3 * This file is part of LyX, the document processor.
4 * License details can be found in the file COPYING.
6 * \author Lars Gullik Bjønnes
9 * \author Alfredo Braunstein
10 * \author Tommaso Cucinotta
11 * \author Kornel Benko
13 * Full author contact details are available in file CREDITS.
21 #include "BufferList.h"
22 #include "BufferParams.h"
23 #include "BufferView.h"
26 #include "CutAndPaste.h"
27 #include "FuncRequest.h"
29 #include "output_latex.h"
30 #include "OutputParams.h"
31 #include "Paragraph.h"
35 #include "frontends/Application.h"
36 #include "frontends/alert.h"
38 #include "mathed/InsetMath.h"
39 #include "mathed/InsetMathHull.h"
40 #include "mathed/MathData.h"
41 #include "mathed/MathStream.h"
42 #include "mathed/MathSupport.h"
44 #include "support/debug.h"
45 #include "support/docstream.h"
46 #include "support/FileName.h"
47 #include "support/gettext.h"
48 #include "support/lassert.h"
49 #include "support/lstrings.h"
50 #include "support/textutils.h"
54 #define USE_QT_FOR_SEARCH
55 #if defined(USE_QT_FOR_SEARCH)
56 #include <QtCore> // sets QT_VERSION
57 #if (QT_VERSION >= 0x050000)
58 #include <QRegularExpression>
68 using namespace lyx::support;
73 // Helper class for deciding what should be ignored
77 IgnoreFormats() = default;
79 bool getFamily() const { return ignoreFamily_; }
81 bool getSeries() const { return ignoreSeries_; }
83 bool getShape() const { return ignoreShape_; }
85 bool getUnderline() const { return ignoreUnderline_; }
87 bool getMarkUp() const { return ignoreMarkUp_; }
89 bool getStrikeOut() const { return ignoreStrikeOut_; }
91 bool getSectioning() const { return ignoreSectioning_; }
93 bool getFrontMatter() const { return ignoreFrontMatter_; }
95 bool getColor() const { return ignoreColor_; }
97 bool getLanguage() const { return ignoreLanguage_; }
99 void setIgnoreFormat(string const & type, bool value);
103 bool ignoreFamily_ = false;
105 bool ignoreSeries_ = false;
107 bool ignoreShape_ = false;
109 bool ignoreUnderline_ = false;
111 bool ignoreMarkUp_ = false;
113 bool ignoreStrikeOut_ = false;
115 bool ignoreSectioning_ = false;
117 bool ignoreFrontMatter_ = false;
119 bool ignoreColor_ = false;
121 bool ignoreLanguage_ = false;
125 void IgnoreFormats::setIgnoreFormat(string const & type, bool value)
127 if (type == "color") {
128 ignoreColor_ = value;
130 else if (type == "language") {
131 ignoreLanguage_ = value;
133 else if (type == "sectioning") {
134 ignoreSectioning_ = value;
135 ignoreFrontMatter_ = value;
137 else if (type == "font") {
138 ignoreSeries_ = value;
139 ignoreShape_ = value;
140 ignoreFamily_ = value;
142 else if (type == "series") {
143 ignoreSeries_ = value;
145 else if (type == "shape") {
146 ignoreShape_ = value;
148 else if (type == "family") {
149 ignoreFamily_ = value;
151 else if (type == "markup") {
152 ignoreMarkUp_ = value;
154 else if (type == "underline") {
155 ignoreUnderline_ = value;
157 else if (type == "strike") {
158 ignoreStrikeOut_ = value;
162 // The global variable that can be changed from outside
163 IgnoreFormats ignoreFormats;
166 void setIgnoreFormat(string const & type, bool value)
168 ignoreFormats.setIgnoreFormat(type, value);
174 bool parse_bool(docstring & howto)
179 howto = split(howto, var, ' ');
187 MatchString(docstring const & s, bool cs, bool mw)
188 : str(s), case_sens(cs), whole_words(mw)
191 // returns true if the specified string is at the specified position
192 // del specifies whether deleted strings in ct mode will be considered
193 int operator()(Paragraph const & par, pos_type pos, bool del = true) const
195 return par.find(str, case_sens, whole_words, pos, del);
203 // match whole words only
208 int findForward(DocIterator & cur, MatchString const & match,
209 bool find_del = true)
211 for (; cur; cur.forwardChar())
212 if (cur.inTexted()) {
213 int len = match(cur.paragraph(), cur.pos(), find_del);
221 int findBackwards(DocIterator & cur, MatchString const & match,
222 bool find_del = true)
226 if (cur.inTexted()) {
227 int len = match(cur.paragraph(), cur.pos(), find_del);
236 bool searchAllowed(docstring const & str)
239 frontend::Alert::error(_("Search error"), _("Search string is empty"));
246 bool findOne(BufferView * bv, docstring const & searchstr,
247 bool case_sens, bool whole, bool forward,
248 bool find_del = true, bool check_wrap = false)
250 if (!searchAllowed(searchstr))
253 DocIterator cur = forward
254 ? bv->cursor().selectionEnd()
255 : bv->cursor().selectionBegin();
257 MatchString const match(searchstr, case_sens, whole);
259 int match_len = forward
260 ? findForward(cur, match, find_del)
261 : findBackwards(cur, match, find_del);
264 bv->putSelectionAt(cur, match_len, !forward);
265 else if (check_wrap) {
266 DocIterator cur_orig(bv->cursor());
269 q = _("End of file reached while searching forward.\n"
270 "Continue searching from the beginning?");
272 q = _("Beginning of file reached while searching backward.\n"
273 "Continue searching from the end?");
274 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
275 q, 0, 1, _("&Yes"), _("&No"));
276 if (wrap_answer == 0) {
278 bv->cursor().clear();
279 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
281 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
282 bv->cursor().backwardPos();
284 bv->clearSelection();
285 if (findOne(bv, searchstr, case_sens, whole, forward, find_del, false))
288 bv->cursor().setCursor(cur_orig);
292 return match_len > 0;
296 int replaceAll(BufferView * bv,
297 docstring const & searchstr, docstring const & replacestr,
298 bool case_sens, bool whole)
300 Buffer & buf = bv->buffer();
302 if (!searchAllowed(searchstr) || buf.isReadonly())
305 DocIterator cur_orig(bv->cursor());
307 MatchString const match(searchstr, case_sens, whole);
310 int const rsize = replacestr.size();
311 int const ssize = searchstr.size();
314 cur.setCursor(doc_iterator_begin(&buf));
315 int match_len = findForward(cur, match, false);
316 while (match_len > 0) {
317 // Backup current cursor position and font.
318 pos_type const pos = cur.pos();
319 Font const font = cur.paragraph().getFontSettings(buf.params(), pos);
321 int striked = ssize -
322 cur.paragraph().eraseChars(pos, pos + match_len,
323 buf.params().track_changes);
324 cur.paragraph().insert(pos, replacestr, font,
325 Change(buf.params().track_changes
327 : Change::UNCHANGED));
328 for (int i = 0; i < rsize + striked; ++i)
331 match_len = findForward(cur, match, false);
334 bv->putSelectionAt(doc_iterator_begin(&buf), 0, false);
336 cur_orig.fixIfBroken();
337 bv->setCursor(cur_orig);
343 // the idea here is that we are going to replace the string that
344 // is selected IF it is the search string.
345 // if there is a selection, but it is not the search string, then
346 // we basically ignore it. (FIXME We ought to replace only within
348 // if there is no selection, then:
349 // (i) if some search string has been provided, then we find it.
350 // (think of how the dialog works when you hit "replace" the
352 // (ii) if no search string has been provided, then we treat the
353 // word the cursor is in as the search string. (why? i have no
354 // idea.) but this only works in text?
356 // returns the number of replacements made (one, if any) and
357 // whether anything at all was done.
358 pair<bool, int> replaceOne(BufferView * bv, docstring searchstr,
359 docstring const & replacestr, bool case_sens,
360 bool whole, bool forward, bool findnext)
362 Cursor & cur = bv->cursor();
363 if (!cur.selection()) {
364 // no selection, non-empty search string: find it
365 if (!searchstr.empty()) {
366 bool const found = findOne(bv, searchstr, case_sens, whole, forward, true, findnext);
367 return make_pair(found, 0);
369 // empty search string
372 return make_pair(false, 0);
373 // select current word and treat it as the search string.
374 // This causes a minor bug as undo will restore this selection,
375 // which the user did not create (#8986).
376 cur.innerText()->selectWord(cur, WHOLE_WORD);
377 searchstr = cur.selectionAsString(false);
380 // if we still don't have a search string, report the error
382 if (!searchAllowed(searchstr))
383 return make_pair(false, 0);
385 bool have_selection = cur.selection();
386 docstring const selected = cur.selectionAsString(false);
389 ? searchstr == selected
390 : compare_no_case(searchstr, selected) == 0;
392 // no selection or current selection is not search word:
393 // just find the search word
394 if (!have_selection || !match) {
395 bool const found = findOne(bv, searchstr, case_sens, whole, forward, true, findnext);
396 return make_pair(found, 0);
399 // we're now actually ready to replace. if the buffer is
400 // read-only, we can't, though.
401 if (bv->buffer().isReadonly())
402 return make_pair(false, 0);
404 cap::replaceSelectionWithString(cur, replacestr);
406 cur.pos() += replacestr.length();
407 LASSERT(cur.pos() <= cur.lastpos(),
408 cur.pos() = cur.lastpos());
411 findOne(bv, searchstr, case_sens, whole, forward, false, findnext);
413 return make_pair(true, 1);
419 docstring const find2string(docstring const & search,
420 bool casesensitive, bool matchword, bool forward)
424 << int(casesensitive) << ' '
425 << int(matchword) << ' '
431 docstring const replace2string(docstring const & replace,
432 docstring const & search,
433 bool casesensitive, bool matchword,
434 bool all, bool forward, bool findnext)
437 ss << replace << '\n'
439 << int(casesensitive) << ' '
440 << int(matchword) << ' '
442 << int(forward) << ' '
448 bool lyxfind(BufferView * bv, FuncRequest const & ev)
450 if (!bv || ev.action() != LFUN_WORD_FIND)
453 //lyxerr << "find called, cmd: " << ev << endl;
455 // data is of the form
457 // <casesensitive> <matchword> <forward>"
459 docstring howto = split(ev.argument(), search, '\n');
461 bool casesensitive = parse_bool(howto);
462 bool matchword = parse_bool(howto);
463 bool forward = parse_bool(howto);
465 return findOne(bv, search, casesensitive, matchword, forward, true, true);
469 bool lyxreplace(BufferView * bv,
470 FuncRequest const & ev, bool has_deleted)
472 if (!bv || ev.action() != LFUN_WORD_REPLACE)
475 // data is of the form
478 // <casesensitive> <matchword> <all> <forward> <findnext>"
481 docstring howto = split(ev.argument(), rplc, '\n');
482 howto = split(howto, search, '\n');
484 bool casesensitive = parse_bool(howto);
485 bool matchword = parse_bool(howto);
486 bool all = parse_bool(howto);
487 bool forward = parse_bool(howto);
488 bool findnext = howto.empty() ? true : parse_bool(howto);
493 int replace_count = 0;
495 replace_count = replaceAll(bv, search, rplc, casesensitive, matchword);
496 update = replace_count > 0;
499 replaceOne(bv, search, rplc, casesensitive, matchword, forward, findnext);
501 replace_count = rv.second;
504 Buffer const & buf = bv->buffer();
506 // emit message signal.
507 buf.message(_("String not found."));
509 if (replace_count == 0) {
510 buf.message(_("String found."));
511 } else if (replace_count == 1) {
512 buf.message(_("String has been replaced."));
514 docstring const str =
515 bformat(_("%1$d strings have been replaced."), replace_count);
519 } else if (findnext) {
520 // if we have deleted characters, we do not replace at all, but
521 // rather search for the next occurence
522 if (findOne(bv, search, casesensitive, matchword, forward, true, findnext))
525 bv->message(_("String not found."));
531 bool findNextChange(BufferView * bv, Cursor & cur, bool const check_wrap)
533 for (; cur; cur.forwardPos())
534 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
538 DocIterator cur_orig(bv->cursor());
539 docstring q = _("End of file reached while searching forward.\n"
540 "Continue searching from the beginning?");
541 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
542 q, 0, 1, _("&Yes"), _("&No"));
543 if (wrap_answer == 0) {
544 bv->cursor().clear();
545 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
546 bv->clearSelection();
547 cur.setCursor(bv->cursor().selectionBegin());
548 if (findNextChange(bv, cur, false))
551 bv->cursor().setCursor(cur_orig);
558 bool findPreviousChange(BufferView * bv, Cursor & cur, bool const check_wrap)
560 for (cur.backwardPos(); cur; cur.backwardPos()) {
561 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
566 DocIterator cur_orig(bv->cursor());
567 docstring q = _("Beginning of file reached while searching backward.\n"
568 "Continue searching from the end?");
569 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
570 q, 0, 1, _("&Yes"), _("&No"));
571 if (wrap_answer == 0) {
572 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
573 bv->cursor().backwardPos();
574 bv->clearSelection();
575 cur.setCursor(bv->cursor().selectionBegin());
576 if (findPreviousChange(bv, cur, false))
579 bv->cursor().setCursor(cur_orig);
586 bool selectChange(Cursor & cur, bool forward)
588 if (!cur.inTexted() || !cur.paragraph().isChanged(cur.pos()))
590 Change ch = cur.paragraph().lookupChange(cur.pos());
592 CursorSlice tip1 = cur.top();
593 for (; tip1.pit() < tip1.lastpit() || tip1.pos() < tip1.lastpos(); tip1.forwardPos()) {
594 Change ch2 = tip1.paragraph().lookupChange(tip1.pos());
595 if (!ch2.isSimilarTo(ch))
598 CursorSlice tip2 = cur.top();
599 for (; tip2.pit() > 0 || tip2.pos() > 0;) {
601 Change ch2 = tip2.paragraph().lookupChange(tip2.pos());
602 if (!ch2.isSimilarTo(ch)) {
603 // take a step forward to correctly set the selection
611 cur.bv().mouseSetCursor(cur, false);
613 cur.bv().mouseSetCursor(cur, true);
621 bool findChange(BufferView * bv, bool forward)
624 cur.setCursor(forward ? bv->cursor().selectionEnd()
625 : bv->cursor().selectionBegin());
626 forward ? findNextChange(bv, cur, true) : findPreviousChange(bv, cur, true);
627 return selectChange(cur, forward);
632 bool findNextChange(BufferView * bv)
634 return findChange(bv, true);
638 bool findPreviousChange(BufferView * bv)
640 return findChange(bv, false);
647 typedef vector<pair<string, string> > Escapes;
649 /// A map of symbols and their escaped equivalent needed within a regex.
