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
10 * Full author contact details are available in file CREDITS.
15 #include "support/lstrings.h"
17 #include "support/convert.h"
18 #include "support/qstring_helpers.h"
19 #include "support/textutils.h"
21 #include <boost/tokenizer.hpp>
22 #include "support/lassert.h"
34 // Using this allows us to have docstring default arguments in headers
35 // without #include "support/docstring" there.
36 docstring const & empty_docstring()
42 // Using this allows us to have string default arguments in headers
43 // without #include <string>
44 string const & empty_string()
51 * Convert a QChar into a UCS4 character.
52 * This is a hack (it does only make sense for the common part of the UCS4
53 * and UTF16 encodings) and should not be used.
54 * This does only exist because of performance reasons (a real conversion
55 * using iconv is too slow on windows).
57 static inline char_type qchar_to_ucs4(QChar const & qchar)
59 LASSERT(is_utf16(static_cast<char_type>(qchar.unicode())), /**/);
60 return static_cast<char_type>(qchar.unicode());
65 * Convert a UCS4 character into a QChar.
66 * This is a hack (it does only make sense for the common part of the UCS4
67 * and UTF16 encodings) and should not be used.
68 * This does only exist because of performance reasons (a real conversion
69 * using iconv is too slow on windows).
71 static inline QChar const ucs4_to_qchar(char_type const ucs4)
73 LASSERT(is_utf16(ucs4), /**/);
74 return QChar(static_cast<unsigned short>(ucs4));
79 /// Maximum valid UCS4 code point
80 char_type const ucs4_max = 0x10ffff;
84 bool isLetterChar(char_type c)
88 // outside the UCS4 range
90 // assume that all non-utf16 characters are letters
93 return ucs4_to_qchar(c).isLetter();
97 bool isLower(char_type c)
101 return ucs4_to_qchar(c).isLower();
105 bool isAlphaASCII(char_type c)
107 return ('a' <= c && c <= 'z') || ('A' <= c && c <= 'Z');
111 bool isPrintable(char_type c)
115 // outside the UCS4 range
117 // assume that all non-utf16 characters are printable
120 return ucs4_to_qchar(c).isPrint();
124 bool isPrintableNonspace(char_type c)
128 // outside the UCS4 range
130 // assume that all non-utf16 characters are printable and
134 QChar const qc = ucs4_to_qchar(c);
135 return qc.isPrint() && !qc.isSpace();
139 bool isSpace(char_type c)
142 // assume that no non-utf16 character is a space
143 // c outside the UCS4 range is catched as well
146 QChar const qc = ucs4_to_qchar(c);
151 bool isDigit(char_type c)
154 // assume that no non-utf16 character is a digit
155 // c outside the UCS4 range is catched as well
157 return ucs4_to_qchar(c).isDigit();
161 bool isDigitASCII(char_type c)
163 return '0' <= c && c <= '9';
168 int compare_no_case(docstring const & s, docstring const & s2)
170 docstring::const_iterator p = s.begin();
171 docstring::const_iterator p2 = s2.begin();
173 while (p != s.end() && p2 != s2.end()) {
174 char_type const lc1 = lowercase(*p);
175 char_type const lc2 = lowercase(*p2);
177 return (lc1 < lc2) ? -1 : 1;
182 if (s.size() == s2.size())
184 if (s.size() < s2.size())
192 template<typename Char>
193 Char ascii_tolower(Char c) {
194 if (c >= 'A' && c <= 'Z')
195 return c - 'A' + 'a';
202 int compare_ascii_no_case(string const & s, string const & s2)
204 string::const_iterator p = s.begin();
205 string::const_iterator p2 = s2.begin();
207 while (p != s.end() && p2 != s2.end()) {
208 int const lc1 = ascii_tolower(*p);
209 int const lc2 = ascii_tolower(*p2);
211 return (lc1 < lc2) ? -1 : 1;
216 if (s.size() == s2.size())
218 if (s.size() < s2.size())
224 int compare_ascii_no_case(docstring const & s, docstring const & s2)
226 docstring::const_iterator p = s.begin();
227 docstring::const_iterator p2 = s2.begin();
229 while (p != s.end() && p2 != s2.end()) {
230 char_type const lc1 = ascii_tolower(*p);
231 char_type const lc2 = ascii_tolower(*p2);
233 return (lc1 < lc2) ? -1 : 1;
238 if (s.size() == s2.size())
240 if (s.size() < s2.size())
246 bool isStrInt(string const & str)
251 // Remove leading and trailing white space chars.
