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 getUnderline() const { return ignoreUnderline_; }
93 bool getMarkUp() const { return ignoreMarkUp_; }
95 bool getStrikeOut() const { return ignoreStrikeOut_; }
97 bool getSectioning() const { return ignoreSectioning_; }
99 bool getFrontMatter() const { return ignoreFrontMatter_; }
101 bool getColor() const { return ignoreColor_; }
103 bool getLanguage() const { return ignoreLanguage_; }
105 bool getDeleted() const { return ignoreDeleted_; }
107 void setIgnoreDeleted(bool value);
109 bool getNonContent() const { return searchNonContent_; }
111 void setIgnoreFormat(string const & type, bool value, bool fromUser = true);
115 bool ignoreFamily_ = false;
117 bool ignoreSeries_ = false;
119 bool ignoreShape_ = false;
121 bool ignoreUnderline_ = false;
123 bool ignoreMarkUp_ = false;
125 bool ignoreStrikeOut_ = false;
127 bool ignoreSectioning_ = false;
129 bool ignoreFrontMatter_ = false;
131 bool ignoreColor_ = false;
133 bool ignoreLanguage_ = false;
134 bool userSelectedIgnoreLanguage_ = false;
136 bool ignoreDeleted_ = true;
138 bool searchNonContent_ = true;
141 void IgnoreFormats::setIgnoreFormat(string const & type, bool value, bool fromUser)
143 if (type == "color") {
144 ignoreColor_ = value;
146 else if (type == "language") {
148 userSelectedIgnoreLanguage_ = value;
149 ignoreLanguage_ = value;
152 ignoreLanguage_ = (value || userSelectedIgnoreLanguage_);
154 else if (type == "sectioning") {
155 ignoreSectioning_ = value;
156 ignoreFrontMatter_ = value;
158 else if (type == "font") {
159 ignoreSeries_ = value;
160 ignoreShape_ = value;
161 ignoreFamily_ = value;
163 else if (type == "series") {
164 ignoreSeries_ = value;
166 else if (type == "shape") {
167 ignoreShape_ = value;
169 else if (type == "family") {
170 ignoreFamily_ = value;
172 else if (type == "markup") {
173 ignoreMarkUp_ = value;
175 else if (type == "underline") {
176 ignoreUnderline_ = value;
178 else if (type == "strike") {
179 ignoreStrikeOut_ = value;
181 else if (type == "deleted") {
182 ignoreDeleted_ = value;
184 else if (type == "non-output-content") {
185 searchNonContent_ = !value;
189 // The global variable that can be changed from outside
190 IgnoreFormats ignoreFormats;
193 void setIgnoreFormat(string const & type, bool value, bool fromUser)
195 ignoreFormats.setIgnoreFormat(type, value, fromUser);
201 bool parse_bool(docstring & howto, bool const defvalue = false)
206 howto = split(howto, var, ' ');
214 MatchString(docstring const & s, bool cs, bool mw)
215 : str(s), case_sens(cs), whole_words(mw)
218 // returns true if the specified string is at the specified position
219 // del specifies whether deleted strings in ct mode will be considered
220 int operator()(Paragraph const & par, pos_type pos, bool del = true) const
222 return par.find(str, case_sens, whole_words, pos, del);
230 // match whole words only
235 int findForward(DocIterator & cur, DocIterator const endcur,
236 MatchString const & match,
237 bool find_del = true, bool onlysel = false)
239 for (; cur; cur.forwardChar()) {
240 if (onlysel && endcur.pit() == cur.pit()
241 && endcur.idx() == cur.idx() && endcur.pos() < cur.pos())
243 if (cur.inTexted()) {
244 int len = match(cur.paragraph(), cur.pos(), find_del);
253 int findBackwards(DocIterator & cur, DocIterator const endcur,
254 MatchString const & match,
255 bool find_del = true, bool onlysel = false)
259 if (onlysel && endcur.pit() == cur.pit()
260 && endcur.idx() == cur.idx() && endcur.pos() > cur.pos())
262 if (cur.inTexted()) {
263 int len = match(cur.paragraph(), cur.pos(), find_del);
272 bool searchAllowed(docstring const & str)
275 frontend::Alert::error(_("Search error"), _("Search string is empty"));
284 bool findOne(BufferView * bv, docstring const & searchstr,
