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"
16 #include "support/convert.h"
17 #include "support/qstring_helpers.h"
18 #include "support/textutils.h"
20 #include <boost/tokenizer.hpp>
21 #include <boost/assert.hpp>
29 // Using this allows us to have docstring default arguments in headers
30 // without #include "support/docstring" there.
31 docstring const & empty_docstring()
37 // Using this allows us to have string default arguments in headers
38 // without #include <string>
39 string const & empty_string()
46 * Convert a QChar into a UCS4 character.
47 * This is a hack (it does only make sense for the common part of the UCS4
48 * and UTF16 encodings) and should not be used.
49 * This does only exist because of performance reasons (a real conversion
50 * using iconv is too slow on windows).
52 static inline char_type qchar_to_ucs4(QChar const & qchar)
54 BOOST_ASSERT(is_utf16(static_cast<char_type>(qchar.unicode())));
55 return static_cast<char_type>(qchar.unicode());
60 * Convert a UCS4 character into a QChar.
61 * This is a hack (it does only make sense for the common part of the UCS4
62 * and UTF16 encodings) and should not be used.
63 * This does only exist because of performance reasons (a real conversion
64 * using iconv is too slow on windows).
66 static inline QChar const ucs4_to_qchar(char_type const ucs4)
68 BOOST_ASSERT(is_utf16(ucs4));
69 return QChar(static_cast<unsigned short>(ucs4));
74 /// Maximum valid UCS4 code point
75 char_type const ucs4_max = 0x10ffff;
79 bool isLetterChar(char_type c)
83 // outside the UCS4 range
85 // assume that all non-utf16 characters are letters
88 return ucs4_to_qchar(c).isLetter();
92 bool isAlphaASCII(char_type c)
94 return ('a' <= c && c <= 'z') || ('A' <= c && c <= 'Z');
98 bool isPrintable(char_type c)
102 // outside the UCS4 range
104 // assume that all non-utf16 characters are printable
107 return ucs4_to_qchar(c).isPrint();
111 bool isPrintableNonspace(char_type c)
115 // outside the UCS4 range
117 // assume that all non-utf16 characters are printable and
121 QChar const qc = ucs4_to_qchar(c);
122 return qc.isPrint() && !qc.isSpace();
126 bool isSpace(char_type c)
129 // assume that no non-utf16 character is a space
130 // c outside the UCS4 range is catched as well
133 QChar const qc = ucs4_to_qchar(c);
138 bool isDigit(char_type c)
141 // assume that no non-utf16 character is a digit
142 // c outside the UCS4 range is catched as well
144 return ucs4_to_qchar(c).isDigit();
148 bool isDigitASCII(char_type c)
150 return '0' <= c && c <= '9';
155 int compare_no_case(docstring const & s, docstring const & s2)
157 docstring::const_iterator p = s.begin();
158 docstring::const_iterator p2 = s2.begin();
160 while (p != s.end() && p2 != s2.end()) {
161 char_type const lc1 = lowercase(*p);
162 char_type const lc2 = lowercase(*p2);
164 return (lc1 < lc2) ? -1 : 1;
169 if (s.size() == s2.size())
171 if (s.size() < s2.size())
179 template<typename Char>
180 Char ascii_tolower(Char c) {
181 if (c >= 'A' && c <= 'Z')
182 return c - 'A' + 'a';
189 int compare_ascii_no_case(string const & s, string const & s2)
191 string::const_iterator p = s.begin();
192 string::const_iterator p2 = s2.begin();
194 while (p != s.end() && p2 != s2.end()) {
195 int const lc1 = ascii_tolower(*p);
196 int const lc2 = ascii_tolower(*p2);
198 return (lc1 < lc2) ? -1 : 1;
203 if (s.size() == s2.size())
205 if (s.size() < s2.size())
211 int compare_ascii_no_case(docstring const & s, docstring const & s2)
213 docstring::const_iterator p = s.begin();
214 docstring::const_iterator p2 = s2.begin();
216 while (p != s.end() && p2 != s2.end()) {
217 char_type const lc1 = ascii_tolower(*p);
218 char_type const lc2 = ascii_tolower(*p2);
220 return (lc1 < lc2) ? -1 : 1;
225 if (s.size() == s2.size())
227 if (s.size() < s2.size())
233 bool isStrInt(string const & str)
238 // Remove leading and trailing white space chars.