252 string const tmpstr = trim(str);
256 string::const_iterator cit = tmpstr.begin();
260 string::const_iterator end = tmpstr.end();
261 for (; cit != end; ++cit)
262 if (!isdigit((*cit)))
269 bool isStrUnsignedInt(string const & str)
274 // Remove leading and trailing white space chars.
275 string const tmpstr = trim(str);
279 string::const_iterator cit = tmpstr.begin();
280 string::const_iterator end = tmpstr.end();
281 for (; cit != end; ++cit)
282 if (!isdigit((*cit)))
289 bool isStrDbl(string const & str)
294 // Remove leading and trailing white space chars.
295 string const tmpstr = trim(str);
298 // if (tmpstr.count('.') > 1) return false;
300 string::const_iterator cit = tmpstr.begin();
301 bool found_dot = false;
304 string::const_iterator end = tmpstr.end();
305 for (; cit != end; ++cit) {
306 if (!isdigit(*cit) && *cit != '.')
318 static bool isHexChar(char_type c)
330 c == 'a' || c == 'A' ||
331 c == 'b' || c == 'B' ||
332 c == 'c' || c == 'C' ||
333 c == 'd' || c == 'D' ||
334 c == 'e' || c == 'E' ||
335 c == 'f' || c == 'F';
339 bool isHex(docstring const & str)
343 if (str.length() > 2 && str[0] == '0' &&
344 (str[1] == 'x' || str[1] == 'X'))
347 int const len = str.length();
349 for (; index < len; ++index) {
350 if (!isHexChar(str[index]))
357 int hexToInt(docstring const & str)
359 string s = to_ascii(str);
361 sscanf(s.c_str(), "%x", &h);
366 bool isAscii(docstring const & str)
368 int const len = str.length();
369 for (int i = 0; i < len; ++i)
376 bool isAscii(string const & str)
378 int const len = str.length();
379 for (int i = 0; i < len; ++i)
380 if (static_cast<unsigned char>(str[i]) >= 0x80)
386 char lowercase(char c)
388 LASSERT(static_cast<unsigned char>(c) < 0x80, /**/);
389 return char(tolower(c));
393 char uppercase(char c)
395 LASSERT(static_cast<unsigned char>(c) < 0x80, /**/);
396 return char(toupper(c));
400 char_type lowercase(char_type c)
403 // We don't know how to lowercase a non-utf16 char
405 return qchar_to_ucs4(ucs4_to_qchar(c).toLower());
409 char_type uppercase(char_type c)
412 // We don't know how to uppercase a non-utf16 char
414 return qchar_to_ucs4(ucs4_to_qchar(c).toUpper());
420 // since we cannot use tolower and toupper directly in the
421 // calls to transform yet, we use these helper clases. (Lgb)
423 struct local_lowercase {
424 char_type operator()(char_type c) const {
426 // We don't know how to lowercase a non-utf16 char
428 return qchar_to_ucs4(ucs4_to_qchar(c).toLower());
432 struct local_uppercase {
433 char_type operator()(char_type c) const {
435 // We don't know how to uppercase a non-utf16 char
437 return qchar_to_ucs4(ucs4_to_qchar(c).toUpper());
441 template<typename Char> struct local_ascii_lowercase {
442 Char operator()(Char c) const { return ascii_tolower(c); }
445 } // end of anon namespace
447 docstring const lowercase(docstring const & a)
450 transform(tmp.begin(), tmp.end(), tmp.begin(), local_lowercase());
455 docstring const uppercase(docstring const & a)
458 transform(tmp.begin(), tmp.end(), tmp.begin(), local_uppercase());
463 string const ascii_lowercase(string const & a)
466 transform(tmp.begin(), tmp.end(), tmp.begin(),
467 local_ascii_lowercase<char>());
472 docstring const ascii_lowercase(docstring const & a)
475 transform(tmp.begin(), tmp.end(), tmp.begin(),
476 local_ascii_lowercase<char_type>());
481 bool prefixIs(docstring const & a, char_type c)
489 bool prefixIs(string const & a, string const & pre)
