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"
36 #include "frontends/Application.h"
37 #include "frontends/alert.h"
39 #include "mathed/InsetMath.h"
40 #include "mathed/InsetMathHull.h"
41 #include "mathed/MathData.h"
42 #include "mathed/MathStream.h"
43 #include "mathed/MathSupport.h"
45 #include "support/debug.h"
46 #include "support/docstream.h"
47 #include "support/FileName.h"
48 #include "support/gettext.h"
49 #include "support/lassert.h"
50 #include "support/lstrings.h"
51 #include "support/textutils.h"
53 #include <unordered_map>
56 //#define ResultsDebug
57 #define USE_QT_FOR_SEARCH
58 #if defined(USE_QT_FOR_SEARCH)
59 #include <QtCore> // sets QT_VERSION
60 #if (QT_VERSION >= 0x050000)
61 #include <QRegularExpression>
71 using namespace lyx::support;
75 typedef unordered_map<string, string> AccentsMap;
76 typedef unordered_map<string,string>::const_iterator AccentsIterator;
77 static AccentsMap accents = unordered_map<string, string>();
79 // Helper class for deciding what should be ignored
83 IgnoreFormats() = default;
85 bool getFamily() const { return ignoreFamily_; }
87 bool getSeries() const { return ignoreSeries_; }
89 bool getShape() const { return ignoreShape_; }
91 bool getSize() const { return ignoreSize_; }
93 bool getUnderline() const { return ignoreUnderline_; }
95 bool getMarkUp() const { return ignoreMarkUp_; }
97 bool getStrikeOut() const { return ignoreStrikeOut_; }
99 bool getSectioning() const { return ignoreSectioning_; }
101 bool getFrontMatter() const { return ignoreFrontMatter_; }
103 bool getColor() const { return ignoreColor_; }
105 bool getLanguage() const { return ignoreLanguage_; }
107 bool getDeleted() const { return ignoreDeleted_; }
109 void setIgnoreDeleted(bool value);
111 bool getNonContent() const { return searchNonContent_; }
113 void setIgnoreFormat(string const & type, bool value, bool fromUser = true);
117 bool ignoreFamily_ = false;
119 bool ignoreSeries_ = false;
121 bool ignoreShape_ = false;
123 bool ignoreSize_ = true;
125 bool ignoreUnderline_ = false;
127 bool ignoreMarkUp_ = false;
129 bool ignoreStrikeOut_ = false;
131 bool ignoreSectioning_ = false;
133 bool ignoreFrontMatter_ = false;
135 bool ignoreColor_ = false;
137 bool ignoreLanguage_ = false;
138 bool userSelectedIgnoreLanguage_ = false;
140 bool ignoreDeleted_ = true;
142 bool searchNonContent_ = true;
145 void IgnoreFormats::setIgnoreFormat(string const & type, bool value, bool fromUser)
147 if (type == "color") {
148 ignoreColor_ = value;
150 else if (type == "language") {
152 userSelectedIgnoreLanguage_ = value;
153 ignoreLanguage_ = value;
156 ignoreLanguage_ = (value || userSelectedIgnoreLanguage_);
158 else if (type == "sectioning") {
159 ignoreSectioning_ = value;
160 ignoreFrontMatter_ = value;
162 else if (type == "font") {
163 ignoreSeries_ = value;
164 ignoreShape_ = value;
165 ignoreFamily_ = value;
167 else if (type == "series") {
168 ignoreSeries_ = value;
170 else if (type == "shape") {
171 ignoreShape_ = value;
173 else if (type == "size") {
176 else if (type == "family") {
177 ignoreFamily_ = value;
179 else if (type == "markup") {
180 ignoreMarkUp_ = value;
182 else if (type == "underline") {
183 ignoreUnderline_ = value;
185 else if (type == "strike") {
186 ignoreStrikeOut_ = value;
188 else if (type == "deleted") {
189 ignoreDeleted_ = value;
191 else if (type == "non-output-content") {
192 searchNonContent_ = !value;
196 // The global variable that can be changed from outside
197 IgnoreFormats ignoreFormats;
200 void setIgnoreFormat(string const & type, bool value, bool fromUser)
202 ignoreFormats.setIgnoreFormat(type, value, fromUser);
208 bool parse_bool(docstring & howto, bool const defvalue = false)
213 howto = split(howto, var, ' ');
221 MatchString(docstring const & s, bool cs, bool mw)
222 : str(s), case_sens(cs), whole_words(mw)
225 // returns true if the specified string is at the specified position
226 // del specifies whether deleted strings in ct mode will be considered
227 int operator()(Paragraph const & par, pos_type pos, bool del = true) const
229 return par.find(str, case_sens, whole_words, pos, del);
237 // match whole words only
242 int findForward(DocIterator & cur, DocIterator const endcur,
243 MatchString const & match,
244 bool find_del = true, bool onlysel = false)
246 for (; cur; cur.forwardChar()) {
247 if (onlysel && endcur.pit() == cur.pit()
248 && endcur.idx() == cur.idx() && endcur.pos() < cur.pos())
250 if (cur.inTexted()) {
251 int len = match(cur.paragraph(), cur.pos(), find_del);
260 int findBackwards(DocIterator & cur, DocIterator const endcur,
261 MatchString const & match,
262 bool find_del = true, bool onlysel = false)
266 if (onlysel && endcur.pit() == cur.pit()
267 && endcur.idx() == cur.idx() && endcur.pos() > cur.pos())
269 if (cur.inTexted()) {
270 int len = match(cur.paragraph(), cur.pos(), find_del);
279 bool searchAllowed(docstring const & str)
282 frontend::Alert::error(_("Search error"), _("Search string is empty"));
291 bool findOne(BufferView * bv, docstring const & searchstr,
292 bool case_sens, bool whole, bool forward,
293 bool find_del, bool check_wrap, bool const auto_wrap,
294 bool instant, bool onlysel)
296 // Clean up previous selections with empty searchstr on instant
297 if (searchstr.empty() && instant) {
298 if (bv->cursor().selection()) {
299 bv->setCursor(bv->cursor().selectionBegin());
300 bv->clearSelection();
305 if (!searchAllowed(searchstr))
308 DocIterator const endcur = forward ? bv->cursor().selectionEnd() : bv->cursor().selectionBegin();
310 if (onlysel && bv->cursor().selection()) {
311 docstring const matchstring = bv->cursor().selectionAsString(false);
312 docstring const lcmatchsting = support::lowercase(matchstring);
313 if (matchstring == searchstr || (!case_sens && lcmatchsting == lowercase(searchstr))) {
314 docstring q = _("The search string matches the selection, and search is limited to selection.\n"
315 "Continue search outside?");
316 int search_answer = frontend::Alert::prompt(_("Search outside selection?"),
317 q, 0, 1, _("&Yes"), _("&No"));
318 if (search_answer == 0) {
319 bv->clearSelection();
320 if (findOne(bv, searchstr, case_sens, whole, forward,
321 find_del, check_wrap, auto_wrap, false, false))
328 DocIterator cur = forward
329 ? ((instant || onlysel) ? bv->cursor().selectionBegin() : bv->cursor().selectionEnd())
330 : ((instant || onlysel) ? bv->cursor().selectionEnd() : bv->cursor().selectionBegin());
332 MatchString const match(searchstr, case_sens, whole);
334 int match_len = forward
335 ? findForward(cur, endcur, match, find_del, onlysel)
336 : findBackwards(cur, endcur, match, find_del, onlysel);
339 bv->putSelectionAt(cur, match_len, !forward);
341 docstring q = _("The search string was not found within the selection.\n"
342 "Continue search outside?");
343 int search_answer = frontend::Alert::prompt(_("Search outside selection?"),
344 q, 0, 1, _("&Yes"), _("&No"));
345 if (search_answer == 0) {
346 bv->clearSelection();
347 if (findOne(bv, searchstr, case_sens, whole, forward,
348 find_del, check_wrap, auto_wrap, false, false))
353 else if (check_wrap) {
354 DocIterator cur_orig(bv->cursor());
355 bool wrap = auto_wrap;
359 q = _("End of file reached while searching forward.\n"
360 "Continue searching from the beginning?");
362 q = _("Beginning of file reached while searching backward.\n"
363 "Continue searching from the end?");
364 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
365 q, 0, 1, _("&Yes"), _("&No"));
366 wrap = wrap_answer == 0;
370 bv->cursor().clear();
371 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
373 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
374 bv->cursor().backwardPos();
377 docstring const msg = forward
378 ? _("Search reached end of document, continuing from beginning.")
379 : _("Search reached beginning of document, continuing from end.");
382 bv->clearSelection();
383 if (findOne(bv, searchstr, case_sens, whole, forward,
384 find_del, false, false, false, false))
387 bv->cursor().setCursor(cur_orig);
391 return match_len > 0;
397 int replaceAll(BufferView * bv,
398 docstring const & searchstr, docstring const & replacestr,
399 bool case_sens, bool whole, bool onlysel)
401 Buffer & buf = bv->buffer();
403 if (!searchAllowed(searchstr) || buf.isReadonly())
406 DocIterator startcur = bv->cursor().selectionBegin();
407 DocIterator endcur = bv->cursor().selectionEnd();
408 bool const had_selection = bv->cursor().selection();
410 MatchString const match(searchstr, case_sens, whole);
413 int const rsize = replacestr.size();
414 int const ssize = searchstr.size();
417 cur.setCursor(doc_iterator_begin(&buf));
418 int match_len = findForward(cur, endcur, match, false, onlysel);
419 while (match_len > 0) {
420 // Backup current cursor position and font.
421 pos_type const pos = cur.pos();
422 Font const font = cur.paragraph().getFontSettings(buf.params(), pos);
424 int ct_deleted_text = ssize -
425 cur.paragraph().eraseChars(pos, pos + match_len,
426 buf.params().track_changes);
427 cur.paragraph().insert(pos, replacestr, font,
428 Change(buf.params().track_changes
430 : Change::UNCHANGED));
431 for (int i = 0; i < rsize + ct_deleted_text
432 && cur.pos() < cur.lastpos(); ++i)
434 if (onlysel && cur.pit() == endcur.pit() && cur.idx() == endcur.idx()) {
435 // Adjust end of selection for replace-all in selection
437 int const offset = rsize - ssize;
438 for (int i = 0; i < offset + ct_deleted_text
439 && endcur.pos() < endcur.lastpos(); ++i)
442 int const offset = ssize - rsize;
443 for (int i = 0; i < offset && endcur.pos() > 0; ++i)
444 endcur.backwardPos();
445 for (int i = 0; i < ct_deleted_text
446 && endcur.pos() < endcur.lastpos(); ++i)
451 match_len = findForward(cur, endcur, match, false, onlysel);
454 bv->putSelectionAt(doc_iterator_begin(&buf), 0, false);
456 startcur.fixIfBroken();
457 bv->setCursor(startcur);
459 // Reset selection, accounting for changes in selection
461 endcur.fixIfBroken();
462 bv->cursor().resetAnchor();
463 bv->setCursorSelectionTo(endcur);
470 // the idea here is that we are going to replace the string that
471 // is selected IF it is the search string.
472 // if there is a selection, but it is not the search string, then
473 // we basically ignore it. (FIXME We ought to replace only within
475 // if there is no selection, then:
476 // (i) if some search string has been provided, then we find it.
477 // (think of how the dialog works when you hit "replace" the
479 // (ii) if no search string has been provided, then we treat the
480 // word the cursor is in as the search string. (why? i have no
481 // idea.) but this only works in text?
483 // returns the number of replacements made (one, if any) and
484 // whether anything at all was done.
485 pair<bool, int> replaceOne(BufferView * bv, docstring searchstr,
486 docstring const & replacestr, bool case_sens,
487 bool whole, bool forward, bool findnext, bool wrap,
490 Cursor & cur = bv->cursor();
491 if (!cur.selection() || onlysel) {
492 // no selection, non-empty search string: find it
493 if (!searchstr.empty()) {
494 bool const found = findOne(bv, searchstr, case_sens, whole,
495 forward, true, findnext, wrap, false, onlysel);
496 return make_pair(found, 0);
498 // empty search string
501 return make_pair(false, 0);
502 // select current word and treat it as the search string.
503 // This causes a minor bug as undo will restore this selection,
504 // which the user did not create (#8986).
505 cur.innerText()->selectWord(cur, WHOLE_WORD);
506 searchstr = cur.selectionAsString(false, true);
509 // if we still don't have a search string, report the error
511 if (!searchAllowed(searchstr))
512 return make_pair(false, 0);
514 bool have_selection = cur.selection();
515 docstring const selected = cur.selectionAsString(false, true);
518 ? searchstr == selected
519 : compare_no_case(searchstr, selected) == 0;
521 // no selection or current selection is not search word:
522 // just find the search word
523 if (!have_selection || !match) {
524 bool const found = findOne(bv, searchstr, case_sens, whole, forward,
525 true, findnext, wrap, false, onlysel);
526 return make_pair(found, 0);
529 // we're now actually ready to replace. if the buffer is
530 // read-only, we can't, though.
531 if (bv->buffer().isReadonly())
532 return make_pair(false, 0);
534 cap::replaceSelectionWithString(cur, replacestr);
536 cur.pos() += replacestr.length();
537 LASSERT(cur.pos() <= cur.lastpos(),
538 cur.pos() = cur.lastpos());
541 findOne(bv, searchstr, case_sens, whole,
542 forward, false, findnext, wrap, false, onlysel);
544 return make_pair(true, 1);
550 docstring const find2string(docstring const & search,
551 bool casesensitive, bool matchword,
552 bool forward, bool wrap, bool instant,
557 << int(casesensitive) << ' '
558 << int(matchword) << ' '
559 << int(forward) << ' '
561 << int(instant) << ' '
567 docstring const replace2string(docstring const & replace,
568 docstring const & search,
569 bool casesensitive, bool matchword,
570 bool all, bool forward, bool findnext,
571 bool wrap, bool onlysel)
574 ss << replace << '\n'
576 << int(casesensitive) << ' '
577 << int(matchword) << ' '
579 << int(forward) << ' '
580 << int(findnext) << ' '
587 docstring const string2find(docstring const & argument,
595 // data is of the form
597 // <casesensitive> <matchword> <forward> <wrap> <onlysel>"
599 docstring howto = split(argument, search, '\n');
601 casesensitive = parse_bool(howto);
602 matchword = parse_bool(howto);
603 forward = parse_bool(howto, true);
604 wrap = parse_bool(howto);
605 instant = parse_bool(howto);
606 onlysel = parse_bool(howto);
612 bool lyxfind(BufferView * bv, FuncRequest const & ev)
614 if (!bv || ev.action() != LFUN_WORD_FIND)
617 //lyxerr << "find called, cmd: " << ev << endl;
625 docstring search = string2find(ev.argument(), casesensitive,
626 matchword, forward, wrap, instant, onlysel);
628 return findOne(bv, search, casesensitive, matchword, forward,
629 false, true, wrap, instant, onlysel);
633 bool lyxreplace(BufferView * bv, FuncRequest const & ev)
635 if (!bv || ev.action() != LFUN_WORD_REPLACE)
638 // data is of the form
641 // <casesensitive> <matchword> <all> <forward> <findnext> <wrap> <onlysel>"
644 docstring howto = split(ev.argument(), rplc, '\n');
645 howto = split(howto, search, '\n');
647 bool casesensitive = parse_bool(howto);
648 bool matchword = parse_bool(howto);
649 bool all = parse_bool(howto);
650 bool forward = parse_bool(howto, true);
651 bool findnext = parse_bool(howto, true);
652 bool wrap = parse_bool(howto);
653 bool onlysel = parse_bool(howto);
655 if (!bv->cursor().selection())
656 // only selection only makes sense with selection
661 int replace_count = 0;
663 replace_count = replaceAll(bv, search, rplc, casesensitive,
665 update = replace_count > 0;
668 replaceOne(bv, search, rplc, casesensitive, matchword,
669 forward, findnext, wrap, onlysel);
671 replace_count = rv.second;
674 Buffer const & buf = bv->buffer();
676 // emit message signal.