239 string const tmpstr = trim(str);
243 string::const_iterator cit = tmpstr.begin();
247 string::const_iterator end = tmpstr.end();
248 for (; cit != end; ++cit)
249 if (!isdigit((*cit)))
256 bool isStrUnsignedInt(string const & str)
261 // Remove leading and trailing white space chars.
262 string const tmpstr = trim(str);
266 string::const_iterator cit = tmpstr.begin();
267 string::const_iterator end = tmpstr.end();
268 for (; cit != end; ++cit)
269 if (!isdigit((*cit)))
276 bool isStrDbl(string const & str)
281 // Remove leading and trailing white space chars.
282 string const tmpstr = trim(str);
285 // if (tmpstr.count('.') > 1) return false;
287 string::const_iterator cit = tmpstr.begin();
288 bool found_dot = false;
291 string::const_iterator end = tmpstr.end();
292 for (; cit != end; ++cit) {
293 if (!isdigit(*cit) && *cit != '.')
305 static bool isHexChar(char_type c)
317 c == 'a' || c == 'A' ||
318 c == 'b' || c == 'B' ||
319 c == 'c' || c == 'C' ||
320 c == 'd' || c == 'D' ||
321 c == 'e' || c == 'E' ||
322 c == 'f' || c == 'F';
326 bool isHex(docstring const & str)
330 if (str.length() > 2 && str[0] == '0' &&
331 (str[1] == 'x' || str[1] == 'X'))
334 int const len = str.length();
336 for (; index < len; ++index) {
337 if (!isHexChar(str[index]))
344 int hexToInt(docstring const & str)
346 string s = to_ascii(str);
348 sscanf(s.c_str(), "%x", &h);
353 bool isAscii(docstring const & str)
355 int const len = str.length();
356 for (int i = 0; i < len; ++i)
363 bool isAscii(string const & str)
365 int const len = str.length();
366 for (int i = 0; i < len; ++i)
367 if (static_cast<unsigned char>(str[i]) >= 0x80)
373 char lowercase(char c)
375 BOOST_ASSERT(static_cast<unsigned char>(c) < 0x80);
376 return char(tolower(c));
380 char uppercase(char c)
382 BOOST_ASSERT(static_cast<unsigned char>(c) < 0x80);
383 return char(toupper(c));
387 char_type lowercase(char_type c)
390 // We don't know how to lowercase a non-utf16 char
392 return qchar_to_ucs4(ucs4_to_qchar(c).toLower());
396 char_type uppercase(char_type c)
399 // We don't know how to uppercase a non-utf16 char
401 return qchar_to_ucs4(ucs4_to_qchar(c).toUpper());
407 // since we cannot use tolower and toupper directly in the
408 // calls to transform yet, we use these helper clases. (Lgb)
410 struct local_lowercase {
411 char_type operator()(char_type c) const {
413 // We don't know how to lowercase a non-utf16 char
415 return qchar_to_ucs4(ucs4_to_qchar(c).toLower());
419 struct local_uppercase {
420 char_type operator()(char_type c) const {
422 // We don't know how to uppercase a non-utf16 char
424 return qchar_to_ucs4(ucs4_to_qchar(c).toUpper());
428 template<typename Char> struct local_ascii_lowercase {
429 Char operator()(Char c) const { return ascii_tolower(c); }
432 } // end of anon namespace
434 docstring const lowercase(docstring const & a)
437 transform(tmp.begin(), tmp.end(), tmp.begin(), local_lowercase());
442 docstring const uppercase(docstring const & a)
445 transform(tmp.begin(), tmp.end(), tmp.begin(), local_uppercase());
450 string const ascii_lowercase(string const & a)
453 transform(tmp.begin(), tmp.end(), tmp.begin(),
454 local_ascii_lowercase<char>());
459 docstring const ascii_lowercase(docstring const & a)