650 /// @note Beware of order
651 Escapes const & get_regexp_escapes()
653 typedef std::pair<std::string, std::string> P;
655 static Escapes escape_map;
656 if (escape_map.empty()) {
657 escape_map.push_back(P("$", "_x_$"));
658 escape_map.push_back(P("{", "_x_{"));
659 escape_map.push_back(P("}", "_x_}"));
660 escape_map.push_back(P("[", "_x_["));
661 escape_map.push_back(P("]", "_x_]"));
662 escape_map.push_back(P("(", "_x_("));
663 escape_map.push_back(P(")", "_x_)"));
664 escape_map.push_back(P("+", "_x_+"));
665 escape_map.push_back(P("*", "_x_*"));
666 escape_map.push_back(P(".", "_x_."));
667 escape_map.push_back(P("\\", "(?:\\\\|\\\\backslash)"));
668 escape_map.push_back(P("~", "(?:\\\\textasciitilde|\\\\sim)"));
669 escape_map.push_back(P("^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\textasciicircum|\\\\mathcircumflex)"));
670 escape_map.push_back(P("_x_", "\\"));
675 /// A map of lyx escaped strings and their unescaped equivalent.
676 Escapes const & get_lyx_unescapes()
678 typedef std::pair<std::string, std::string> P;
680 static Escapes escape_map;
681 if (escape_map.empty()) {
682 escape_map.push_back(P("\\%", "%"));
683 escape_map.push_back(P("\\{", "{"));
684 escape_map.push_back(P("\\}", "}"));
685 escape_map.push_back(P("\\mathcircumflex ", "^"));
686 escape_map.push_back(P("\\mathcircumflex", "^"));
687 escape_map.push_back(P("\\backslash ", "\\"));
688 escape_map.push_back(P("\\backslash", "\\"));
689 escape_map.push_back(P("\\sim ", "~"));
690 escape_map.push_back(P("\\sim", "~"));
695 /// A map of escapes turning a regexp matching text to one matching latex.
696 Escapes const & get_regexp_latex_escapes()
698 typedef std::pair<std::string, std::string> P;
700 static Escapes escape_map;
701 if (escape_map.empty()) {
702 escape_map.push_back(P("\\\\", "(?:\\\\\\\\|\\\\backslash|\\\\textbackslash\\{\\}|\\\\textbackslash)"));
703 escape_map.push_back(P("(<?!\\\\\\\\textbackslash)\\{", "\\\\\\{"));
704 escape_map.push_back(P("(<?!\\\\\\\\textbackslash\\\\\\{)\\}", "\\\\\\}"));
705 escape_map.push_back(P("\\[", "\\{\\[\\}"));
706 escape_map.push_back(P("\\]", "\\{\\]\\}"));
707 escape_map.push_back(P("\\^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\textasciicircum|\\\\mathcircumflex)"));
708 escape_map.push_back(P("%", "\\\\\\%"));
709 escape_map.push_back(P("#", "\\\\#"));
714 /** @todo Probably the maps need to be migrated to regexps, in order to distinguish if
715 ** the found occurrence were escaped.
717 string apply_escapes(string s, Escapes const & escape_map)
719 LYXERR(Debug::FIND, "Escaping: '" << s << "'");
720 Escapes::const_iterator it;
721 for (it = escape_map.begin(); it != escape_map.end(); ++it) {
722 // LYXERR(Debug::FIND, "Escaping " << it->first << " as " << it->second);
723 unsigned int pos = 0;
724 while (pos < s.length() && (pos = s.find(it->first, pos)) < s.length()) {
725 s.replace(pos, it->first.length(), it->second);
726 LYXERR(Debug::FIND, "After escape: " << s);
727 pos += it->second.length();
728 // LYXERR(Debug::FIND, "pos: " << pos);
731 LYXERR(Debug::FIND, "Escaped : '" << s << "'");
736 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
737 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
738 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
739 string escape_for_regex(string s, bool match_latex)
742 while (pos < s.size()) {
743 size_t new_pos = s.find("\\regexp{", pos);
744 if (new_pos == string::npos)
748 LYXERR(Debug::FIND, "new_pos: " << new_pos);
749 t = apply_escapes(s.substr(pos, new_pos - pos), get_lyx_unescapes());
750 LYXERR(Debug::FIND, "t [lyx]: " << t);
751 t = apply_escapes(t, get_regexp_escapes());
752 LYXERR(Debug::FIND, "t [rxp]: " << t);
753 s.replace(pos, new_pos - pos, t);
754 new_pos = pos + t.size();
755 LYXERR(Debug::FIND, "Regexp after escaping: " << s);
756 LYXERR(Debug::FIND, "new_pos: " << new_pos);
757 if (new_pos == s.size())
760 // Might fail if \\endregexp{} is preceeded by unexpected stuff (weird escapes)
761 size_t end_pos = s.find("\\endregexp{}}", new_pos + 8);
762 LYXERR(Debug::FIND, "end_pos: " << end_pos);
763 t = s.substr(new_pos + 8, end_pos - (new_pos + 8));
764 LYXERR(Debug::FIND, "t in regexp : " << t);
765 t = apply_escapes(t, get_lyx_unescapes());
766 LYXERR(Debug::FIND, "t in regexp after unescapes [lyx]: " << t);
768 t = apply_escapes(t, get_regexp_latex_escapes());
769 LYXERR(Debug::FIND, "t in regexp after latex_escapes [ltx]: " << t);
771 if (end_pos == s.size()) {
772 s.replace(new_pos, end_pos - new_pos, t);
773 LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s);
776 s.replace(new_pos, end_pos + 13 - new_pos, t);
777 LYXERR(Debug::FIND, "Regexp after \\regexp{...\\endregexp{}} removal: " << s);
778 pos = new_pos + t.size();
779 LYXERR(Debug::FIND, "pos: " << pos);
785 /// Wrapper for lyx::regex_replace with simpler interface
786 bool regex_replace(string const & s, string & t, string const & searchstr,
787 string const & replacestr)
789 regex e(searchstr, regex_constants::ECMAScript);
791 ostream_iterator<char, char> it(oss);
792 regex_replace(it, s.begin(), s.end(), e, replacestr);
793 // tolerate t and s be references to the same variable
794 bool rv = (s != oss.str());
800 /** Checks if supplied string segment is well-formed from the standpoint of matching open-closed braces.
802 ** Verify that closed braces exactly match open braces. This avoids that, for example,
803 ** \frac{.*}{x} matches \frac{x+\frac{y}{x}}{z} with .* being 'x+\frac{y'.
806 ** Number of open braces that must remain open at the end for the verification to succeed.
809 bool braces_match(QString const & beg,
812 bool braces_match(string const & beg,
818 LYXERR(Debug::FIND, "Checking " << unmatched << " unmatched braces in '" << beg.toStdString() << "'");
820 LYXERR(Debug::FIND, "Checking " << unmatched << " unmatched braces in '" << beg << "'");
822 int lastidx = beg.size();
823 for (int i=0; i < lastidx; ++i) {
824 // Skip escaped braces in the count
834 } else if (c == '{') {
836 } else if (c == '}') {
837 if (open_pars == 0) {
838 LYXERR(Debug::FIND, "Found unmatched closed brace");
844 if (open_pars != unmatched) {
845 LYXERR(Debug::FIND, "Found " << open_pars
846 << " instead of " << unmatched
847 << " unmatched open braces at the end of count");
850 LYXERR(Debug::FIND, "Braces match as expected");
867 MatchResult(): match_len(0),match_prefix(0),match2end(0), pos(0),leadsize(0) {};
870 static MatchResult::range interpretMatch(MatchResult &oldres, MatchResult &newres)
872 if (newres.match2end < oldres.match2end)
873 return MatchResult::newIsTooFar;
874 if (newres.match_len < oldres.match_len)
875 return MatchResult::newIsTooFar;
876 if ((newres.match_len == oldres.match_len) && (newres.match2end == oldres.match2end))
877 return MatchResult::newIsBetter;
878 return MatchResult::newIsInvalid;
881 /** The class performing a match between a position in the document and the FindAdvOptions.
884 class MatchStringAdv {
886 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt);
888 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
889 ** constructor as opt.search, under the opt.* options settings.
892 ** If set, then match is searched only against beginning of text starting at cur.
893 ** If unset, then match is searched anywhere in text starting at cur.
896 ** The length of the matching text, or zero if no match was found.
898 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
907 /// first buffer on which search was started
908 lyx::Buffer * const p_first_buf;
910 FindAndReplaceOptions const & opt;
913 /// Auxiliary find method (does not account for opt.matchword)
914 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
916 /** Normalize a stringified or latexified LyX paragraph.
920 ** <li>if search is not casesensitive, then lowercase the string;
921 ** <li>remove any newline at begin or end of the string;
922 ** <li>replace any newline in the middle of the string with a simple space;
923 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
926 ** @todo Normalization should also expand macros, if the corresponding
927 ** search option was checked.
929 string normalize(docstring const & s, bool hack_braces) const;
930 // normalized string to search
931 string par_as_string;
932 // regular expression to use for searching
933 // regexp2 is same as regexp, but prefixed with a ".*?"
935 QRegularExpression regexp;
936 QRegularExpression regexp2;
941 // leading format material as string
942 string lead_as_string;
943 // par_as_string after removal of lead_as_string
944 string par_as_string_nolead;
945 // unmatched open braces in the search string/regexp
947 // number of (.*?) subexpressions added at end of search regexp for closing
948 // environments, math mode, styles, etc...
950 // Are we searching with regular expressions ?
955 static docstring buffer_to_latex(Buffer & buffer)
957 //OutputParams runparams(&buffer.params().encoding());
958 OutputParams runparams(encodings.fromLyXName("utf8"));
959 odocstringstream ods;
961 runparams.nice = true;
962 runparams.flavor = Flavor::XeTeX;
963 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
964 // No side effect of file copying and image conversion
965 runparams.dryrun = true;
966 runparams.for_search = true;
967 pit_type const endpit = buffer.paragraphs().size();
968 for (pit_type pit = 0; pit != endpit; ++pit) {
969 TeXOnePar(buffer, buffer.text(), pit, os, runparams);
970 LYXERR(Debug::FIND, "searchString up to here: " << ods.str());
976 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
979 if (!opt.ignoreformat) {
980 str = buffer_to_latex(buffer);
982 // OutputParams runparams(&buffer.params().encoding());
983 OutputParams runparams(encodings.fromLyXName("utf8"));
984 runparams.nice = true;
985 runparams.flavor = Flavor::XeTeX;
986 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
987 runparams.dryrun = true;
988 runparams.for_search = true;
989 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
990 Paragraph const & par = buffer.paragraphs().at(pit);
991 LYXERR(Debug::FIND, "Adding to search string: '"
992 << par.asString(pos_type(0), par.size(),
993 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
996 str += par.asString(pos_type(0), par.size(),
997 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
1000 // Even in ignore-format we have to remove "\text{}, \lyxmathsym{}" parts
1001 string t = to_utf8(str);
1002 while (regex_replace(t, t, "\\\\(text|lyxmathsym)\\{([^\\}]*)\\}", "$2"));
1009 /// Return separation pos between the leading material and the rest
1010 static size_t identifyLeading(string const & s)
1013 // @TODO Support \item[text]
1014 // Kornel: Added textsl, textsf, textit, texttt and noun
1015 // + allow to search for colored text too
1016 while (regex_replace(t, t, "^\\\\(("
1017 "(author|title|subtitle|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|"
1018 "lyxaddress|lyxrightaddress|"
1019 "footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1020 "emph|noun|minisec|text(bf|md|sl|sf|it|tt))|"
1021 "((textcolor|foreignlanguage|latexenvironment)\\{[a-z]+\\*?\\})|"
1022 "(u|uu)line|(s|x)out|uwave)|((sub)?(((sub)?section)|paragraph)|part|chapter)\\*?)\\{", "")
1023 || regex_replace(t, t, "^\\$", "")
1024 || regex_replace(t, t, "^\\\\\\[", "")
1025 || regex_replace(t, t, "^ ?\\\\item\\{[a-z]+\\}", "")
1026 || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\}", ""))
1028 LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
1033 * Given a latexified string, retrieve some handled features
1034 * The features of the regex will later be compared with the features
1035 * of the searched text. If the regex features are not a
1036 * subset of the analized, then, in not format ignoring search
1037 * we can early stop the search in the relevant inset.
