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"
17 #include "support/qstring_helpers.h"
21 #include <boost/tokenizer.hpp>
22 #include <boost/assert.hpp>
24 #ifndef I_AM_NOT_AFRAID_OF_HEADER_LIBRARIES
26 #include <boost/format.hpp>
41 #ifndef CXX_GLOBAL_CSTD
51 int compare_no_case(docstring const & s, docstring const & s2)
53 docstring::const_iterator p = s.begin();
54 docstring::const_iterator p2 = s2.begin();
56 while (p != s.end() && p2 != s2.end()) {
57 char_type const lc1 = lowercase(*p);
58 char_type const lc2 = lowercase(*p2);
60 return (lc1 < lc2) ? -1 : 1;
65 if (s.size() == s2.size())
67 if (s.size() < s2.size())
75 template<typename Char>
76 Char ascii_tolower(Char c) {
77 if (c >= 'A' && c <= 'Z')
85 int compare_ascii_no_case(string const & s, string const & s2)
87 string::const_iterator p = s.begin();
88 string::const_iterator p2 = s2.begin();
90 while (p != s.end() && p2 != s2.end()) {
91 int const lc1 = ascii_tolower(*p);
92 int const lc2 = ascii_tolower(*p2);
94 return (lc1 < lc2) ? -1 : 1;
99 if (s.size() == s2.size())
101 if (s.size() < s2.size())
107 int compare_ascii_no_case(docstring const & s, docstring const & s2)
109 docstring::const_iterator p = s.begin();
110 docstring::const_iterator p2 = s2.begin();
112 while (p != s.end() && p2 != s2.end()) {
113 char_type const lc1 = ascii_tolower(*p);
114 char_type const lc2 = ascii_tolower(*p2);
116 return (lc1 < lc2) ? -1 : 1;
121 if (s.size() == s2.size())
123 if (s.size() < s2.size())
129 bool isStrInt(string const & str)
131 if (str.empty()) return false;
133 // Remove leading and trailing white space chars.
134 string const tmpstr = trim(str);
135 if (tmpstr.empty()) return false;
137 string::const_iterator cit = tmpstr.begin();
138 if ((*cit) == '-') ++cit;
139 string::const_iterator end = tmpstr.end();
140 for (; cit != end; ++cit) {
141 if (!isdigit((*cit))) return false;
147 bool isStrUnsignedInt(string const & str)
149 if (str.empty()) return false;
151 // Remove leading and trailing white space chars.
152 string const tmpstr = trim(str);
153 if (tmpstr.empty()) return false;
155 string::const_iterator cit = tmpstr.begin();
156 string::const_iterator end = tmpstr.end();
157 for (; cit != end; ++cit) {
158 if (!isdigit((*cit))) return false;
164 bool isStrDbl(string const & str)
166 if (str.empty()) return false;
168 // Remove leading and trailing white space chars.