285 bool case_sens, bool whole, bool forward,
286 bool find_del, bool check_wrap, bool const auto_wrap,
287 bool instant, bool onlysel)
289 // Clean up previous selections with empty searchstr on instant
290 if (searchstr.empty() && instant) {
291 if (bv->cursor().selection()) {
292 bv->setCursor(bv->cursor().selectionBegin());
293 bv->clearSelection();
298 if (!searchAllowed(searchstr))
301 DocIterator const endcur = forward ? bv->cursor().selectionEnd() : bv->cursor().selectionBegin();
303 if (onlysel && bv->cursor().selection()) {
304 docstring const matchstring = bv->cursor().selectionAsString(false);
305 docstring const lcmatchsting = support::lowercase(matchstring);
306 if (matchstring == searchstr || (!case_sens && lcmatchsting == lowercase(searchstr))) {
307 docstring q = _("The search string matches the selection, and search is limited to selection.\n"
308 "Continue search outside?");
309 int search_answer = frontend::Alert::prompt(_("Search outside selection?"),
310 q, 0, 1, _("&Yes"), _("&No"));
311 if (search_answer == 0) {
312 bv->clearSelection();
313 if (findOne(bv, searchstr, case_sens, whole, forward,
314 find_del, check_wrap, auto_wrap, false, false))
321 DocIterator cur = forward
322 ? ((instant || onlysel) ? bv->cursor().selectionBegin() : bv->cursor().selectionEnd())
323 : ((instant || onlysel) ? bv->cursor().selectionEnd() : bv->cursor().selectionBegin());
325 MatchString const match(searchstr, case_sens, whole);
327 int match_len = forward
328 ? findForward(cur, endcur, match, find_del, onlysel)
329 : findBackwards(cur, endcur, match, find_del, onlysel);
332 bv->putSelectionAt(cur, match_len, !forward);
334 docstring q = _("The search string was not found within the selection.\n"
335 "Continue search outside?");
336 int search_answer = frontend::Alert::prompt(_("Search outside selection?"),
337 q, 0, 1, _("&Yes"), _("&No"));
338 if (search_answer == 0) {
339 bv->clearSelection();
340 if (findOne(bv, searchstr, case_sens, whole, forward,
341 find_del, check_wrap, auto_wrap, false, false))
346 else if (check_wrap) {
347 DocIterator cur_orig(bv->cursor());
348 bool wrap = auto_wrap;
352 q = _("End of file reached while searching forward.\n"
353 "Continue searching from the beginning?");
355 q = _("Beginning of file reached while searching backward.\n"
356 "Continue searching from the end?");
357 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
358 q, 0, 1, _("&Yes"), _("&No"));
359 wrap = wrap_answer == 0;
363 bv->cursor().clear();
364 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
366 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
367 bv->cursor().backwardPos();
370 docstring const msg = forward
371 ? _("Search reached end of document, continuing from beginning.")
372 : _("Search reached beginning of document, continuing from end.");
375 bv->clearSelection();
376 if (findOne(bv, searchstr, case_sens, whole, forward,
377 find_del, false, false, false, false))
380 bv->cursor().setCursor(cur_orig);
384 return match_len > 0;
390 int replaceAll(BufferView * bv,
391 docstring const & searchstr, docstring const & replacestr,
392 bool case_sens, bool whole, bool onlysel)
394 Buffer & buf = bv->buffer();
396 if (!searchAllowed(searchstr) || buf.isReadonly())
399 DocIterator startcur = bv->cursor().selectionBegin();
400 DocIterator endcur = bv->cursor().selectionEnd();
401 bool const had_selection = bv->cursor().selection();
403 MatchString const match(searchstr, case_sens, whole);
406 int const rsize = replacestr.size();
407 int const ssize = searchstr.size();
410 cur.setCursor(doc_iterator_begin(&buf));
411 int match_len = findForward(cur, endcur, match, false, onlysel);
412 while (match_len > 0) {
413 // Backup current cursor position and font.