1039 typedef map<string, bool> Features;
1041 static Features identifyFeatures(string const & s)
1043 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
1044 static regex const valid("^("
1046 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1047 "emph|noun|text(bf|md|sl|sf|it|tt)|"
1048 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
1049 "(u|uu)line|(s|x)out|uwave|"
1050 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
1051 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
1056 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1059 if (sub.str(1).compare("regexp") == 0) {
1063 string token = sub.str(1);
1065 if (regex_match(token, sub2, valid)) {
1073 if (sub.str(1).compare("endregexp") == 0) {
1083 * defines values features of a key "\\[a-z]+{"
1088 /* Char type with content discarded
1089 * like \hspace{1cm} */
1091 /* Char, like \backslash */
1093 /* replace starting backslash with '#' */
1095 /* \part, \section*, ... */
1097 /* title, author etc */
1099 /* \foreignlanguage{ngerman}, ... */
1102 * to discard language in content */
1105 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1107 /* fonts, colors, markups, ... */
1109 /* footnotesize, ... large, ...
1110 * Ignore all of them */
1113 /* inputencoding, ...
1114 * Discard also content, because they do not help in search */
1117 * like remove, but also all arguments */
1119 /* item, listitem */
1121 /* tex, latex, ... like isChar */
1123 /* like \lettrine[lines=5]{}{} */
1125 // like isStandard, but always remove head
1127 /* End of arguments marker for lettrine,
1128 * so that they can be ignored */
1131 KeyInfo() = default;
1132 KeyInfo(KeyType type, int parcount, bool disable)
1134 parenthesiscount(parcount),
1135 disabled(disable) {}
1136 KeyType keytype = invalid;
1138 int _tokensize = -1;
1139 int _tokenstart = -1;
1140 int _dataStart = -1;
1142 int parenthesiscount = 1;
1143 bool disabled = false;
1144 bool used = false; /* by pattern */
1149 Border(int l=0, int u=0) : low(l), upper(u) {};
1154 #define MAXOPENED 30
1156 bool isPatternString_;
1158 explicit Intervall(bool isPattern, string const & p) :
1159 isPatternString_(isPattern), par(p), ignoreidx(-1), actualdeptindex(0),
1160 hasTitle(false), langcount(0)
1168 static vector<Border> borders;
1169 int depts[MAXOPENED];
1170 int closes[MAXOPENED];
1171 int actualdeptindex;
1172 int previousNotIgnored(int) const;
1173 int nextNotIgnored(int) const;
1174 void handleOpenP(int i);
1175 void handleCloseP(int i, bool closingAllowed);
1176 void resetOpenedP(int openPos);
1177 void addIntervall(int upper);
1178 void addIntervall(int low, int upper); /* if explicit */
1179 void removeAccents();
1180 void setForDefaultLang(KeyInfo const & defLang) const;
1181 int findclosing(int start, int end, char up, char down, int repeat);
1182 void handleParentheses(int lastpos, bool closingAllowed);
1184 int langcount; // Number of disabled language specs up to current position in actual interval
1185 int isOpeningPar(int pos) const;
1187 void output(ostringstream &os, int lastpos);
1188 // string show(int lastpos);
1191 vector<Border> Intervall::borders = vector<Border>(30);
1193 int Intervall::isOpeningPar(int pos) const
1195 if ((pos < 0) || (size_t(pos) >= par.size()))
1197 if (par[pos] != '{')
1199 if (size_t(pos) + 2 >= par.size())
1201 if (par[pos+2] != '}')
1203 if (par[pos+1] == '[' || par[pos+1] == ']')
1208 void Intervall::setForDefaultLang(KeyInfo const & defLang) const
1210 // Enable the use of first token again
1211 if (ignoreidx >= 0) {
1212 int value = defLang._tokenstart + defLang._tokensize;
1218 if (borders[borderidx].low < value)
1219 borders[borderidx].low = value;
1220 if (borders[borderidx].upper < value)
1221 borders[borderidx].upper = value;
1226 static void checkDepthIndex(int val)
1228 static int maxdepthidx = MAXOPENED-2;
1229 static int lastmaxdepth = 0;
1230 if (val > lastmaxdepth) {
1231 LYXERR(Debug::INFO, "Depth reached " << val);
1234 if (val > maxdepthidx) {
1236 LYXERR(Debug::INFO, "maxdepthidx now " << val);
1241 // Not needed, because borders are now dynamically expanded
1242 static void checkIgnoreIdx(int val)
1244 static int lastmaxignore = -1;
1245 if ((lastmaxignore < val) && (size_t(val+1) >= borders.size())) {
1246 LYXERR(Debug::INFO, "IgnoreIdx reached " << val);
1247 lastmaxignore = val;
1253 * Expand the region of ignored parts of the input latex string
1254 * The region is only relevant in output()
1256 void Intervall::addIntervall(int low, int upper)
1259 if (low == upper) return;
1260 for (idx = ignoreidx+1; idx > 0; --idx) {
1261 if (low > borders[idx-1].upper) {
1265 Border br(low, upper);
1266 if (idx > ignoreidx) {
1267 if (borders.size() <= size_t(idx)) {
1268 borders.push_back(br);
1274 // checkIgnoreIdx(ignoreidx);
1278 // Expand only if one of the new bound is inside the interwall
1279 // We know here that br.low > borders[idx-1].upper
1280 if (br.upper < borders[idx].low) {
1281 // We have to insert at this pos
1282 if (size_t(ignoreidx+1) >= borders.size()) {
1283 borders.push_back(borders[ignoreidx]);
1286 borders[ignoreidx+1] = borders[ignoreidx];
1288 for (int i = ignoreidx; i > idx; --i) {
1289 borders[i] = borders[i-1];
1293 // checkIgnoreIdx(ignoreidx);
1296 // Here we know, that we are overlapping
1297 if (br.low > borders[idx].low)
1298 br.low = borders[idx].low;
1299 // check what has to be concatenated
1301 for (int i = idx; i <= ignoreidx; i++) {
1302 if (br.upper >= borders[i].low) {
1304 if (br.upper < borders[i].upper)
1305 br.upper = borders[i].upper;
1311 // count should be >= 1 here
1314 for (int i = idx + count; i <= ignoreidx; i++) {
1315 borders[i-count+1] = borders[i];
1317 ignoreidx -= count - 1;
1323 typedef map<string, string> AccentsMap;
1324 static AccentsMap accents = map<string, string>();
1326 static void buildaccent(string n, string param, string values)
1330 const char delim = '|';
1331 while (getline(s, name, delim)) {
1333 for (char c : param) {
1334 string key = name + "{" + c + "}";
1335 // get the corresponding utf8-value
1336 if ((values[start] & 0xc0) != 0xc0) {
1337 // should not happen, utf8 encoding starts at least with 11xxxxxx
1338 // but value for '\dot{i}' is 'i', which is ascii
1339 if ((values[start] & 0x80) == 0) {
1341 accents[key] = values.substr(start, 1);
1342 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1347 for (int j = 1; ;j++) {
1348 if (start + j >= values.size()) {
1349 accents[key] = values.substr(start, j);
1350 start = values.size() - 1;
1353 else if ((values[start+j] & 0xc0) != 0x80) {
1354 // This is the first byte of following utf8 char
1355 accents[key] = values.substr(start, j);
1357 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1366 static string getutf8(unsigned uchar)
1369 string ret = string();
1371 if (uchar <= 0x7f) {
1372 c[maxc-1] = uchar & 0x7f;
1375 unsigned char rest = 0x40;
1376 unsigned char first = 0x80;
1378 for (int i = start; i >=0; --i) {
1380 c[i] = first + uchar;
1383 c[i] = 0x80 | (uchar & 0x3f);
1390 for (int i = 0; i < maxc; i++) {
1391 if (c[i] == 0) continue;
1397 static void buildAccentsMap()
1399 accents["imath"] = "ı";
1401 accents["jmath"] = "ȷ";
1402 accents["cdot"] = "·";
1403 accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
1404 accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
1405 accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
1406 accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
1407 accents["negmedspace"] = getutf8(0xf0005);
1408 accents["thickspace"] = getutf8(0xf0006);
1409 accents["negthickspace"] = getutf8(0xf0007);
1410 accents["lyx"] = getutf8(0xf0010); // Used logos
1411 accents["LyX"] = getutf8(0xf0010);
1412 accents["tex"] = getutf8(0xf0011);
1413 accents["TeX"] = getutf8(0xf0011);
1414 accents["latex"] = getutf8(0xf0012);
1415 accents["LaTeX"] = getutf8(0xf0012);
1416 accents["latexe"] = getutf8(0xf0013);
1417 accents["LaTeXe"] = getutf8(0xf0013);
1418 accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash
1419 accents["backslash LyX"] = getutf8(0xf0010);
1420 accents["backslash tex"] = getutf8(0xf0011);
1421 accents["backslash TeX"] = getutf8(0xf0011);
1422 accents["backslash latex"] = getutf8(0xf0012);
1423 accents["backslash LaTeX"] = getutf8(0xf0012);
1424 accents["backslash latexe"] = getutf8(0xf0013);
1425 accents["backslash LaTeXe"] = getutf8(0xf0013);
1426 accents["ddot{\\imath}"] = "ï";
1427 buildaccent("ddot", "aAeEhHiIioOtuUwWxXyY",
1428 "äÄëËḧḦïÏïöÖẗüÜẅẄẍẌÿŸ"); // umlaut
1429 buildaccent("dot|.", "aAbBcCdDeEfFGghHIimMnNoOpPrRsStTwWxXyYzZ",
1430 "ȧȦḃḂċĊḋḊėĖḟḞĠġḣḢİİṁṀṅṄȯȮṗṖṙṘṡṠṫṪẇẆẋẊẏẎżŻ"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'İ'
1431 accents["acute{\\imath}"] = "í";
1432 buildaccent("acute", "aAcCeEgGkKlLmMoOnNpPrRsSuUwWyYzZiI",
1433 "áÁćĆéÉǵǴḱḰĺĹḿḾóÓńŃṕṔŕŔśŚúÚẃẂýÝźŹíÍ");
1434 buildaccent("dacute|H|h", "oOuU", "őŐűŰ"); // double acute
1435 buildaccent("mathring|r", "aAuUwy",
1437 accents["check{\\imath}"] = "ǐ";
1438 accents["check{\\jmath}"] = "ǰ";
1439 buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
1440 "čČďĎǎǍěĚǐǏǒǑǔǓǧǦǩǨȟȞľĽňŇřŘšŠŤťžŽ"); // caron
1441 accents["hat{\\imath}"] = "î";
1442 accents["hat{\\jmath}"] = "ĵ";
1443 buildaccent("hat|^", "aAcCeEgGhHiIjJoOsSuUwWyYzZ",
1444 "âÂĉĈêÊĝĜĥĤîÎĵĴôÔŝŜûÛŵŴŷŶẑẐ"); // circ
1445 accents["bar{\\imath}"] = "ī";
1446 buildaccent("bar|=", "aAeEiIoOuUyY",
1447 "āĀēĒīĪōŌūŪȳȲ"); // macron
1448 accents["tilde{\\imath}"] = "ĩ";
1449 buildaccent("tilde", "aAeEiInNoOuUvVyY",
1450 "ãÃẽẼĩĨñÑõÕũŨṽṼỹỸ"); // tilde
1451 accents["breve{\\imath}"] = "ĭ";
1452 buildaccent("breve|u", "aAeEgGiIoOuU",
1453 "ăĂĕĔğĞĭĬŏŎŭŬ"); // breve
1454 accents["grave{\\imath}"] = "ì";
1455 buildaccent("grave|`", "aAeEiIoOuUnNwWyY",
1456 "àÀèÈìÌòÒùÙǹǸẁẀỳỲ"); // grave
1457 buildaccent("subdot|d", "BbDdHhKkLlMmNnRrSsTtVvWwZzAaEeIiOoUuYy",
1458 "ḄḅḌḍḤḥḲḳḶḷṂṃṆṇṚṛṢṣṬṭṾṿẈẉẒẓẠạẸẹỊịỌọỤụỴỵ"); // dot below
1459 buildaccent("ogonek|k", "AaEeIiUuOo",
1460 "ĄąĘęĮįŲųǪǫ"); // ogonek
1461 buildaccent("cedilla|c", "CcGgKkLlNnRrSsTtEeDdHh",
1462 "ÇçĢģĶķĻļŅņŖŗŞşŢţȨȩḐḑḨḩ"); // cedilla
1463 buildaccent("subring|textsubring", "Aa",
1465 buildaccent("subhat|textsubcircum", "DdEeLlNnTtUu",
1466 "ḒḓḘḙḼḽṊṋṰṱṶṷ"); // subcircum
1467 buildaccent("subtilde|textsubtilde", "EeIiUu",
1468 "ḚḛḬḭṴṵ"); // subtilde
1469 accents["dgrave{\\imath}"] = "ȉ";
1470 accents["textdoublegrave{\\i}"] = "ȉ";
1471 buildaccent("dgrave|textdoublegrave", "AaEeIiOoRrUu",
1472 "ȀȁȄȅȈȉȌȍȐȑȔȕ"); // double grave
1473 accents["rcap{\\imath}"] = "ȉ";
1474 accents["textroundcap{\\i}"] = "ȉ";
1475 buildaccent("rcap|textroundcap", "AaEeIiOoRrUu",
1476 "ȂȃȆȇȊȋȎȏȒȓȖȗ"); // inverted breve
1477 buildaccent("slashed", "oO",
1482 * Created accents in math or regexp environment
1483 * are macros, but we need the utf8 equivalent
1485 void Intervall::removeAccents()
1487 if (accents.empty())
1489 static regex const accre("\\\\(([\\S]|grave|breve|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot|ogonek|"
1490 "cedilla|subring|textsubring|subhat|textsubcircum|subtilde|textsubtilde|dgrave|textdoublegrave|rcap|textroundcap|slashed)\\{[^\\{\\}]+\\}"
1491 "|((i|imath|jmath|cdot|[a-z]+space)|((backslash )?([lL]y[xX]|[tT]e[xX]|[lL]a[tT]e[xX]e?)))(?![a-zA-Z]))");
1493 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
1495 string key = sub.str(1);
1496 if (accents.find(key) != accents.end()) {
1497 string val = accents[key];
1498 size_t pos = sub.position(size_t(0));
1499 for (size_t i = 0; i < val.size(); i++) {
1500 par[pos+i] = val[i];
1502 // Remove possibly following space too
1503 if (par[pos+sub.str(0).size()] == ' ')
1504 addIntervall(pos+val.size(), pos + sub.str(0).size()+1);
1506 addIntervall(pos+val.size(), pos + sub.str(0).size());
1507 for (size_t i = pos+val.size(); i < pos + sub.str(0).size(); i++) {
1508 // remove traces of any remaining chars
1513 LYXERR(Debug::INFO, "Not added accent for \"" << key << "\"");
1518 void Intervall::handleOpenP(int i)
1521 depts[actualdeptindex] = i+1;
1522 closes[actualdeptindex] = -1;
1523 checkDepthIndex(actualdeptindex);
1526 void Intervall::handleCloseP(int i, bool closingAllowed)
1528 if (actualdeptindex <= 0) {
1529 if (! closingAllowed)
1530 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1531 // if we are at the very end
1532 addIntervall(i, i+1);
1535 closes[actualdeptindex] = i+1;
1540 void Intervall::resetOpenedP(int openPos)
1542 // Used as initializer for foreignlanguage entry
1543 actualdeptindex = 1;
1544 depts[1] = openPos+1;
1548 int Intervall::previousNotIgnored(int start) const
1550 int idx = 0; /* int intervalls */
1551 for (idx = ignoreidx; idx >= 0; --idx) {
1552 if (start > borders[idx].upper)
1554 if (start >= borders[idx].low)
1555 start = borders[idx].low-1;
1560 int Intervall::nextNotIgnored(int start) const
1562 int idx = 0; /* int intervalls */