169 string const tmpstr = trim(str);
170 if (tmpstr.empty()) return false;
171 // if (1 < tmpstr.count('.')) return false;
173 string::const_iterator cit = tmpstr.begin();
174 bool found_dot(false);
175 if ((*cit) == '-') ++cit;
176 string::const_iterator end = tmpstr.end();
177 for (; cit != end; ++cit) {
197 bool isHexChar(char_type c)
209 c == 'a' || c == 'A' ||
210 c == 'b' || c == 'B' ||
211 c == 'c' || c == 'C' ||
212 c == 'd' || c == 'D' ||
213 c == 'e' || c == 'E' ||
214 c == 'f' || c == 'F';
220 bool isHex(docstring const & str)
224 if (str.length() > 2 && str[0] == '0' &&
225 (str[1] == 'x' || str[1] == 'X'))
228 int const len = str.length();
230 for (; index < len; ++index) {
231 if (!isHexChar(str[index]))
238 int hexToInt(docstring const & str)
240 string s = to_ascii(str);
242 sscanf(s.c_str(), "%x", &h);
247 bool isAscii(docstring const & str)
249 int const len = str.length();
250 for (int i = 0; i < len; ++i)
257 char lowercase(char c)
259 BOOST_ASSERT(static_cast<unsigned char>(c) < 0x80);
260 return char(tolower(c));
264 char uppercase(char c)
266 BOOST_ASSERT(static_cast<unsigned char>(c) < 0x80);
267 return char(toupper(c));
271 char_type lowercase(char_type c)
274 // We don't know how to lowercase a non-utf16 char
276 return qchar_to_ucs4(ucs4_to_qchar(c).toLower());
280 char_type uppercase(char_type c)
283 // We don't know how to uppercase a non-utf16 char
285 return qchar_to_ucs4(ucs4_to_qchar(c).toUpper());
291 // since we cannot use std::tolower and std::toupper directly in the
292 // calls to std::transform yet, we use these helper clases. (Lgb)
294 struct local_lowercase {
295 char_type operator()(char_type c) const {
297 // We don't know how to lowercase a non-utf16 char
299 return qchar_to_ucs4(ucs4_to_qchar(c).toLower());
303 struct local_uppercase {
304 char_type operator()(char_type c) const {
306 // We don't know how to uppercase a non-utf16 char
308 return qchar_to_ucs4(ucs4_to_qchar(c).toUpper());
312 template<typename Char> struct local_ascii_lowercase {
313 Char operator()(Char c) const {
314 return ascii_tolower(c);
318 } // end of anon namespace
320 docstring const lowercase(docstring const & a)
323 transform(tmp.begin(), tmp.end(), tmp.begin(), local_lowercase());
328 docstring const uppercase(docstring const & a)
331 transform(tmp.begin(), tmp.end(), tmp.begin(), local_uppercase());
336 string const ascii_lowercase(string const & a)
339 transform(tmp.begin(), tmp.end(), tmp.begin(),
340 local_ascii_lowercase<char>());
345 docstring const ascii_lowercase(docstring const & a)
348 transform(tmp.begin(), tmp.end(), tmp.begin(),
349 local_ascii_lowercase<char_type>());
354 bool prefixIs(docstring const & a, char_type c)
362 bool prefixIs(string const & a, string const & pre)
364 string::size_type const prelen = pre.length();
365 string::size_type const alen = a.length();
367 if (prelen > alen || a.empty())
370 #if defined(STD_STRING_IS_GOOD)
371 return a.compare(0, prelen, pre) == 0;
373 return ::strncmp(a.c_str(), pre.c_str(), prelen) == 0;
379 bool prefixIs(docstring const & a, docstring const & pre)
381 docstring::size_type const prelen = pre.length();
382 docstring::size_type const alen = a.length();
384 if (prelen > alen || a.empty())
387 return a.compare(0, prelen, pre) == 0;
391 bool suffixIs(string const & a, char c)
393 if (a.empty()) return false;
394 return a[a.length() - 1] == c;
398 bool suffixIs(docstring const & a, char_type c)
402 return a[a.length() - 1] == c;
406 bool suffixIs(string const & a, string const & suf)
408 string::size_type const suflen = suf.length();
409 string::size_type const alen = a.length();
414 #if !defined(USE_INCLUDED_STRING) && !defined(STD_STRING_IS_GOOD)
415 string tmp(a, alen - suflen);
416 return ::strncmp(tmp.c_str(), suf.c_str(), suflen) == 0;
418 return a.compare(alen - suflen, suflen, suf) == 0;
424 bool containsOnly(string const & s, string const & cset)
426 return s.find_first_not_of(cset) == string::npos;
430 // ale970405+lasgoutt-970425
431 // rewritten to use new string (Lgb)
432 string const token(string const & a, char delim, int n)
434 if (a.empty()) return string();
436 string::size_type k = 0;
437 string::size_type i = 0;
439 // Find delimiter or end of string
441 if ((i = a.find(delim, i)) == string::npos)