678 buf.message(_("String not found in selection."));
680 buf.message(_("String not found."));
682 if (replace_count == 0) {
683 buf.message(_("String found."));
684 } else if (replace_count == 1) {
685 buf.message(_("String has been replaced."));
687 docstring const str = onlysel
688 ? bformat(_("%1$d strings have been replaced in the selection."), replace_count)
689 : bformat(_("%1$d strings have been replaced."), replace_count);
697 bool findNextChange(BufferView * bv, Cursor & cur, bool const check_wrap)
699 for (; cur; cur.forwardPos())
700 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
704 DocIterator cur_orig(bv->cursor());
705 docstring q = _("End of file reached while searching forward.\n"
706 "Continue searching from the beginning?");
707 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
708 q, 0, 1, _("&Yes"), _("&No"));
709 if (wrap_answer == 0) {
710 bv->cursor().clear();
711 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
712 bv->clearSelection();
713 cur.setCursor(bv->cursor().selectionBegin());
714 if (findNextChange(bv, cur, false))
717 bv->cursor().setCursor(cur_orig);
724 bool findPreviousChange(BufferView * bv, Cursor & cur, bool const check_wrap)
726 for (cur.backwardPos(); cur; cur.backwardPos()) {
727 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
732 DocIterator cur_orig(bv->cursor());
733 docstring q = _("Beginning of file reached while searching backward.\n"
734 "Continue searching from the end?");
735 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
736 q, 0, 1, _("&Yes"), _("&No"));
737 if (wrap_answer == 0) {
738 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
739 bv->cursor().backwardPos();
740 bv->clearSelection();
741 cur.setCursor(bv->cursor().selectionBegin());
742 if (findPreviousChange(bv, cur, false))
745 bv->cursor().setCursor(cur_orig);
752 bool selectChange(Cursor & cur, bool forward)
754 if (!cur.inTexted() || !cur.paragraph().isChanged(cur.pos()))
756 Change ch = cur.paragraph().lookupChange(cur.pos());
758 CursorSlice tip1 = cur.top();
759 for (; tip1.pit() < tip1.lastpit() || tip1.pos() < tip1.lastpos(); tip1.forwardPos()) {
760 Change ch2 = tip1.paragraph().lookupChange(tip1.pos());
761 if (!ch2.isSimilarTo(ch))
764 CursorSlice tip2 = cur.top();
765 for (; tip2.pit() > 0 || tip2.pos() > 0;) {
767 Change ch2 = tip2.paragraph().lookupChange(tip2.pos());
768 if (!ch2.isSimilarTo(ch)) {
769 // take a step forward to correctly set the selection
777 cur.bv().mouseSetCursor(cur, false);
779 cur.bv().mouseSetCursor(cur, true);
787 bool findChange(BufferView * bv, bool forward)
790 cur.setCursor(forward ? bv->cursor().selectionEnd()
791 : bv->cursor().selectionBegin());
792 forward ? findNextChange(bv, cur, true) : findPreviousChange(bv, cur, true);
793 return selectChange(cur, forward);
798 bool findNextChange(BufferView * bv)
800 return findChange(bv, true);
804 bool findPreviousChange(BufferView * bv)
806 return findChange(bv, false);
813 typedef vector<pair<string, string> > Escapes;
815 static string getRegexSpaceCount(int count)
819 return "\\s{" + std::to_string(count) + "}";
826 string string2regex(string in)
828 static std::regex specialChars { R"([-[\]{}()*+?.,\^$|#\$\\])" };
829 string tempx = std::regex_replace(in, specialChars, R"(\$&)" );
830 // Special handling for ' '
833 for (unsigned i = 0; i < tempx.size(); i++) {
834 if (tempx[i] == ' ' || tempx[i] == '~' ) {
838 else if ((tempx[i] == '\302' && tempx[i+1] == '\240')
839 || (tempx[i] == '\342' && tempx[i+1] == '\200')) {
847 temp += getRegexSpaceCount(blanks);
854 temp += getRegexSpaceCount(blanks);
861 while (fl_pos < temp.size()) {
862 fl_pos = temp.find("\\\\foreignlanguage", lastpos + offset);
863 if (fl_pos == string::npos)
866 temp2 += temp.substr(lastpos, fl_pos - lastpos);
872 if (lastpos < temp.size()) {
873 temp2 += temp.substr(lastpos, temp.size() - lastpos);
878 static void buildAccentsMap();
880 string correctRegex(string t, bool withformat)
882 /* Convert \backslash => \
883 * and \{, \}, \[, \] => {, }, [, ]
886 static std::regex wordre("(\\\\)*(\\\\(( |[A-Za-z]+|[\\{\\}%])( |\\{\\})?|[\\[\\]\\{\\}]))");
887 static std::regex protectedSpace { R"(~)" };
890 bool backslashed = false;
894 //LYXERR0("correctRegex input '" << t << "'");
896 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
899 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
903 if (sub.str(4) == "backslash") {
906 // transforms '\backslash \{' into '\{'
907 string next = t.substr(sub.position(2) + sub.str(2).length(), 2);
908 if ((next == "\\{") || (next == "\\}") || (next == "\\ ")) {
912 else if (withformat && next[0] == '$') {
913 replace = accents["lyxdollar"];
914 skip = 1; // Skip following '$'
918 else if (sub.str(4) == "mathcircumflex")
920 else if (backslashed) {
923 if (sub.str(3) == "{")
924 replace = accents["braceleft"];
925 else if (sub.str(3) == "}")
926 replace = accents["braceright"];
927 else if (sub.str(3) == " ")
930 // else part should not exist
935 if (sub.str(3) == "{")
937 else if (sub.str(3) == "}")
939 else if (sub.str(3) == " ")
942 // else part should not exist
947 else if (sub.str(4) == "{") // transforms '\{' into '{'
949 else if (sub.str(4) == "}")
951 else if (sub.str(4) == "%")
953 else if (sub.str(4) == " ")
956 AccentsIterator it_ac = accents.find(sub.str(4));
957 if (it_ac == accents.end()) {
958 replace = sub.str(2);
961 replace = it_ac->second;
965 if (lastpos < (size_t) sub.position(2))
966 s += std::regex_replace(t.substr(lastpos, sub.position(2) - lastpos), protectedSpace, R"( )");
968 lastpos = sub.position(2) + sub.length(2) + skip;
972 s = std::regex_replace(t, protectedSpace, R"( )");
973 else if (lastpos < t.length())
974 s += std::regex_replace(t.substr(lastpos, t.length() - lastpos), protectedSpace, R"( )");
975 // Handle quotes in regex
976 // substitute all '„', '“', '»', '«' with '"'
977 // and all '‚', '‘', '›', '‹' with "\'"
978 static std::regex plainquotes { R"(„|“|»|«)" };
979 static std::regex innerquotes { R"(‚|‘|›|‹)" };
980 t = std::regex_replace(s, plainquotes, R"(")");
981 s = std::regex_replace(t, innerquotes, R"(')");
982 //LYXERR0("correctRegex output '" << s << "'");
986 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
987 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
988 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
989 string escape_for_regex(string s, bool withformat)
993 while (lastpos < s.size()) {
994 size_t regex_pos = s.find("\\regexp{", lastpos);
995 if (regex_pos == string::npos) {
996 regex_pos = s.size();
998 if (regex_pos > lastpos) {
999 result += string2regex(s.substr(lastpos, regex_pos-lastpos));
1000 lastpos = regex_pos;
1001 if (lastpos == s.size())
1004 size_t end_pos = s.find("\\endregexp", regex_pos + 8);
1005 result += correctRegex(s.substr(regex_pos + 8, end_pos -(regex_pos + 8)), withformat);
1006 lastpos = end_pos + 13;
1012 /// Wrapper for lyx::regex_replace with simpler interface
1013 bool regex_replace(string const & s, string & t, string const & searchstr,
1014 string const & replacestr)
1016 regex e(searchstr, regex_constants::ECMAScript);
1018 ostream_iterator<char, char> it(oss);
1019 regex_replace(it, s.begin(), s.end(), e, replacestr);
1020 // tolerate t and s be references to the same variable
1021 bool rv = (s != oss.str());
1040 vector <string> result = vector <string>();
1041 MatchResult(int len = 0): match_len(len),match_prefix(0),match2end(0), pos(0),leadsize(0),pos_len(-1),searched_size(0) {}
1044 static MatchResult::range interpretMatch(MatchResult &oldres, MatchResult &newres)
1046 if (newres.match2end < oldres.match2end)
1047 return MatchResult::newIsTooFar;
1048 if (newres.match_len < oldres.match_len)
1049 return MatchResult::newIsTooFar;
1051 if (newres.match_len == oldres.match_len) {
1052 if (newres.match2end == oldres.match2end)
1053 return MatchResult::newIsBetter;
1055 return MatchResult::newIsInvalid;
1058 /** The class performing a match between a position in the document and the FindAdvOptions.
1061 class MatchStringAdv {
1063 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt);
1065 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
1066 ** constructor as opt.search, under the opt.* options settings.
1069 ** If set, then match is searched only against beginning of text starting at cur.
1070 ** If unset, then match is searched anywhere in text starting at cur.
1073 ** The length of the matching text, or zero if no match was found.
1075 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
1083 lyx::Buffer * p_buf;
1084 /// first buffer on which search was started
1085 lyx::Buffer * const p_first_buf;
1087 FindAndReplaceOptions const & opt;
1090 /// Auxiliary find method (does not account for opt.matchword)
1091 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
1092 void CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string = "");
1094 /** Normalize a stringified or latexified LyX paragraph.
1098 ** <li>if search is not casesensitive, then lowercase the string;
1099 ** <li>remove any newline at begin or end of the string;
1100 ** <li>replace any newline in the middle of the string with a simple space;
1101 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
1104 ** @todo Normalization should also expand macros, if the corresponding
1105 ** search option was checked.
1107 string normalize(docstring const & s, bool ignore_fomat) const;
1108 // normalized string to search
1109 string par_as_string;
1110 // regular expression to use for searching
1111 // regexp2 is same as regexp, but prefixed with a ".*?"
1113 QRegularExpression regexp;
1114 QRegularExpression regexp2;
1119 // leading format material as string
1120 string lead_as_string;
1121 // par_as_string after removal of lead_as_string
1122 string par_as_string_nolead;
1123 // unmatched open braces in the search string/regexp
1125 // number of (.*?) subexpressions added at end of search regexp for closing
1126 // environments, math mode, styles, etc...
1127 int close_wildcards;
1129 // Are we searching with regular expressions ?
1130 bool use_regexp = false;
1131 static int valid_matches;
1132 static vector <string> matches;
1133 void FillResults(MatchResult &found_mr);
1136 int MatchStringAdv::valid_matches = 0;
1137 vector <string> MatchStringAdv::matches = vector <string>(10);
1139 void MatchStringAdv::FillResults(MatchResult &found_mr)
1141 if (found_mr.match_len > 0) {
1142 valid_matches = found_mr.result.size();
1143 for (size_t i = 0; i < found_mr.result.size(); i++)
1144 matches[i] = found_mr.result[i];
1149 static void setFindParams(OutputParams &runparams)
1151 runparams.flavor = Flavor::XeTeX;
1152 //runparams.use_polyglossia = true;
1153 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
1154 // No side effect of file copying and image conversion
1155 runparams.dryrun = true;
1158 static docstring buffer_to_latex(Buffer & buffer)
1160 //OutputParams runparams(&buffer.params().encoding());
1161 OutputParams runparams(encodings.fromLyXName("utf8"));
1162 odocstringstream ods;
1164 runparams.nice = true;
1165 setFindParams(runparams);
1166 if (ignoreFormats.getDeleted())
1167 runparams.find_set_feature(OutputParams::SearchWithoutDeleted);
1169 runparams.find_set_feature(OutputParams::SearchWithDeleted);
1170 if (ignoreFormats.getNonContent()) {
1171 runparams.find_add_feature(OutputParams::SearchNonOutput);
1173 pit_type const endpit = buffer.paragraphs().size();
1174 for (pit_type pit = 0; pit != endpit; ++pit) {
1175 TeXOnePar(buffer, buffer.text(), pit, os, runparams, string(), -1, -1, true);
1176 LYXERR(Debug::FINDVERBOSE, "searchString up to here: " << ods.str());
1181 static string latexNamesToUtf8(docstring strIn, bool withformat)
1183 string addtmp = to_utf8(strIn);
1184 static regex const rmAcc("(\\\\)*("
1185 "\\\\([A-Za-z]+\\{.\\})" // e.g. "ddot{A}" == sub.str(3)
1186 "|\\\\([A-Za-z]+)( |\\{\\})?" // e.g. "LyX", "LyX{}", "LyX " == sub.str(4)
1193 if (accents.empty())
1195 for (sregex_iterator it_add(addtmp.begin(), addtmp.end(), rmAcc), end; it_add != end; ++it_add) {
1197 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
1202 if (sub.length(3) > 0)
1206 AccentsIterator it_ac = accents.find(key);
1207 if (it_ac == accents.end()) {
1208 replace = sub.str(2);
1211 replace = it_ac->second;
1214 if (lastpos < (size_t) sub.position(2))
1215 add += addtmp.substr(lastpos, sub.position(2) - lastpos);
1217 lastpos = sub.position(2) + sub.length(2);
1221 else if (addtmp.length() > lastpos)
1222 add += addtmp.substr(lastpos, addtmp.length() - lastpos);
1224 static std::regex repltilde { R"(~)" };
1225 add = std::regex_replace(add, repltilde, accents["lyxtilde"]);
1227 LYXERR(Debug::FINDVERBOSE, "Adding to search string: '"
1232 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
1235 if (!opt.ignoreformat) {
1236 str = buffer_to_latex(buffer);
1238 // OutputParams runparams(&buffer.params().encoding());
1239 OutputParams runparams(encodings.fromLyXName("utf8"));
1240 runparams.nice = true;
1241 setFindParams(runparams);
1242 int option = AS_STR_INSETS |AS_STR_PLAINTEXT;
1243 if (ignoreFormats.getDeleted()) {
1244 option |= AS_STR_SKIPDELETE;
1245 runparams.find_set_feature(OutputParams::SearchWithoutDeleted);
1248 runparams.find_set_feature(OutputParams::SearchWithDeleted);
1250 if (ignoreFormats.getNonContent()) {
1251 runparams.find_add_feature(OutputParams::SearchNonOutput);
1254 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
1255 Paragraph const & par = buffer.paragraphs().at(pit);
1256 string add = latexNamesToUtf8(par.asString(pos_type(0), par.size(),
1258 &runparams), !opt.ignoreformat);
1259 LYXERR(Debug::FINDVERBOSE, "Adding to search string: '"
1263 // Even in ignore-format we have to remove "\text{}, \lyxmathsym{}" parts
1264 while (regex_replace(t, t, "\\\\(text|lyxmathsym|ensuremath)\\{([^\\}]*)\\}", "$2"));
1271 /// Return separation pos between the leading material and the rest
1272 static size_t identifyLeading(string const & s)
1275 // @TODO Support \item[text]
1276 // Kornel: Added textsl, textsf, textit, texttt and noun
1277 // + allow to search for colored text too
1278 while (regex_replace(t, t, "^\\\\(("
1279 "(author|title|subtitle|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|"
1280 "lyxaddress|lyxrightaddress|"
1281 "footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1282 "emph|noun|minisec|text(bf|md|sl|sf|it|tt))|"
1283 "((textcolor|foreignlanguage|latexenvironment)\\{[a-z]+\\*?\\})|"
1284 "(u|uu)line|(s|x)out|uwave)|((sub)?(((sub)?section)|paragraph)|part|chapter)\\*?)\\{", "")
1285 || regex_replace(t, t, "^\\$", "")
1286 || regex_replace(t, t, "^\\\\\\[", "")
1287 || regex_replace(t, t, "^ ?\\\\item\\{[a-z]+\\}", "")
1288 || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\}", ""))
1290 LYXERR(Debug::FINDVERBOSE, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
1291 return s.size() - t.size();
1295 * Given a latexified string, retrieve some handled features
1296 * The features of the regex will later be compared with the features
1297 * of the searched text. If the regex features are not a
1298 * subset of the analized, then, in not format ignoring search
1299 * we can early stop the search in the relevant inset.
1301 typedef map<string, bool> Features;
1303 static Features identifyFeatures(string const & s)
1305 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
1306 static regex const valid("^("
1308 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1309 "emph|noun|text(bf|md|sl|sf|it|tt)|"
1310 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
1311 "(u|uu)line|(s|x)out|uwave|"
1312 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
1313 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
1318 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1321 if (sub.str(1).compare("regexp") == 0) {
1325 string token = sub.str(1);
1327 if (regex_match(token, sub2, valid)) {
1335 if (sub.str(1).compare("endregexp") == 0) {
1345 * defines values features of a key "\\[a-z]+{"
1350 /* Char type with content discarded
1351 * like \hspace{1cm} */
1353 /* Char, like \backslash */
1355 /* replace starting backslash with '#' */
1357 /* \part, \section*, ... */
1359 /* title, author etc */
1361 /* \foreignlanguage{ngerman}, ... */
1364 * to discard language in content */
1367 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1369 /* fonts, colors, markups, ... */
1371 /* footnotesize, ... large, ...
1372 * Ignore all of them */
1375 /* inputencoding, ...