1563 for (idx = 0; idx <= ignoreidx; idx++) {
1564 if (start < borders[idx].low)
1566 if (start < borders[idx].upper)
1567 start = borders[idx].upper;
1572 typedef map<string, KeyInfo> KeysMap;
1573 typedef vector< KeyInfo> Entries;
1574 static KeysMap keys = map<string, KeyInfo>();
1580 Intervall interval_;
1581 void buildKeys(bool);
1582 void buildEntries(bool);
1583 void makeKey(const string &, KeyInfo, bool isPatternString);
1584 void processRegion(int start, int region_end); /* remove {} parts */
1585 void removeHead(KeyInfo const &, int count=0);
1588 LatexInfo(string const & par, bool isPatternString)
1589 : entidx_(-1), interval_(isPatternString, par)
1591 buildKeys(isPatternString);
1592 entries_ = vector<KeyInfo>();
1593 buildEntries(isPatternString);
1597 if (entries_.empty()) {
1600 if (entries_[0].keytype == KeyInfo::isTitle) {
1601 interval_.hasTitle = true;
1602 if (! entries_[0].disabled) {
1603 interval_.titleValue = entries_[0].head;
1606 interval_.titleValue = "";
1608 removeHead(entries_[0]);
1609 if (entries_.size() > 1)
1618 if (int(entries_.size()) > entidx_) {
1625 bool setNextKey(int idx) {
1626 if ((idx == entidx_) && (entidx_ >= 0)) {
1633 int find(int start, KeyInfo::KeyType keytype) const {
1637 while (tmpIdx < int(entries_.size())) {
1638 if (entries_[tmpIdx].keytype == keytype)
1644 int process(ostringstream & os, KeyInfo const & actual);
1645 int dispatch(ostringstream & os, int previousStart, KeyInfo & actual);
1646 // string show(int lastpos) { return interval.show(lastpos);};
1647 int nextNotIgnored(int start) { return interval_.nextNotIgnored(start);};
1648 KeyInfo &getKeyInfo(int keyinfo) {
1649 static KeyInfo invalidInfo = KeyInfo();
1650 if ((keyinfo < 0) || ( keyinfo >= int(entries_.size())))
1653 return entries_[keyinfo];
1655 void setForDefaultLang(KeyInfo const & defLang) {interval_.setForDefaultLang(defLang);};
1656 void addIntervall(int low, int up) { interval_.addIntervall(low, up); };
1660 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
1664 for (int i = start; i < end; i += 1 + skip) {
1668 if (c == '\\') skip = 1;
1672 else if (c == down) {
1675 if ((repeat <= 0) || (par[i+1] != up))
1693 vector<MathEntry> entries_;
1698 void insert(string const & wait, size_t start, size_t end) {
1699 MathEntry m = MathEntry();
1701 m.mathStart = start;
1703 m.mathSize = end - start;
1704 entries_.push_back(m);
1706 bool empty() const { return entries_.empty(); };
1707 size_t getEndPos() const {
1708 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1711 return entries_[actualIdx_].mathEnd;
1713 size_t getStartPos() const {
1714 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1715 return 100000; /* definitely enough? */
1717 return entries_[actualIdx_].mathStart;
1719 size_t getFirstPos() {
1721 return getStartPos();
1723 size_t getSize() const {
1724 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1727 return entries_[actualIdx_].mathSize;
1729 void incrEntry() { actualIdx_++; };
1732 void LatexInfo::buildEntries(bool isPatternString)
1734 static regex const rmath("\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|alignat)\\*?)\\}");
1735 static regex const rkeys("\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?))");
1736 static bool disableLanguageOverride = false;
1737 smatch sub, submath;
1738 bool evaluatingRegexp = false;
1740 bool evaluatingMath = false;
1741 bool evaluatingCode = false;
1743 bool evaluatingOptional = false;
1744 size_t optionalEnd = 0;
1747 bool math_end_waiting = false;
1748 size_t math_pos = 10000;
1750 static vector<string> usedText = vector<string>();
1752 interval_.removeAccents();
1754 for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) {
1756 if (math_end_waiting) {
1757 size_t pos = submath.position(size_t(0));
1758 if ((math_end == "$") &&
1759 (submath.str(0) == "$") &&
1760 (interval_.par[pos-1] != '\\')) {
1761 mi.insert("$", math_pos, pos + 1);
1762 math_end_waiting = false;
1764 else if ((math_end == "\\]") &&
1765 (submath.str(0) == "\\]")) {
1766 mi.insert("\\]", math_pos, pos + 2);
1767 math_end_waiting = false;
1769 else if ((submath.str(1).compare("end") == 0) &&
1770 (submath.str(2).compare(math_end) == 0)) {
1771 mi.insert(math_end, math_pos, pos + submath.str(0).length());
1772 math_end_waiting = false;
1778 if (submath.str(1).compare("begin") == 0) {
1779 math_end_waiting = true;
1780 math_end = submath.str(2);
1781 math_pos = submath.position(size_t(0));
1783 else if (submath.str(0).compare("\\[") == 0) {
1784 math_end_waiting = true;
1786 math_pos = submath.position(size_t(0));
1788 else if (submath.str(0) == "$") {
1789 size_t pos = submath.position(size_t(0));
1790 if ((pos == 0) || (interval_.par[pos-1] != '\\')) {
1791 math_end_waiting = true;
1798 // Ignore language if there is math somewhere in pattern-string
1799 if (isPatternString) {
1800 for (auto s: usedText) {
1801 // Remove entries created in previous search runs
1804 usedText = vector<string>();
1807 keys["foreignlanguage"].disabled = true;
1808 disableLanguageOverride = true;
1811 disableLanguageOverride = false;
1814 if (disableLanguageOverride) {
1815 keys["foreignlanguage"].disabled = true;
1818 math_pos = mi.getFirstPos();
1819 for (sregex_iterator it(interval_.par.begin(), interval_.par.end(), rkeys), end; it != end; ++it) {
1821 string key = sub.str(3);
1823 if (sub.str(0)[0] == '\\')
1824 key = sub.str(0)[1];
1828 size_t k_pos = sub.position(size_t(0));
1829 if ((k_pos > 0) && (interval_.par[k_pos - 1] == '\\')) {
1830 // Escaped '$', ignoring
1836 if (keys.find(key) != keys.end()) {
1837 if (keys[key].keytype == KeyInfo::headRemove) {
1838 KeyInfo found1 = keys[key];
1839 found1.disabled = true;
1840 found1.head = "\\" + key + "{";
1841 found1._tokenstart = sub.position(size_t(0));
1842 found1._tokensize = found1.head.length();
1843 found1._dataStart = found1._tokenstart + found1.head.length();
1844 int endpos = interval_.findclosing(found1._dataStart, interval_.par.length(), '{', '}', 1);
1845 found1._dataEnd = endpos;
1850 if (evaluatingRegexp) {
1851 if (sub.str(1).compare("endregexp") == 0) {
1852 evaluatingRegexp = false;
1853 // found._tokenstart already set
1854 found._dataEnd = sub.position(size_t(0)) + 13;
1855 found._dataStart = found._dataEnd;
1856 found._tokensize = found._dataEnd - found._tokenstart;
1857 found.parenthesiscount = 0;
1858 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1865 if (evaluatingMath) {
1866 if (size_t(sub.position(size_t(0))) < mi.getEndPos())
1868 evaluatingMath = false;
1870 math_pos = mi.getStartPos();
1872 if (keys.find(key) == keys.end()) {
1873 found = KeyInfo(KeyInfo::isStandard, 0, true);
1874 LYXERR(Debug::INFO, "Undefined key " << key << " ==> will be used as text");
1875 found = KeyInfo(KeyInfo::isText, 0, false);
1876 if (isPatternString) {
1877 found.keytype = KeyInfo::isChar;
1878 found.disabled = false;
1882 usedText.push_back(key);
1886 if (key.compare("regexp") == 0) {
1887 evaluatingRegexp = true;
1888 found._tokenstart = sub.position(size_t(0));
1889 found._tokensize = 0;
1893 // Handle the other params of key
1894 if (found.keytype == KeyInfo::isIgnored)
1896 else if (found.keytype == KeyInfo::isMath) {
1897 if (size_t(sub.position(size_t(0))) == math_pos) {
1899 found._tokenstart = sub.position(size_t(0));
1900 found._tokensize = mi.getSize();
1901 found._dataEnd = found._tokenstart + found._tokensize;
1902 found._dataStart = found._dataEnd;
1903 found.parenthesiscount = 0;
1904 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1905 evaluatingMath = true;
1908 // begin|end of unknown env, discard
1909 // First handle tables
1910 // longtable|tabular
1911 bool discardComment;
1913 found.keytype = KeyInfo::doRemove;
1914 if ((sub.str(5).compare("longtable") == 0) ||
1915 (sub.str(5).compare("tabular") == 0)) {
1916 discardComment = true; /* '%' */
1919 discardComment = false;
1920 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
1922 string token = sub.str(5);
1923 if (regex_match(token, sub2, removeArgs)) {
1924 found.keytype = KeyInfo::removeWithArg;
1927 // discard spaces before pos(0)
1928 int pos = sub.position(size_t(0));
1930 for (count = 0; pos - count > 0; count++) {
1931 char c = interval_.par[pos-count-1];
1932 if (discardComment) {
1933 if ((c != ' ') && (c != '%'))
1939 found._tokenstart = pos - count;
1940 if (sub.str(1).compare(0, 5, "begin") == 0) {
1941 size_t pos1 = pos + sub.str(0).length();
1942 if (sub.str(5).compare("cjk") == 0) {
1943 pos1 = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
1944 if ((interval_.par[pos1] == '{') && (interval_.par[pos1+1] == '}'))
1946 found.keytype = KeyInfo::isMain;
1947 found._dataStart = pos1;
1948 found._dataEnd = interval_.par.length();
1949 found.disabled = keys["foreignlanguage"].disabled;
1950 found.used = keys["foreignlanguage"].used;
1951 found._tokensize = pos1 - found._tokenstart;
1952 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1955 // Swallow possible optional params
1956 while (interval_.par[pos1] == '[') {
1957 pos1 = interval_.findclosing(pos1+1, interval_.par.length(), '[', ']')+1;
1959 // Swallow also the eventual parameter
1960 if (interval_.par[pos1] == '{') {
1961 found._dataEnd = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
1964 found._dataEnd = pos1;
1966 found._dataStart = found._dataEnd;
1967 found._tokensize = count + found._dataEnd - pos;
1968 found.parenthesiscount = 0;
1969 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1970 found.disabled = true;
1974 // Handle "\end{...}"
1975 found._dataStart = pos + sub.str(0).length();
1976 found._dataEnd = found._dataStart;
1977 found._tokensize = count + found._dataEnd - pos;
1978 found.parenthesiscount = 0;
1979 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1980 found.disabled = true;
1984 else if (found.keytype != KeyInfo::isRegex) {
1985 found._tokenstart = sub.position(size_t(0));
1986 if (found.parenthesiscount == 0) {
1987 // Probably to be discarded
1988 size_t following_pos = sub.position(size_t(0)) + sub.str(3).length() + 1;
1989 char following = interval_.par[following_pos];
1990 if (following == ' ')
1991 found.head = "\\" + sub.str(3) + " ";
1992 else if (following == '=') {
1993 // like \uldepth=1000pt
1994 found.head = sub.str(0);
1997 found.head = "\\" + key;
1998 found._tokensize = found.head.length();
1999 found._dataEnd = found._tokenstart + found._tokensize;
2000 found._dataStart = found._dataEnd;
2003 int params = found._tokenstart + key.length() + 1;
2004 if (evaluatingOptional) {
2005 if (size_t(found._tokenstart) > optionalEnd) {
2006 evaluatingOptional = false;
2009 found.disabled = true;
2012 int optend = params;
2013 while (interval_.par[optend] == '[') {
2014 // discard optional parameters
2015 optend = interval_.findclosing(optend+1, interval_.par.length(), '[', ']') + 1;
2017 if (optend > params) {
2018 key += interval_.par.substr(params, optend-params);
2019 evaluatingOptional = true;
2020 optionalEnd = optend;
2022 string token = sub.str(5);
2024 if (interval_.par[optend] != '{') {
2026 found.parenthesiscount = 0;
2029 closings = found.parenthesiscount;
2030 if (found.parenthesiscount == 1) {
2031 found.head = "\\" + key + "{";
2033 else if (found.parenthesiscount > 1) {
2035 found.head = sub.str(0) + "{";
2036 closings = found.parenthesiscount - 1;
2039 found.head = "\\" + key + "{";
2042 found._tokensize = found.head.length();
2043 found._dataStart = found._tokenstart + found.head.length();
2044 if (found.keytype == KeyInfo::doRemove) {
2045 int endpar = 2 + interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2046 found._dataStart = endpar;
2047 found._tokensize = found._dataStart - found._tokenstart;
2050 if (interval_.par.substr(found._dataStart-1, 15).compare("\\endarguments{}") == 0) {
2051 found._dataStart += 15;
2055 endpos = found._dataStart - 1;
2057 endpos = interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2058 if (found.keytype == KeyInfo::isList) {
2059 // Check if it really is list env
2060 static regex const listre("^([a-z]+)$");
2062 if (!regex_match(token, sub2, listre)) {
2063 // Change the key of this entry. It is not in a list/item environment
2064 found.keytype = KeyInfo::endArguments;
2067 if (found.keytype == KeyInfo::noMain) {
2068 evaluatingCode = true;
2070 codeStart = found._dataStart;
2072 else if (evaluatingCode) {
2073 if (size_t(found._dataStart) > codeEnd)
2074 evaluatingCode = false;
2075 else if (found.keytype == KeyInfo::isMain) {
2076 // Disable this key, treate it as standard
2077 found.keytype = KeyInfo::isStandard;
2078 found.disabled = true;
2079 if ((codeEnd +1 >= interval_.par.length()) &&
2080 (found._tokenstart == codeStart)) {
2081 // trickery, because the code inset starts
2082 // with \selectlanguage ...