445 // i is now the n'th delim (or string::npos)
446 if (i == string::npos) return string();
447 k = a.find(delim, i);
448 // k is now the n'th + 1 delim (or string::npos)
450 return a.substr(i, k - i);
454 docstring const token(docstring const & a, char_type delim, int n)
456 if (a.empty()) return docstring();
458 string::size_type k = 0;
459 string::size_type i = 0;
461 // Find delimiter or end of string
463 if ((i = a.find(delim, i)) == docstring::npos)
467 // i is now the n'th delim (or string::npos)
468 if (i == docstring::npos) return docstring();
469 k = a.find(delim, i);
470 // k is now the n'th + 1 delim (or string::npos)
472 return a.substr(i, k - i);
476 // this could probably be faster and/or cleaner, but it seems to work (JMarc)
477 // rewritten to use new string (Lgb)
478 int tokenPos(string const & a, char delim, string const & tok)
484 while (!str.empty()) {
485 str = split(str, tmptok, delim);
496 /// Substitute all \a oldchar with \a newchar
497 template<typename Ch> inline
498 std::basic_string<Ch> const subst_char(std::basic_string<Ch> const & a,
499 Ch oldchar, Ch newchar)
501 typedef std::basic_string<Ch> String;
503 typename String::iterator lit = tmp.begin();
504 typename String::iterator end = tmp.end();
505 for (; lit != end; ++lit)
506 if ((*lit) == oldchar)
512 /// substitutes all instances of \a oldstr with \a newstr
513 template<typename String> inline
514 String const subst_string(String const & a,
515 String const & oldstr, String const & newstr)
517 BOOST_ASSERT(!oldstr.empty());
519 typename String::size_type i = 0;
520 typename String::size_type const olen = oldstr.length();
521 while ((i = lstr.find(oldstr, i)) != string::npos) {
522 lstr.replace(i, olen, newstr);
523 i += newstr.length(); // We need to be sure that we dont
524 // use the same i over and over again.
532 string const subst(string const & a, char oldchar, char newchar)
534 return subst_char(a, oldchar, newchar);
538 docstring const subst(docstring const & a,
539 char_type oldchar, char_type newchar)
541 return subst_char(a, oldchar, newchar);
545 string const subst(string const & a,
546 string const & oldstr, string const & newstr)
548 return subst_string(a, oldstr, newstr);
552 docstring const subst(docstring const & a,
553 docstring const & oldstr, docstring const & newstr)
555 return subst_string(a, oldstr, newstr);
559 docstring const trim(docstring const & a, char const * p)
563 if (a.empty() || !*p)
566 docstring s = lyx::from_ascii(p);
567 docstring::size_type r = a.find_last_not_of(s);
568 docstring::size_type l = a.find_first_not_of(s);
570 // Is this the minimal test? (lgb)
571 if (r == docstring::npos && l == docstring::npos)
574 return a.substr(l, r - l + 1);
578 string const trim(string const & a, char const * p)
582 if (a.empty() || !*p)
585 string::size_type r = a.find_last_not_of(p);
586 string::size_type l = a.find_first_not_of(p);
588 // Is this the minimal test? (lgb)
589 if (r == string::npos && l == string::npos)
592 return a.substr(l, r - l + 1);
596 string const rtrim(string const & a, char const * p)
600 if (a.empty() || !*p)
603 string::size_type r = a.find_last_not_of(p);
605 // Is this test really needed? (Lgb)
606 if (r == string::npos)
609 return a.substr(0, r + 1);
613 docstring const rtrim(docstring const & a, char const * p)
617 if (a.empty() || !*p)
620 docstring::size_type r = a.find_last_not_of(from_ascii(p));
622 // Is this test really needed? (Lgb)
623 if (r == docstring::npos)
626 return a.substr(0, r + 1);
630 string const ltrim(string const & a, char const * p)
633 if (a.empty() || !*p)
635 string::size_type l = a.find_first_not_of(p);
636 if (l == string::npos)
638 return a.substr(l, string::npos);
642 docstring const ltrim(docstring const & a, char const * p)
645 if (a.empty() || !*p)
647 size_t l = a.find_first_not_of(from_ascii(p));
648 if (l == docstring::npos)
650 return a.substr(l, docstring::npos);
655 template<typename String, typename Char> inline
656 String const doSplit(String const & a, String & piece, Char delim)
659 typename String::size_type i = a.find(delim);
660 if (i == a.length() - 1) {
661 piece = a.substr(0, i);
662 } else if (i != String::npos) {
663 piece = a.substr(0, i);
664 tmp = a.substr(i + 1);
667 tmp = a.substr(i + 1);
