstring arg1, docstring arg2, docstring arg3, docstring arg4)
938 return (boost::basic_format<char_type>(fmt) % arg1 % arg2 % arg3 % arg4).str();
944 docstring bformat(docstring const & fmt, 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, long arg1)
955 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
956 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
957 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
962 docstring bformat(docstring const & fmt, unsigned int arg1)
964 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
965 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
966 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
971 docstring bformat(docstring const & fmt, docstring arg1)
973 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
974 docstring const str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
975 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
980 docstring bformat(docstring const & fmt, char * arg1)
982 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
983 docstring const str = subst(fmt, lyx::from_ascii("%1$s"), lyx::from_ascii(arg1));
984 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
989 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2)
991 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
992 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
993 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
994 str = subst(str, lyx::from_ascii("%2$s"), arg2);
995 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1000 docstring bformat(docstring const & fmt, char const * arg1, docstring arg2)
1002 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1003 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1004 docstring str = subst(fmt, lyx::from_ascii("%1$s"), lyx::from_ascii(arg1));
1005 str = subst(fmt, lyx::from_ascii("%2$s"), arg2);
1006 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1011 docstring bformat(docstring const & fmt, int arg1, int arg2)
1013 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
1014 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$d")));
1015 docstring str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
1016 str = subst(str, lyx::from_ascii("%2$d"), convert<docstring>(arg2));
1017 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1022 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2, docstring arg3)
1024 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1025 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1026 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%3$s")));
1027 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
1028 str = subst(str, lyx::from_ascii("%2$s"), arg2);
1029 str = subst(str, lyx::from_ascii("%3$s"), arg3);
1030 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1035 docstring bformat(docstring const & fmt,
1036 docstring arg1, docstring arg2, docstring arg3, docstring arg4)
1038 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1039 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1040 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%3$s")));
1041 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%4$s")));
1042 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
1043 str = subst(str, lyx::from_ascii("%2$s"), arg2);
1044 str = subst(str, lyx::from_ascii("%3$s"), arg3);
1045 str = subst(str, lyx::from_ascii("%4$s"), arg4);
1046 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1052 } // namespace support