2084 if (entries_.size() > 1) {
2085 entries_[entries_.size()-1]._dataEnd = codeEnd;
2090 if ((endpos == interval_.par.length()) &&
2091 (found.keytype == KeyInfo::doRemove)) {
2092 // Missing closing => error in latex-input?
2093 // therefore do not delete remaining data
2094 found._dataStart -= 1;
2095 found._dataEnd = found._dataStart;
2098 found._dataEnd = endpos;
2100 if (isPatternString) {
2101 keys[key].used = true;
2104 entries_.push_back(found);
2108 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
2110 stringstream s(keysstring);
2112 const char delim = '|';
2113 while (getline(s, key, delim)) {
2114 KeyInfo keyII(keyI);
2115 if (isPatternString) {
2118 else if ( !keys[key].used)
2119 keyII.disabled = true;
2124 void LatexInfo::buildKeys(bool isPatternString)
2127 static bool keysBuilt = false;
2128 if (keysBuilt && !isPatternString) return;
2130 // Keys to ignore in any case
2131 makeKey("text|textcyrillic|lyxmathsym", KeyInfo(KeyInfo::headRemove, 1, true), true);
2132 // Known standard keys with 1 parameter.
2133 // Split is done, if not at start of region
2134 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
2135 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
2136 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
2137 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
2138 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
2139 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
2141 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
2142 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2143 makeKey("section*|subsection*|subsubsection*|paragraph*",
2144 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2145 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2146 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
2148 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
2150 // Split is done, if not at start of region
2151 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
2152 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
2154 // Split is done always.
2155 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
2159 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2160 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2161 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2162 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2163 makeKey("guillemotright|guillemotleft", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2165 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2166 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2167 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2168 makeKey("thickspace|medspace|thinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2170 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2171 // Custom space/skip, remove the content (== length value)
2172 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
2173 // Found in fr/UserGuide.lyx
2174 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2176 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2177 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2178 // Known macros to remove (including their parameter)
2180 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
2181 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
2182 // handle like standard keys with 1 parameter.
2183 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
2185 // Ignore deleted text
2186 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString);
2187 // but preserve added text
2188 makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2190 // Macros to remove, but let the parameter survive
2192 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2194 // Remove language spec from content of these insets
2195 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
2197 // Same effect as previous, parameter will survive (because there is no one anyway)
2199 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2200 // Remove table decorations
2201 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
2202 // Discard shape-header.
2203 // For footnote or shortcut too, because of lang settings
2204 // and wrong handling if used 'KeyInfo::noMain'
2205 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2206 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2207 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2208 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2209 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2210 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
2211 // like ('tiny{}' or '\tiny ' ... )
2212 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
2214 // Survives, like known character
2215 // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2216 makeKey("tableofcontents", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2217 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
2219 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2220 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2221 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2223 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2224 // Remove RTL/LTR marker
2225 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2226 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
2227 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
2228 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
2229 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
2230 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
2231 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2232 if (isPatternString) {
2233 // Allow the first searched string to rebuild the keys too
2237 // no need to rebuild again
2243 * Keep the list of actual opened parentheses actual
2244 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2246 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2249 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2253 if (c == '\\') skip = 1;
2254 else if (c == '{') {
2257 else if (c == '}') {
2258 handleCloseP(i, closingAllowed);
2264 string Intervall::show(int lastpos)
2266 int idx = 0; /* int intervalls */
2269 for (idx = 0; idx <= ignoreidx; idx++) {
2270 while (i < lastpos) {
2272 if (i <= borders[idx].low) {
2273 if (borders[idx].low > lastpos)
2274 printsize = lastpos - i;
2276 printsize = borders[idx].low - i;
2277 s += par.substr(i, printsize);
2279 if (i >= borders[idx].low)
2280 i = borders[idx].upper;
2283 i = borders[idx].upper;
2289 s += par.substr(i, lastpos-i);
2295 void Intervall::output(ostringstream &os, int lastpos)
2297 // get number of chars to output
2298 int idx = 0; /* int intervalls */
2301 string startTitle = titleValue;
2302 for (idx = 0; idx <= ignoreidx; idx++) {
2304 if (i <= borders[idx].low) {
2306 if (borders[idx].low > lastpos)
2307 printsize = lastpos - i;
2309 printsize = borders[idx].low - i;
2310 if (printsize > 0) {
2311 os << startTitle << par.substr(i, printsize);
2313 printed += printsize;
2316 handleParentheses(i, false);
2317 if (i >= borders[idx].low)
2318 i = borders[idx].upper;
2321 i = borders[idx].upper;
2328 os << startTitle << par.substr(i, lastpos-i);
2329 printed += lastpos-i;
2331 handleParentheses(lastpos, false);
2333 if (keys["foreignlanguage"].disabled)
2334 startindex = actualdeptindex-langcount;
2336 startindex = actualdeptindex;
2337 for (int i = startindex; i > 0; --i) {
2340 if (hasTitle && (printed > 0))
2342 if (! isPatternString_)
2344 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2347 void LatexInfo::processRegion(int start, int region_end)
2349 while (start < region_end) { /* Let {[} and {]} survive */
2350 int cnt = interval_.isOpeningPar(start);
2352 // Closing is allowed past the region
2353 int closing = interval_.findclosing(start+1, interval_.par.length());
2354 interval_.addIntervall(start, start+1);
2355 interval_.addIntervall(closing, closing+1);
2359 start = interval_.nextNotIgnored(start+1);
2363 void LatexInfo::removeHead(KeyInfo const & actual, int count)
2365 if (actual.parenthesiscount == 0) {
2366 // "{\tiny{} ...}" ==> "{{} ...}"
2367 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2370 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2371 interval_.addIntervall(actual._tokenstart - count, actual._dataStart);
2372 interval_.addIntervall(actual._dataEnd, actual._dataEnd+1);
2376 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2379 switch (actual.keytype)
2381 case KeyInfo::isTitle: {
2383 nextKeyIdx = getNextKey();
2386 case KeyInfo::cleanToStart: {
2387 actual._dataEnd = actual._dataStart;
2388 nextKeyIdx = getNextKey();
2389 // Search for end of arguments
2390 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2392 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2393 entries_[i].disabled = true;
2395 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2397 while (interval_.par[actual._dataEnd] == ' ')
2399 interval_.addIntervall(0, actual._dataEnd+1);
2400 interval_.actualdeptindex = 0;
2401 interval_.depts[0] = actual._dataEnd+1;
2402 interval_.closes[0] = -1;
2405 case KeyInfo::isText:
2406 interval_.par[actual._tokenstart] = '#';
2407 //interval_.addIntervall(actual._tokenstart, actual._tokenstart+1);
2408 nextKeyIdx = getNextKey();
2410 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2411 if (actual.disabled)
2412 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2414 interval_.addIntervall(actual._dataStart, actual._dataEnd);
2417 case KeyInfo::isChar: {
2418 nextKeyIdx = getNextKey();
2421 case KeyInfo::isSize: {
2422 if (actual.disabled || (interval_.par[actual._dataStart] != '{') || (interval_.par[actual._dataStart-1] == ' ')) {
2423 processRegion(actual._dataEnd, actual._dataEnd+1); /* remove possibly following {} */
2424 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2425 nextKeyIdx = getNextKey();
2427 // Here _dataStart points to '{', so correct it
2428 actual._dataStart += 1;
2429 actual._tokensize += 1;
2430 actual.parenthesiscount = 1;
2431 if (interval_.par[actual._dataStart] == '}') {
2432 // Determine the end if used like '{\tiny{}...}'
2433 actual._dataEnd = interval_.findclosing(actual._dataStart+1, interval_.par.length()) + 1;
2434 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2437 // Determine the end if used like '\tiny{...}'
2438 actual._dataEnd = interval_.findclosing(actual._dataStart, interval_.par.length()) + 1;
2440 // Split on this key if not at start
2441 int start = interval_.nextNotIgnored(previousStart);
2442 if (start < actual._tokenstart) {
2443 interval_.output(os, actual._tokenstart);
2444 interval_.addIntervall(start, actual._tokenstart);
2446 // discard entry if at end of actual
2447 nextKeyIdx = process(os, actual);
2451 case KeyInfo::endArguments: {
2452 // Remove trailing '{}' too
2453 actual._dataStart += 1;
2454 actual._dataEnd += 1;
2455 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2456 nextKeyIdx = getNextKey();
2459 case KeyInfo::noMain:
2461 case KeyInfo::isStandard: {
2462 if (actual.disabled) {
2464 processRegion(actual._dataStart, actual._dataStart+1);
2465 nextKeyIdx = getNextKey();
2467 // Split on this key if not at datastart of calling entry
2468 int start = interval_.nextNotIgnored(previousStart);
2469 if (start < actual._tokenstart) {
2470 interval_.output(os, actual._tokenstart);
2471 interval_.addIntervall(start, actual._tokenstart);
2473 // discard entry if at end of actual
2474 nextKeyIdx = process(os, actual);
2478 case KeyInfo::removeWithArg: {
2479 nextKeyIdx = getNextKey();
2480 // Search for end of arguments
2481 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2483 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2484 entries_[i].disabled = true;
2486 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2488 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2491 case KeyInfo::doRemove: {
2492 // Remove the key with all parameters and following spaces
2494 for (pos = actual._dataEnd+1; pos < interval_.par.length(); pos++) {
2495 if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
2498 // Remove also enclosing parentheses [] and {}
2501 while (actual._tokenstart > numpars) {
2502 if (pos+numpars >= interval_.par.size())
2504 else if (interval_.par[pos+numpars] == ']' && interval_.par[actual._tokenstart-numpars-1] == '[')
2506 else if (interval_.par[pos+numpars] == '}' && interval_.par[actual._tokenstart-numpars-1] == '{')
2512 if (interval_.par[pos+numpars] == ' ')
2516 interval_.addIntervall(actual._tokenstart-numpars, pos+numpars+spaces);
2517 nextKeyIdx = getNextKey();
2520 case KeyInfo::isList: {
2521 // Discard space before _tokenstart
2523 for (count = 0; count < actual._tokenstart; count++) {
2524 if (interval_.par[actual._tokenstart-count-1] != ' ')
2527 nextKeyIdx = getNextKey();
2528 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2530 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2532 // How else can we catch this one?