491 size_t const prelen = pre.length();
492 size_t const alen = a.length();
493 return prelen <= alen && !a.empty() && a.compare(0, prelen, pre) == 0;
497 bool prefixIs(docstring const & a, docstring const & pre)
499 size_t const prelen = pre.length();
500 size_t const alen = a.length();
501 return prelen <= alen && !a.empty() && a.compare(0, prelen, pre) == 0;
505 bool suffixIs(string const & a, char c)
509 return a[a.length() - 1] == c;
513 bool suffixIs(docstring const & a, char_type c)
517 return a[a.length() - 1] == c;
521 bool suffixIs(string const & a, string const & suf)
523 size_t const suflen = suf.length();
524 size_t const alen = a.length();
525 return suflen <= alen && a.compare(alen - suflen, suflen, suf) == 0;
529 bool suffixIs(docstring const & a, docstring const & suf)
531 size_t const suflen = suf.length();
532 size_t const alen = a.length();
533 return suflen <= alen && a.compare(alen - suflen, suflen, suf) == 0;
537 bool containsOnly(string const & s, string const & cset)
539 return s.find_first_not_of(cset) == string::npos;
543 // ale970405+lasgoutt-970425
544 // rewritten to use new string (Lgb)
545 string const token(string const & a, char delim, int n)
553 // Find delimiter or end of string
555 if ((i = a.find(delim, i)) == string::npos)
561 // i is now the n'th delim (or string::npos)
562 if (i == string::npos)
565 k = a.find(delim, i);
566 // k is now the n'th + 1 delim (or string::npos)
568 return a.substr(i, k - i);
572 docstring const token(docstring const & a, char_type delim, int n)
580 // Find delimiter or end of string
582 if ((i = a.find(delim, i)) == docstring::npos)
588 // i is now the n'th delim (or string::npos)
589 if (i == docstring::npos)
592 k = a.find(delim, i);
593 // k is now the n'th + 1 delim (or string::npos)
595 return a.substr(i, k - i);
599 // this could probably be faster and/or cleaner, but it seems to work (JMarc)
600 // rewritten to use new string (Lgb)
601 int tokenPos(string const & a, char delim, string const & tok)
607 while (!str.empty()) {
608 str = split(str, tmptok, delim);
617 // this could probably be faster and/or cleaner, but it seems to work (JMarc)
618 // rewritten to use new string (Lgb)
619 int tokenPos(docstring const & a, char_type delim, docstring const & tok)
625 while (!str.empty()) {
626 str = split(str, tmptok, delim);
637 /// Substitute all \a oldchar with \a newchar
638 template<typename Ch> inline
639 basic_string<Ch> const subst_char(basic_string<Ch> const & a,
640 Ch oldchar, Ch newchar)
642 typedef basic_string<Ch> String;
644 typename String::iterator lit = tmp.begin();
645 typename String::iterator end = tmp.end();
646 for (; lit != end; ++lit)
647 if ((*lit) == oldchar)
652 /// Substitute all \a oldchar with \a newchar
653 docstring const subst_char(docstring const & a,
654 docstring::value_type oldchar, docstring::value_type newchar)
657 docstring::iterator lit = tmp.begin();
658 docstring::iterator end = tmp.end();
659 for (; lit != end; ++lit)
660 if ((*lit) == oldchar)
666 /// substitutes all instances of \a oldstr with \a newstr
667 template<typename String> inline
668 String const subst_string(String const & a,
669 String const & oldstr, String const & newstr)
671 LASSERT(!oldstr.empty(), /**/);
674 size_t const olen = oldstr.length();
675 while ((i = lstr.find(oldstr, i)) != string::npos) {
676 lstr.replace(i, olen, newstr);
677 i += newstr.length(); // We need to be sure that we dont
678 // use the same i over and over again.