1376 * Discard also content, because they do not help in search */
1379 * like remove, but also all arguments */
1381 /* item, listitem */
1383 /* tex, latex, ... like isChar */
1385 /* like \lettrine[lines=5]{}{} */
1387 // like isStandard, but always remove head
1389 /* End of arguments marker for lettrine,
1390 * so that they can be ignored */
1393 KeyInfo() = default;
1394 KeyInfo(KeyType type, int parcount, bool disable)
1396 parenthesiscount(parcount),
1397 disabled(disable) {}
1398 KeyType keytype = invalid;
1400 int _tokensize = -1;
1401 int _tokenstart = -1;
1402 int _dataStart = -1;
1404 int parenthesiscount = 1;
1405 bool disabled = false;
1406 bool used = false; /* by pattern */
1411 Border(int l=0, int u=0) : low(l), upper(u) {}
1416 #define MAXOPENED 30
1418 bool isPatternString_;
1420 explicit Intervall(bool isPattern, string const & p)
1421 : isPatternString_(isPattern), par(p), ignoreidx(-1),
1422 actualdeptindex(0), hasTitle(false), langcount(0)
1430 static vector<Border> borders;
1431 int depts[MAXOPENED];
1432 int closes[MAXOPENED];
1433 int actualdeptindex;
1434 int previousNotIgnored(int) const;
1435 int nextNotIgnored(int) const;
1436 void handleOpenP(int i);
1437 void handleCloseP(int i, bool closingAllowed);
1438 void resetOpenedP(int openPos);
1439 void addIntervall(int upper);
1440 void addIntervall(int low, int upper); /* if explicit */
1441 void removeAccents();
1442 void setForDefaultLang(KeyInfo const & defLang) const;
1443 int findclosing(int start, int end, char up, char down, int repeat);
1444 void removeInvalidClosings(void);
1445 void handleParentheses(int lastpos, bool closingAllowed);
1447 // Number of disabled language specs up
1448 // to current position in actual interval
1450 int isOpeningPar(int pos) const;
1452 void output(ostringstream &os, int lastpos);
1453 // string show(int lastpos);
1456 vector<Border> Intervall::borders = vector<Border>(30);
1458 int Intervall::isOpeningPar(int pos) const
1460 if ((pos < 0) || (size_t(pos) >= par.size()))
1462 if (par[pos] != '{')
1464 if (size_t(pos) + 2 >= par.size())
1466 if (par[pos+2] != '}')
1468 if (par[pos+1] == '[' || par[pos+1] == ']')
1473 void Intervall::setForDefaultLang(KeyInfo const & defLang) const
1475 // Enable the use of first token again
1476 if (ignoreidx >= 0) {
1477 int value = defLang._tokenstart + defLang._tokensize;
1482 if (borders[borderidx].low < value)
1483 borders[borderidx].low = value;
1484 if (borders[borderidx].upper < value)
1485 borders[borderidx].upper = value;
1490 static void checkDepthIndex(int val)
1492 static int maxdepthidx = MAXOPENED-2;
1493 static int lastmaxdepth = 0;
1494 if (val > lastmaxdepth) {
1495 LYXERR(Debug::INFO, "Depth reached " << val);
1498 if (val > maxdepthidx) {
1500 LYXERR(Debug::INFO, "maxdepthidx now " << val);
1505 // Not needed, because borders are now dynamically expanded
1506 static void checkIgnoreIdx(int val)
1508 static int lastmaxignore = -1;
1509 if ((lastmaxignore < val) && (size_t(val+1) >= borders.size())) {
1510 LYXERR(Debug::INFO, "IgnoreIdx reached " << val);
1511 lastmaxignore = val;
1517 * Expand the region of ignored parts of the input latex string
1518 * The region is only relevant in output()
1520 void Intervall::addIntervall(int low, int upper)
1523 if (low == upper) return;
1524 for (idx = ignoreidx+1; idx > 0; --idx) {
1525 if (low > borders[idx-1].upper) {
1529 Border br(low, upper);
1530 if (idx > ignoreidx) {
1531 if (borders.size() <= size_t(idx)) {
1532 borders.push_back(br);
1538 // checkIgnoreIdx(ignoreidx);
1542 // Expand only if one of the new bound is inside the interwall
1543 // We know here that br.low > borders[idx-1].upper
1544 if (br.upper < borders[idx].low) {
1545 // We have to insert at this pos
1546 if (size_t(ignoreidx+1) >= borders.size()) {
1547 borders.push_back(borders[ignoreidx]);
1550 borders[ignoreidx+1] = borders[ignoreidx];
1552 for (int i = ignoreidx; i > idx; --i) {
1553 borders[i] = borders[i-1];
1557 // checkIgnoreIdx(ignoreidx);
1560 // Here we know, that we are overlapping
1561 if (br.low > borders[idx].low)
1562 br.low = borders[idx].low;
1563 // check what has to be concatenated
1565 for (int i = idx; i <= ignoreidx; i++) {
1566 if (br.upper >= borders[i].low) {
1568 if (br.upper < borders[i].upper)
1569 br.upper = borders[i].upper;
1575 // count should be >= 1 here
1578 for (int i = idx + count; i <= ignoreidx; i++) {
1579 borders[i-count+1] = borders[i];
1581 ignoreidx -= count - 1;
1587 static void buildaccent(string n, string param, string values)
1591 const char delim = '|';
1592 while (getline(s, name, delim)) {
1594 for (char c : param) {
1595 string key = name + "{" + c + "}";
1596 // get the corresponding utf8-value
1597 if ((values[start] & 0xc0) != 0xc0) {
1598 // should not happen, utf8 encoding starts at least with 11xxxxxx
1599 // but value for '\dot{i}' is 'i', which is ascii
1600 if ((values[start] & 0x80) == 0) {
1602 accents[key] = values.substr(start, 1);
1603 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1608 for (int j = 1; ;j++) {
1609 if (start + j >= values.size()) {
1610 accents[key] = values.substr(start, j);
1611 start = values.size() - 1;
1614 else if ((values[start+j] & 0xc0) != 0x80) {
1615 // This is the first byte of following utf8 char
1616 accents[key] = values.substr(start, j);
1618 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1627 static string getutf8(unsigned uchar)
1630 string ret = string();
1632 if (uchar <= 0x7f) {
1633 c[maxc-1] = uchar & 0x7f;
1636 unsigned char rest = 0x40;
1637 unsigned char first = 0x80;
1639 for (int i = start; i >=0; --i) {
1641 c[i] = first + uchar;
1644 c[i] = 0x80 | (uchar & 0x3f);
1651 for (int i = 0; i < maxc; i++) {
1652 if (c[i] == 0) continue;
1658 static void addAccents(string latex_in, string unicode_out)
1660 latex_in = latex_in.substr(1);
1661 AccentsIterator it_ac = accents.find(latex_in);
1662 if (it_ac == accents.end()) {
1663 accents[latex_in] = unicode_out;
1666 LYXERR0("Key " << latex_in << " already set");
1670 void static fillMissingUnicodesymbols()
1672 addAccents("\\pounds", getutf8(0x00a3));
1673 addAccents("\\textsterling", getutf8(0x00a3));
1674 addAccents("\\textyen", getutf8(0x00a5));
1675 addAccents("\\yen", getutf8(0x00a5));
1676 addAccents("\\textsection", getutf8(0x00a7));
1677 addAccents("\\mathsection", getutf8(0x00a7));
1678 addAccents("\\textcopyright", getutf8(0x00a9));
1679 addAccents("\\copyright", getutf8(0x00a9));
1680 addAccents("\\textlnot", getutf8(0x00ac));
1681 addAccents("\\neg", getutf8(0x00ac));
1682 addAccents("\\textregistered", getutf8(0x00ae));
1683 addAccents("\\circledR", getutf8(0x00ae));
1684 addAccents("\\textpm", getutf8(0x00b1));
1685 addAccents("\\pm", getutf8(0x00b1));
1686 addAccents("\\textparagraph", getutf8(0x00b6));
1687 addAccents("\\mathparagraph", getutf8(0x00b6));
1688 addAccents("\\textperiodcentered", getutf8(0x00b7));
1689 addAccents("\\texttimes", getutf8(0x00d7));
1690 addAccents("\\times", getutf8(0x00d7));
1691 addAccents("\\O", getutf8(0x00d8));
1692 addAccents("\\dh", getutf8(0x00f0));
1693 addAccents("\\eth", getutf8(0x00f0));
1694 addAccents("\\textdiv", getutf8(0x00f7));
1695 addAccents("\\div", getutf8(0x00f7));
1696 addAccents("\\o", getutf8(0x00f8));
1697 addAccents("\\textcrlambda", getutf8(0x019b));
1698 addAccents("\\j", getutf8(0x0237));
1699 addAccents("\\textrevepsilon", getutf8(0x025c));
1700 addAccents("\\textbaru", getutf8(0x0289));
1701 addAccents("\\textquoteleft", getutf8(0x02bb));
1702 addAccents("\\textGamma", getutf8(0x0393));
1703 addAccents("\\Gamma", getutf8(0x0393));
1704 addAccents("\\textDelta", getutf8(0x0394));
1705 addAccents("\\Delta", getutf8(0x0394));
1706 addAccents("\\textTheta", getutf8(0x0398));
1707 addAccents("\\Theta", getutf8(0x0398));
1708 addAccents("\\textLambda", getutf8(0x039b));
1709 addAccents("\\Lambda", getutf8(0x039b));
1710 addAccents("\\textXi", getutf8(0x039e));
1711 addAccents("\\Xi", getutf8(0x039e));
1712 addAccents("\\textPi", getutf8(0x03a0));
1713 addAccents("\\Pi", getutf8(0x03a0));
1714 addAccents("\\textSigma", getutf8(0x03a3));
1715 addAccents("\\Sigma", getutf8(0x03a3));
1716 addAccents("\\textUpsilon", getutf8(0x03a5));
1717 addAccents("\\Upsilon", getutf8(0x03a5));
1718 addAccents("\\textPhi", getutf8(0x03a6));
1719 addAccents("\\Phi", getutf8(0x03a6));
1720 addAccents("\\textPsi", getutf8(0x03a8));
1721 addAccents("\\Psi", getutf8(0x03a8));
1722 addAccents("\\textOmega", getutf8(0x03a9));
1723 addAccents("\\Omega", getutf8(0x03a9));
1724 addAccents("\\textalpha", getutf8(0x03b1));
1725 addAccents("\\alpha", getutf8(0x03b1));
1726 addAccents("\\textbeta", getutf8(0x03b2));
1727 addAccents("\\beta", getutf8(0x03b2));
1728 addAccents("\\textgamma", getutf8(0x03b3));
1729 addAccents("\\gamma", getutf8(0x03b3));
1730 addAccents("\\textdelta", getutf8(0x03b4));
1731 addAccents("\\delta", getutf8(0x03b4));
1732 addAccents("\\textepsilon", getutf8(0x03b5));
1733 addAccents("\\varepsilon", getutf8(0x03b5));
1734 addAccents("\\textzeta", getutf8(0x03b6));
1735 addAccents("\\zeta", getutf8(0x03b6));
1736 addAccents("\\texteta", getutf8(0x03b7));
1737 addAccents("\\eta", getutf8(0x03b7));
1738 addAccents("\\texttheta", getutf8(0x03b8));
1739 addAccents("\\theta", getutf8(0x03b8));
1740 addAccents("\\textiota", getutf8(0x03b9));
1741 addAccents("\\iota", getutf8(0x03b9));
1742 addAccents("\\textkappa", getutf8(0x03ba));
1743 addAccents("\\kappa", getutf8(0x03ba));
1744 addAccents("\\textlambda", getutf8(0x03bb));
1745 addAccents("\\lambda", getutf8(0x03bb));
1746 addAccents("\\textmu", getutf8(0x03bc));
1747 addAccents("\\mu", getutf8(0x03bc));
1748 addAccents("\\textnu", getutf8(0x03bd));
1749 addAccents("\\nu", getutf8(0x03bd));
1750 addAccents("\\textxi", getutf8(0x03be));
1751 addAccents("\\xi", getutf8(0x03be));
1752 addAccents("\\textpi", getutf8(0x03c0));
1753 addAccents("\\pi", getutf8(0x03c0));
1754 addAccents("\\textrho", getutf8(0x03c1));
1755 addAccents("\\rho", getutf8(0x03c1));
1756 addAccents("\\textfinalsigma", getutf8(0x03c2));
1757 addAccents("\\varsigma", getutf8(0x03c2));
1758 addAccents("\\textsigma", getutf8(0x03c3));
1759 addAccents("\\sigma", getutf8(0x03c3));
1760 addAccents("\\texttau", getutf8(0x03c4));
1761 addAccents("\\tau", getutf8(0x03c4));
1762 addAccents("\\textupsilon", getutf8(0x03c5));
1763 addAccents("\\upsilon", getutf8(0x03c5));
1764 addAccents("\\textphi", getutf8(0x03c6));
1765 addAccents("\\varphi", getutf8(0x03c6));
1766 addAccents("\\textchi", getutf8(0x03c7));
1767 addAccents("\\chi", getutf8(0x03c7));
1768 addAccents("\\textpsi", getutf8(0x03c8));
1769 addAccents("\\psi", getutf8(0x03c8));
1770 addAccents("\\textomega", getutf8(0x03c9));
1771 addAccents("\\omega", getutf8(0x03c9));
1772 addAccents("\\textdigamma", getutf8(0x03dd));
1773 addAccents("\\digamma", getutf8(0x03dd));
1774 addAccents("\\hebalef", getutf8(0x05d0));
1775 addAccents("\\aleph", getutf8(0x05d0));
1776 addAccents("\\hebbet", getutf8(0x05d1));
1777 addAccents("\\beth", getutf8(0x05d1));
1778 addAccents("\\hebgimel", getutf8(0x05d2));
1779 addAccents("\\gimel", getutf8(0x05d2));
1780 addAccents("\\hebdalet", getutf8(0x05d3));
1781 addAccents("\\daleth", getutf8(0x05d3));
1782 addAccents("\\hebhe", getutf8(0x05d4));
1783 addAccents("\\hebvav", getutf8(0x05d5));
1784 addAccents("\\hebzayin", getutf8(0x05d6));
1785 addAccents("\\hebhet", getutf8(0x05d7));
1786 addAccents("\\hebtet", getutf8(0x05d8));
1787 addAccents("\\hebyod", getutf8(0x05d9));
1788 addAccents("\\hebfinalkaf", getutf8(0x05da));
1789 addAccents("\\hebkaf", getutf8(0x05db));
1790 addAccents("\\heblamed", getutf8(0x05dc));
1791 addAccents("\\hebfinalmem", getutf8(0x05dd));
1792 addAccents("\\hebmem", getutf8(0x05de));
1793 addAccents("\\hebfinalnun", getutf8(0x05df));
1794 addAccents("\\hebnun", getutf8(0x05e0));
1795 addAccents("\\hebsamekh", getutf8(0x05e1));
1796 addAccents("\\hebayin", getutf8(0x05e2));
1797 addAccents("\\hebfinalpe", getutf8(0x05e3));
1798 addAccents("\\hebpe", getutf8(0x05e4));
1799 addAccents("\\hebfinaltsadi", getutf8(0x05e5));
1800 addAccents("\\hebtsadi", getutf8(0x05e6));
1801 addAccents("\\hebqof", getutf8(0x05e7));
1802 addAccents("\\hebresh", getutf8(0x05e8));
1803 addAccents("\\hebshin", getutf8(0x05e9));
1804 addAccents("\\hebtav", getutf8(0x05ea));
1807 addAccents("\\thaiKoKai", getutf8(0x0e01));
1808 addAccents("\\thaiKhoKhai", getutf8(0x0e02));
1809 addAccents("\\thaiKhoKhuat", getutf8(0x0e03));
1810 addAccents("\\thaiKhoKhwai", getutf8(0x0e04));
1811 addAccents("\\thaiKhoKhon", getutf8(0x0e05));
1812 addAccents("\\thaiKhoRakhang", getutf8(0x0e06));
1813 addAccents("\\thaiNgoNgu", getutf8(0x0e07));
1814 addAccents("\\thaiChoChan", getutf8(0x0e08));
1815 addAccents("\\thaiChoChing", getutf8(0x0e09));
1816 addAccents("\\thaiChoChang", getutf8(0x0e0a));
1817 addAccents("\\thaiSoSo", getutf8(0x0e0b));
1818 addAccents("\\thaiChoChoe", getutf8(0x0e0c));
1819 addAccents("\\thaiYoYing", getutf8(0x0e0d));
1820 addAccents("\\thaiDoChada", getutf8(0x0e0e));
1821 addAccents("\\thaiToPatak", getutf8(0x0e0f));
1822 addAccents("\\thaiThoThan", getutf8(0x0e10));
1823 addAccents("\\thaiThoNangmontho", getutf8(0x0e11));
1824 addAccents("\\thaiThoPhuthao", getutf8(0x0e12));
1825 addAccents("\\thaiNoNen", getutf8(0x0e13));
1826 addAccents("\\thaiDoDek", getutf8(0x0e14));
1827 addAccents("\\thaiToTao", getutf8(0x0e15));
1828 addAccents("\\thaiThoThung", getutf8(0x0e16));
1829 addAccents("\\thaiThoThahan", getutf8(0x0e17));
1830 addAccents("\\thaiThoThong", getutf8(0x0e18));
1831 addAccents("\\thaiNoNu", getutf8(0x0e19));
1832 addAccents("\\thaiBoBaimai", getutf8(0x0e1a));
1833 addAccents("\\thaiPoPla", getutf8(0x0e1b));
1834 addAccents("\\thaiPhoPhung", getutf8(0x0e1c));
1835 addAccents("\\thaiFoFa", getutf8(0x0e1d));
1836 addAccents("\\thaiPhoPhan", getutf8(0x0e1e));
1837 addAccents("\\thaiFoFan", getutf8(0x0e1f));
1838 addAccents("\\thaiPhoSamphao", getutf8(0x0e20));
1839 addAccents("\\thaiMoMa", getutf8(0x0e21));
1840 addAccents("\\thaiYoYak", getutf8(0x0e22));
1841 addAccents("\\thaiRoRua", getutf8(0x0e23));
1842 addAccents("\\thaiRu", getutf8(0x0e24));
1843 addAccents("\\thaiLoLing", getutf8(0x0e25));
1844 addAccents("\\thaiLu", getutf8(0x0e26));
1845 addAccents("\\thaiWoWaen", getutf8(0x0e27));
1846 addAccents("\\thaiSoSala", getutf8(0x0e28));
1847 addAccents("\\thaiSoRusi", getutf8(0x0e29));
1848 addAccents("\\thaiSoSua", getutf8(0x0e2a));
1849 addAccents("\\thaiHoHip", getutf8(0x0e2b));
1850 addAccents("\\thaiLoChula", getutf8(0x0e2c));
1851 addAccents("\\thaiOAng", getutf8(0x0e2d));
1852 addAccents("\\thaiHoNokhuk", getutf8(0x0e2e));
1853 addAccents("\\thaiPaiyannoi", getutf8(0x0e2f));
1854 addAccents("\\thaiSaraA", getutf8(0x0e30));
1855 addAccents("\\thaiMaiHanakat", getutf8(0x0e31));
1856 addAccents("\\thaiSaraAa", getutf8(0x0e32));
1857 addAccents("\\thaiSaraAm", getutf8(0x0e33));
1858 addAccents("\\thaiSaraI", getutf8(0x0e34));
1859 addAccents("\\thaiSaraIi", getutf8(0x0e35));
1860 addAccents("\\thaiSaraUe", getutf8(0x0e36));
1861 addAccents("\\thaiSaraUee", getutf8(0x0e37));
1862 addAccents("\\thaiSaraU", getutf8(0x0e38));
1863 addAccents("\\thaiSaraUu", getutf8(0x0e39));
1864 addAccents("\\thaiPhinthu", getutf8(0x0e3a));
1865 addAccents("\\thaiSaraE", getutf8(0x0e40));
1866 addAccents("\\thaiSaraAe", getutf8(0x0e41));
1867 addAccents("\\thaiSaraO", getutf8(0x0e42));
1868 addAccents("\\thaiSaraAiMaimuan", getutf8(0x0e43));
1869 addAccents("\\thaiSaraAiMaimalai", getutf8(0x0e44));
1870 addAccents("\\thaiLakkhangyao", getutf8(0x0e45));
1871 addAccents("\\thaiMaiyamok", getutf8(0x0e46));
1872 addAccents("\\thaiMaitaikhu", getutf8(0x0e47));
1873 addAccents("\\thaiMaiEk", getutf8(0x0e48));
1874 addAccents("\\thaiMaiTho", getutf8(0x0e49));
1875 addAccents("\\thaiMaiTri", getutf8(0x0e4a));
1876 addAccents("\\thaiMaiChattawa", getutf8(0x0e4b));
1877 addAccents("\\thaiThanthakhat", getutf8(0x0e4c));
1878 addAccents("\\thaiNikhahit", getutf8(0x0e4d));
1879 addAccents("\\thaiYamakkan", getutf8(0x0e4e));
1880 addAccents("\\thaiFongman", getutf8(0x0e4f));
1881 addAccents("\\thaizero", getutf8(0x0e50));
1882 addAccents("\\thaione", getutf8(0x0e51));
1883 addAccents("\\thaitwo", getutf8(0x0e52));
1884 addAccents("\\thaithree", getutf8(0x0e53));
1885 addAccents("\\thaifour", getutf8(0x0e54));
1886 addAccents("\\thaifive", getutf8(0x0e55));
1887 addAccents("\\thaisix", getutf8(0x0e56));
1888 addAccents("\\thaiseven", getutf8(0x0e57));
1889 addAccents("\\thaieight", getutf8(0x0e58));
1890 addAccents("\\thainine", getutf8(0x0e59));
1891 addAccents("\\thaiAngkhankhu", getutf8(0x0e5a));
1892 addAccents("\\thaiKhomut", getutf8(0x0e5b));
1893 addAccents("\\dag", getutf8(0x2020));
1894 addAccents("\\dagger", getutf8(0x2020));
1895 addAccents("\\textdagger", getutf8(0x2020));
1896 addAccents("\\ddag", getutf8(0x2021));
1897 addAccents("\\ddagger", getutf8(0x2021));
1898 addAccents("\\textdaggerdbl", getutf8(0x2021));
1899 addAccents("\\textbullet", getutf8(0x2022));
1900 addAccents("\\bullet", getutf8(0x2022));
1901 addAccents("\\dots", getutf8(0x2026));
1902 addAccents("\\ldots", getutf8(0x2026));
1903 addAccents("\\textellipsis", getutf8(0x2026));
1904 addAccents("\\textasciiacute", getutf8(0x2032));
1905 addAccents("\\prime", getutf8(0x2032));
1906 addAccents("\\textacutedbl", getutf8(0x2033));
1907 addAccents("\\dprime", getutf8(0x2033));
1908 addAccents("\\textasciigrave", getutf8(0x2035));
1909 addAccents("\\backprime", getutf8(0x2035));
1910 addAccents("\\textsubcircum{ }", getutf8(0x2038));
1911 addAccents("\\caretinsert", getutf8(0x2038));
1912 addAccents("\\textasteriskcentered", getutf8(0x204e));
1913 addAccents("\\ast", getutf8(0x204e));
1914 addAccents("\\textmho", getutf8(0x2127));
1915 addAccents("\\mho", getutf8(0x2127));
1916 addAccents("\\textleftarrow", getutf8(0x2190));
