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
19 #include "BoostFormat.h"
21 #include <boost/regex.hpp>
22 #include <boost/tokenizer.hpp>
33 #ifndef CXX_GLOBAL_CSTD
42 int compare_no_case(string const & s, string const & s2)
44 string::const_iterator p = s.begin();
45 string::const_iterator p2 = s2.begin();
47 while (p != s.end() && p2 != s2.end()) {
48 int const lc1 = tolower(*p);
49 int const lc2 = tolower(*p2);
51 return (lc1 < lc2) ? -1 : 1;
56 if (s.size() == s2.size())
58 if (s.size() < s2.size())
65 int ascii_tolower(int c) {
66 if (c >= 'A' && c <= 'Z')
73 int compare_ascii_no_case(string const & s, string const & s2)
75 string::const_iterator p = s.begin();
76 string::const_iterator p2 = s2.begin();
78 while (p != s.end() && p2 != s2.end()) {
79 int const lc1 = ascii_tolower(*p);
80 int const lc2 = ascii_tolower(*p2);
82 return (lc1 < lc2) ? -1 : 1;
87 if (s.size() == s2.size())
89 if (s.size() < s2.size())
95 int compare_no_case(string const & s, string const & s2, unsigned int len)
97 string::const_iterator p = s.begin();
98 string::const_iterator p2 = s2.begin();
100 while (i < len && p != s.end() && p2 != s2.end()) {
101 int const lc1 = tolower(*p);
102 int const lc2 = tolower(*p2);
104 return (lc1 < lc2) ? -1 : 1;
110 if (s.size() >= len && s2.size() >= len)
112 if (s.size() < s2.size())
118 bool isStrInt(string const & str)
120 if (str.empty()) return false;
122 // Remove leading and trailing white space chars.
123 string const tmpstr = trim(str);
124 if (tmpstr.empty()) return false;
126 string::const_iterator cit = tmpstr.begin();
127 if ((*cit) == '-') ++cit;
128 string::const_iterator end = tmpstr.end();
129 for (; cit != end; ++cit) {
130 if (!isdigit((*cit))) return false;
136 bool isStrUnsignedInt(string const & str)
138 if (str.empty()) return false;
140 // Remove leading and trailing white space chars.
141 string const tmpstr = trim(str);
142 if (tmpstr.empty()) return false;
144 string::const_iterator cit = tmpstr.begin();
145 string::const_iterator end = tmpstr.end();
146 for (; cit != end; ++cit) {
147 if (!isdigit((*cit))) return false;
153 int strToInt(string const & str)
156 // Remove leading and trailing white space chars.
157 string const tmpstr = trim(str);
158 // Do the conversion proper.
159 return lyx::atoi(tmpstr);
166 unsigned int strToUnsignedInt(string const & str)
168 if (isStrUnsignedInt(str)) {
169 // Remove leading and trailing white space chars.
170 string const tmpstr = trim(str);
171 // Do the conversion proper.
172 return lyx::atoi(tmpstr);
179 bool isStrDbl(string const & str)
181 if (str.empty()) return false;
183 // Remove leading and trailing white space chars.
184 string const tmpstr = trim(str);
185 if (tmpstr.empty()) return false;
186 // if (1 < tmpstr.count('.')) return false;
188 string::const_iterator cit = tmpstr.begin();
189 bool found_dot(false);
190 if ((*cit) == '-') ++cit;
191 string::const_iterator end = tmpstr.end();
192 for (; cit != end; ++cit) {
209 double strToDbl(string const & str)
212 // Remove leading and trailing white space chars.
213 string const tmpstr = trim(str);
214 // Do the conversion proper.
215 return ::atof(tmpstr.c_str());
222 char lowercase(char c)
224 return char(tolower(c));
228 char uppercase(char c)
230 return char(toupper(c));
236 // since we cannot use std::tolower and std::toupper directly in the
237 // calls to std::transform yet, we use these helper clases. (Lgb)
239 struct local_lowercase {
240 char operator()(char c) const {
245 struct local_uppercase {
246 char operator()(char c) const {
251 struct local_ascii_lowercase {
252 char operator()(char c) const {
253 return ascii_tolower(c);
257 } // end of anon namespace
259 string const lowercase(string const & a)
262 transform(tmp.begin(), tmp.end(), tmp.begin(), local_lowercase());
266 string const uppercase(string const & a)
269 transform(tmp.begin(), tmp.end(), tmp.begin(), local_uppercase());
274 string const ascii_lowercase(string const & a)
277 transform(tmp.begin(), tmp.end(), tmp.begin(),
278 local_ascii_lowercase());
283 bool prefixIs(string const & a, char const * pre)
287 size_t const l = strlen(pre);
288 string::size_type const alen = a.length();
290 if (l > alen || a.empty())
293 #if !defined(USE_INCLUDED_STRING) && !defined(STD_STRING_IS_GOOD)
294 // Delete this code when the compilers get a bit better.