683 docstring const subst_string(docstring const & a,
684 docstring const & oldstr, docstring const & newstr)
686 LASSERT(!oldstr.empty(), /**/);
689 size_t const olen = oldstr.length();
690 while ((i = lstr.find(oldstr, i)) != string::npos) {
691 lstr.replace(i, olen, newstr);
692 i += newstr.length(); // We need to be sure that we dont
693 // use the same i over and over again.
701 string const subst(string const & a, char oldchar, char newchar)
703 return subst_char(a, oldchar, newchar);
707 docstring const subst(docstring const & a,
708 char_type oldchar, char_type newchar)
710 return subst_char(a, oldchar, newchar);
714 string const subst(string const & a,
715 string const & oldstr, string const & newstr)
717 return subst_string(a, oldstr, newstr);
721 docstring const subst(docstring const & a,
722 docstring const & oldstr, docstring const & newstr)
724 return subst_string(a, oldstr, newstr);
728 docstring const trim(docstring const & a, char const * p)
732 if (a.empty() || !*p)
735 docstring s = from_ascii(p);
736 size_t r = a.find_last_not_of(s);
737 size_t l = a.find_first_not_of(s);
739 // Is this the minimal test? (lgb)
740 if (r == docstring::npos && l == docstring::npos)
743 return a.substr(l, r - l + 1);
747 string const trim(string const & a, char const * p)
751 if (a.empty() || !*p)
754 size_t r = a.find_last_not_of(p);
755 size_t l = a.find_first_not_of(p);
757 // Is this the minimal test? (lgb)
758 if (r == string::npos && l == string::npos)
761 return a.substr(l, r - l + 1);
765 string const rtrim(string const & a, char const * p)
769 if (a.empty() || !*p)
772 size_t r = a.find_last_not_of(p);
774 // Is this test really needed? (Lgb)
775 if (r == string::npos)
778 return a.substr(0, r + 1);
782 docstring const rtrim(docstring const & a, char const * p)
786 if (a.empty() || !*p)
789 size_t r = a.find_last_not_of(from_ascii(p));
791 // Is this test really needed? (Lgb)
792 if (r == docstring::npos)
795 return a.substr(0, r + 1);
799 string const ltrim(string const & a, char const * p)
802 if (a.empty() || !*p)
804 size_t l = a.find_first_not_of(p);
805 if (l == string::npos)
807 return a.substr(l, string::npos);
811 docstring const ltrim(docstring const & a, char const * p)
814 if (a.empty() || !*p)
816 size_t l = a.find_first_not_of(from_ascii(p));
817 if (l == docstring::npos)
819 return a.substr(l, docstring::npos);
824 template<typename String, typename Char> inline
825 String const doSplit(String const & a, String & piece, Char delim)
828 size_t i = a.find(delim);
829 if (i == a.length() - 1) {
830 piece = a.substr(0, i);
831 } else if (i != String::npos) {
832 piece = a.substr(0, i);
833 tmp = a.substr(i + 1);
836 tmp = a.substr(i + 1);
843 template<typename Char> inline
844 docstring const doSplit(docstring const & a, docstring & piece, Char delim)
847 size_t i = a.find(delim);
848 if (i == a.length() - 1) {
849 piece = a.substr(0, i);
850 } else if (i != docstring::npos) {
851 piece = a.substr(0, i);
852 tmp = a.substr(i + 1);
855 tmp = a.substr(i + 1);
865 string const split(string const & a, string & piece, char delim)
867 return doSplit(a, piece, delim);