1917 addAccents("\\leftarrow", getutf8(0x2190));
1918 addAccents("\\textuparrow", getutf8(0x2191));
1919 addAccents("\\uparrow", getutf8(0x2191));
1920 addAccents("\\textrightarrow", getutf8(0x2192));
1921 addAccents("\\rightarrow", getutf8(0x2192));
1922 addAccents("\\textdownarrow", getutf8(0x2193));
1923 addAccents("\\downarrow", getutf8(0x2193));
1924 addAccents("\\textglobrise", getutf8(0x2197));
1925 addAccents("\\nearrow", getutf8(0x2197));
1926 addAccents("\\textglobfall", getutf8(0x2198));
1927 addAccents("\\searrow", getutf8(0x2198));
1928 addAccents("\\textsurd", getutf8(0x221a));
1929 addAccents("\\surd", getutf8(0x221a));
1930 addAccents("\\textbigcircle", getutf8(0x25ef));
1931 addAccents("\\bigcirc", getutf8(0x25ef));
1932 addAccents("\\FiveStar", getutf8(0x2605));
1933 addAccents("\\bigstar", getutf8(0x2605));
1934 addAccents("\\FiveStarOpen", getutf8(0x2606));
1935 addAccents("\\bigwhitestar", getutf8(0x2606));
1936 addAccents("\\Checkmark", getutf8(0x2713));
1937 addAccents("\\checkmark", getutf8(0x2713));
1938 addAccents("\\CrossMaltese", getutf8(0x2720));
1939 addAccents("\\maltese", getutf8(0x2720));
1940 addAccents("\\textlangle", getutf8(0x27e8));
1941 addAccents("\\langle", getutf8(0x27e8));
1942 addAccents("\\textrangle", getutf8(0x27e9));
1943 addAccents("\\rangle", getutf8(0x27e9));
1946 static void buildAccentsMap()
1948 accents["imath"] = "ı";
1950 accents["jmath"] = "ȷ";
1951 accents["cdot"] = "·";
1952 accents["textasciicircum"] = "^";
1953 accents["mathcircumflex"] = "^";
1954 accents["guillemotright"] = "»";
1955 accents["guillemotleft"] = "«";
1956 accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
1957 accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
1958 accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
1959 accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
1960 accents["negmedspace"] = getutf8(0xf0005);
1961 accents["thickspace"] = getutf8(0xf0006);
1962 accents["negthickspace"] = getutf8(0xf0007);
1963 accents["lyx"] = getutf8(0xf0010); // Used logos
1964 accents["LyX"] = getutf8(0xf0010);
1965 accents["tex"] = getutf8(0xf0011);
1966 accents["TeX"] = getutf8(0xf0011);
1967 accents["latex"] = getutf8(0xf0012);
1968 accents["LaTeX"] = getutf8(0xf0012);
1969 accents["latexe"] = getutf8(0xf0013);
1970 accents["LaTeXe"] = getutf8(0xf0013);
1971 accents["lyxarrow"] = getutf8(0xf0020);
1972 accents["braceleft"] = getutf8(0xf0030);
1973 accents["braceright"] = getutf8(0xf0031);
1974 accents["lyxtilde"] = getutf8(0xf0032);
1975 accents["sim"] = getutf8(0xf0032);
1976 accents["lyxdollar"] = getutf8(0xf0033);
1977 accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash
1978 accents["backslash LyX"] = getutf8(0xf0010);
1979 accents["backslash tex"] = getutf8(0xf0011);
1980 accents["backslash TeX"] = getutf8(0xf0011);
1981 accents["backslash latex"] = getutf8(0xf0012);
1982 accents["backslash LaTeX"] = getutf8(0xf0012);
1983 accents["backslash latexe"] = getutf8(0xf0013);
1984 accents["backslash LaTeXe"] = getutf8(0xf0013);
1985 accents["backslash lyxarrow"] = getutf8(0xf0020);
1986 accents["ddot{\\imath}"] = "ï";
1987 buildaccent("ddot", "aAeEhHiIoOtuUwWxXyY",
1988 "äÄëËḧḦïÏöÖẗüÜẅẄẍẌÿŸ"); // umlaut
1989 buildaccent("dot|.", "aAbBcCdDeEfFGghHIimMnNoOpPrRsStTwWxXyYzZ",
1990 "ȧȦḃḂċĊḋḊėĖḟḞĠġḣḢİİṁṀṅṄȯȮṗṖṙṘṡṠṫṪẇẆẋẊẏẎżŻ"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'İ'
1991 accents["acute{\\imath}"] = "í";
1992 buildaccent("acute", "aAcCeEgGkKlLmMoOnNpPrRsSuUwWyYzZiI",
1993 "áÁćĆéÉǵǴḱḰĺĹḿḾóÓńŃṕṔŕŔśŚúÚẃẂýÝźŹíÍ");
1994 buildaccent("dacute|H|h", "oOuU", "őŐűŰ"); // double acute
1995 buildaccent("mathring|r", "aAuUwy",
1997 accents["check{\\imath}"] = "ǐ";
1998 accents["check{\\jmath}"] = "ǰ";
1999 buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
2000 "čČďĎǎǍěĚǐǏǒǑǔǓǧǦǩǨȟȞľĽňŇřŘšŠŤťžŽ"); // caron
2001 accents["hat{\\imath}"] = "î";
2002 accents["hat{\\jmath}"] = "ĵ";
2003 buildaccent("hat|^", "aAcCeEgGhHiIjJoOsSuUwWyYzZ",
2004 "âÂĉĈêÊĝĜĥĤîÎĵĴôÔŝŜûÛŵŴŷŶẑẐ"); // circ
2005 accents["bar{\\imath}"] = "ī";
2006 buildaccent("bar|=", "aAeEiIoOuUyY",
2007 "āĀēĒīĪōŌūŪȳȲ"); // macron
2008 accents["tilde{\\imath}"] = "ĩ";
2009 buildaccent("tilde", "aAeEiInNoOuUvVyY",
2010 "ãÃẽẼĩĨñÑõÕũŨṽṼỹỸ"); // tilde
2011 accents["breve{\\imath}"] = "ĭ";
2012 buildaccent("breve|u", "aAeEgGiIoOuU",
2013 "ăĂĕĔğĞĭĬŏŎŭŬ"); // breve
2014 accents["grave{\\imath}"] = "ì";
2015 buildaccent("grave|`", "aAeEiIoOuUnNwWyY",
2016 "àÀèÈìÌòÒùÙǹǸẁẀỳỲ"); // grave
2017 buildaccent("subdot|d", "BbDdHhKkLlMmNnRrSsTtVvWwZzAaEeIiOoUuYy",
2018 "ḄḅḌḍḤḥḲḳḶḷṂṃṆṇṚṛṢṣṬṭṾṿẈẉẒẓẠạẸẹỊịỌọỤụỴỵ"); // dot below
2019 buildaccent("ogonek|k", "AaEeIiUuOo",
2020 "ĄąĘęĮįŲųǪǫ"); // ogonek
2021 buildaccent("cedilla|c", "CcGgKkLlNnRrSsTtEeDdHh",
2022 "ÇçĢģĶķĻļŅņŖŗŞşŢţȨȩḐḑḨḩ"); // cedilla
2023 buildaccent("subring|textsubring", "Aa",
2025 buildaccent("subhat|textsubcircum", "DdEeLlNnTtUu",
2026 "ḒḓḘḙḼḽṊṋṰṱṶṷ"); // subcircum
2027 buildaccent("subtilde|textsubtilde", "EeIiUu",
2028 "ḚḛḬḭṴṵ"); // subtilde
2029 accents["dgrave{\\imath}"] = "ȉ";
2030 accents["textdoublegrave{\\i}"] = "ȉ";
2031 buildaccent("dgrave|textdoublegrave", "AaEeIiOoRrUu",
2032 "ȀȁȄȅȈȉȌȍȐȑȔȕ"); // double grave
2033 accents["rcap{\\imath}"] = "ȋ";
2034 accents["textroundcap{\\i}"] = "ȋ";
2035 buildaccent("rcap|textroundcap", "AaEeIiOoRrUu",
2036 "ȂȃȆȇȊȋȎȏȒȓȖȗ"); // inverted breve
2037 buildaccent("slashed", "oO",
2039 fillMissingUnicodesymbols(); // Add some still not handled entries contained in 'unicodesynbols'
2040 // LYXERR0("Number of accents " << accents.size());
2044 * Created accents in math or regexp environment
2045 * are macros, but we need the utf8 equivalent
2047 void Intervall::removeAccents()
2049 if (accents.empty())
2051 static regex const accre("\\\\("
2052 "([\\S]|[A-Za-z]+)\\{[^\\{\\}]+\\}"
2054 "(backslash ([lL]y[xX]|[tT]e[xX]|[lL]a[tT]e[xX]e?|lyxarrow))"
2059 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
2061 string key = sub.str(1);
2062 AccentsIterator it_ac = accents.find(key);
2063 if (it_ac != accents.end()) {
2064 string val = it_ac->second;
2065 size_t pos = sub.position(size_t(0));
2066 for (size_t i = 0; i < val.size(); i++) {
2067 par[pos+i] = val[i];
2069 // Remove possibly following space too
2070 if (par[pos+sub.str(0).size()] == ' ')
2071 addIntervall(pos+val.size(), pos + sub.str(0).size()+1);
2073 addIntervall(pos+val.size(), pos + sub.str(0).size());
2074 for (size_t i = pos+val.size(); i < pos + sub.str(0).size(); i++) {
2075 // remove traces of any remaining chars
2080 LYXERR(Debug::INFO, "Not added accent for \"" << key << "\"");
2085 void Intervall::handleOpenP(int i)
2088 depts[actualdeptindex] = i+1;
2089 closes[actualdeptindex] = -1;
2090 checkDepthIndex(actualdeptindex);
2093 void Intervall::handleCloseP(int i, bool closingAllowed)
2095 if (actualdeptindex <= 0) {
2096 if (! closingAllowed)
2097 LYXERR(Debug::FINDVERBOSE, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
2098 // if we are at the very end
2099 addIntervall(i, i+1);
2102 closes[actualdeptindex] = i+1;
2107 void Intervall::resetOpenedP(int openPos)
2109 // Used as initializer for foreignlanguage entry
2110 actualdeptindex = 1;
2111 depts[1] = openPos+1;
2115 int Intervall::previousNotIgnored(int start) const
2117 int idx = 0; /* int intervalls */
2118 for (idx = ignoreidx; idx >= 0; --idx) {
2119 if (start > borders[idx].upper)
2121 if (start >= borders[idx].low)
2122 start = borders[idx].low-1;
2127 int Intervall::nextNotIgnored(int start) const
2129 int idx = 0; /* int intervalls */
2130 for (idx = 0; idx <= ignoreidx; idx++) {
2131 if (start < borders[idx].low)
2133 if (start < borders[idx].upper)
2134 start = borders[idx].upper;
2139 typedef unordered_map<string, KeyInfo> KeysMap;
2140 typedef unordered_map<string, KeyInfo>::const_iterator KeysIterator;
2141 typedef vector< KeyInfo> Entries;
2142 static KeysMap keys = unordered_map<string, KeyInfo>();
2148 Intervall interval_;
2149 void buildKeys(bool);
2150 void buildEntries(bool);
2151 void makeKey(const string &, KeyInfo, bool isPatternString);
2152 void processRegion(int start, int region_end); /* remove {} parts */
2153 void removeHead(KeyInfo const &, int count=0);
2156 LatexInfo(string const & par, bool isPatternString)
2157 : entidx_(-1), interval_(isPatternString, par)
2159 buildKeys(isPatternString);
2160 entries_ = vector<KeyInfo>();
2161 buildEntries(isPatternString);
2165 if (entries_.empty()) {
2168 if (entries_[0].keytype == KeyInfo::isTitle) {
2169 interval_.hasTitle = true;
2170 if (! entries_[0].disabled) {
2171 interval_.titleValue = entries_[0].head;
2174 interval_.titleValue = "";
2176 removeHead(entries_[0]);
2177 if (entries_.size() > 1)
2186 if (int(entries_.size()) > entidx_) {
2193 bool setNextKey(int idx) {
2194 if ((idx == entidx_) && (entidx_ >= 0)) {
2201 int find(int start, KeyInfo::KeyType keytype) const {
2205 while (tmpIdx < int(entries_.size())) {
2206 if (entries_[tmpIdx].keytype == keytype)
2212 int process(ostringstream & os, KeyInfo const & actual);
2213 int dispatch(ostringstream & os, int previousStart, KeyInfo & actual);
2214 // string show(int lastpos) { return interval.show(lastpos);}
2215 int nextNotIgnored(int start) { return interval_.nextNotIgnored(start);}
2216 KeyInfo &getKeyInfo(int keyinfo) {
2217 static KeyInfo invalidInfo = KeyInfo();
2218 if ((keyinfo < 0) || ( keyinfo >= int(entries_.size())))
2221 return entries_[keyinfo];
2223 void setForDefaultLang(KeyInfo const & defLang) {interval_.setForDefaultLang(defLang);}
2224 void addIntervall(int low, int up) { interval_.addIntervall(low, up); }
2228 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
2232 for (int i = start; i < end; i += 1 + skip) {
2236 if (c == '\\') skip = 1;
2240 else if (c == down) {
2243 if ((repeat <= 0) || (par[i+1] != up))
2252 void Intervall::removeInvalidClosings(void)
2254 // this can happen, if there are deleted parts
2257 for (unsigned i = 0; i < par.size(); i += 1 + skip) {
2260 if (c == '\\') skip = 1;
2263 else if (c == '}') {
2265 addIntervall(i, i+1);
2266 LYXERR(Debug::FINDVERBOSE, "removed invalid closing '}' at " << i);
2278 size_t mathpostfixsize;
2280 size_t mathprefixsize;
2284 vector<MathEntry> entries_;
2289 void insert(string const & wait, size_t start, size_t prefixsize, size_t end, size_t postfixsize) {
2290 MathEntry m = MathEntry();
2292 m.mathStart = start;
2293 m.mathprefixsize = prefixsize;
2294 m.mathEnd = end + postfixsize;
2295 m.mathpostfixsize = postfixsize;
2296 m.mathSize = m.mathEnd - m.mathStart;
2297 entries_.push_back(m);
2299 bool empty() const { return entries_.empty(); }
2300 size_t getEndPos() const {
2301 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2304 return entries_[actualIdx_].mathEnd;
2306 size_t getStartPos() const {
2307 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2308 return 100000; /* definitely enough? */
2310 return entries_[actualIdx_].mathStart;
2312 size_t getPrefixSize() const {
2313 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2316 return entries_[actualIdx_].mathprefixsize;
2318 size_t getPostfixSize() const {
2319 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2322 return entries_[actualIdx_].mathpostfixsize;
2324 size_t getFirstPos() {
2326 return getStartPos();
2328 size_t getSize() const {
2329 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2332 return entries_[actualIdx_].mathSize;
2334 void incrEntry() { actualIdx_++; }
2337 void LatexInfo::buildEntries(bool isPatternString)
2339 static regex const rmath("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|x?x?alignat)\\*?\\})(\\{[0-9]+\\})?)");
2340 static regex const rkeys("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?)))");
2341 static bool disableLanguageOverride = false;
2342 smatch sub, submath;
2343 bool evaluatingRegexp = false;
2345 bool evaluatingMath = false;
2346 bool evaluatingCode = false;
2348 bool evaluatingOptional = false;
2349 size_t optionalEnd = 0;
2352 bool math_end_waiting = false;
2353 size_t math_pos = 10000;
2354 size_t math_prefix_size = 1;
2356 static vector<string> usedText = vector<string>();
2357 static bool removeMathHull = false;
2359 interval_.removeAccents();
2360 interval_.removeInvalidClosings();
2362 for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) {
2364 if ((submath.position(2) - submath.position(0)) %2 == 1) {
2365 // prefixed by odd count of '\\'
2368 if (math_end_waiting) {
2369 size_t pos = submath.position(size_t(2));
2370 if ((math_end == "$") &&
2371 (submath.str(2) == "$")) {
2372 mi.insert("$", math_pos, 1, pos, 1);
2373 math_end_waiting = false;
2375 else if ((math_end == "\\]") &&
2376 (submath.str(2) == "\\]")) {
2377 mi.insert("\\]", math_pos, 2, pos, 2);
2378 math_end_waiting = false;
2380 else if ((submath.str(3).compare("end") == 0) &&
2381 (submath.str(5).compare(math_end) == 0)) {
2382 mi.insert(math_end, math_pos, math_prefix_size, pos, submath.str(2).length());
2383 math_end_waiting = false;
2389 if (submath.str(3).compare("begin") == 0) {
2390 math_end_waiting = true;
2391 math_end = submath.str(5);
2392 math_pos = submath.position(size_t(2));
2393 math_prefix_size = submath.str(2).length();
2395 else if (submath.str(2).compare("\\[") == 0) {
2396 math_end_waiting = true;
2398 math_pos = submath.position(size_t(2));
2400 else if (submath.str(2) == "$") {
2401 size_t pos = submath.position(size_t(2));
2402 math_end_waiting = true;
2408 // Ignore language if there is math somewhere in pattern-string
2409 if (isPatternString) {
2410 for (auto s: usedText) {
2411 // Remove entries created in previous search runs
2414 usedText = vector<string>();
2417 keys["foreignlanguage"].disabled = true;
2418 disableLanguageOverride = true;
2419 removeMathHull = false;
2422 removeMathHull = true; // used later if not isPatternString
2423 disableLanguageOverride = false;
2427 if (disableLanguageOverride) {
2428 keys["foreignlanguage"].disabled = true;
2431 math_pos = mi.getFirstPos();
2432 for (sregex_iterator it(interval_.par.begin(), interval_.par.end(), rkeys), end; it != end; ++it) {
2434 if ((sub.position(2) - sub.position(0)) %2 == 1) {
2435 // prefixed by odd count of '\\'
2438 string key = sub.str(5);
2440 if (sub.str(2)[0] == '\\')
2441 key = sub.str(2)[1];
2446 KeysIterator it_key = keys.find(key);
2447 if (it_key != keys.end()) {
2448 if (it_key->second.keytype == KeyInfo::headRemove) {
2449 KeyInfo found1 = it_key->second;
2450 found1.disabled = true;
2451 found1.head = "\\" + key + "{";
2452 found1._tokenstart = sub.position(size_t(2));
2453 found1._tokensize = found1.head.length();
2454 found1._dataStart = found1._tokenstart + found1.head.length();
2455 int endpos = interval_.findclosing(found1._dataStart, interval_.par.length(), '{', '}', 1);
2456 found1._dataEnd = endpos;
2461 if (evaluatingRegexp) {
2462 if (sub.str(3).compare("endregexp") == 0) {
2463 evaluatingRegexp = false;
2464 // found._tokenstart already set
2465 found._dataEnd = sub.position(size_t(2)) + 13;
2466 found._dataStart = found._dataEnd;
2467 found._tokensize = found._dataEnd - found._tokenstart;
2468 found.parenthesiscount = 0;
2469 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2476 if (evaluatingMath) {
2477 if (size_t(sub.position(size_t(2))) < mi.getEndPos())
2479 evaluatingMath = false;
2481 math_pos = mi.getStartPos();
2483 if (it_key == keys.end()) {
2484 found = KeyInfo(KeyInfo::isStandard, 0, true);
2485 LYXERR(Debug::INFO, "Undefined key " << key << " ==> will be used as text");
2486 found = KeyInfo(KeyInfo::isText, 0, false);
2487 if (isPatternString) {
2488 found.keytype = KeyInfo::isChar;
2489 found.disabled = false;
2493 usedText.push_back(key);
2497 if (key.compare("regexp") == 0) {
2498 evaluatingRegexp = true;
2499 found._tokenstart = sub.position(size_t(2));
2500 found._tokensize = 0;
2504 // Handle the other params of key
2505 if (found.keytype == KeyInfo::isIgnored)
2507 else if (found.keytype == KeyInfo::isMath) {
2508 if (size_t(sub.position(size_t(2))) == math_pos) {
2510 found._tokenstart = sub.position(size_t(2));
2511 found._tokensize = mi.getSize();
2512 found._dataEnd = found._tokenstart + found._tokensize;
2513 found._dataStart = found._dataEnd;
2514 found.parenthesiscount = 0;
2515 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2516 if (removeMathHull) {
2517 interval_.addIntervall(found._tokenstart, found._tokenstart + mi.getPrefixSize());
2518 interval_.addIntervall(found._dataEnd - mi.getPostfixSize(), found._dataEnd);
2521 // Treate all math constructs as simple math
2522 interval_.par[found._tokenstart] = '$';
2523 interval_.par[found._dataEnd - mi.getPostfixSize()] = '$';
2524 interval_.addIntervall(found._tokenstart + 1, found._tokenstart + mi.getPrefixSize());
2525 interval_.addIntervall(found._dataEnd - mi.getPostfixSize() + 1, found._dataEnd);
2527 evaluatingMath = true;
2530 // begin|end of unknown env, discard
2531 // First handle tables
2532 // longtable|tabular
2533 bool discardComment;
2535 found.keytype = KeyInfo::doRemove;
2536 if ((sub.str(7).compare("longtable") == 0) ||
2537 (sub.str(7).compare("tabular") == 0)) {
2538 discardComment = true; /* '%' */
2541 discardComment = false;
2542 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
2544 string token = sub.str(7);
2545 if (regex_match(token, sub2, removeArgs)) {
2546 found.keytype = KeyInfo::removeWithArg;
2549 // discard spaces before pos(2)
2550 int pos = sub.position(size_t(2));
2552 for (count = 0; pos - count > 0; count++) {
2553 char c = interval_.par[pos-count-1];
2554 if (discardComment) {
2555 if ((c != ' ') && (c != '%'))
2561 found._tokenstart = pos - count;
2562 if (sub.str(3).compare(0, 5, "begin") == 0) {
2563 size_t pos1 = pos + sub.str(2).length();
2564 if (sub.str(7).compare("cjk") == 0) {
2565 pos1 = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2566 if ((interval_.par[pos1] == '{') && (interval_.par[pos1+1] == '}'))