462 transform(tmp.begin(), tmp.end(), tmp.begin(),
463 local_ascii_lowercase<char_type>());
468 bool prefixIs(docstring const & a, char_type c)
476 bool prefixIs(string const & a, string const & pre)
478 size_t const prelen = pre.length();
479 size_t const alen = a.length();
481 if (prelen > alen || a.empty())
484 return a.compare(0, prelen, pre) == 0;
488 bool prefixIs(docstring const & a, docstring const & pre)
490 size_t const prelen = pre.length();
491 size_t const alen = a.length();
493 if (prelen > alen || a.empty())
496 return a.compare(0, prelen, pre) == 0;
500 bool suffixIs(string const & a, char c)
502 if (a.empty()) return false;
503 return a[a.length() - 1] == c;
507 bool suffixIs(docstring const & a, char_type c)
511 return a[a.length() - 1] == c;
515 bool suffixIs(string const & a, string const & suf)
517 size_t const suflen = suf.length();
518 size_t const alen = a.length();
522 return a.compare(alen - suflen, suflen, suf) == 0;
526 bool containsOnly(string const & s, string const & cset)
528 return s.find_first_not_of(cset) == string::npos;
532 // ale970405+lasgoutt-970425
533 // rewritten to use new string (Lgb)
534 string const token(string const & a, char delim, int n)
542 // Find delimiter or end of string
544 if ((i = a.find(delim, i)) == string::npos)
550 // i is now the n'th delim (or string::npos)
551 if (i == string::npos)
554 k = a.find(delim, i);
555 // k is now the n'th + 1 delim (or string::npos)
557 return a.substr(i, k - i);
561 docstring const token(docstring const & a, char_type delim, int n)
569 // Find delimiter or end of string
571 if ((i = a.find(delim, i)) == docstring::npos)
577 // i is now the n'th delim (or string::npos)
578 if (i == docstring::npos)
581 k = a.find(delim, i);
582 // k is now the n'th + 1 delim (or string::npos)
584 return a.substr(i, k - i);
588 // this could probably be faster and/or cleaner, but it seems to work (JMarc)
589 // rewritten to use new string (Lgb)
590 int tokenPos(string const & a, char delim, string const & tok)
596 while (!str.empty()) {
597 str = split(str, tmptok, delim);
608 /// Substitute all \a oldchar with \a newchar
609 template<typename Ch> inline
610 basic_string<Ch> const subst_char(basic_string<Ch> const & a,
611 Ch oldchar, Ch newchar)
613 typedef basic_string<Ch> String;
615 typename String::iterator lit = tmp.begin();
616 typename String::iterator end = tmp.end();
617 for (; lit != end; ++lit)
618 if ((*lit) == oldchar)
623 /// Substitute all \a oldchar with \a newchar
624 docstring const subst_char(docstring const & a,
625 docstring::value_type oldchar, docstring::value_type newchar)
628 docstring::iterator lit = tmp.begin();
629 docstring::iterator end = tmp.end();
630 for (; lit != end; ++lit)
631 if ((*lit) == oldchar)
637 /// substitutes all instances of \a oldstr with \a newstr
638 template<typename String> inline
639 String const subst_string(String const & a,
640 String const & oldstr, String const & newstr)
642 BOOST_ASSERT(!oldstr.empty());
645 size_t const olen = oldstr.length();
646 while ((i = lstr.find(oldstr, i)) != string::npos) {
647 lstr.replace(i, olen, newstr);
648 i += newstr.length(); // We need to be sure that we dont
649 // use the same i over and over again.