414 pos_type const pos = cur.pos();
415 Font const font = cur.paragraph().getFontSettings(buf.params(), pos);
417 int ct_deleted_text = ssize -
418 cur.paragraph().eraseChars(pos, pos + match_len,
419 buf.params().track_changes);
420 cur.paragraph().insert(pos, replacestr, font,
421 Change(buf.params().track_changes
423 : Change::UNCHANGED));
424 for (int i = 0; i < rsize + ct_deleted_text
425 && cur.pos() < cur.lastpos(); ++i)
427 if (onlysel && cur.pit() == endcur.pit() && cur.idx() == endcur.idx()) {
428 // Adjust end of selection for replace-all in selection
430 int const offset = rsize - ssize;
431 for (int i = 0; i < offset + ct_deleted_text
432 && endcur.pos() < endcur.lastpos(); ++i)
435 int const offset = ssize - rsize;
436 for (int i = 0; i < offset && endcur.pos() > 0; ++i)
437 endcur.backwardPos();
438 for (int i = 0; i < ct_deleted_text
439 && endcur.pos() < endcur.lastpos(); ++i)
444 match_len = findForward(cur, endcur, match, false, onlysel);
447 bv->putSelectionAt(doc_iterator_begin(&buf), 0, false);
449 startcur.fixIfBroken();
450 bv->setCursor(startcur);
452 // Reset selection, accounting for changes in selection
454 endcur.fixIfBroken();
455 bv->cursor().resetAnchor();
456 bv->setCursorSelectionTo(endcur);
463 // the idea here is that we are going to replace the string that
464 // is selected IF it is the search string.
465 // if there is a selection, but it is not the search string, then
466 // we basically ignore it. (FIXME We ought to replace only within
468 // if there is no selection, then:
469 // (i) if some search string has been provided, then we find it.
470 // (think of how the dialog works when you hit "replace" the
472 // (ii) if no search string has been provided, then we treat the
473 // word the cursor is in as the search string. (why? i have no
474 // idea.) but this only works in text?
476 // returns the number of replacements made (one, if any) and
477 // whether anything at all was done.
478 pair<bool, int> replaceOne(BufferView * bv, docstring searchstr,
479 docstring const & replacestr, bool case_sens,
480 bool whole, bool forward, bool findnext, bool wrap,
483 Cursor & cur = bv->cursor();
484 if (!cur.selection() || onlysel) {
485 // no selection, non-empty search string: find it
486 if (!searchstr.empty()) {
487 bool const found = findOne(bv, searchstr, case_sens, whole,
488 forward, true, findnext, wrap, false, onlysel);
489 return make_pair(found, 0);
491 // empty search string
494 return make_pair(false, 0);
495 // select current word and treat it as the search string.
496 // This causes a minor bug as undo will restore this selection,
497 // which the user did not create (#8986).
498 cur.innerText()->selectWord(cur, WHOLE_WORD);
499 searchstr = cur.selectionAsString(false, true);
502 // if we still don't have a search string, report the error
504 if (!searchAllowed(searchstr))
505 return make_pair(false, 0);
507 bool have_selection = cur.selection();
508 docstring const selected = cur.selectionAsString(false, true);
511 ? searchstr == selected
512 : compare_no_case(searchstr, selected) == 0;
514 // no selection or current selection is not search word:
515 // just find the search word
516 if (!have_selection || !match) {
517 bool const found = findOne(bv, searchstr, case_sens, whole, forward,
518 true, findnext, wrap, false, onlysel);
519 return make_pair(found, 0);
522 // we're now actually ready to replace. if the buffer is
523 // read-only, we can't, though.