2568 found.keytype = KeyInfo::isMain;
2569 found._dataStart = pos1;
2570 found._dataEnd = interval_.par.length();
2571 found.disabled = keys["foreignlanguage"].disabled;
2572 found.used = keys["foreignlanguage"].used;
2573 found._tokensize = pos1 - found._tokenstart;
2574 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2577 // Swallow possible optional params
2578 while (interval_.par[pos1] == '[') {
2579 pos1 = interval_.findclosing(pos1+1, interval_.par.length(), '[', ']')+1;
2581 // Swallow also the eventual parameter
2582 if (interval_.par[pos1] == '{') {
2583 found._dataEnd = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2586 found._dataEnd = pos1;
2588 found._dataStart = found._dataEnd;
2589 found._tokensize = count + found._dataEnd - pos;
2590 found.parenthesiscount = 0;
2591 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2592 found.disabled = true;
2596 // Handle "\end{...}"
2597 found._dataStart = pos + sub.str(2).length();
2598 found._dataEnd = found._dataStart;
2599 found._tokensize = count + found._dataEnd - pos;
2600 found.parenthesiscount = 0;
2601 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2602 found.disabled = true;
2606 else if (found.keytype != KeyInfo::isRegex) {
2607 found._tokenstart = sub.position(size_t(2));
2608 if (found.parenthesiscount == 0) {
2609 // Probably to be discarded
2610 size_t following_pos = sub.position(size_t(2)) + sub.str(5).length() + 1;
2611 char following = interval_.par[following_pos];
2612 if (following == ' ')
2613 found.head = "\\" + sub.str(5) + " ";
2614 else if (following == '=') {
2615 // like \uldepth=1000pt
2616 found.head = sub.str(2);
2619 found.head = "\\" + key;
2620 found._tokensize = found.head.length();
2621 found._dataEnd = found._tokenstart + found._tokensize;
2622 found._dataStart = found._dataEnd;
2625 int params = found._tokenstart + key.length() + 1;
2626 if (evaluatingOptional) {
2627 if (size_t(found._tokenstart) > optionalEnd) {
2628 evaluatingOptional = false;
2631 found.disabled = true;
2634 int optend = params;
2635 while (interval_.par[optend] == '[') {
2636 // discard optional parameters
2637 optend = interval_.findclosing(optend+1, interval_.par.length(), '[', ']') + 1;
2639 if (optend > params) {
2640 key += interval_.par.substr(params, optend-params);
2641 evaluatingOptional = true;
2642 optionalEnd = optend;
2643 if (found.keytype == KeyInfo::isSectioning) {
2644 // Remove optional values (but still keep in header)
2645 interval_.addIntervall(params, optend);
2648 string token = sub.str(7);
2650 if (interval_.par[optend] != '{') {
2652 found.parenthesiscount = 0;
2653 found.head = "\\" + key;
2656 closings = found.parenthesiscount;
2657 if (found.parenthesiscount == 1) {
2658 found.head = "\\" + key + "{";
2660 else if (found.parenthesiscount > 1) {
2662 found.head = sub.str(2) + "{";
2663 closings = found.parenthesiscount - 1;
2666 found.head = "\\" + key + "{";
2669 found._tokensize = found.head.length();
2670 found._dataStart = found._tokenstart + found.head.length();
2671 if (found.keytype == KeyInfo::doRemove) {
2673 size_t endpar = 2 + interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2674 if (endpar >= interval_.par.length())
2675 found._dataStart = interval_.par.length();
2677 found._dataStart = endpar;
2678 found._tokensize = found._dataStart - found._tokenstart;
2681 found._dataStart = found._tokenstart + found._tokensize;
2685 if (interval_.par.substr(found._dataStart, 15).compare("\\endarguments{}") == 0) {
2686 found._dataStart += 15;
2690 endpos = found._dataStart - 1;
2692 endpos = interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2693 if (found.keytype == KeyInfo::isList) {
2694 // Check if it really is list env
2695 static regex const listre("^([a-z]+)$");
2697 if (!regex_match(token, sub2, listre)) {
2698 // Change the key of this entry. It is not in a list/item environment
2699 found.keytype = KeyInfo::endArguments;
2702 if (found.keytype == KeyInfo::noMain) {
2703 evaluatingCode = true;
2705 codeStart = found._dataStart;
2707 else if (evaluatingCode) {
2708 if (size_t(found._dataStart) > codeEnd)
2709 evaluatingCode = false;
2710 else if (found.keytype == KeyInfo::isMain) {
2711 // Disable this key, treate it as standard
2712 found.keytype = KeyInfo::isStandard;
2713 found.disabled = true;
2714 if ((codeEnd +1 >= interval_.par.length()) &&
2715 (found._tokenstart == codeStart)) {
2716 // trickery, because the code inset starts
2717 // with \selectlanguage ...
2719 if (entries_.size() > 1) {
2720 entries_[entries_.size()-1]._dataEnd = codeEnd;
2725 if ((endpos == interval_.par.length()) &&
2726 (found.keytype == KeyInfo::doRemove)) {
2727 // Missing closing => error in latex-input?
2728 // therefore do not delete remaining data
2729 found._dataStart -= 1;
2730 found._dataEnd = found._dataStart;
2733 found._dataEnd = endpos;
2735 if (isPatternString) {
2736 keys[key].used = true;
2739 entries_.push_back(found);
2743 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
2745 stringstream s(keysstring);
2747 const char delim = '|';
2748 while (getline(s, key, delim)) {
2749 KeyInfo keyII(keyI);
2750 if (isPatternString) {
2753 else if ( !keys[key].used)
2754 keyII.disabled = true;
2759 void LatexInfo::buildKeys(bool isPatternString)
2762 static bool keysBuilt = false;
2763 if (keysBuilt && !isPatternString) return;
2765 // Keys to ignore in any case
2766 makeKey("text|textcyrillic|lyxmathsym|ensuremath", KeyInfo(KeyInfo::headRemove, 1, true), true);
2767 // Known standard keys with 1 parameter.
2768 // Split is done, if not at start of region
2769 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
2770 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
2771 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
2772 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
2773 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
2774 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
2776 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
2777 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2778 makeKey("section*|subsection*|subsubsection*|paragraph*",
2779 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2780 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2781 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
2783 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
2785 // Split is done, if not at start of region
2786 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
2787 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
2789 // Split is done always.
2790 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
2794 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2795 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2796 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2797 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2799 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2800 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2801 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2802 makeKey("thickspace|medspace|thinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2804 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2805 // Custom space/skip, remove the content (== length value)
2806 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
2807 // Found in fr/UserGuide.lyx
2808 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2810 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2811 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2812 // Known macros to remove (including their parameter)
2814 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
2815 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
2816 // handle like standard keys with 1 parameter.
2817 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
2819 if (ignoreFormats.getDeleted()) {
2820 // Ignore deleted text
2821 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString);
2824 // but preserve added text
2825 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2827 makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2829 // Macros to remove, but let the parameter survive
2831 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2833 // Remove language spec from content of these insets
2834 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
2836 // Same effect as previous, parameter will survive (because there is no one anyway)
2838 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2839 // Remove table decorations
2840 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
2841 // Discard shape-header.
2842 // For footnote or shortcut too, because of lang settings
2843 // and wrong handling if used 'KeyInfo::noMain'
2844 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2845 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2846 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2847 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2848 makeKey("hphantom|vphantom|note|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2849 makeKey("textgreek|textcyrillic", KeyInfo(KeyInfo::isStandard, 1, true), false);
2850 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
2851 // like ('tiny{}' or '\tiny ' ... )
2852 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, ignoreFormats.getSize()), isPatternString);
2854 // Survives, like known character
2855 // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2856 makeKey("tableofcontents", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2857 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
2859 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2860 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2861 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2863 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2864 // Remove RTL/LTR marker
2865 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2866 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
2867 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
2868 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
2869 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
2870 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
2871 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2872 if (isPatternString) {
2873 // Allow the first searched string to rebuild the keys too
2877 // no need to rebuild again
2883 * Keep the list of actual opened parentheses actual
2884 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2886 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2889 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2893 if (c == '\\') skip = 1;
2894 else if (c == '{') {
2897 else if (c == '}') {
2898 handleCloseP(i, closingAllowed);
2904 string Intervall::show(int lastpos)
2906 int idx = 0; /* int intervalls */
2909 if ((unsigned) lastpos > par.size())
2910 lastpos = par.size();
2911 for (idx = 0; idx <= ignoreidx; idx++) {
2912 while (i < lastpos) {
2914 if (i <= borders[idx].low) {
2915 if (borders[idx].low > lastpos)
2916 printsize = lastpos - i;
2918 printsize = borders[idx].low - i;
2919 s += par.substr(i, printsize);
2921 if (i >= borders[idx].low)
2922 i = borders[idx].upper;
2925 i = borders[idx].upper;
2931 s += par.substr(i, lastpos-i);
2937 void Intervall::output(ostringstream &os, int lastpos)
2939 // get number of chars to output
2940 int idx = 0; /* int intervalls */
2943 string startTitle = titleValue;
2944 for (idx = 0; idx <= ignoreidx; idx++) {
2946 if (i <= borders[idx].low) {
2948 if (borders[idx].low > lastpos)
2949 printsize = lastpos - i;
2951 printsize = borders[idx].low - i;
2952 if (printsize > 0) {
2953 os << startTitle << par.substr(i, printsize);
2955 printed += printsize;
2958 handleParentheses(i, false);
2959 if (i >= borders[idx].low)
2960 i = borders[idx].upper;
2963 i = borders[idx].upper;
2970 os << startTitle << par.substr(i, lastpos-i);
2971 printed += lastpos-i;
2973 handleParentheses(lastpos, false);
2975 if (keys["foreignlanguage"].disabled)
2976 startindex = actualdeptindex-langcount;
2978 startindex = actualdeptindex;
2979 for (int i = startindex; i > 0; --i) {
2982 if (hasTitle && (printed > 0))
2984 if (! isPatternString_)
2986 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2989 void LatexInfo::processRegion(int start, int region_end)
2991 while (start < region_end) { /* Let {[} and {]} survive */
2992 int cnt = interval_.isOpeningPar(start);
2994 // Closing is allowed past the region
2995 int closing = interval_.findclosing(start+1, interval_.par.length());
2996 interval_.addIntervall(start, start+1);
2997 interval_.addIntervall(closing, closing+1);
3001 start = interval_.nextNotIgnored(start+1);
3005 void LatexInfo::removeHead(KeyInfo const & actual, int count)
3007 if (actual.parenthesiscount == 0) {
3008 // "{\tiny{} ...}" ==> "{{} ...}"
3009 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
3012 // Remove header hull, that is "\url{abcd}" ==> "abcd"
3013 interval_.addIntervall(actual._tokenstart - count, actual._dataStart);
3014 interval_.addIntervall(actual._dataEnd, actual._dataEnd+1);
3018 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
3021 switch (actual.keytype)
3023 case KeyInfo::isTitle: {
3025 nextKeyIdx = getNextKey();
3028 case KeyInfo::cleanToStart: {
3029 actual._dataEnd = actual._dataStart;
3030 nextKeyIdx = getNextKey();
3031 // Search for end of arguments
3032 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
3034 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
3035 entries_[i].disabled = true;
3037 actual._dataEnd = entries_[tmpIdx]._dataEnd;
3039 while (interval_.par[actual._dataEnd] == ' ')
3041 interval_.addIntervall(0, actual._dataEnd+1);
3042 interval_.actualdeptindex = 0;
3043 interval_.depts[0] = actual._dataEnd+1;
3044 interval_.closes[0] = -1;
3047 case KeyInfo::isText:
3048 interval_.par[actual._tokenstart] = '#';
3049 //interval_.addIntervall(actual._tokenstart, actual._tokenstart+1);
3050 nextKeyIdx = getNextKey();
3052 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
3053 if (actual.disabled)
3054 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
3056 interval_.addIntervall(actual._dataStart, actual._dataEnd);
3059 case KeyInfo::isChar: {
3060 nextKeyIdx = getNextKey();
3063 case KeyInfo::isSize: {
3064 if (actual.disabled || (interval_.par[actual._dataStart] != '{') || (interval_.par[actual._dataStart-1] == ' ')) {
3065 if (actual.parenthesiscount == 0)
3066 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
3068 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
3070 nextKeyIdx = getNextKey();
3072 // Here _dataStart points to '{', so correct it
3073 actual._dataStart += 1;
3074 actual._tokensize += 1;
3075 actual.parenthesiscount = 1;
3076 if (interval_.par[actual._dataStart] == '}') {
3077 // Determine the end if used like '{\tiny{}...}'
3078 actual._dataEnd = interval_.findclosing(actual._dataStart+1, interval_.par.length()) + 1;
3079 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
3082 // Determine the end if used like '\tiny{...}'
3083 actual._dataEnd = interval_.findclosing(actual._dataStart, interval_.par.length()) + 1;
3085 // Split on this key if not at start
3086 int start = interval_.nextNotIgnored(previousStart);
3087 if (start < actual._tokenstart) {
3088 interval_.output(os, actual._tokenstart);
3089 interval_.addIntervall(start, actual._tokenstart);
3091 // discard entry if at end of actual
3092 nextKeyIdx = process(os, actual);
3096 case KeyInfo::endArguments: {
3097 // Remove trailing '{}' too
3098 actual._dataStart += 1;
3099 actual._dataEnd += 1;
3100 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
3101 nextKeyIdx = getNextKey();
3104 case KeyInfo::noMain:
3106 case KeyInfo::isStandard: {
3107 if (actual.disabled) {
3109 processRegion(actual._dataStart, actual._dataStart+1);
3110 nextKeyIdx = getNextKey();
3112 // Split on this key if not at datastart of calling entry
3113 int start = interval_.nextNotIgnored(previousStart);
3114 if (start < actual._tokenstart) {
3115 interval_.output(os, actual._tokenstart);
3116 interval_.addIntervall(start, actual._tokenstart);
3118 // discard entry if at end of actual
3119 nextKeyIdx = process(os, actual);
3123 case KeyInfo::removeWithArg: {
3124 nextKeyIdx = getNextKey();
3125 // Search for end of arguments
3126 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
3128 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
3129 entries_[i].disabled = true;
3131 actual._dataEnd = entries_[tmpIdx]._dataEnd;
3133 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
3136 case KeyInfo::doRemove: {
3137 // Remove the key with all parameters and following spaces
3140 if (interval_.par[actual._dataEnd-1] == ' ' || interval_.par[actual._dataEnd-1] == '}')
3141 start = actual._dataEnd;
3143 start = actual._dataEnd+1;
3144 for (pos = start; pos < interval_.par.length(); pos++) {
3145 if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
3148 // Remove also enclosing parentheses [] and {}
3151 while (actual._tokenstart > numpars) {
3152 if (pos+numpars >= interval_.par.size())
3154 else if (interval_.par[pos+numpars] == ']' && interval_.par[actual._tokenstart-numpars-1] == '[')
3156 else if (interval_.par[pos+numpars] == '}' && interval_.par[actual._tokenstart-numpars-1] == '{')
3162 if (interval_.par[pos+numpars] == ' ')
3166 interval_.addIntervall(actual._tokenstart-numpars, pos+numpars+spaces);
3167 nextKeyIdx = getNextKey();
3170 case KeyInfo::isList: {
3171 // Discard space before _tokenstart
3173 for (count = 0; count < actual._tokenstart; count++) {
3174 if (interval_.par[actual._tokenstart-count-1] != ' ')
3177 nextKeyIdx = getNextKey();
3178 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
3180 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
3182 // How else can we catch this one?