654 docstring const subst_string(docstring const & a,
655 docstring const & oldstr, docstring const & newstr)
657 BOOST_ASSERT(!oldstr.empty());
660 size_t const olen = oldstr.length();
661 while ((i = lstr.find(oldstr, i)) != string::npos) {
662 lstr.replace(i, olen, newstr);
663 i += newstr.length(); // We need to be sure that we dont
664 // use the same i over and over again.
672 string const subst(string const & a, char oldchar, char newchar)
674 return subst_char(a, oldchar, newchar);
678 docstring const subst(docstring const & a,
679 char_type oldchar, char_type newchar)
681 return subst_char(a, oldchar, newchar);
685 string const subst(string const & a,
686 string const & oldstr, string const & newstr)
688 return subst_string(a, oldstr, newstr);
692 docstring const subst(docstring const & a,
693 docstring const & oldstr, docstring const & newstr)
695 return subst_string(a, oldstr, newstr);
699 docstring const trim(docstring const & a, char const * p)
703 if (a.empty() || !*p)
706 docstring s = from_ascii(p);
707 size_t r = a.find_last_not_of(s);
708 size_t l = a.find_first_not_of(s);
710 // Is this the minimal test? (lgb)
711 if (r == docstring::npos && l == docstring::npos)
714 return a.substr(l, r - l + 1);
718 string const trim(string const & a, char const * p)
722 if (a.empty() || !*p)
725 size_t r = a.find_last_not_of(p);
726 size_t l = a.find_first_not_of(p);
728 // Is this the minimal test? (lgb)
729 if (r == string::npos && l == string::npos)
732 return a.substr(l, r - l + 1);
736 string const rtrim(string const & a, char const * p)
740 if (a.empty() || !*p)
743 size_t r = a.find_last_not_of(p);
745 // Is this test really needed? (Lgb)
746 if (r == string::npos)
749 return a.substr(0, r + 1);
753 docstring const rtrim(docstring const & a, char const * p)
757 if (a.empty() || !*p)
760 size_t r = a.find_last_not_of(from_ascii(p));
762 // Is this test really needed? (Lgb)
763 if (r == docstring::npos)
766 return a.substr(0, r + 1);
770 string const ltrim(string const & a, char const * p)
773 if (a.empty() || !*p)
775 size_t l = a.find_first_not_of(p);
776 if (l == string::npos)
778 return a.substr(l, string::npos);
782 docstring const ltrim(docstring const & a, char const * p)
785 if (a.empty() || !*p)
787 size_t l = a.find_first_not_of(from_ascii(p));
788 if (l == docstring::npos)
790 return a.substr(l, docstring::npos);
795 template<typename String, typename Char> inline
796 String const doSplit(String const & a, String & piece, Char delim)
799 size_t i = a.find(delim);
800 if (i == a.length() - 1) {
801 piece = a.substr(0, i);
802 } else if (i != String::npos) {
803 piece = a.substr(0, i);
804 tmp = a.substr(i + 1);
807 tmp = a.substr(i + 1);
814 template<typename Char> inline
815 docstring const doSplit(docstring const & a, docstring & piece, Char delim)
818 size_t i = a.find(delim);
819 if (i == a.length() - 1) {
820 piece = a.substr(0, i);
821 } else if (i != docstring::npos) {
822 piece = a.substr(0, i);
823 tmp = a.substr(i + 1);
826 tmp = a.substr(i + 1);
836 string const split(string const & a, string & piece, char delim)
838 return doSplit(a, piece, delim);
842 docstring const split(docstring const & a, docstring & piece, char_type delim)
844 return doSplit(a, piece, delim);
848 string const split(string const & a, char delim)
851 size_t i = a.find(delim);
852 if (i != string::npos) // found delim
853 tmp = a.substr(i + 1);
859 string const rsplit(string const & a, string & piece, char delim)
862 size_t i = a.rfind(delim);
863 if (i != string::npos) { // delimiter was found
864 piece = a.substr(0, i);
865 tmp = a.substr(i + 1);
866 } else { // delimiter was not found
873 docstring const escape(docstring const & lab)
875 char_type hexdigit[16] = { '0', '1', '2', '3', '4', '5', '6', '7',
876 '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' };
878 for (size_t i = 0; i < lab.length(); ++i) {
879 char_type c = lab[i];
880 if (c >= 128 || c == '=' || c == '%') {
881 // Although char_type is a 32 bit type we know that
882 // UCS4 occupies only 21 bits, so we don't need to
883 // encode bigger values. Test for 2^24 because we
884 // can encode that with the 6 hex digits that are
885 // needed for 21 bits anyway.