871 docstring const split(docstring const & a, docstring & piece, char_type delim)
873 return doSplit(a, piece, delim);
877 string const split(string const & a, char delim)
880 size_t i = a.find(delim);
881 if (i != string::npos) // found delim
882 tmp = a.substr(i + 1);
888 string const rsplit(string const & a, string & piece, char delim)
891 size_t i = a.rfind(delim);
892 if (i != string::npos) { // delimiter was found
893 piece = a.substr(0, i);
894 tmp = a.substr(i + 1);
895 } else { // delimiter was not found
902 docstring const escape(docstring const & lab)
904 char_type hexdigit[16] = { '0', '1', '2', '3', '4', '5', '6', '7',
905 '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' };
907 for (size_t i = 0; i < lab.length(); ++i) {
908 char_type c = lab[i];
909 if (c >= 128 || c == '=' || c == '%') {
910 // Although char_type is a 32 bit type we know that
911 // UCS4 occupies only 21 bits, so we don't need to
912 // encode bigger values. Test for 2^24 because we
913 // can encode that with the 6 hex digits that are
914 // needed for 21 bits anyway.
915 LASSERT(c < (1 << 24), /**/);
917 enc += hexdigit[(c>>20) & 15];
918 enc += hexdigit[(c>>16) & 15];
919 enc += hexdigit[(c>>12) & 15];
920 enc += hexdigit[(c>> 8) & 15];
921 enc += hexdigit[(c>> 4) & 15];
922 enc += hexdigit[ c & 15];
933 template<typename String> vector<String> const
934 getVectorFromStringT(String const & str, String const & delim, bool keepempty)
936 // Lars would like this code to go, but for now his replacement (below)
937 // doesn't fullfil the same function. I have, therefore, reactivated the
938 // old code for now. Angus 11 Nov 2002.
943 String keys = rtrim(str);
945 size_t const idx = keys.find(delim);
946 if (idx == String::npos) {
947 vec.push_back(ltrim(keys));
950 String const key = trim(keys.substr(0, idx));
951 if (!key.empty() || keepempty)
953 size_t const start = idx + delim.size();
954 keys = keys.substr(start);
958 typedef boost::char_separator<typename String::value_type> Separator;
959 typedef boost::tokenizer<Separator, typename String::const_iterator, String> Tokenizer;
960 Separator sep(delim.c_str());
961 Tokenizer tokens(str, sep);
962 return vector<String>(tokens.begin(), tokens.end());
969 vector<string> const getVectorFromString(string const & str,
970 string const & delim,
973 return getVectorFromStringT<string>(str, delim, keepempty);
977 vector<docstring> const getVectorFromString(docstring const & str,
978 docstring const & delim,
981 return getVectorFromStringT<docstring>(str, delim, keepempty);
985 // the same vice versa
986 string const getStringFromVector(vector<string> const & vec,
987 string const & delim)
991 for (vector<string>::const_iterator it = vec.begin();
992 it != vec.end(); ++it) {
993 string item = trim(*it);
1004 int findToken(char const * const str[], string const & search_token)
1008 while (str[i][0] && str[i] != search_token)
1017 docstring bformat(docstring const & fmt, int arg1)
1019 LASSERT(contains(fmt, from_ascii("%1$d")), /**/);
1020 docstring const str = subst(fmt, from_ascii("%1$d"), convert<docstring>(arg1));
1021 return subst(str, from_ascii("%%"), from_ascii("%"));
1026 docstring bformat(docstring const & fmt, long arg1)