3183 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
3184 entries_[i].disabled = true;
3186 actual._dataEnd = entries_[tmpIdx]._dataEnd;
3188 else if (nextKeyIdx > 0) {
3189 // Ignore any lang entries inside data region
3190 for (int i = nextKeyIdx; i < int(entries_.size()) && entries_[i]._tokenstart < actual._dataEnd; i++) {
3191 if (entries_[i].keytype == KeyInfo::isMain)
3192 entries_[i].disabled = true;
3195 if (actual.disabled) {
3196 interval_.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
3199 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
3201 if (interval_.par[actual._dataEnd+1] == '[') {
3202 int posdown = interval_.findclosing(actual._dataEnd+2, interval_.par.length(), '[', ']');
3203 if ((interval_.par[actual._dataEnd+2] == '{') &&
3204 (interval_.par[posdown-1] == '}')) {
3205 interval_.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
3206 interval_.addIntervall(posdown-1, posdown+1);
3209 interval_.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
3210 interval_.addIntervall(posdown, posdown+1);
3212 int blk = interval_.nextNotIgnored(actual._dataEnd+1);
3213 if (blk > posdown) {
3214 // Discard at most 1 space after empty item
3216 for (count = 0; count < 1; count++) {
3217 if (interval_.par[blk+count] != ' ')
3221 interval_.addIntervall(blk, blk+count);
3226 case KeyInfo::isSectioning: {
3227 // Discard spaces before _tokenstart
3229 int val = actual._tokenstart;
3230 for (count = 0; count < actual._tokenstart;) {
3231 val = interval_.previousNotIgnored(val-1);
3232 if (val < 0 || interval_.par[val] != ' ')
3235 count = actual._tokenstart - val;
3238 if (actual.disabled) {
3239 removeHead(actual, count);
3240 nextKeyIdx = getNextKey();
3242 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
3243 nextKeyIdx = process(os, actual);
3247 case KeyInfo::isMath: {
3248 // Same as regex, use the content unchanged
3249 nextKeyIdx = getNextKey();
3252 case KeyInfo::isRegex: {
3253 // DO NOT SPLIT ON REGEX
3255 nextKeyIdx = getNextKey();
3258 case KeyInfo::isIgnored: {
3259 // Treat like a character for now
3260 nextKeyIdx = getNextKey();
3263 case KeyInfo::isMain: {
3264 if (interval_.par.substr(actual._dataStart, 2) == "% ")
3265 interval_.addIntervall(actual._dataStart, actual._dataStart+2);
3266 if (actual._tokenstart > 0) {
3267 int prev = interval_.previousNotIgnored(actual._tokenstart - 1);
3268 if ((prev >= 0) && interval_.par[prev] == '%')
3269 interval_.addIntervall(prev, prev+1);
3271 if (actual.disabled) {
3273 interval_.langcount++;
3274 if ((interval_.par.substr(actual._dataStart, 3) == " \\[") ||
3275 (interval_.par.substr(actual._dataStart, 8) == " \\begin{")) {
3276 // Discard also the space before math-equation
3277 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
3279 nextKeyIdx = getNextKey();
3280 // interval.resetOpenedP(actual._dataStart-1);
3283 if (actual._tokenstart < 26) {
3284 // for the first (and maybe dummy) language
3285 interval_.setForDefaultLang(actual);
3287 interval_.resetOpenedP(actual._dataStart-1);
3291 case KeyInfo::invalid:
3292 case KeyInfo::headRemove:
3293 // These two cases cannot happen, already handled
3296 // LYXERR(Debug::INFO, "Unhandled keytype");
3297 nextKeyIdx = getNextKey();
3304 int LatexInfo::process(ostringstream & os, KeyInfo const & actual )
3306 int end = interval_.nextNotIgnored(actual._dataEnd);
3307 int oldStart = actual._dataStart;
3308 int nextKeyIdx = getNextKey();
3310 if ((nextKeyIdx < 0) ||
3311 (entries_[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
3312 (entries_[nextKeyIdx].keytype == KeyInfo::invalid)) {
3313 if (oldStart <= end) {
3314 processRegion(oldStart, end);
3319 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
3321 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
3322 (void) dispatch(os, actual._dataStart, nextKey);
3323 end = nextKey._tokenstart;
3326 processRegion(oldStart, nextKey._tokenstart);
3327 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
3329 oldStart = nextKey._dataEnd+1;
3331 // now nextKey is either invalid or is outside of actual._dataEnd
3332 // output the remaining and discard myself
3333 if (oldStart <= end) {
3334 processRegion(oldStart, end);
3336 if (interval_.par.size() > (size_t) end && interval_.par[end] == '}') {
3338 // This is the normal case.
3339 // But if using the firstlanguage, the closing may be missing
3341 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
3343 if (actual._dataEnd < end)
3344 output_end = interval_.nextNotIgnored(actual._dataEnd);
3345 else if (interval_.par.size() > (size_t) end)
3346 output_end = interval_.nextNotIgnored(end);
3348 output_end = interval_.par.size();
3349 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
3350 interval_.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
3352 // Remove possible empty data
3353 int dstart = interval_.nextNotIgnored(actual._dataStart);
3354 while (interval_.isOpeningPar(dstart) == 1) {
3355 interval_.addIntervall(dstart, dstart+1);
3356 int dend = interval_.findclosing(dstart+1, output_end);
3357 interval_.addIntervall(dend, dend+1);
3358 dstart = interval_.nextNotIgnored(dstart+1);
3360 if (dstart < output_end)
3361 interval_.output(os, output_end);
3363 interval_.addIntervall(0, end);
3365 interval_.addIntervall(actual._tokenstart, end);
3369 string splitOnKnownMacros(string par, bool isPatternString)
3372 LatexInfo li(par, isPatternString);
3373 // LYXERR(Debug::INFO, "Berfore split: " << par);
3374 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
3376 DummyKey._tokensize = 0;
3377 DummyKey._dataStart = 0;
3378 DummyKey._dataEnd = par.length();
3379 DummyKey.disabled = true;
3380 int firstkeyIdx = li.getFirstKey();
3382 if (firstkeyIdx >= 0) {
3383 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
3384 DummyKey._tokenstart = firstKey._tokenstart;
3386 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
3387 // Use dummy firstKey
3388 firstKey = DummyKey;
3389 (void) li.setNextKey(firstkeyIdx);
3392 if (par.substr(firstKey._dataStart, 2) == "% ")
3393 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
3395 nextkeyIdx = li.process(os, firstKey);
3396 while (nextkeyIdx >= 0) {
3397 // Check for a possible gap between the last
3398 // entry and this one
3399 int datastart = li.nextNotIgnored(firstKey._dataStart);
3400 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
3401 if ((nextKey._tokenstart > datastart)) {
3403 firstKey._dataStart = datastart;
3404 firstKey._dataEnd = par.length();
3405 (void) li.setNextKey(nextkeyIdx);
3406 // Fake the last opened parenthesis
3407 li.setForDefaultLang(firstKey);
3408 nextkeyIdx = li.process(os, firstKey);
3411 if (nextKey.keytype != KeyInfo::isMain) {
3412 firstKey._dataStart = datastart;
3413 firstKey._dataEnd = nextKey._dataEnd+1;
3414 (void) li.setNextKey(nextkeyIdx);
3415 li.setForDefaultLang(firstKey);
3416 nextkeyIdx = li.process(os, firstKey);
3419 nextkeyIdx = li.process(os, nextKey);
3423 // Handle the remaining
3424 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
3425 firstKey._dataEnd = par.length();
3427 if ((firstKey._dataStart < firstKey._dataEnd) &&
3428 (par[firstKey._dataStart] != '}')) {
3429 li.setForDefaultLang(firstKey);
3430 (void) li.process(os, firstKey);
3433 // return string definitelly impossible to match, but should be known
3436 s = par; /* no known macros found */
3437 // LYXERR(Debug::INFO, "After split: " << s);
3442 * Try to unify the language specs in the latexified text.
3443 * Resulting modified string is set to "", if
3444 * the searched tex does not contain all the features in the search pattern
3446 static string correctlanguagesetting(string par, bool isPatternString, bool withformat, lyx::Buffer *pbuf = nullptr)
3448 static Features regex_f;
3449 static int missed = 0;
3450 static bool regex_with_format = false;
3452 int parlen = par.length();
3454 while ((parlen > 0) && (par[parlen-1] == '\n')) {
3458 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
3459 // Happens to be there in case of description or labeling environment
3465 // Split the latex input into pieces which
3466 // can be digested by our search engine
3467 LYXERR(Debug::FINDVERBOSE, "input: \"" << par << "\"");
3468 if (isPatternString && (pbuf != nullptr)) { // Check if we should disable/enable test for language
3469 // We check for polyglossia, because in runparams.flavor we use Flavor::XeTeX
3470 string doclang = pbuf->params().language->polyglossia();
3471 static regex langre("\\\\(foreignlanguage)\\{([^\\}]+)\\}");
3473 bool toIgnoreLang = true;
3474 for (sregex_iterator it(par.begin(), par.end(), langre), end; it != end; ++it) {
3476 if (sub.str(2) != doclang) {
3477 toIgnoreLang = false;
3481 setIgnoreFormat("language", toIgnoreLang, false);
3484 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
3485 LYXERR(Debug::FINDVERBOSE, "After splitOnKnownMacros:\n\"" << result << "\"");
3488 result = par.substr(0, parlen);
3489 if (isPatternString) {
3492 regex_f = identifyFeatures(result);
3493 string features = "";
3494 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3495 string a = it->first;
3496 regex_with_format = true;
3497 features += " " + a;
3498 // LYXERR(Debug::INFO, "Identified regex format:" << a);
3500 LYXERR(Debug::FINDVERBOSE, "Identified Features" << features);
3503 } else if (regex_with_format) {
3504 Features info = identifyFeatures(result);
3505 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3506 string a = it->first;
3507 bool b = it->second;
3508 if (b && ! info[a]) {
3510 LYXERR(Debug::FINDVERBOSE, "Missed(" << missed << " " << a <<", srclen = " << parlen );
3517 // LYXERR(Debug::INFO, "No regex formats");
3523 // Remove trailing closure of math, macros and environments, so to catch parts of them.