2533 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2534 entries_[i].disabled = true;
2536 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2538 else if (nextKeyIdx > 0) {
2539 // Ignore any lang entries inside data region
2540 for (int i = nextKeyIdx; i < int(entries_.size()) && entries_[i]._tokenstart < actual._dataEnd; i++) {
2541 if (entries_[i].keytype == KeyInfo::isMain)
2542 entries_[i].disabled = true;
2545 if (actual.disabled) {
2546 interval_.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2549 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2551 if (interval_.par[actual._dataEnd+1] == '[') {
2552 int posdown = interval_.findclosing(actual._dataEnd+2, interval_.par.length(), '[', ']');
2553 if ((interval_.par[actual._dataEnd+2] == '{') &&
2554 (interval_.par[posdown-1] == '}')) {
2555 interval_.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
2556 interval_.addIntervall(posdown-1, posdown+1);
2559 interval_.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
2560 interval_.addIntervall(posdown, posdown+1);
2562 int blk = interval_.nextNotIgnored(actual._dataEnd+1);
2563 if (blk > posdown) {
2564 // Discard at most 1 space after empty item
2566 for (count = 0; count < 1; count++) {
2567 if (interval_.par[blk+count] != ' ')
2571 interval_.addIntervall(blk, blk+count);
2576 case KeyInfo::isSectioning: {
2577 // Discard spaces before _tokenstart
2579 int val = actual._tokenstart;
2580 for (count = 0; count < actual._tokenstart;) {
2581 val = interval_.previousNotIgnored(val-1);
2582 if (val < 0 || interval_.par[val] != ' ')
2585 count = actual._tokenstart - val;
2588 if (actual.disabled) {
2589 removeHead(actual, count);
2590 nextKeyIdx = getNextKey();
2592 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2593 nextKeyIdx = process(os, actual);
2597 case KeyInfo::isMath: {
2598 // Same as regex, use the content unchanged
2599 nextKeyIdx = getNextKey();
2602 case KeyInfo::isRegex: {
2603 // DO NOT SPLIT ON REGEX
2605 nextKeyIdx = getNextKey();
2608 case KeyInfo::isIgnored: {
2609 // Treat like a character for now
2610 nextKeyIdx = getNextKey();
2613 case KeyInfo::isMain: {
2614 if (interval_.par.substr(actual._dataStart, 2) == "% ")
2615 interval_.addIntervall(actual._dataStart, actual._dataStart+2);
2616 if (actual._tokenstart > 0) {
2617 int prev = interval_.previousNotIgnored(actual._tokenstart - 1);
2618 if ((prev >= 0) && interval_.par[prev] == '%')
2619 interval_.addIntervall(prev, prev+1);
2621 if (actual.disabled) {
2623 interval_.langcount++;
2624 if ((interval_.par.substr(actual._dataStart, 3) == " \\[") ||
2625 (interval_.par.substr(actual._dataStart, 8) == " \\begin{")) {
2626 // Discard also the space before math-equation
2627 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2629 nextKeyIdx = getNextKey();
2630 // interval.resetOpenedP(actual._dataStart-1);
2633 if (actual._tokenstart < 26) {
2634 // for the first (and maybe dummy) language
2635 interval_.setForDefaultLang(actual);
2637 interval_.resetOpenedP(actual._dataStart-1);
2641 case KeyInfo::invalid:
2642 case KeyInfo::headRemove:
2643 // These two cases cannot happen, already handled
2646 // LYXERR(Debug::INFO, "Unhandled keytype");
2647 nextKeyIdx = getNextKey();
2654 int LatexInfo::process(ostringstream & os, KeyInfo const & actual )
2656 int end = interval_.nextNotIgnored(actual._dataEnd);
2657 int oldStart = actual._dataStart;
2658 int nextKeyIdx = getNextKey();
2660 if ((nextKeyIdx < 0) ||
2661 (entries_[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
2662 (entries_[nextKeyIdx].keytype == KeyInfo::invalid)) {
2663 if (oldStart <= end) {
2664 processRegion(oldStart, end);
2669 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
2671 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
2672 (void) dispatch(os, actual._dataStart, nextKey);
2673 end = nextKey._tokenstart;
2676 processRegion(oldStart, nextKey._tokenstart);
2677 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
2679 oldStart = nextKey._dataEnd+1;
2681 // now nextKey is either invalid or is outside of actual._dataEnd
2682 // output the remaining and discard myself
2683 if (oldStart <= end) {
2684 processRegion(oldStart, end);
2686 if (interval_.par.size() > (size_t) end && interval_.par[end] == '}') {
2688 // This is the normal case.
2689 // But if using the firstlanguage, the closing may be missing
2691 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
2693 if (actual._dataEnd < end)
2694 output_end = interval_.nextNotIgnored(actual._dataEnd);
2695 else if (interval_.par.size() > (size_t) end)
2696 output_end = interval_.nextNotIgnored(end);
2698 output_end = interval_.par.size();
2699 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
2700 interval_.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
2702 // Remove possible empty data
2703 int dstart = interval_.nextNotIgnored(actual._dataStart);
2704 while (interval_.isOpeningPar(dstart) == 1) {
2705 interval_.addIntervall(dstart, dstart+1);
2706 int dend = interval_.findclosing(dstart+1, output_end);
2707 interval_.addIntervall(dend, dend+1);
2708 dstart = interval_.nextNotIgnored(dstart+1);
2710 if (dstart < output_end)
2711 interval_.output(os, output_end);
2712 interval_.addIntervall(actual._tokenstart, end);
2716 string splitOnKnownMacros(string par, bool isPatternString)
2719 LatexInfo li(par, isPatternString);
2720 // LYXERR(Debug::INFO, "Berfore split: " << par);
2721 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
2723 DummyKey._tokensize = 0;
2724 DummyKey._dataStart = 0;
2725 DummyKey._dataEnd = par.length();
2726 DummyKey.disabled = true;
2727 int firstkeyIdx = li.getFirstKey();
2729 if (firstkeyIdx >= 0) {
2730 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
2731 DummyKey._tokenstart = firstKey._tokenstart;
2733 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
2734 // Use dummy firstKey
2735 firstKey = DummyKey;
2736 (void) li.setNextKey(firstkeyIdx);
2739 if (par.substr(firstKey._dataStart, 2) == "% ")
2740 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
2742 nextkeyIdx = li.process(os, firstKey);
2743 while (nextkeyIdx >= 0) {
2744 // Check for a possible gap between the last
2745 // entry and this one
2746 int datastart = li.nextNotIgnored(firstKey._dataStart);
2747 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
2748 if ((nextKey._tokenstart > datastart)) {
2750 firstKey._dataStart = datastart;
2751 firstKey._dataEnd = par.length();
2752 (void) li.setNextKey(nextkeyIdx);
2753 // Fake the last opened parenthesis
2754 li.setForDefaultLang(firstKey);
2755 nextkeyIdx = li.process(os, firstKey);
2758 if (nextKey.keytype != KeyInfo::isMain) {
2759 firstKey._dataStart = datastart;
2760 firstKey._dataEnd = nextKey._dataEnd+1;
2761 (void) li.setNextKey(nextkeyIdx);
2762 li.setForDefaultLang(firstKey);
2763 nextkeyIdx = li.process(os, firstKey);
2766 nextkeyIdx = li.process(os, nextKey);
2770 // Handle the remaining
2771 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
2772 firstKey._dataEnd = par.length();
2774 if ((firstKey._dataStart < firstKey._dataEnd) &&
2775 (par[firstKey._dataStart] != '}')) {
2776 li.setForDefaultLang(firstKey);
2777 (void) li.process(os, firstKey);
2781 // return string definitelly impossible to match
2782 s = "\\foreignlanguage{ignore}{ }";
2786 s = par; /* no known macros found */
2787 // LYXERR(Debug::INFO, "After split: " << s);
2792 * Try to unify the language specs in the latexified text.
2793 * Resulting modified string is set to "", if
2794 * the searched tex does not contain all the features in the search pattern
2796 static string correctlanguagesetting(string par, bool isPatternString, bool withformat)
2798 static Features regex_f;
2799 static int missed = 0;
2800 static bool regex_with_format = false;
2802 int parlen = par.length();
2804 while ((parlen > 0) && (par[parlen-1] == '\n')) {
2807 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
2808 // Happens to be there in case of description or labeling environment
2813 // Split the latex input into pieces which
2814 // can be digested by our search engine
2815 LYXERR(Debug::FIND, "input: \"" << par << "\"");
2816 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
2817 LYXERR(Debug::FIND, "After split: \"" << result << "\"");
2820 result = par.substr(0, parlen);
2821 if (isPatternString) {
2824 regex_f = identifyFeatures(result);
2825 string features = "";
2826 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2827 string a = it->first;
2828 regex_with_format = true;
2829 features += " " + a;
2830 // LYXERR(Debug::INFO, "Identified regex format:" << a);
2832 LYXERR(Debug::FIND, "Identified Features" << features);
2835 } else if (regex_with_format) {
2836 Features info = identifyFeatures(result);
2837 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2838 string a = it->first;
2839 bool b = it->second;
2840 if (b && ! info[a]) {
2842 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
2848 // LYXERR(Debug::INFO, "No regex formats");
2854 // Remove trailing closure of math, macros and environments, so to catch parts of them.
2855 static int identifyClosing(string & t)
2857 int open_braces = 0;
2859 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
2860 if (regex_replace(t, t, "(.*[^\\\\])\\$$", "$1"))
2862 if (regex_replace(t, t, "(.*[^\\\\])\\\\\\]$", "$1"))
2864 if (regex_replace(t, t, "(.*[^\\\\])\\\\end\\{[a-zA-Z_]*\\*?\\}$", "$1"))
2866 if (regex_replace(t, t, "(.*[^\\\\])\\}$", "$1")) {
2875 static int num_replaced = 0;
2876 static bool previous_single_replace = true;
2878 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt)
2879 : p_buf(&buf), p_first_buf(&buf), opt(opt)
2881 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
2882 docstring const & ds = stringifySearchBuffer(find_buf, opt);
2883 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
2884 if (opt.replace_all && previous_single_replace) {
2885 previous_single_replace = false;
2888 else if (!opt.replace_all) {
2889 num_replaced = 0; // count number of replaced strings
2890 previous_single_replace = true;
2892 // When using regexp, braces are hacked already by escape_for_regex()
2893 par_as_string = normalize(ds, !use_regexp);
2895 close_wildcards = 0;
2897 size_t lead_size = 0;
2898 // correct the language settings
2899 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat);
2900 if (opt.ignoreformat) {
2902 // if par_as_string_nolead were emty,
2903 // the following call to findAux will always *find* the string
2904 // in the checked data, and thus always using the slow
2905 // examining of the current text part.
2906 par_as_string_nolead = par_as_string;
2909 lead_size = identifyLeading(par_as_string);
2910 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
2911 lead_as_string = par_as_string.substr(0, lead_size);
2912 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
2916 open_braces = identifyClosing(par_as_string);
2917 identifyClosing(par_as_string_nolead);
2918 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2919 LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'");
2921 string lead_as_regexp;
2922 if (lead_size > 0) {
2923 // @todo No need to search for \regexp{} insets in leading material
2924 lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat);
2925 par_as_string = par_as_string_nolead;
2926 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
2927 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2929 LYXERR(Debug::FIND, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
2930 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
2931 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
2932 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2934 // Insert .* before trailing '\$' ('$' has been escaped by escape_for_regex)
2935 regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\$)\\'", "$1(.*?)$2")
2936 // Insert .* before trailing '\\\]' ('\]' has been escaped by escape_for_regex)
2937 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])( \\\\\\\\\\\\\\])\\'", "$1(.*?)$2")
2938 // Insert .* before trailing '\\end\{...}' ('\end{...}' has been escaped by escape_for_regex)
2939 || regex_replace(par_as_string, par_as_string,
2940 "(.*[^\\\\])( \\\\\\\\end\\\\\\{[a-zA-Z_]*)(\\\\\\*)?(\\\\\\})\\'", "$1(.*?)$2$3$4")
2941 // Insert .* before trailing '\}' ('}' has been escaped by escape_for_regex)
2942 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\})\\'", "$1(.*?)$2")
2946 if (!opt.ignoreformat) {
2947 // Remove extra '\}' at end if not part of \{\.\}
2948 size_t lng = par_as_string.size();
2950 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
2952 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
2961 if (lng < par_as_string.size())
2962 par_as_string = par_as_string.substr(0,lng);
2965 regex_replace(par_as_string, par_as_string, "\\\\\\.", "_xxbdotxx_");
2966 // handle '.' -> '[^]', replace later as '[^\}\{\\]'
2967 regex_replace(par_as_string, par_as_string, "\\.", "[^]");
2968 // replace '[^...]' with '[^...\}\{\\]'
2969 regex_replace(par_as_string, par_as_string, "\\[\\^([^\\\\\\]]*)\\]", "_xxbrlxx_$1\\}\\{\\\\_xxbrrxx_");
2970 regex_replace(par_as_string, par_as_string, "_xxbrlxx_", "[^");
2971 regex_replace(par_as_string, par_as_string, "_xxbrrxx_", "]");
2973 regex_replace(par_as_string, par_as_string, "_xxbdotxx_", "\\.");
2976 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2977 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2978 LYXERR(Debug::FIND, "Close .*? : " << close_wildcards);
2979 LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
2981 // If entered regexp must match at begin of searched string buffer
2982 // Kornel: Added parentheses to use $1 for size of the leading string
2986 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
2987 // Unfortunately is '\1', '\2', etc not working for strings with extra format
2988 // so the convert has no effect in that case
2989 for (int i = 7; i > 0; --i) {
2990 string orig = "\\\\" + std::to_string(i);
2991 string dest = "\\" + std::to_string(i+2);
2992 while (regex_replace(par_as_string, par_as_string, orig, dest));
2994 regexp_str = "(" + lead_as_regexp + ")()" + par_as_string;
2995 regexp2_str = "(" + lead_as_regexp + ")(.*?)" + par_as_string;
2997 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
2998 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3000 // Handle \w properly
3001 QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
3002 if (! opt.casesensitive) {
3003 popts |= QRegularExpression::CaseInsensitiveOption;
3005 regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
3006 regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
3008 if (regexp.isValid() && regexp2.isValid()) {
3009 regexIsValid = true;
3010 // Check '{', '}' pairs inside the regex
3013 for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
3014 char c = par_as_string[i];
3021 else if (c == '}') {
3028 if (balanced != 0) {
3029 regexIsValid = false;
3030 regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
3034 regexIsValid = false;
3035 if (!regexp.isValid())
3036 regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
3037 if (!regexp2.isValid())
3038 regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
3041 if (opt.casesensitive) {
3042 regexp = regex(regexp_str);
3043 regexp2 = regex(regexp2_str);
3046 regexp = regex(regexp_str, std::regex_constants::icase);
3047 regexp2 = regex(regexp2_str, std::regex_constants::icase);
3054 // Count number of characters in string
3058 // \\[a-zA-Z]+ ==> 1
3060 static int computeSize(QStringRef s, int len)
3061 #define isLyxAlpha(arg) arg.isLetter()
3063 static int computeSize(string s, int len)
3064 #define isLyxAlpha(arg) isalpha(arg)
3071 for (int i = 0; i < len; i += skip, count++) {
3072 if (s.at(i) == '\\') {
3074 if (i + 1 < len && isLyxAlpha(s.at(i+1))) {
3075 for (int j = 2; i+j < len; j++) {
3076 if (! isLyxAlpha(s.at(i+j))) {
3077 if (s.at(i+j) == ' ')
3079 else if (s.at(i+j) == '{') {
3080 if (i+j+1 < len && s.at(i+j+1) == '}')
3082 else if (i + j + 1 >= len)
3091 else if (s.at(i) == '{') {
3092 if (i + 1 < len && s.at(i+1) == '}')
3097 else if (s.at(i) == '-') {
3098 if (i+1 < len && s.at(i+1) == '-') {
3099 if (i + 2 < len && s.at(i+2) == '-')
3114 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
3119 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
3122 docstring docstr = stringifyFromForSearch(opt, cur, len);
3124 if (use_regexp || opt.casesensitive)
3125 str = normalize(docstr, true);
3127 str = normalize(lowercase(docstr), true);
3128 if (!opt.ignoreformat) {
3129 str = correctlanguagesetting(str, false, !opt.ignoreformat);
3132 mres.match_len = -1;
3135 LYXERR(Debug::FIND, "Matching against '" << lyx::to_utf8(docstr) << "'");
3136 LYXERR(Debug::FIND, "After normalization: '" << str << "'");
3139 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
3141 QString qstr = QString::fromStdString(str);
3142 QRegularExpression const *p_regexp;
3143 QRegularExpression::MatchType flags = QRegularExpression::NormalMatch;
3147 p_regexp = ®exp2;
3149 QRegularExpressionMatch match = p_regexp->match(qstr, 0, flags);
3150 if (!match.hasMatch())
3152 // Check braces on segments that matched all (.*?) subexpressions,
3153 // except the last "padding" one inserted by lyx.