295 return ::strncmp(a.c_str(), pre, l) == 0;
297 // This is the code that we really want to use
298 // but until gcc ships with a basic_string that
299 // implements std::string correctly we have to
300 // use the code above.
301 return a.compare(0, l, pre, l) == 0;
307 bool prefixIs(string const & a, string const & pre)
309 string::size_type const prelen = pre.length();
310 string::size_type const alen = a.length();
312 if (prelen > alen || a.empty())
315 #if !defined(USE_INCLUDED_STRING) && !defined(STD_STRING_IS_GOOD)
316 return ::strncmp(a.c_str(), pre.c_str(), prelen) == 0;
318 return a.compare(0, prelen, pre) == 0;
324 bool suffixIs(string const & a, char c)
326 if (a.empty()) return false;
327 return a[a.length() - 1] == c;
331 bool suffixIs(string const & a, char const * suf)
335 size_t const suflen = strlen(suf);
336 string::size_type const alen = a.length();
341 #if !defined(USE_INCLUDED_STRING) && !defined(STD_STRING_IS_GOOD)
342 // Delete this code when the compilers get a bit better.
343 string tmp(a, alen - suflen);
344 return ::strncmp(tmp.c_str(), suf, suflen) == 0;
346 // This is the code that we really want to use
347 // but until gcc ships with a basic_string that
348 // implements std::string correctly we have to
349 // use the code above.
350 return a.compare(alen - suflen, suflen, suf) == 0;
356 bool suffixIs(string const & a, string const & suf)
358 string::size_type const suflen = suf.length();
359 string::size_type const alen = a.length();
364 #if !defined(USE_INCLUDED_STRING) && !defined(STD_STRING_IS_GOOD)
365 string tmp(a, alen - suflen);
366 return ::strncmp(tmp.c_str(), suf.c_str(), suflen) == 0;
368 return a.compare(alen - suflen, suflen, suf) == 0;
374 bool contains(string const & a, string const & b)
378 return a.find(b) != string::npos;
382 bool contains(string const & a, char b)
386 return a.find(b) != string::npos;
390 bool containsOnly(string const & s, string const & cset)
392 return s.find_first_not_of(cset) == string::npos;
396 // ale970405+lasgoutt-970425
397 // rewritten to use new string (Lgb)
398 string const token(string const & a, char delim, int n)
400 if (a.empty()) return string();
402 string::size_type k = 0;
403 string::size_type i = 0;
405 // Find delimiter or end of string
407 if ((i = a.find(delim, i)) == string::npos)
411 // i is now the n'th delim (or string::npos)
412 if (i == string::npos) return string();
413 k = a.find(delim, i);
414 // k is now the n'th + 1 delim (or string::npos)
416 return a.substr(i, k - i);
420 // this could probably be faster and/or cleaner, but it seems to work (JMarc)
421 // rewritten to use new string (Lgb)
422 int tokenPos(string const & a, char delim, string const & tok)
428 while (!str.empty()) {
429 str = split(str, tmptok, delim);
438 bool regexMatch(string const & a, string const & pattern)
440 // We massage the pattern a bit so that the usual
441 // shell pattern we all are used to will work.
442 // One nice thing about using a real regex is that
443 // things like "*.*[^~]" will work also.
444 // build the regex string.
445 string regex(pattern);
446 regex = subst(regex, ".", "\\.");
447 regex = subst(regex, "*", ".*");
448 boost::regex reg(STRCONV(regex));
449 return boost::regex_match(STRCONV(a), reg);
453 string const subst(string const & a, char oldchar, char newchar)
456 string::iterator lit = tmp.begin();
457 string::iterator end = tmp.end();
458 for (; lit != end; ++lit)
459 if ((*lit) == oldchar)
465 string const subst(string const & a,
466 char const * oldstr, string const & newstr)
471 string::size_type i = 0;
472 string::size_type olen = strlen(oldstr);
473 while ((i = lstr.find(oldstr, i)) != string::npos) {
474 lstr.replace(i, olen, newstr);
475 i += newstr.length(); // We need to be sure that we dont
476 // use the same i over and over again.
482 string const subst(string const & a,
483 string const & oldstr, string const & newstr)
486 string::size_type i = 0;
487 string::size_type const olen = oldstr.length();
488 while ((i = lstr.find(oldstr, i)) != string::npos) {
489 lstr.replace(i, olen, newstr);
490 i += newstr.length(); // We need to be sure that we dont
491 // use the same i over and over again.