524 if (bv->buffer().isReadonly())
525 return make_pair(false, 0);
527 cap::replaceSelectionWithString(cur, replacestr);
529 cur.pos() += replacestr.length();
530 LASSERT(cur.pos() <= cur.lastpos(),
531 cur.pos() = cur.lastpos());
534 findOne(bv, searchstr, case_sens, whole,
535 forward, false, findnext, wrap, false, onlysel);
537 return make_pair(true, 1);
543 docstring const find2string(docstring const & search,
544 bool casesensitive, bool matchword,
545 bool forward, bool wrap, bool instant,
550 << int(casesensitive) << ' '
551 << int(matchword) << ' '
552 << int(forward) << ' '
554 << int(instant) << ' '
560 docstring const replace2string(docstring const & replace,
561 docstring const & search,
562 bool casesensitive, bool matchword,
563 bool all, bool forward, bool findnext,
564 bool wrap, bool onlysel)
567 ss << replace << '\n'
569 << int(casesensitive) << ' '
570 << int(matchword) << ' '
572 << int(forward) << ' '
573 << int(findnext) << ' '
580 docstring const string2find(docstring const & argument,
588 // data is of the form
590 // <casesensitive> <matchword> <forward> <wrap> <onlysel>"
592 docstring howto = split(argument, search, '\n');
594 casesensitive = parse_bool(howto);
595 matchword = parse_bool(howto);
596 forward = parse_bool(howto, true);
597 wrap = parse_bool(howto);
598 instant = parse_bool(howto);
599 onlysel = parse_bool(howto);
605 bool lyxfind(BufferView * bv, FuncRequest const & ev)
607 if (!bv || ev.action() != LFUN_WORD_FIND)
610 //lyxerr << "find called, cmd: " << ev << endl;
618 docstring search = string2find(ev.argument(), casesensitive,
619 matchword, forward, wrap, instant, onlysel);
621 return findOne(bv, search, casesensitive, matchword, forward,
622 false, true, wrap, instant, onlysel);
626 bool lyxreplace(BufferView * bv, FuncRequest const & ev)
628 if (!bv || ev.action() != LFUN_WORD_REPLACE)
631 // data is of the form
634 // <casesensitive> <matchword> <all> <forward> <findnext> <wrap> <onlysel>"
637 docstring howto = split(ev.argument(), rplc, '\n');
638 howto = split(howto, search, '\n');
640 bool casesensitive = parse_bool(howto);
641 bool matchword = parse_bool(howto);
642 bool all = parse_bool(howto);
643 bool forward = parse_bool(howto, true);
644 bool findnext = parse_bool(howto, true);
645 bool wrap = parse_bool(howto);
646 bool onlysel = parse_bool(howto);
648 if (!bv->cursor().selection())
649 // only selection only makes sense with selection
654 int replace_count = 0;
656 replace_count = replaceAll(bv, search, rplc, casesensitive,
658 update = replace_count > 0;
661 replaceOne(bv, search, rplc, casesensitive, matchword,
662 forward, findnext, wrap, onlysel);
664 replace_count = rv.second;
667 Buffer const & buf = bv->buffer();
669 // emit message signal.
671 buf.message(_("String not found in selection."));
673 buf.message(_("String not found."));
675 if (replace_count == 0) {
676 buf.message(_("String found."));
677 } else if (replace_count == 1) {
678 buf.message(_("String has been replaced."));
680 docstring const str = onlysel
681 ? bformat(_("%1$d strings have been replaced in the selection."), replace_count)
682 : bformat(_("%1$d strings have been replaced."), replace_count);
690 bool findNextChange(BufferView * bv, Cursor & cur, bool const check_wrap)
692 for (; cur; cur.forwardPos())
693 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
697 DocIterator cur_orig(bv->cursor());
698 docstring q = _("End of file reached while searching forward.\n"
699 "Continue searching from the beginning?");
700 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
701 q, 0, 1, _("&Yes"), _("&No"));
702 if (wrap_answer == 0) {
703 bv->cursor().clear();
704 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
705 bv->clearSelection();
706 cur.setCursor(bv->cursor().selectionBegin());