3154 for (int i = 3; i < match.lastCapturedIndex(); ++i)
3155 if (!braces_match(match.captured(i), open_braces))
3158 regex const *p_regexp;
3159 regex_constants::match_flag_type flags;
3161 flags = regex_constants::match_continuous;
3164 flags = regex_constants::match_default;
3165 p_regexp = ®exp2;
3167 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
3168 if (re_it == sregex_iterator())
3170 match_results<string::const_iterator> const & m = *re_it;
3171 // Check braces on segments that matched all (.*?) subexpressions,
3172 // except the last "padding" one inserted by lyx.
3173 for (size_t i = 3; i < m.size() - 1; ++i)
3174 if (!braces_match(m[i], open_braces))
3177 // Exclude from the returned match length any length
3178 // due to close wildcards added at end of regexp
3179 // and also the length of the leading (e.g. '\emph{}')
3181 // Whole found string, including the leading: m[0].second - m[0].first
3182 // Size of the leading string: m[1].second - m[1].first
3183 int leadingsize = 0;
3186 if (match.lastCapturedIndex() > 0) {
3187 leadingsize = match.capturedEnd(1) - match.capturedStart(1);
3189 int lastidx = match.lastCapturedIndex();
3190 for (int i = 0; i <= lastidx; i++) {
3191 LYXERR(Debug::FIND, "Match " << i << " is " << match.capturedEnd(i) - match.capturedStart(i) << " long");
3193 if (close_wildcards == 0)
3194 result = match.capturedEnd(0) - match.capturedStart(0);
3196 result = match.capturedStart(lastidx + 1 - close_wildcards) - match.capturedStart(0);
3200 leadingsize = m[1].second - m[1].first;
3202 for (size_t i = 0; i < m.size(); i++) {
3203 LYXERR(Debug::FIND, "Match " << i << " is " << m[i].second - m[i].first << " long");
3205 if (close_wildcards == 0)
3206 result = m[0].second - m[0].first;
3208 result = m[m.size() - close_wildcards].first - m[0].first;
3210 if (result > leadingsize)
3211 result -= leadingsize;
3215 mres.match_prefix = match.capturedEnd(2) - match.capturedStart(2);
3216 // mres.match_len = computeSize(QStringRef(&qstr, pos+leadingsize,result), result) - mres.match_prefix;
3217 mres.match_len = match.capturedEnd(0) - match.capturedEnd(2);
3218 // mres.match2end = qstr.size() - pos - leadingsize - mres.match_prefix;
3219 mres.match2end = qstr.size() - match.capturedEnd(0);
3220 mres.pos = match.capturedStart(2);
3222 mres.match_prefix = m[2].second - m[2].first;
3223 // mres.match_len = computeSize(str.substr(pos+leadingsize,result), result) - mres.match_prefix;
3224 mres.match_len = m[0].second - m[2].second;
3225 // mres.match2end = str.size() - pos - leadingsize - mres.match_prefix;
3226 mres.match2end = str.size() - m[0].second;
3227 mres.pos = m[2].first;
3229 mres.leadsize = leadingsize;
3233 // else !use_regexp: but all code paths above return
3234 LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='"
3235 << par_as_string << "', str='" << str << "'");
3236 LYXERR(Debug::FIND, "Searching in normal mode: lead_as_string='"
3237 << lead_as_string << "', par_as_string_nolead='"
3238 << par_as_string_nolead << "'");
3241 LYXERR(Debug::FIND, "size=" << par_as_string.size()
3242 << ", substr='" << str.substr(0, par_as_string.size()) << "'");
3243 if (str.substr(0, par_as_string.size()) == par_as_string) {
3244 mres.match_len = par_as_string.size();
3245 mres.match2end = str.size();
3250 // Start the search _after_ the leading part
3251 size_t pos = str.find(par_as_string_nolead, lead_as_string.size());
3252 if (pos != string::npos) {
3253 mres.match_len = par_as_string.size();
3254 mres.match2end = str.size() - pos;
3263 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
3265 MatchResult mres = findAux(cur, len, at_begin);
3266 int res = mres.match_len;
3268 "res=" << res << ", at_begin=" << at_begin
3269 << ", matchword=" << opt.matchword
3270 << ", inTexted=" << cur.inTexted());
3271 if (res == 0 || !at_begin || !opt.matchword || !cur.inTexted())
3273 if ((len > 0) && (res < len)) {
3277 Paragraph const & par = cur.paragraph();
3278 bool ws_left = (cur.pos() > 0)
3279 ? par.isWordSeparator(cur.pos() - 1)
3285 ws_right = (cur.pos() + len < par.size())
3286 ? par.isWordSeparator(cur.pos() + len)
3290 "cur.pos()=" << cur.pos() << ", res=" << res
3291 << ", separ: " << ws_left << ", " << ws_right
3294 if (ws_left && ws_right) {
3295 // Check for word separators inside the found 'word'
3296 for (int i = 0; i < len; i++) {
3297 if (par.isWordSeparator(cur.pos() + i)) {
3309 string MatchStringAdv::normalize(docstring const & s, bool hack_braces) const
3312 t = lyx::to_utf8(s);
3313 // Remove \n at begin
3314 while (!t.empty() && t[0] == '\n')
3317 while (!t.empty() && t[t.size() - 1] == '\n')
3318 t = t.substr(0, t.size() - 1);
3320 // Handle all other '\n'
3321 while ((pos = t.find("\n")) != string::npos) {
3322 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
3324 if (isAlnumASCII(t[pos+1])) {
3325 t.replace(pos-2, 3, " ");
3328 t.replace(pos-2, 3, "");
3331 else if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
3332 // '\n' adjacent to non-alpha-numerics, discard
3333 t.replace(pos, 1, "");
3336 // Replace all other \n with spaces
3337 t.replace(pos, 1, " ");
3340 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
3341 // Kornel: Added textsl, textsf, textit, texttt and noun
3342 // + allow to seach for colored text too
3343 LYXERR(Debug::FIND, "Removing stale empty \\emph{}, \\textbf{}, \\*section{} macros from: " << t);
3344 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
3345 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3346 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
3347 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3348 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
3349 // FIXME - check what preceeds the brace
3351 if (opt.ignoreformat)
3352 while (regex_replace(t, t, "\\{", "_x_<")
3353 || regex_replace(t, t, "\\}", "_x_>"))
3354 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
3356 while (regex_replace(t, t, "\\\\\\{", "_x_<")
3357 || regex_replace(t, t, "\\\\\\}", "_x_>"))
3358 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
3365 docstring stringifyFromCursor(DocIterator const & cur, int len)
3367 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
3368 if (cur.inTexted()) {
3369 Paragraph const & par = cur.paragraph();
3370 // TODO what about searching beyond/across paragraph breaks ?
3371 // TODO Try adding a AS_STR_INSERTS as last arg
3372 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
3373 int(par.size()) : cur.pos() + len;
3374 // OutputParams runparams(&cur.buffer()->params().encoding());
3375 OutputParams runparams(encodings.fromLyXName("utf8"));
3376 runparams.nice = true;
3377 runparams.flavor = Flavor::XeTeX;
3378 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
3379 // No side effect of file copying and image conversion
3380 runparams.dryrun = true;
3381 runparams.for_search = true;
3382 LYXERR(Debug::FIND, "Stringifying with cur: "
3383 << cur << ", from pos: " << cur.pos() << ", end: " << end);
3384 return par.asString(cur.pos(), end,
3385 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
3387 } else if (cur.inMathed()) {
3388 CursorSlice cs = cur.top();
3389 MathData md = cs.cell();
3390 MathData::const_iterator it_end =
3391 (( len == -1 || cs.pos() + len > int(md.size()))
3393 : md.begin() + cs.pos() + len );
3395 for (MathData::const_iterator it = md.begin() + cs.pos();
3398 docstring s = asString(md2);
3399 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
3402 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3407 /** Computes the LaTeX export of buf starting from cur and ending len positions
3408 * after cur, if len is positive, or at the paragraph or innermost inset end
3411 docstring latexifyFromCursor(DocIterator const & cur, int len)
3413 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
3414 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
3415 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
3416 Buffer const & buf = *cur.buffer();
3418 odocstringstream ods;
3420 //OutputParams runparams(&buf.params().encoding());
3421 OutputParams runparams(encodings.fromLyXName("utf8"));
3422 runparams.nice = false;
3423 runparams.flavor = Flavor::XeTeX;
3424 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3425 // No side effect of file copying and image conversion
3426 runparams.dryrun = true;
3427 runparams.for_search = true;
3429 if (cur.inTexted()) {
3430 // @TODO what about searching beyond/across paragraph breaks ?