677 string const split(string const & a, string & piece, char delim)
679 return doSplit(a, piece, delim);
683 docstring const split(docstring const & a, docstring & piece, char_type delim)
685 return doSplit(a, piece, delim);
689 string const split(string const & a, char delim)
692 string::size_type i = a.find(delim);
693 if (i != string::npos) // found delim
694 tmp = a.substr(i + 1);
700 string const rsplit(string const & a, string & piece, char delim)
703 string::size_type i = a.rfind(delim);
704 if (i != string::npos) { // delimiter was found
705 piece = a.substr(0, i);
706 tmp = a.substr(i + 1);
707 } else { // delimiter was not found
714 docstring const escape(docstring const & lab)
716 char_type hexdigit[16] = { '0', '1', '2', '3', '4', '5', '6', '7',
717 '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' };
719 for (docstring::size_type i = 0; i < lab.length(); ++i) {
720 char_type c = lab[i];
721 if (c >= 128 || c == '=' || c == '%') {
722 // Although char_type is a 32 bit type we know that
723 // UCS4 occupies only 21 bits, so we don't need to
724 // encode bigger values. Test for 2^24 because we
725 // can encode that with the 6 hex digits that are
726 // needed for 21 bits anyway.
727 BOOST_ASSERT(c < (1 << 24));
729 enc += hexdigit[(c>>20) & 15];
730 enc += hexdigit[(c>>16) & 15];
731 enc += hexdigit[(c>>12) & 15];
732 enc += hexdigit[(c>> 8) & 15];
733 enc += hexdigit[(c>> 4) & 15];
734 enc += hexdigit[ c & 15];
745 template<typename String> vector<String> const
746 getVectorFromStringT(String const & str, String const & delim)
748 // Lars would like this code to go, but for now his replacement (below)
749 // doesn't fullfil the same function. I have, therefore, reactivated the
750 // old code for now. Angus 11 Nov 2002.
755 String keys = rtrim(str);
757 typename String::size_type const idx = keys.find(delim);
758 if (idx == String::npos) {
759 vec.push_back(ltrim(keys));
762 String const key = trim(keys.substr(0, idx));
765 typename String::size_type const start = idx + delim.size();
766 keys = keys.substr(start);
770 typedef boost::char_separator<typename String::value_type> Separator;
771 typedef boost::tokenizer<Separator, typename String::const_iterator, String> Tokenizer;
772 Separator sep(delim.c_str());
773 Tokenizer tokens(str, sep);
774 return vector<String>(tokens.begin(), tokens.end());
781 vector<string> const getVectorFromString(string const & str,
782 string const & delim)
784 return getVectorFromStringT<string>(str, delim);
788 vector<docstring> const getVectorFromString(docstring const & str,
789 docstring const & delim)
791 return getVectorFromStringT<docstring>(str, delim);
795 // the same vice versa
796 string const getStringFromVector(vector<string> const & vec,
797 string const & delim)
801 for (vector<string>::const_iterator it = vec.begin();
802 it != vec.end(); ++it) {
803 string item = trim(*it);
814 int findToken(char const * const str[], string const & search_token)
818 while (str[i][0] && str[i] != search_token)
826 docstring const externalLineEnding(docstring const & str)
828 #if defined(__APPLE__)
829 // The MAC clipboard uses \r for lineendings, and we use \n
830 return subst(str, '\n', '\r');
831 #elif defined (_WIN32) || (defined (__CYGWIN__) && defined (X_DISPLAY_MISSING))
832 // Windows clipboard uses \r\n for lineendings, and we use \n
833 return subst(str, lyx::from_ascii("\n"), lyx::from_ascii("\r\n"));
840 docstring const internalLineEnding(docstring const & str)
842 docstring const s = subst(str,
843 lyx::from_ascii("\r\n"), lyx::from_ascii("\n"));
844 return subst(s, '\r', '\n');
848 #ifndef I_AM_NOT_AFRAID_OF_HEADER_LIBRARIES
852 docstring bformat(docstring const & fmt, int arg1)
854 return (boost::basic_format<char_type>(fmt) % arg1).str();
859 docstring bformat(docstring const & fmt, long arg1)
861 return (boost::basic_format<char_type>(fmt) % arg1).str();
866 docstring bformat(docstring const & fmt, unsigned int arg1)
868 return (boost::basic_format<char_type>(fmt) % arg1).str();
873 docstring bformat<docstring>(docstring const & fmt, docstring arg1)
875 return (boost::basic_format<char_type>(fmt) % arg1).str();
880 docstring bformat(docstring const & fmt, char * arg1)
882 return (boost::basic_format<char_type>(fmt) % arg1).str();