3524 static void identifyClosing(string & t, bool ignoreformat)
3527 LYXERR(Debug::FINDVERBOSE, "identifyClosing(): t now is '" << t << "'");
3528 if (regex_replace(t, t, "(.*[^\\\\])\\$$", "$1"))
3530 if (regex_replace(t, t, "(.*[^\\\\])\\\\\\]$", "$1"))
3532 if (regex_replace(t, t, "(.*[^\\\\])\\\\end\\{[a-zA-Z_]+\\*?\\}$", "$1"))
3534 if (! ignoreformat) {
3535 if (regex_replace(t, t, "(.*[^\\\\])\\}$", "$1"))
3543 static int num_replaced = 0;
3544 static bool previous_single_replace = true;
3546 void MatchStringAdv::CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string)
3549 if (regexp_str.empty() || regexp2_str.empty()) {
3550 regexIsValid = false;
3551 regexError = "Invalid empty regex";
3554 // Handle \w properly
3555 QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
3556 if (! opt.casesensitive) {
3557 popts |= QRegularExpression::CaseInsensitiveOption;
3559 regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
3560 regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
3562 if (regexp.isValid() && regexp2.isValid()) {
3563 regexIsValid = true;
3564 // Check '{', '}' pairs inside the regex
3567 for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
3568 char c = par_as_string[i];
3575 else if (c == '}') {
3582 if (balanced != 0) {
3583 regexIsValid = false;
3584 regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
3588 regexIsValid = false;
3589 if (!regexp.isValid())
3590 regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
3592 regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
3595 (void)par_as_string;
3596 if (opt.casesensitive) {
3597 regexp = regex(regexp_str);
3598 regexp2 = regex(regexp2_str);
3601 regexp = regex(regexp_str, std::regex_constants::icase);
3602 regexp2 = regex(regexp2_str, std::regex_constants::icase);
3607 static void modifyRegexForMatchWord(string &t)
3610 regex wordre("(\\\\)*((\\.|\\\\b))");
3613 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
3615 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
3618 else if (sub.str(2) == "\\\\b")
3620 if (lastpos < (size_t) sub.position(2))
3621 s += t.substr(lastpos, sub.position(2) - lastpos);
3623 lastpos = sub.position(2) + sub.length(2);
3626 s = "\\b" + t + "\\b";
3630 else if (lastpos < t.length())
3631 s += t.substr(lastpos, t.length() - lastpos);
3632 t = "\\b" + s + "\\b";
3635 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt)
3636 : p_buf(&buf), p_first_buf(&buf), opt(opt)
3638 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
3639 docstring const & ds = stringifySearchBuffer(find_buf, opt);
3641 CreateRegexp(opt, "", "", "");
3644 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
3645 if (opt.replace_all && previous_single_replace) {
3646 previous_single_replace = false;
3649 else if (!opt.replace_all) {
3650 num_replaced = 0; // count number of replaced strings
3651 previous_single_replace = true;
3653 // When using regexp, braces are hacked already by escape_for_regex()
3654 par_as_string = normalize(ds, opt.ignoreformat);
3656 close_wildcards = 0;
3658 size_t lead_size = 0;
3659 // correct the language settings
3660 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat, &buf);
3661 if (par_as_string.empty()) {
3662 CreateRegexp(opt, "", "", "");
3665 opt.matchAtStart = false;
3667 identifyClosing(par_as_string, opt.ignoreformat); // Removes math closings ($, ], ...) at end of string
3668 if (opt.ignoreformat) {
3672 lead_size = identifyLeading(par_as_string);
3674 lead_as_string = par_as_string.substr(0, lead_size);
3675 string lead_as_regex_string = string2regex(lead_as_string);
3676 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3677 string par_as_regex_string_nolead = string2regex(par_as_string_nolead);
3678 /* Handle whole words too in this case
3680 if (opt.matchword) {
3681 par_as_regex_string_nolead = "\\b" + par_as_regex_string_nolead + "\\b";
3682 opt.matchword = false;
3684 string regexp_str = "(" + lead_as_regex_string + ")()" + par_as_regex_string_nolead;
3685 string regexp2_str = "(" + lead_as_regex_string + ")(.*?)" + par_as_regex_string_nolead;
3686 CreateRegexp(opt, regexp_str, regexp2_str);
3688 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3689 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3693 if (!opt.ignoreformat) {
3694 lead_size = identifyLeading(par_as_string);
3695 LYXERR(Debug::FINDVERBOSE, "Lead_size: " << lead_size);
3696 lead_as_string = par_as_string.substr(0, lead_size);
3697 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3700 // Here we are using regexp
3701 LASSERT(use_regexp, /**/);
3703 string lead_as_regexp;
3704 if (lead_size > 0) {
3705 lead_as_regexp = string2regex(par_as_string.substr(0, lead_size));
3706 (void)regex_replace(par_as_string_nolead, par_as_string_nolead, "\\$$", "");
3707 (void)regex_replace(par_as_string_nolead, par_as_string_nolead, "}$", "");
3708 par_as_string = par_as_string_nolead;
3709 LYXERR(Debug::FINDVERBOSE, "lead_as_regexp is '" << lead_as_regexp << "'");
3710 LYXERR(Debug::FINDVERBOSE, "par_as_string now is '" << par_as_string << "'");
3712 // LYXERR(Debug::FINDVERBOSE, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
3713 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
3714 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
3715 // LYXERR(Debug::FINDVERBOSE, "par_as_string now is '" << par_as_string << "'");
3717 size_t lng = par_as_string.size();
3718 if (!opt.ignoreformat) {
3719 // Remove extra '\}' at end if not part of \{\.\}
3721 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
3723 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
3732 if (lng < par_as_string.size())
3733 par_as_string = par_as_string.substr(0,lng);
3735 LYXERR(Debug::FINDVERBOSE, "par_as_string after correctRegex is '" << par_as_string << "'");
3736 if ((lng > 0) && (par_as_string[0] == '^')) {
3737 par_as_string = par_as_string.substr(1);
3739 opt.matchAtStart = true;
3741 // LYXERR(Debug::FINDVERBOSE, "par_as_string now is '" << par_as_string << "'");
3742 // LYXERR(Debug::FINDVERBOSE, "Open braces: " << open_braces);
3743 // LYXERR(Debug::FINDVERBOSE, "Replaced text (to be used as regex): " << par_as_string);
3745 // If entered regexp must match at begin of searched string buffer
3746 // Kornel: Added parentheses to use $1 for size of the leading string
3750 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
3751 // Unfortunately is '\1', '\2', etc not working for strings with extra format
3752 // so the convert has no effect in that case
3753 for (int i = 7; i > 0; --i) {
3754 string orig = "\\\\" + std::to_string(i);
3755 string dest = "\\" + std::to_string(i+2);
3756 while (regex_replace(par_as_string, par_as_string, orig, dest));
3758 if (opt.matchword) {
3759 modifyRegexForMatchWord(par_as_string);
3760 opt.matchword = false;
3762 regexp_str = "(" + lead_as_regexp + ")()" + par_as_string;
3763 regexp2_str = "(" + lead_as_regexp + ")(.*?)" + par_as_string;
3765 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3766 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3767 CreateRegexp(opt, regexp_str, regexp2_str, par_as_string);
3771 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
3775 mres.searched_size = len;
3777 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
3780 docstring docstr = stringifyFromForSearch(opt, cur, len);
3782 str = normalize(docstr, opt.ignoreformat);
3783 if (!opt.ignoreformat) {
3784 str = correctlanguagesetting(str, false, !opt.ignoreformat);
3785 // remove closing '}' and '\n' to allow for use of '$' in regex
3786 size_t lng = str.size();
3787 while ((lng > 1) && ((str[lng -1] == '}') || (str[lng -1] == '\n')))
3789 if (lng != str.size()) {
3790 str = str.substr(0, lng);
3792 // Replace occurences of '~' to ' '
3793 static std::regex specialChars { R"(~)" };
3794 str = std::regex_replace(str, specialChars, R"( )" );
3797 mres.match_len = -1;
3800 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "After normalization: Matching against:\n'" << str << "'");
3802 LASSERT(use_regexp, /**/);
3804 // use_regexp always true
3805 LYXERR(Debug::FINDVERBOSE, "Searching in regexp mode: at_begin=" << at_begin);
3807 QString qstr = QString::fromStdString(str);
3808 QRegularExpression const *p_regexp;
3809 QRegularExpression::MatchType flags = QRegularExpression::NormalMatch;
3813 p_regexp = ®exp2;
3815 QRegularExpressionMatch match = p_regexp->match(qstr, 0, flags);
3816 if (!match.hasMatch())
3819 regex const *p_regexp;
3820 regex_constants::match_flag_type flags;
3822 flags = regex_constants::match_continuous;
3825 flags = regex_constants::match_default;
3826 p_regexp = ®exp2;
3828 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
3829 if (re_it == sregex_iterator())
3831 match_results<string::const_iterator> const & m = *re_it;
3833 // Whole found string, including the leading
3834 // std: m[0].second - m[0].first
3835 // Qt: match.capturedEnd(0) - match.capturedStart(0)
3837 // Size of the leading string
3838 // std: m[1].second - m[1].first
3839 // Qt: match.capturedEnd(1) - match.capturedStart(1)
3840 int leadingsize = 0;
3842 if (match.lastCapturedIndex() > 0) {
3843 leadingsize = match.capturedEnd(1) - match.capturedStart(1);
3848 leadingsize = m[1].second - m[1].first;
3852 mres.match_prefix = match.capturedEnd(2) - match.capturedStart(2);
3853 mres.match_len = match.capturedEnd(0) - match.capturedEnd(2);
3854 // because of different number of closing at end of string
3855 // we have to 'unify' the length of the post-match.
3856 // Done by ignoring closing parenthesis and linefeeds at string end
3857 int matchend = match.capturedEnd(0);
3858 size_t strsize = qstr.size();
3859 if (!opt.ignoreformat) {
3860 while (mres.match_len > 1) {
3861 QChar c = qstr.at(matchend - 1);
3862 if ((c == '\n') || (c == '}') || (c == '{')) {
3869 while (strsize > (size_t) match.capturedEnd(0)) {
3870 QChar c = qstr.at(strsize-1);
3871 if ((c == '\n') || (c == '}')) {
3878 // LYXERR0(qstr.toStdString());
3879 mres.match2end = strsize - matchend;
3880 mres.pos = match.capturedStart(2);
3882 mres.match_prefix = m[2].second - m[2].first;
3883 mres.match_len = m[0].second - m[2].second;
3884 // ignore closing parenthesis and linefeeds at string end
3885 size_t strend = m[0].second - m[0].first;
3886 int matchend = strend;
3887 size_t strsize = str.size();
3888 if (!opt.ignoreformat) {
3889 while (mres.match_len > 1) {
3890 char c = str.at(matchend - 1);
3891 if ((c == '\n') || (c == '}') || (c == '{')) {
3898 while (strsize > strend) {
3899 if ((str.at(strsize-1) == '}') || (str.at(strsize-1) == '\n')) {
3907 mres.match2end = strsize - matchend;
3908 mres.pos = m[2].first - m[0].first;;
3910 if (mres.match2end < 0)
3912 mres.leadsize = leadingsize;
3914 if (mres.match_len > 0) {
3915 string a0 = match.captured(0).mid(mres.pos + mres.match_prefix, mres.match_len).toStdString();
3916 mres.result.push_back(a0);
3917 for (int i = 3; i <= match.lastCapturedIndex(); i++) {
3918 mres.result.push_back(match.captured(i).toStdString());
3922 if (mres.match_len > 0) {
3923 string a0 = m[0].str().substr(mres.pos + mres.match_prefix, mres.match_len);
3924 mres.result.push_back(a0);
3925 for (size_t i = 3; i < m.size(); i++) {
3926 mres.result.push_back(m[i]);
3935 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
3937 MatchResult mres = findAux(cur, len, at_begin);
3938 int res = mres.match_len;
3939 LYXERR(Debug::FINDVERBOSE,
3940 "res=" << res << ", at_begin=" << at_begin
3941 << ", matchAtStart=" << opt.matchAtStart
3942 << ", inTexted=" << cur.inTexted());
3943 if (opt.matchAtStart) {
3946 else if (mres.match_prefix > 0)
3955 static bool simple_replace(string &t, string from, string to)
3957 regex repl("(\\\\)*(" + from + ")");
3961 for (sregex_iterator it(t.begin(), t.end(), repl), end; it != end; ++it) {
3963 if ((sub.position(2) - sub.position(0)) % 2 == 1)
3965 if (lastpos < (size_t) sub.position(2))
3966 s += t.substr(lastpos, sub.position(2) - lastpos);
3968 lastpos = sub.position(2) + sub.length(2);
3972 else if (lastpos < t.length())
3973 s += t.substr(lastpos, t.length() - lastpos);
3980 static string convertLF2Space(docstring const &s, bool ignore_format)
3982 // Using original docstring to handle '\n'
3984 if (s.size() == 0) return "";
3988 size_t end = s.size() - 1;
3989 while (s[start] == '\n' && start <= end)
3991 while (end >= start && s[end] == '\n')
3993 if (start >= end + 1)
3996 bool dospace = true;
3998 pos = s.find('\n', start);
4000 t << lyx::to_utf8(s.substr(start, end + 1 - start));
4003 if (!ignore_format) {
4004 if ((pos > start + 1) &&
4008 if ((pos > start + 2) &&
4009 (s[pos+1] == '~' || isSpace(s[pos+1]) ||
4010 s[pos-3] == '~' || isSpace(s[pos-3]))) {
4015 else if ((pos > start) &&
4018 while ((pos > start+skip) && (s[pos-1-skip] == '%'))
4020 if ((pos > start+skip) &&
4021 (s[pos+1] == '~' || isSpace(s[pos+1]) ||
4022 s[pos-1-skip] == '~' || isSpace(s[pos-1-skip]))) {
4023 // discard '%%%%%\n'
4032 t << lyx::to_utf8(s.substr(start, pos-skip-start));
4036 } while (start <= end);
4041 static string convertLF2Space(docstring const & s, bool ignore_format)
4043 // Using utf8-converted string to handle '\n'
4046 t = lyx::to_utf8(s);
4047 // Remove \n at begin
4048 while (!t.empty() && t[0] == '\n')
4050 // Remove \n* at end
4051 while (!t.empty() && t[t.size() - 1] == '\n') {
4052 t = t.substr(0, t.size() - 1);
4055 // Handle all other '\n'
4056 while ((pos = t.find("\n")) != string::npos) {
4057 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
4059 if (isPrintableNonspace(t[pos+1]) && ((pos < 3) || isPrintableNonspace(t[pos-3]))) {
4060 t.replace(pos-2, 3, " ");
4063 // Already a space there
4064 t.replace(pos-2, 3, "");
4068 if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
4069 // '\n' adjacent to non-alpha-numerics, discard
4070 t.replace(pos, 1, "");
4073 // Replace all other \n with spaces
4074 t.replace(pos, 1, " ");
4076 if (!ignore_format) {
4078 while ((pos > count + 1) && (t[pos - 1 -count] == '%')) {
4082 t.replace(pos - count, count, "");
4092 string MatchStringAdv::normalize(docstring const & s, bool ignore_format) const
4094 string t = convertLF2Space(s, ignore_format);
4096 // The following replaces are not appropriate in non-format-search mode
4097 if (!ignore_format) {
4098 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
4099 // Kornel: Added textsl, textsf, textit, texttt and noun
4100 // + allow to seach for colored text too
4101 LYXERR(Debug::FINDVERBOSE, "Removing stale empty macros from: " << t);
4102 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
4103 LYXERR(Debug::FINDVERBOSE, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
4104 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
4105 LYXERR(Debug::FINDVERBOSE, " further removing stale empty \\section{}, \\part{}, \\paragraph{} macros from: " << t);
4106 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
4111 docstring stringifyFromCursor(DocIterator const & cur, int len)
4113 LYXERR(Debug::FINDVERBOSE, "Stringifying with len=" << len << " from cursor at pos: " << cur);
4114 if (cur.inTexted()) {
4115 Paragraph const & par = cur.paragraph();
4116 // TODO what about searching beyond/across paragraph breaks ?
4117 // TODO Try adding a AS_STR_INSERTS as last arg
4118 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
4119 int(par.size()) : cur.pos() + len;
4120 // OutputParams runparams(&cur.buffer()->params().encoding());
4121 OutputParams runparams(encodings.fromLyXName("utf8"));
4122 runparams.nice = true;
4123 setFindParams(runparams);
4124 int option = AS_STR_INSETS | AS_STR_PLAINTEXT;
4125 if (ignoreFormats.getDeleted()) {
4126 option |= AS_STR_SKIPDELETE;
4127 runparams.find_set_feature(OutputParams::SearchWithoutDeleted);
4130 runparams.find_set_feature(OutputParams::SearchWithDeleted);
4132 if (ignoreFormats.getNonContent()) {
4133 runparams.find_add_feature(OutputParams::SearchNonOutput);
4135 LYXERR(Debug::FINDVERBOSE, "Stringifying with cur: "
4136 << cur << ", from pos: " << cur.pos() << ", end: " << end);
4137 docstring res = from_utf8(latexNamesToUtf8(par.asString(cur.pos(), end,
4139 &runparams), false));
4140 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Stringified text from pos(" << cur.pos() << ") len(" << len << "): " << res);
4142 } else if (cur.inMathed()) {
4143 CursorSlice cs = cur.top();
4144 MathData md = cs.cell();
4145 MathData::const_iterator it_end =
4146 (( len == -1 || cs.pos() + len > int(md.size()))
4148 : md.begin() + cs.pos() + len );
4150 for (MathData::const_iterator it = md.begin() + cs.pos(); it != it_end; ++it)
4152 docstring res = from_utf8(latexNamesToUtf8(asString(md2), false));
4153 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Stringified math from pos(" << cur.pos() << ") len(" << len << "): " << res);
4156 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Don't know how to stringify from here: " << cur);
4160 /** Computes the LaTeX export of buf starting from cur and ending len positions
4161 * after cur, if len is positive, or at the paragraph or innermost inset end
4164 docstring latexifyFromCursor(DocIterator const & cur, int len)
4167 LYXERR(Debug::FINDVERBOSE, "Latexifying with len=" << len << " from cursor at pos: " << cur);
4168 LYXERR(Debug::FINDVERBOSE, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
4169 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
4171 Buffer const & buf = *cur.buffer();
4173 odocstringstream ods;
4175 //OutputParams runparams(&buf.params().encoding());
4176 OutputParams runparams(encodings.fromLyXName("utf8"));
4177 runparams.nice = false;
4178 setFindParams(runparams);
4179 if (ignoreFormats.getDeleted()) {
4180 runparams.find_set_feature(OutputParams::SearchWithoutDeleted);
4183 runparams.find_set_feature(OutputParams::SearchWithDeleted);
4185 if (ignoreFormats.getNonContent()) {
4186 runparams.find_add_feature(OutputParams::SearchNonOutput);
4189 if (cur.inTexted()) {
4190 // @TODO what about searching beyond/across paragraph breaks ?