3431 pos_type endpos = cur.paragraph().size();
3432 if (len != -1 && endpos > cur.pos() + len)
3433 endpos = cur.pos() + len;
3434 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
3435 string(), cur.pos(), endpos);
3436 string s = lyx::to_utf8(ods.str());
3437 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
3438 return(lyx::from_utf8(s));
3439 } else if (cur.inMathed()) {
3440 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
3441 for (int s = cur.depth() - 1; s >= 0; --s) {
3442 CursorSlice const & cs = cur[s];
3443 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
3445 cs.asInsetMath()->asHullInset()->header_write(ws);
3450 CursorSlice const & cs = cur.top();
3451 MathData md = cs.cell();
3452 MathData::const_iterator it_end =
3453 ((len == -1 || cs.pos() + len > int(md.size()))
3455 : md.begin() + cs.pos() + len);
3457 for (MathData::const_iterator it = md.begin() + cs.pos();
3461 ods << asString(md2);
3462 // Retrieve the math environment type, and add '$' or '$]'
3463 // or others (\end{equation}) accordingly
3464 for (int s = cur.depth() - 1; s >= 0; --s) {
3465 CursorSlice const & cs2 = cur[s];
3466 InsetMath * inset = cs2.asInsetMath();
3467 if (inset && inset->asHullInset()) {
3469 inset->asHullInset()->footer_write(ws);
3473 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3475 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3481 /** Finalize an advanced find operation, advancing the cursor to the innermost
3482 ** position that matches, plus computing the length of the matching text to
3485 int findAdvFinalize(DocIterator & cur, MatchStringAdv const & match)
3487 // Search the foremost position that matches (avoids find of entire math
3488 // inset when match at start of it)
3490 DocIterator old_cur(cur.buffer());
3492 LYXERR(Debug::FIND, "Forwarding one step (searching for innermost match)");
3496 } while (cur && cur.depth() > d && match(cur).match_len > 0);
3498 int max_match = match(cur).match_len; /* match valid only if not searching whole words */
3499 if (max_match <= 0) return 0;
3500 LYXERR(Debug::FIND, "Ok");
3502 // Compute the match length
3504 if (cur.pos() + len > cur.lastpos())
3506 if (match.opt.matchword) {
3507 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3508 while (cur.pos() + len <= cur.lastpos() && match(cur, len).match_len <= 0) {
3510 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3512 // Length of matched text (different from len param)
3513 int old_match = match(cur, len).match_len;
3517 // Greedy behaviour while matching regexps
3518 while ((new_match = match(cur, len + 1).match_len) > old_match) {
3520 old_match = new_match;
3521 LYXERR(Debug::FIND, "verifying match with len = " << len);
3528 int maxl = cur.lastpos() - cur.pos();
3529 // Greedy behaviour while matching regexps
3530 while (maxl > minl) {
3531 int actual_match = match(cur, len).match_len;
3532 if (actual_match >= max_match) {
3533 // actual_match > max_match _can_ happen,
3534 // if the search area splits
3535 // some following word so that the regex
3536 // (e.g. 'r.*r\b' matches 'r' from the middle of the
3538 // This means, the len value is too big
3540 len = (int)((maxl + minl)/2);
3543 // (actual_match < max_match)
3545 len = (int)((maxl + minl)/2);
3549 // Search for real start of matched characters
3554 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
3555 if (cur.depth() < old_cur.depth()) {
3557 LYXERR(Debug::INFO, "cur.depth() < old_cur.depth(), this should never happen");
3560 if (cur.pos() != old_cur.pos()) {
3561 // OK, forwarded 1 pos in actual inset
3562 actual_match = match(cur, len-1).match_len;
3563 if (actual_match == max_match) {
3564 // Ha, got it! The shorter selection has the same match length
3569 // OK, the shorter selection matches less chars, revert to previous value
3575 LYXERR(Debug::INFO, "cur.pos() == old_cur.pos(), this should never happen");
3576 actual_match = match(cur, len).match_len;
3577 if (actual_match == max_match)
3586 static void displayMResult(MatchResult &mres)
3588 LYXERR0( "pos: " << mres.pos);
3589 LYXERR0( "leadsize: " << mres.leadsize);
3590 LYXERR0( "match_len: " << mres.match_len);
3591 LYXERR0( "match_prefix: " << mres.match_prefix);
3592 LYXERR0( "match2end: " << mres.match2end);
3594 #define displayMres(s) displayMResult(s);
3596 #define displayMres(s)
3600 int findForwardAdv(DocIterator & cur, MatchStringAdv const & match)
3604 while (!theApp()->longOperationCancelled() && cur) {
3605 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
3606 MatchResult mres = match(cur, -1, false);
3608 int match_len = mres.match_len;
3609 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
3610 LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
3613 if (match_len > 0) {
3614 // Try to find the begin of searched string
3616 int firstInvalid = 100000;
3617 if (mres.match_prefix + mres.pos - mres.leadsize > 0)
3618 increment = (mres.match_prefix + mres.pos - mres.leadsize)/2;
3621 LYXERR(Debug::FIND, "Set increment to " << increment);
3622 while (increment > 0) {
3623 DocIterator old_cur = cur;
3624 for (int i = 0; i < increment && cur; cur.forwardPos(), i++) {
3626 if (! cur || (cur.pit() > old_cur.pit())) {
3627 // Are we outside of the paragraph?
3628 // This can happen if moving past some UTF8-encoded chars
3633 MatchResult mres2 = match(cur, -1, false);
3635 switch (interpretMatch(mres, mres2)) {
3636 case MatchResult::newIsTooFar:
3637 // behind the expected match
3638 firstInvalid = increment;
3642 case MatchResult::newIsBetter:
3645 firstInvalid -= increment;
3646 if (increment > firstInvalid/2)
3647 increment = firstInvalid/2;
3650 // Handle not like MatchResult::newIsTooFar
3651 LYXERR0( "Something is wrong: Increment = " << increment << " match_prefix = " << mres.match_prefix);
3653 increment = firstInvalid -1;
3659 // LYXERR0("Leaving first loop");
3660 int match_len_zero_count = 0;
3662 for (int i = 0; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) {
3664 mres3 = match(cur, -1, false);
3666 int remaining_len = mres3.match_len;
3667 if (remaining_len <= 0) {
3668 // Apparently the searched string is not in the remaining part
3675 LYXERR(Debug::FIND, "Advancing cur: " << cur);
3676 mres3 = match(cur, 1);
3678 int match_len3 = mres3.match_len;
3683 int match_len2 = mres3.match_len;
3684 LYXERR(Debug::FIND, "match_len2: " << match_len2);
3685 if (match_len2 > 0) {
3686 // Sometimes in finalize we understand it wasn't a match
3687 // and we need to continue the outest loop
3688 LYXERR(Debug::FIND, "Finalizing");
3689 int len = findAdvFinalize(cur, match);
3695 match_len_zero_count = 0;
3696 else if (match_len2 == 0)
3697 match_len_zero_count++;
3698 if (match_len2 < 0) {
3699 if (++match_len_zero_count > 3) {
3700 LYXERR(Debug::FIND, "match_len2_zero_count: " << match_len_zero_count << ", match_len was " << match_len);
3708 if (match_len >= 0 && cur.pit() < cur.lastpit()) {
3709 LYXERR(Debug::FIND, "Advancing par: cur=" << cur);
3712 // This should exit nested insets, if any, or otherwise undefine the currsor.
3713 cur.pos() = cur.lastpos();
3714 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
3722 /// Find the most backward consecutive match within same paragraph while searching backwards.
3723 int findMostBackwards(DocIterator & cur, MatchStringAdv const & match)
3725 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3726 DocIterator tmp_cur = cur;
3727 int len = findAdvFinalize(tmp_cur, match);
3728 Inset & inset = cur.inset();
3729 for (; cur != cur_begin; cur.backwardPos()) {
3730 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
3731 DocIterator new_cur = cur;
3732 new_cur.backwardPos();
3733 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
3735 int new_len = findAdvFinalize(new_cur, match);
3740 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
3746 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
3750 // Backup of original position
3751 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3752 if (cur == cur_begin)
3755 DocIterator cur_orig(cur);
3756 bool pit_changed = false;
3759 bool found_match = (match(cur, -1, false).match_len > 0);
3763 cur.pos() = cur.lastpos();
3765 cur.pos() = cur_orig.pos();
3766 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
3767 DocIterator cur_prev_iter;
3769 found_match = (match(cur).match_len > 0);
3770 LYXERR(Debug::FIND, "findBackAdv3: found_match="
3771 << found_match << ", cur: " << cur);
3773 return findMostBackwards(cur, match);
3775 // Stop if begin of document reached
3776 if (cur == cur_begin)
3778 cur_prev_iter = cur;
3782 if (cur == cur_begin)
3789 } while (!theApp()->longOperationCancelled());
3797 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
3798 DocIterator const & cur, int len)
3800 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
3802 if (!opt.ignoreformat)
3803 return latexifyFromCursor(cur, len);
3805 return stringifyFromCursor(cur, len);
3809 FindAndReplaceOptions::FindAndReplaceOptions(
3810 docstring const & _find_buf_name, bool _casesensitive,
3811 bool _matchword, bool _forward, bool _expandmacros, bool _ignoreformat,
3812 docstring const & _repl_buf_name, bool _keep_case,
3813 SearchScope _scope, SearchRestriction _restr, bool _replace_all)
3814 : find_buf_name(_find_buf_name), casesensitive(_casesensitive), matchword(_matchword),
3815 forward(_forward), expandmacros(_expandmacros), ignoreformat(_ignoreformat),
3816 repl_buf_name(_repl_buf_name), keep_case(_keep_case), scope(_scope), restr(_restr), replace_all(_replace_all)
3824 /** Check if 'len' letters following cursor are all non-lowercase */
3825 static bool allNonLowercase(Cursor const & cur, int len)
3827 pos_type beg_pos = cur.selectionBegin().pos();
3828 pos_type end_pos = cur.selectionBegin().pos() + len;
3829 if (len > cur.lastpos() + 1 - beg_pos) {
3830 LYXERR(Debug::FIND, "This should not happen, more debug needed");
3831 len = cur.lastpos() + 1 - beg_pos;
3832 end_pos = beg_pos + len;
3834 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
3835 if (isLowerCase(cur.paragraph().getChar(pos)))
3841 /** Check if first letter is upper case and second one is lower case */
3842 static bool firstUppercase(Cursor const & cur)
3845 pos_type pos = cur.selectionBegin().pos();
3846 if (pos >= cur.lastpos() - 1) {
3847 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
3850 ch1 = cur.paragraph().getChar(pos);
3851 ch2 = cur.paragraph().getChar(pos + 1);
3852 bool result = isUpperCase(ch1) && isLowerCase(ch2);
3853 LYXERR(Debug::FIND, "firstUppercase(): "
3854 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
3855 << ch2 << "(" << char(ch2) << ")"
3856 << ", result=" << result << ", cur=" << cur);
3861 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
3863 ** \fixme What to do with possible further paragraphs in replace buffer ?
3865 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
3867 ParagraphList::iterator pit = buffer.paragraphs().begin();
3868 LASSERT(!pit->empty(), /**/);
3869 pos_type right = pos_type(1);
3870 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
3871 right = pit->size();
3872 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
3878 static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
3880 Cursor & cur = bv->cursor();
3881 if (opt.repl_buf_name.empty()
3882 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
3883 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3886 DocIterator sel_beg = cur.selectionBegin();
3887 DocIterator sel_end = cur.selectionEnd();
3888 if (&sel_beg.inset() != &sel_end.inset()
3889 || sel_beg.pit() != sel_end.pit()
3890 || sel_beg.idx() != sel_end.idx())
3892 int sel_len = sel_end.pos() - sel_beg.pos();
3893 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
3894 << ", sel_len: " << sel_len << endl);
3897 LASSERT(sel_len > 0, return 0);
3899 if (!matchAdv(sel_beg, sel_len).match_len)
3902 // Build a copy of the replace buffer, adapted to the KeepCase option
3903 Buffer const & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
3905 repl_buffer_orig.write(oss);
3906 string lyx = oss.str();
3907 Buffer repl_buffer("", false);
3908 repl_buffer.setUnnamed(true);
3909 LASSERT(repl_buffer.readString(lyx), return 0);
3910 if (opt.keep_case && sel_len >= 2) {
3911 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
3912 if (cur.inTexted()) {
3913 if (firstUppercase(cur))
3914 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
3915 else if (allNonLowercase(cur, sel_len))
3916 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
3919 cap::cutSelection(cur, false);
3920 if (cur.inTexted()) {
3921 repl_buffer.changeLanguage(
3922 repl_buffer.language(),
3923 cur.getFont().language());
3924 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
3925 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
3926 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
3927 repl_buffer.params().documentClassPtr(),
3928 bv->buffer().errorList("Paste"));
3929 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
3930 sel_len = repl_buffer.paragraphs().begin()->size();
3931 } else if (cur.inMathed()) {
3932 odocstringstream ods;
3934 // OutputParams runparams(&repl_buffer.params().encoding());
3935 OutputParams runparams(encodings.fromLyXName("utf8"));
3936 runparams.nice = false;
3937 runparams.flavor = Flavor::XeTeX;
3938 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3939 runparams.dryrun = true;
3940 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
3941 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
3942 docstring repl_latex = ods.str();
3943 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
3945 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
3946 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
3947 repl_latex = from_utf8(s);
3948 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
3949 MathData ar(cur.buffer());
3950 asArray(repl_latex, ar, Parse::NORMAL);
3952 sel_len = ar.size();
3953 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3955 if (cur.pos() >= sel_len)
3956 cur.pos() -= sel_len;
3959 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3960 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
3961 bv->processUpdateFlags(Update::Force);
3966 /// Perform a FindAdv operation.
3967 bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt)
3972 // e.g., when invoking word-findadv from mini-buffer wither with
3973 // wrong options syntax or before ever opening advanced F&R pane
3974 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3978 MatchStringAdv matchAdv(bv->buffer(), opt);
3980 if (!matchAdv.regexIsValid) {
3981 bv->message(lyx::from_utf8(matchAdv.regexError));
3985 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
3987 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
3988 num_replaced += findAdvReplace(bv, opt, matchAdv);
3991 match_len = findForwardAdv(cur, matchAdv);
3993 match_len = findBackwardsAdv(cur, matchAdv);
3995 // This may only be raised by lyx::regex()
3996 bv->message(_("Invalid regular expression!"));
4000 if (match_len == 0) {
4001 if (num_replaced > 0) {
4002 switch (num_replaced)
4005 bv->message(_("One match has been replaced."));
4008 bv->message(_("Two matches have been replaced."));
4011 bv->message(bformat(_("%1$d matches have been replaced."), num_replaced));
4017 bv->message(_("Match not found."));
4022 if (num_replaced > 0)
4023 bv->message(_("Match has been replaced."));
4025 bv->message(_("Match found."));
4027 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << match_len);
4028 bv->putSelectionAt(cur, match_len, !opt.forward);
4034 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
4036 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
4037 << opt.casesensitive << ' '
4038 << opt.matchword << ' '
4039 << opt.forward << ' '
4040 << opt.expandmacros << ' '
4041 << opt.ignoreformat << ' '
4042 << opt.replace_all << ' '
4043 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
4044 << opt.keep_case << ' '
4045 << int(opt.scope) << ' '
4048 LYXERR(Debug::FIND, "built: " << os.str());
4054 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
4056 LYXERR(Debug::FIND, "parsing");
4060 while (line != "EOSS") {
4064 if (is.eof()) // Tolerate malformed request
4068 LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
4069 opt.find_buf_name = from_utf8(s);
4070 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.replace_all;
4071 is.get(); // Waste space before replace string
4074 while (line != "EOSS") {
4078 if (is.eof()) // Tolerate malformed request
4082 LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
4083 opt.repl_buf_name = from_utf8(s);
4084 is >> opt.keep_case;
4087 opt.scope = FindAndReplaceOptions::SearchScope(i);
4089 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
4091 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
4092 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
4093 << opt.scope << ' ' << opt.restr);