887 docstring bformat(docstring const & fmt, int arg1, int arg2)
889 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
894 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2)
896 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
901 docstring bformat(docstring const & fmt, char const * arg1, docstring arg2)
903 return (boost::basic_format<char_type>(fmt) % arg1 % arg2).str();
908 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2, docstring arg3)
910 return (boost::basic_format<char_type>(fmt) % arg1 % arg2 % arg3).str();
915 docstring bformat(docstring const & fmt,
916 docstring arg1, docstring arg2, docstring arg3, docstring arg4)
918 return (boost::basic_format<char_type>(fmt) % arg1 % arg2 % arg3 % arg4).str();
924 docstring bformat(docstring const & fmt, int arg1)
926 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
927 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
928 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
933 docstring bformat(docstring const & fmt, long arg1)
935 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
936 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
937 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
942 docstring bformat(docstring const & fmt, unsigned int arg1)
944 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
945 docstring const str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
946 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
951 docstring bformat(docstring const & fmt, docstring arg1)
953 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
954 docstring const str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
955 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
960 docstring bformat(docstring const & fmt, char * arg1)
962 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
963 docstring const str = subst(fmt, lyx::from_ascii("%1$s"), lyx::from_ascii(arg1));
964 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
969 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2)
971 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
972 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
973 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
974 str = subst(str, lyx::from_ascii("%2$s"), arg2);
975 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
980 docstring bformat(docstring const & fmt, char const * arg1, docstring arg2)
982 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
983 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
984 docstring str = subst(fmt, lyx::from_ascii("%1$s"), lyx::from_ascii(arg1));
985 str = subst(fmt, lyx::from_ascii("%2$s"), arg2);
986 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
991 docstring bformat(docstring const & fmt, int arg1, int arg2)
993 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$d")));
994 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$d")));
995 docstring str = subst(fmt, lyx::from_ascii("%1$d"), convert<docstring>(arg1));
996 str = subst(str, lyx::from_ascii("%2$d"), convert<docstring>(arg2));
997 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1002 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2, docstring arg3)
1004 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1005 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1006 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%3$s")));
1007 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
1008 str = subst(str, lyx::from_ascii("%2$s"), arg2);
1009 str = subst(str, lyx::from_ascii("%3$s"), arg3);
1010 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1015 docstring bformat(docstring const & fmt,
1016 docstring arg1, docstring arg2, docstring arg3, docstring arg4)
1018 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%1$s")));
1019 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%2$s")));
1020 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%3$s")));
1021 BOOST_ASSERT(contains(fmt, lyx::from_ascii("%4$s")));
1022 docstring str = subst(fmt, lyx::from_ascii("%1$s"), arg1);
1023 str = subst(str, lyx::from_ascii("%2$s"), arg2);
1024 str = subst(str, lyx::from_ascii("%3$s"), arg3);
1025 str = subst(str, lyx::from_ascii("%4$s"), arg4);
1026 return subst(str, lyx::from_ascii("%%"), lyx::from_ascii("%"));
1032 } // namespace support