886 BOOST_ASSERT(c < (1 << 24));
888 enc += hexdigit[(c>>20) & 15];
889 enc += hexdigit[(c>>16) & 15];
890 enc += hexdigit[(c>>12) & 15];
891 enc += hexdigit[(c>> 8) & 15];
892 enc += hexdigit[(c>> 4) & 15];
893 enc += hexdigit[ c & 15];
904 template<typename String> vector<String> const
905 getVectorFromStringT(String const & str, String const & delim)
907 // Lars would like this code to go, but for now his replacement (below)
908 // doesn't fullfil the same function. I have, therefore, reactivated the
909 // old code for now. Angus 11 Nov 2002.
914 String keys = rtrim(str);
916 size_t const idx = keys.find(delim);
917 if (idx == String::npos) {
918 vec.push_back(ltrim(keys));
921 String const key = trim(keys.substr(0, idx));
924 size_t const start = idx + delim.size();
925 keys = keys.substr(start);
929 typedef boost::char_separator<typename String::value_type> Separator;
930 typedef boost::tokenizer<Separator, typename String::const_iterator, String> Tokenizer;
931 Separator sep(delim.c_str());
932 Tokenizer tokens(str, sep);
933 return vector<String>(tokens.begin(), tokens.end());
940 vector<string> const getVectorFromString(string const & str,
941 string const & delim)
943 return getVectorFromStringT<string>(str, delim);
947 vector<docstring> const getVectorFromString(docstring const & str,
948 docstring const & delim)
950 return getVectorFromStringT<docstring>(str, delim);
954 // the same vice versa
955 string const getStringFromVector(vector<string> const & vec,
956 string const & delim)
960 for (vector<string>::const_iterator it = vec.begin();
961 it != vec.end(); ++it) {
962 string item = trim(*it);
973 int findToken(char const * const str[], string const & search_token)
977 while (str[i][0] && str[i] != search_token)
985 docstring const externalLineEnding(docstring const & str)
987 #if defined(__APPLE__)
988 // The MAC clipboard uses \r for lineendings, and we use \n
989 return subst(str, '\n', '\r');
990 #elif defined (_WIN32) || (defined (__CYGWIN__) && defined (X_DISPLAY_MISSING))
991 // Windows clipboard uses \r\n for lineendings, and we use \n
992 return subst(str, from_ascii("\n"), from_ascii("\r\n"));
999 docstring const internalLineEnding(docstring const & str)
1001 docstring const s = subst(str, from_ascii("\r\n"), from_ascii("\n"));
1002 return subst(s, '\r', '\n');
1007 docstring bformat(docstring const & fmt, int arg1)
1009 BOOST_ASSERT(contains(fmt, from_ascii("%1$d")));
1010 docstring const str = subst(fmt, from_ascii("%1$d"), convert<docstring>(arg1));
1011 return subst(str, from_ascii("%%"), from_ascii("%"));
1016 docstring bformat(docstring const & fmt, long arg1)
1018 BOOST_ASSERT(contains(fmt, from_ascii("%1$d")));
1019 docstring const str = subst(fmt, from_ascii("%1$d"), convert<docstring>(arg1));
1020 return subst(str, from_ascii("%%"), from_ascii("%"));
1025 docstring bformat(docstring const & fmt, unsigned int arg1)