4191 pos_type endpos = cur.paragraph().size();
4192 if (len != -1 && endpos > cur.pos() + len)
4193 endpos = cur.pos() + len;
4194 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
4195 string(), cur.pos(), endpos, true);
4196 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Latexified text from pos(" << cur.pos() << ") len(" << len << "): " << ods.str());
4198 } else if (cur.inMathed()) {
4199 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
4200 for (int s = cur.depth() - 1; s >= 0; --s) {
4201 CursorSlice const & cs = cur[s];
4202 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
4203 TeXMathStream ws(os);
4204 cs.asInsetMath()->asHullInset()->header_write(ws);
4209 CursorSlice const & cs = cur.top();
4210 MathData md = cs.cell();
4211 MathData::const_iterator it_end =
4212 ((len == -1 || cs.pos() + len > int(md.size()))
4214 : md.begin() + cs.pos() + len);
4216 for (MathData::const_iterator it = md.begin() + cs.pos();
4220 ods << asString(md2);
4221 // Retrieve the math environment type, and add '$' or '$]'
4222 // or others (\end{equation}) accordingly
4223 for (int s = cur.depth() - 1; s >= 0; --s) {
4224 CursorSlice const & cs2 = cur[s];
4225 InsetMath * inset = cs2.asInsetMath();
4226 if (inset && inset->asHullInset()) {
4227 TeXMathStream ws(os);
4228 inset->asHullInset()->footer_write(ws);
4232 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Latexified math from pos(" << cur.pos() << ") len(" << len << "): " << ods.str());
4234 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Don't know how to stringify from here: " << cur);
4239 #if defined(ResultsDebug)
4241 static void displayMResult(MatchResult &mres, string from, DocIterator & cur)
4243 LYXERR0( "from:\t\t\t" << from);
4245 if (mres.pos_len > 0) {
4247 status = "FINALSEARCH";
4250 if (mres.match_len > 0) {
4251 if ((mres.match_prefix == 0) && (mres.pos == mres.leadsize))
4252 status = "Good Match";
4254 status = "Matched in";
4257 status = "MissedSearch";
4260 LYXERR0( status << "(" << cur.pos() << " ... " << mres.searched_size + cur.pos() << ") cur.lastpos(" << cur.lastpos() << ")");
4261 if ((mres.leadsize > 0) || (mres.match_len > 0) || (mres.match2end > 0))
4262 LYXERR0( "leadsize(" << mres.leadsize << ") match_len(" << mres.match_len << ") match2end(" << mres.match2end << ")");
4263 if ((mres.pos > 0) || (mres.match_prefix > 0))
4264 LYXERR0( "pos(" << mres.pos << ") match_prefix(" << mres.match_prefix << ")");
4265 for (size_t i = 0; i < mres.result.size(); i++)
4266 LYXERR0( "Match " << i << " = \"" << mres.result[i] << "\"");
4268 #define displayMres(s, txt, cur) displayMResult(s, txt, cur);
4270 #define displayMres(s, txt, cur)
4273 /** Finalize an advanced find operation, advancing the cursor to the innermost
4274 ** position that matches, plus computing the length of the matching text to
4276 ** Return the cur.pos() difference between start and end of found match
4278 MatchResult findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, MatchResult const & expected = MatchResult(-1))
4280 // Search the foremost position that matches (avoids find of entire math
4281 // inset when match at start of it)
4282 DocIterator old_cur(cur.buffer());
4284 static MatchResult fail = MatchResult();
4285 MatchResult max_match;
4286 // If (prefix_len > 0) means that forwarding 1 position will remove the complete entry
4287 // Happens with e.g. hyperlinks
4288 // either one sees "http://www.bla.bla" or nothing
4289 // so the search for "www" gives prefix_len = 7 (== sizeof("http://")
4290 // and although we search for only 3 chars, we find the whole hyperlink inset
4291 bool at_begin = (expected.match_prefix == 0);
4292 if (!match.opt.forward && match.opt.ignoreformat) {
4293 if (expected.pos > 0)
4296 LASSERT(at_begin, /**/);
4297 if (expected.match_len > 0 && at_begin) {
4298 // Search for deepest match
4300 max_match = expected;
4302 size_t d = cur.depth();
4306 if (cur.depth() < d)
4308 if (cur.depth() == d)
4311 while (cur && cur.depth() > lastd) {
4312 lastd = cur.depth();
4313 mres = match(cur, -1, at_begin);
4314 displayMres(mres, "Checking innermost", cur);
4315 if (mres.match_len > 0)
4320 if (mres.match_len < expected.match_len)
4328 // (expected.match_len <= 0)
4329 mres = match(cur); /* match valid only if not searching whole words */
4330 displayMres(mres, "Start with negative match", cur);
4333 if (max_match.match_len <= 0) return fail;
4334 LYXERR(Debug::FINDVERBOSE, "Ok");
4336 // Compute the match length
4338 if (cur.pos() + len > cur.lastpos())
4341 LASSERT(match.use_regexp, /**/);
4344 int maxl = cur.lastpos() - cur.pos();
4345 // Greedy behaviour while matching regexps
4346 while (maxl > minl) {
4348 mres2 = match(cur, len, at_begin);
4349 displayMres(mres2, "Finalize loop", cur);
4350 int actual_match_len = mres2.match_len;
4351 if (actual_match_len >= max_match.match_len) {
4352 // actual_match_len > max_match _can_ happen,
4353 // if the search area splits
4354 // some following word so that the regex
4355 // (e.g. 'r.*r\b' matches 'r' from the middle of the
4357 // This means, the len value is too big
4358 actual_match_len = max_match.match_len;
4360 max_match.match_len = actual_match_len;
4362 if (maxl - minl < 4)
4363 len = (maxl + minl)/2;
4365 len = minl + (maxl - minl + 3)/4;
4368 // (actual_match_len < max_match.match_len)
4370 len = (maxl + minl)/2;
4375 // Search for real start of matched characters
4377 MatchResult actual_match;
4380 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
4381 if (cur.depth() < old_cur.depth()) {
4383 LYXERR(Debug::INFO, "cur.depth() < old_cur.depth(), this should never happen");
4386 if (cur.pos() != old_cur.pos()) {
4387 // OK, forwarded 1 pos in actual inset
4388 actual_match = match(cur, len-1, at_begin);
4389 if (actual_match.match_len == max_match.match_len) {
4390 // Ha, got it! The shorter selection has the same match length
4393 max_match = actual_match;
4396 // OK, the shorter selection matches less chars, revert to previous value
4402 LYXERR(Debug::INFO, "cur.pos() == old_cur.pos(), this should never happen");
4403 actual_match = match(cur, len, at_begin);
4404 if (actual_match.match_len == max_match.match_len) {
4406 max_match = actual_match;
4413 max_match.pos_len = len;
4414 displayMres(max_match, "SEARCH RESULT", cur)
4421 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
4425 bool repeat = false;
4426 DocIterator orig_cur; // to be used if repeat not successful
4427 MatchResult orig_mres;
4428 while (!theApp()->longOperationCancelled() && cur) {
4429 //(void) findAdvForwardInnermost(cur);
4430 LYXERR(Debug::FINDVERBOSE, "findForwardAdv() cur: " << cur);
4431 MatchResult mres = match(cur, -1, false);
4432 string msg = "Starting";
4435 displayMres(mres, msg + " findForwardAdv", cur)
4436 int match_len = mres.match_len;
4437 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
4438 LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
4441 if (match_len <= 0) {
4442 // This should exit nested insets, if any, or otherwise undefine the currsor.
4443 cur.pos() = cur.lastpos();
4444 LYXERR(Debug::FINDVERBOSE, "Advancing pos: cur=" << cur);
4447 else { // match_len > 0
4448 // Try to find the begin of searched string
4450 int firstInvalid = cur.lastpos() - cur.pos();
4452 int incrmatch = (mres.match_prefix + mres.pos - mres.leadsize + 1)*3/4;
4453 int incrcur = (firstInvalid + 1 )*3/4;
4454 if (incrcur < incrmatch)
4455 increment = incrcur;
4457 increment = incrmatch;
4461 LYXERR(Debug::FINDVERBOSE, "Set increment to " << increment);
4462 while (increment > 0) {
4463 DocIterator old_cur = cur;
4464 if (cur.pos() + increment >= cur.lastpos()) {
4468 cur.pos() = cur.pos() + increment;
4469 MatchResult mres2 = match(cur, -1, false);
4470 displayMres(mres2, "findForwardAdv loop", cur)
4471 switch (interpretMatch(mres, mres2)) {
4472 case MatchResult::newIsTooFar:
4473 // behind the expected match
4474 firstInvalid = increment;
4478 case MatchResult::newIsBetter:
4479 // not reached yet, but cur.pos()+increment is bettert
4481 firstInvalid -= increment;
4482 if (increment > firstInvalid*3/4)
4483 increment = firstInvalid*3/4;
4484 if ((mres2.pos == mres2.leadsize) && (increment >= mres2.match_prefix)) {
4485 if (increment >= mres2.match_prefix)
4486 increment = (mres2.match_prefix+1)*3/4;
4491 // Handle not like MatchResult::newIsTooFar
4492 LYXERR0( "Probably too far: Increment = " << increment << " match_prefix = " << mres.match_prefix);
4494 increment = increment*3/4;
4499 if (mres.match_len > 0) {
4500 if (mres.match_prefix + mres.pos - mres.leadsize > 0) {
4501 // The match seems to indicate some deeper level
4510 // should never be reached.
4514 // LYXERR0("Leaving first loop");
4515 LYXERR(Debug::FINDVERBOSE, "Finalizing 1");
4516 MatchResult found_match = findAdvFinalize(cur, match, mres);
4517 if (found_match.match_len > 0) {
4518 match.FillResults(found_match);
4519 return found_match.pos_len;
4522 // try next possible match
4533 /// Find the most backward consecutive match within same paragraph while searching backwards.
4534 MatchResult findMostBackwards(DocIterator & cur, MatchStringAdv const & match, MatchResult &expected)
4536 DocIterator cur_begin = cur;
4537 cur_begin.pos() = 0;
4538 DocIterator tmp_cur = cur;
4539 MatchResult mr = findAdvFinalize(tmp_cur, match, expected);
4540 Inset & inset = cur.inset();
4541 for (; cur != cur_begin; cur.backwardPos()) {
4542 LYXERR(Debug::FINDVERBOSE, "findMostBackwards(): cur=" << cur);
4543 DocIterator new_cur = cur;
4544 new_cur.backwardPos();
4545 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
4547 MatchResult new_mr = findAdvFinalize(new_cur, match, expected);
4548 if (new_mr.match_len == mr.match_len)
4552 LYXERR(Debug::FINDVERBOSE, "findMostBackwards(): exiting with cur=" << cur);
4558 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
4562 // Backup of original position
4563 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
4564 if (cur == cur_begin)
4567 DocIterator cur_orig(cur);
4568 bool pit_changed = false;
4571 MatchResult found_match = match(cur, -1, false);
4573 if (found_match.match_len > 0) {
4575 cur.pos() = cur.lastpos();
4577 cur.pos() = cur_orig.pos();
4578 LYXERR(Debug::FINDVERBOSE, "findBackAdv2: cur: " << cur);
4579 DocIterator cur_prev_iter;
4581 found_match = match(cur);
4582 LYXERR(Debug::FINDVERBOSE, "findBackAdv3: found_match="
4583 << (found_match.match_len > 0) << ", cur: " << cur);
4584 if (found_match.match_len > 0) {
4585 MatchResult found_mr = findMostBackwards(cur, match, found_match);
4586 if (found_mr.pos_len > 0) {
4587 match.FillResults(found_mr);
4588 return found_mr.pos_len;
4592 // Stop if begin of document reached
4593 if (cur == cur_begin)
4595 cur_prev_iter = cur;
4599 if (cur == cur_begin)
4606 } while (!theApp()->longOperationCancelled());
4614 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
4615 DocIterator const & cur, int len)
4617 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
4619 if (!opt.ignoreformat)
4620 return latexifyFromCursor(cur, len);
4622 return stringifyFromCursor(cur, len);
4626 FindAndReplaceOptions::FindAndReplaceOptions(
4627 docstring const & _find_buf_name, bool _casesensitive,
4628 bool _matchword, bool _forward, bool _expandmacros, bool _ignoreformat,
4629 docstring const & _repl_buf_name, bool _keep_case,
4630 SearchScope _scope, SearchRestriction _restr, bool _replace_all)
4631 : find_buf_name(_find_buf_name), casesensitive(_casesensitive), matchword(_matchword),
4632 forward(_forward), expandmacros(_expandmacros), ignoreformat(_ignoreformat),
4633 repl_buf_name(_repl_buf_name), keep_case(_keep_case), scope(_scope), restr(_restr), replace_all(_replace_all)
4641 /** Check if 'len' letters following cursor are all non-lowercase */
4642 static bool allNonLowercase(Cursor const & cur, int len)
4644 pos_type beg_pos = cur.selectionBegin().pos();
4645 pos_type end_pos = cur.selectionBegin().pos() + len;
4646 if (len > cur.lastpos() + 1 - beg_pos) {
4647 LYXERR(Debug::FINDVERBOSE, "This should not happen, more debug needed");
4648 len = cur.lastpos() + 1 - beg_pos;
4649 end_pos = beg_pos + len;
4651 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
4652 if (isLowerCase(cur.paragraph().getChar(pos)))
4658 /** Check if first letter is upper case and second one is lower case */
4659 static bool firstUppercase(Cursor const & cur)
4662 pos_type pos = cur.selectionBegin().pos();
4663 if (pos >= cur.lastpos() - 1) {
4664 LYXERR(Debug::FINDVERBOSE, "No upper-case at cur: " << cur);
4667 ch1 = cur.paragraph().getChar(pos);
4668 ch2 = cur.paragraph().getChar(pos + 1);
4669 bool result = isUpperCase(ch1) && isLowerCase(ch2);
4670 LYXERR(Debug::FINDVERBOSE, "firstUppercase(): "
4671 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
4672 << ch2 << "(" << char(ch2) << ")"
4673 << ", result=" << result << ", cur=" << cur);
4678 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
4680 ** \fixme What to do with possible further paragraphs in replace buffer ?
4682 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
4684 ParagraphList::iterator pit = buffer.paragraphs().begin();
4685 LASSERT(!pit->empty(), /**/);
4686 pos_type right = pos_type(1);
4687 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
4688 right = pit->size();
4689 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
4693 static bool replaceMatches(string &t, int maxmatchnum, vector <string> const & replacements)
4695 // Should replace the string "$" + std::to_string(matchnum) with replacement
4696 // if the char '$' is not prefixed with odd number of char '\\'
4697 static regex const rematch("(\\\\)*(\\$\\$([0-9]))");
4701 for (sregex_iterator it(t.begin(), t.end(), rematch), end; it != end; ++it) {
4703 if ((sub.position(2) - sub.position(0)) % 2 == 1)
4705 int num = stoi(sub.str(3), nullptr, 10);
4706 if (num >= maxmatchnum)
4708 if (lastpos < (size_t) sub.position(2))
4709 s += t.substr(lastpos, sub.position(2) - lastpos);
4710 s += replacements[num];
4711 lastpos = sub.position(2) + sub.length(2);
4715 else if (lastpos < t.length())
4716 s += t.substr(lastpos, t.length() - lastpos);
4722 static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
4724 Cursor & cur = bv->cursor();
4725 if (opt.repl_buf_name.empty()
4726 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
4727 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4730 DocIterator sel_beg = cur.selectionBegin();
4731 DocIterator sel_end = cur.selectionEnd();
4732 if (&sel_beg.inset() != &sel_end.inset()
4733 || sel_beg.pit() != sel_end.pit()
4734 || sel_beg.idx() != sel_end.idx())
4736 int sel_len = sel_end.pos() - sel_beg.pos();
4737 LYXERR(Debug::FINDVERBOSE, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
4738 << ", sel_len: " << sel_len << endl);
4741 LASSERT(sel_len > 0, return 0);
4743 if (!matchAdv(sel_beg, sel_len).match_len)
4746 // Build a copy of the replace buffer, adapted to the KeepCase option
4747 Buffer const & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
4749 repl_buffer_orig.write(oss);
4750 string lyx = oss.str();
4751 if (matchAdv.valid_matches > 0)
4752 replaceMatches(lyx, matchAdv.valid_matches, matchAdv.matches);
4753 Buffer repl_buffer(string(), false);
4754 repl_buffer.setInternal(true);
4755 repl_buffer.setUnnamed(true);
4756 LASSERT(repl_buffer.readString(lyx), return 0);
4757 if (opt.keep_case && sel_len >= 2) {
4758 LYXERR(Debug::FINDVERBOSE, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
4759 if (cur.inTexted()) {
4760 if (firstUppercase(cur))
4761 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
4762 else if (allNonLowercase(cur, sel_len))
4763 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
4766 cap::cutSelection(cur, false);
4767 if (cur.inTexted()) {
4768 repl_buffer.changeLanguage(
4769 repl_buffer.language(),
4770 cur.getFont().language());
4771 LYXERR(Debug::FINDVERBOSE, "Replacing by pasteParagraphList()ing repl_buffer");
4772 LYXERR(Debug::FINDVERBOSE, "Before pasteParagraphList() cur=" << cur << endl);
4773 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
4774 repl_buffer.params().documentClassPtr(),
4775 repl_buffer.params().authors(),
4776 bv->buffer().errorList("Paste"));
4777 LYXERR(Debug::FINDVERBOSE, "After pasteParagraphList() cur=" << cur << endl);
4778 sel_len = repl_buffer.paragraphs().begin()->size();
4779 } else if (cur.inMathed()) {
4780 odocstringstream ods;
4782 // OutputParams runparams(&repl_buffer.params().encoding());
4783 OutputParams runparams(encodings.fromLyXName("utf8"));
4784 runparams.nice = false;
4785 setFindParams(runparams);
4786 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams, string(), -1, -1, true);
4787 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
4788 docstring repl_latex = ods.str();
4789 LYXERR(Debug::FINDVERBOSE, "Latexified replace_buffer: '" << repl_latex << "'");
4791 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
4792 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
4793 repl_latex = from_utf8(s);
4794 LYXERR(Debug::FINDVERBOSE, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
4795 MathData ar(cur.buffer());
4796 asArray(repl_latex, ar, Parse::NORMAL);
4798 sel_len = ar.size();
4799 LYXERR(Debug::FINDVERBOSE, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4801 if (cur.pos() >= sel_len)
4802 cur.pos() -= sel_len;
4805 LYXERR(Debug::FINDVERBOSE, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4806 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
4807 bv->processUpdateFlags(Update::Force);
4812 /// Perform a FindAdv operation.
4813 bool findAdv(BufferView * bv, FindAndReplaceOptions & opt)
4818 // e.g., when invoking word-findadv from mini-buffer wither with
4819 // wrong options syntax or before ever opening advanced F&R pane
4820 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4824 MatchStringAdv matchAdv(bv->buffer(), opt);
4826 if (!matchAdv.regexIsValid) {
4827 bv->message(lyx::from_utf8(matchAdv.regexError));
4831 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
4833 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
4834 num_replaced += findAdvReplace(bv, opt, matchAdv);
4837 pos_len = findForwardAdv(cur, matchAdv);
4839 pos_len = findBackwardsAdv(cur, matchAdv);
4840 } catch (exception & ex) {
4841 bv->message(from_utf8(ex.what()));
4846 if (num_replaced > 0) {
4847 switch (num_replaced)
4850 bv->message(_("One match has been replaced."));
4853 bv->message(_("Two matches have been replaced."));
4856 bv->message(bformat(_("%1$d matches have been replaced."), num_replaced));
4862 bv->message(_("Match not found."));
4867 if (num_replaced > 0)
4868 bv->message(_("Match has been replaced."));
4870 bv->message(_("Match found."));
4872 if (cur.pos() + pos_len > cur.lastpos()) {
4873 // Prevent crash in bv->putSelectionAt()
4874 // Should never happen, maybe LASSERT() here?
4875 pos_len = cur.lastpos() - cur.pos();
4877 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Putting selection at cur=" << cur << " with len: " << pos_len);
4878 bv->putSelectionAt(cur, pos_len, !opt.forward);
4884 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
4886 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
4887 << opt.casesensitive << ' '
4888 << opt.matchword << ' '
4889 << opt.forward << ' '
4890 << opt.expandmacros << ' '
4891 << opt.ignoreformat << ' '
4892 << opt.replace_all << ' '
4893 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
4894 << opt.keep_case << ' '
4895 << int(opt.scope) << ' '
4898 LYXERR(Debug::FINDVERBOSE, "built: " << os.str());
4904 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
4906 // LYXERR(Debug::FINDVERBOSE, "parsing");
4910 while (line != "EOSS") {
4914 if (is.eof()) // Tolerate malformed request
4918 // LYXERR(Debug::FINDVERBOSE, "file_buf_name: '" << s << "'");
4919 opt.find_buf_name = from_utf8(s);
4920 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.replace_all;
4921 is.get(); // Waste space before replace string
4924 while (line != "EOSS") {
4928 if (is.eof()) // Tolerate malformed request
4932 // LYXERR(Debug::FINDVERBOSE, "repl_buf_name: '" << s << "'");
4933 opt.repl_buf_name = from_utf8(s);
4934 is >> opt.keep_case;
4937 opt.scope = FindAndReplaceOptions::SearchScope(i);
4939 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
4942 LYXERR(Debug::FINDVERBOSE, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
4943 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
4944 << opt.scope << ' ' << opt.restr);