1028 LASSERT(contains(fmt, from_ascii("%1$d")), /**/);
1029 docstring const str = subst(fmt, from_ascii("%1$d"), convert<docstring>(arg1));
1030 return subst(str, from_ascii("%%"), from_ascii("%"));
1035 docstring bformat(docstring const & fmt, unsigned int arg1)
1037 LASSERT(contains(fmt, from_ascii("%1$d")), /**/);
1038 docstring const str = subst(fmt, from_ascii("%1$d"), convert<docstring>(arg1));
1039 return subst(str, from_ascii("%%"), from_ascii("%"));
1044 docstring bformat(docstring const & fmt, docstring arg1)
1046 LASSERT(contains(fmt, from_ascii("%1$s")), /**/);
1047 docstring const str = subst(fmt, from_ascii("%1$s"), arg1);
1048 return subst(str, from_ascii("%%"), from_ascii("%"));
1053 docstring bformat(docstring const & fmt, char * arg1)
1055 LASSERT(contains(fmt, from_ascii("%1$s")), /**/);
1056 docstring const str = subst(fmt, from_ascii("%1$s"), from_ascii(arg1));
1057 return subst(str, from_ascii("%%"), from_ascii("%"));
1062 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2)
1064 LASSERT(contains(fmt, from_ascii("%1$s")), /**/);
1065 LASSERT(contains(fmt, from_ascii("%2$s")), /**/);
1066 docstring str = subst(fmt, from_ascii("%1$s"), arg1);
1067 str = subst(str, from_ascii("%2$s"), arg2);
1068 return subst(str, from_ascii("%%"), from_ascii("%"));
1073 docstring bformat(docstring const & fmt, char const * arg1, docstring arg2)
1075 LASSERT(contains(fmt, from_ascii("%1$s")), /**/);
1076 LASSERT(contains(fmt, from_ascii("%2$s")), /**/);
1077 docstring str = subst(fmt, from_ascii("%1$s"), from_ascii(arg1));
1078 str = subst(fmt, from_ascii("%2$s"), arg2);
1079 return subst(str, from_ascii("%%"), from_ascii("%"));
1084 docstring bformat(docstring const & fmt, int arg1, int arg2)
1086 LASSERT(contains(fmt, from_ascii("%1$d")), /**/);
1087 LASSERT(contains(fmt, from_ascii("%2$d")), /**/);
1088 docstring str = subst(fmt, from_ascii("%1$d"), convert<docstring>(arg1));
1089 str = subst(str, from_ascii("%2$d"), convert<docstring>(arg2));
1090 return subst(str, from_ascii("%%"), from_ascii("%"));
1095 docstring bformat(docstring const & fmt, docstring arg1, docstring arg2, docstring arg3)
1097 LASSERT(contains(fmt, from_ascii("%1$s")), /**/);
1098 LASSERT(contains(fmt, from_ascii("%2$s")), /**/);
1099 LASSERT(contains(fmt, from_ascii("%3$s")), /**/);
1100 docstring str = subst(fmt, from_ascii("%1$s"), arg1);
1101 str = subst(str, from_ascii("%2$s"), arg2);
1102 str = subst(str, from_ascii("%3$s"), arg3);
1103 return subst(str, from_ascii("%%"), from_ascii("%"));
1108 docstring bformat(docstring const & fmt,
1109 docstring arg1, docstring arg2, docstring arg3, docstring arg4)
1111 LASSERT(contains(fmt, from_ascii("%1$s")), /**/);
1112 LASSERT(contains(fmt, from_ascii("%2$s")), /**/);
1113 LASSERT(contains(fmt, from_ascii("%3$s")), /**/);
1114 LASSERT(contains(fmt, from_ascii("%4$s")), /**/);
1115 docstring str = subst(fmt, from_ascii("%1$s"), arg1);
1116 str = subst(str, from_ascii("%2$s"), arg2);
1117 str = subst(str, from_ascii("%3$s"), arg3);
1118 str = subst(str, from_ascii("%4$s"), arg4);
1119 return subst(str, from_ascii("%%"), from_ascii("%"));
1122 } // namespace support