497 string const trim(string const & a, char const * p)
501 if (a.empty() || !*p)
504 string::size_type r = a.find_last_not_of(p);
505 string::size_type l = a.find_first_not_of(p);
507 // Is this the minimal test? (lgb)
508 if (r == string::npos && l == string::npos)
511 return a.substr(l, r - l + 1);
515 string const rtrim(string const & a, char const * p)
519 if (a.empty() || !*p)
522 string::size_type r = a.find_last_not_of(p);
524 // Is this test really needed? (Lgb)
525 if (r == string::npos)
528 return a.substr(0, r + 1);
532 string const ltrim(string const & a, char const * p)
536 if (a.empty() || !*p)
539 string::size_type l = a.find_first_not_of(p);
541 if (l == string::npos)
544 return a.substr(l, string::npos);
548 string const split(string const & a, string & piece, char delim)
551 string::size_type i = a.find(delim);
552 if (i == a.length() - 1) {
553 piece = a.substr(0, i);
554 } else if (i != string::npos) {
555 piece = a.substr(0, i);
556 tmp = a.substr(i + 1);
559 tmp = a.substr(i + 1);
567 string const split(string const & a, char delim)
570 string::size_type i = a.find(delim);
571 if (i != string::npos) // found delim
572 tmp = a.substr(i + 1);
578 string const rsplit(string const & a, string & piece, char delim)
581 string::size_type i = a.rfind(delim);
582 if (i != string::npos) { // delimiter was found
583 piece = a.substr(0, i);
584 tmp = a.substr(i + 1);
585 } else { // delimiter was not found
592 // This function escapes 8-bit characters and other problematic
593 // characters that cause problems in latex labels.
594 string const escape(string const & lab)
596 char hexdigit[16] = { '0', '1', '2', '3', '4', '5', '6', '7',
597 '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' };
599 for (string::size_type i = 0; i < lab.length(); ++i) {
600 unsigned char c= lab[i];
601 if (c >= 128 || c == '=' || c == '%') {
603 enc += hexdigit[c>>4];
604 enc += hexdigit[c & 15];
613 /// gives a vector of stringparts which have the delimiter delim
614 vector<string> const getVectorFromString(string const & str,
615 string const & delim)
617 // Lars would like this code to go, but for now his replacement (below)
618 // doesn't fullfil the same function. I have, therefore, reactivated the
619 // old code for now. Angus 11 Nov 2002.
624 string keys(rtrim(str));
626 string::size_type const idx = keys.find(delim);
627 if (idx == string::npos) {
628 vec.push_back(ltrim(keys));
631 string const key = trim(keys.substr(0, idx));
634 string::size_type const start = idx + delim.size();
635 keys = keys.substr(start);
639 boost::char_separator<char> sep(delim.c_str());
640 boost::tokenizer<boost::char_separator<char> > tokens(str, sep);
641 #ifndef USE_INCLUDED_STRING
642 return vector<string>(tokens.begin(), tokens.end());
645 using boost::tokenizer;
646 using boost::char_separator;
648 tokenizer<char_separator<char> >::iterator it = tokens.begin();
649 tokenizer<char_separator<char> >::iterator end = tokens.end();
650 for (; it != end; ++it) {
651 vec.push_back(STRCONV((*it)));
659 // the same vice versa
660 string const getStringFromVector(vector<string> const & vec,
661 string const & delim)
665 for (vector<string>::const_iterator it = vec.begin();
666 it != vec.end(); ++it) {
667 string item = trim(*it);
668 if (item.empty()) continue;
670 if (i++ > 0) str += delim;
679 string bformat(char const * fmt, string const & arg1)
681 return STRCONV((boost::format(fmt) % STRCONV(arg1)).str());
685 string bformat(char const * fmt, string const & arg1, string const & arg2)
687 return STRCONV((boost::format(fmt) % STRCONV(arg1) % STRCONV(arg2)).str());
690 string bformat(char const * fmt, string const & arg1, string const & arg2,
691 string const & arg3, string const & arg4)
693 return STRCONV((boost::format(fmt) % STRCONV(arg1) % STRCONV(arg2)
694 % STRCONV(arg3) % STRCONV(arg4)).str());
699 string bformat(char const * fmt, string const & arg1)
701 lyx::Assert(contains(fmt, "%1$s"));
702 string const str = subst(fmt, "%1$s", arg1);
703 return subst(str, "%%", "%");
707 string bformat(char const * fmt, string const & arg1, string const & arg2)
709 lyx::Assert(contains(fmt, "%1$s"));
710 lyx::Assert(contains(fmt, "%2$s"));
711 string str = subst(fmt, "%1$s", arg1);
712 str = subst(str, "%2$s", arg2);
713 return subst(str, "%%", "%");
717 string bformat(char const * fmt, string const & arg1, string const & arg2,
718 string const & arg3, string const & arg4)
720 lyx::Assert(contains(fmt, "%1$s"));
721 lyx::Assert(contains(fmt, "%2$s"));
722 lyx::Assert(contains(fmt, "%3$s"));
723 lyx::Assert(contains(fmt, "%4$s"));
724 string str = subst(fmt, "%1$s", arg1);
725 str = subst(str, "%2$s", arg2);
726 str = subst(str, "%3$s", arg3);
727 str = subst(str, "%4$s", arg4);
728 return subst(str, "%%", "%");