1027 BOOST_ASSERT(contains(fmt, from_ascii("%1$d")));
1028 docstring const str = subst(fmt, from_ascii("%1$d"), convert<docstring>(arg1));
1029 return subst(str, from_ascii("%%"), from_ascii("%"));
1034 docstring bformat(docstring const & fmt, docstring arg1)
1036 BOOST_ASSERT(contains(fmt, from_ascii("%1$s")));
1037 docstring const str = subst(fmt, from_ascii("%1$s"), arg1);
1038 return subst(str, from_ascii("%%"), from_ascii("%"));
1043 docstring bformat(docstring const & fmt, char * arg1)
1045 BOOST_ASSERT(contains(fmt, from_ascii("%1$s")));
1046 docstring const str = subst(fmt, from_ascii("%1$s"), from_ascii(arg1));
1047 return subst(str, from_ascii("%%"), from_ascii("%"));
1052 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2)
1054 BOOST_ASSERT(contains(fmt, from_ascii("%1$s")));
1055 BOOST_ASSERT(contains(fmt, from_ascii("%2$s")));
1056 docstring str = subst(fmt, from_ascii("%1$s"), arg1);
1057 str = subst(str, from_ascii("%2$s"), arg2);
1058 return subst(str, from_ascii("%%"), from_ascii("%"));
1063 docstring bformat(docstring const & fmt, char const * arg1, docstring arg2)
1065 BOOST_ASSERT(contains(fmt, from_ascii("%1$s")));
1066 BOOST_ASSERT(contains(fmt, from_ascii("%2$s")));
1067 docstring str = subst(fmt, from_ascii("%1$s"), from_ascii(arg1));
1068 str = subst(fmt, from_ascii("%2$s"), arg2);
1069 return subst(str, from_ascii("%%"), from_ascii("%"));
1074 docstring bformat(docstring const & fmt, int arg1, int arg2)
1076 BOOST_ASSERT(contains(fmt, from_ascii("%1$d")));
1077 BOOST_ASSERT(contains(fmt, from_ascii("%2$d")));
1078 docstring str = subst(fmt, from_ascii("%1$d"), convert<docstring>(arg1));
1079 str = subst(str, from_ascii("%2$d"), convert<docstring>(arg2));
1080 return subst(str, from_ascii("%%"), from_ascii("%"));
1085 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2, docstring arg3)
1087 BOOST_ASSERT(contains(fmt, from_ascii("%1$s")));
1088 BOOST_ASSERT(contains(fmt, from_ascii("%2$s")));
1089 BOOST_ASSERT(contains(fmt, from_ascii("%3$s")));
1090 docstring str = subst(fmt, from_ascii("%1$s"), arg1);
1091 str = subst(str, from_ascii("%2$s"), arg2);
1092 str = subst(str, from_ascii("%3$s"), arg3);
1093 return subst(str, from_ascii("%%"), from_ascii("%"));
1098 docstring bformat(docstring const & fmt,
1099 docstring arg1, docstring arg2, docstring arg3, docstring arg4)
1101 BOOST_ASSERT(contains(fmt, from_ascii("%1$s")));
1102 BOOST_ASSERT(contains(fmt, from_ascii("%2$s")));
1103 BOOST_ASSERT(contains(fmt, from_ascii("%3$s")));
1104 BOOST_ASSERT(contains(fmt, from_ascii("%4$s")));
1105 docstring str = subst(fmt, from_ascii("%1$s"), arg1);
1106 str = subst(str, from_ascii("%2$s"), arg2);
1107 str = subst(str, from_ascii("%3$s"), arg3);
1108 str = subst(str, from_ascii("%4$s"), arg4);
1109 return subst(str, from_ascii("%%"), from_ascii("%"));
1112 } // namespace support