707 if (findNextChange(bv, cur, false))
710 bv->cursor().setCursor(cur_orig);
717 bool findPreviousChange(BufferView * bv, Cursor & cur, bool const check_wrap)
719 for (cur.backwardPos(); cur; cur.backwardPos()) {
720 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
725 DocIterator cur_orig(bv->cursor());
726 docstring q = _("Beginning of file reached while searching backward.\n"
727 "Continue searching from the end?");
728 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
729 q, 0, 1, _("&Yes"), _("&No"));
730 if (wrap_answer == 0) {
731 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
732 bv->cursor().backwardPos();
733 bv->clearSelection();
734 cur.setCursor(bv->cursor().selectionBegin());
735 if (findPreviousChange(bv, cur, false))
738 bv->cursor().setCursor(cur_orig);
745 bool selectChange(Cursor & cur, bool forward)
747 if (!cur.inTexted() || !cur.paragraph().isChanged(cur.pos()))
749 Change ch = cur.paragraph().lookupChange(cur.pos());
751 CursorSlice tip1 = cur.top();
752 for (; tip1.pit() < tip1.lastpit() || tip1.pos() < tip1.lastpos(); tip1.forwardPos()) {
753 Change ch2 = tip1.paragraph().lookupChange(tip1.pos());
754 if (!ch2.isSimilarTo(ch))
757 CursorSlice tip2 = cur.top();
758 for (; tip2.pit() > 0 || tip2.pos() > 0;) {
760 Change ch2 = tip2.paragraph().lookupChange(tip2.pos());
761 if (!ch2.isSimilarTo(ch)) {
762 // take a step forward to correctly set the selection
770 cur.bv().mouseSetCursor(cur, false);
772 cur.bv().mouseSetCursor(cur, true);
780 bool findChange(BufferView * bv, bool forward)
783 cur.setCursor(forward ? bv->cursor().selectionEnd()
784 : bv->cursor().selectionBegin());
785 forward ? findNextChange(bv, cur, true) : findPreviousChange(bv, cur, true);
786 return selectChange(cur, forward);
791 bool findNextChange(BufferView * bv)
793 return findChange(bv, true);
797 bool findPreviousChange(BufferView * bv)
799 return findChange(bv, false);
806 typedef vector<pair<string, string> > Escapes;
808 string string2regex(string in)
810 static std::regex specialChars { R"([-[\]{}()*+?.,\^$|#\s\$\\])" };
811 string temp = std::regex_replace(in, specialChars, R"(\$&)" );
816 while (fl_pos < temp.size()) {
817 fl_pos = temp.find("\\\\foreignlanguage", lastpos + offset);
818 if (fl_pos == string::npos)
821 temp2 += temp.substr(lastpos, fl_pos - lastpos);
827 if (lastpos < temp.size()) {
828 temp2 += temp.substr(lastpos, temp.size() - lastpos);
833 static void buildAccentsMap();
835 string correctRegex(string t, bool withformat)
837 /* Convert \backslash => \
838 * and \{, \}, \[, \] => {, }, [, ]
841 regex wordre("(\\\\)*(\\\\(([A-Za-z]+|[\\{\\}%])( |\\{\\})?|[\\[\\]\\{\\}]))");
844 bool backslashed = false;
848 //LYXERR0("correctRegex input '" << t << "'");
849 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
852 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
856 if (sub.str(4) == "backslash") {
859 // transforms '\backslash \{' into '\{'
860 string next = t.substr(sub.position(2) + sub.str(2).length(), 2);
861 if ((next == "\\{") || (next == "\\}")) {
867 else if (sub.str(4) == "mathcircumflex")
869 else if (backslashed) {
872 if (sub.str(3) == "{")
873 replace = accents["braceleft"];
874 else if (sub.str(3) == "}")
875 replace = accents["braceright"];
877 // else part should not exist
882 if (sub.str(3) == "{")
884 else if (sub.str(3) == "}")
887 // else part should not exist
892 else if (sub.str(4) == "{") // transforms '\{' into '{'
894 else if (sub.str(4) == "}")
896 else if (sub.str(4) == "%")
899 AccentsIterator it_ac = accents.find(sub.str(4));
900 if (it_ac == accents.end()) {
901 replace = sub.str(2);
904 replace = it_ac->second;
908 if (lastpos < (size_t) sub.position(2))
909 s += t.substr(lastpos, sub.position(2) - lastpos);
911 lastpos = sub.position(2) + sub.length(2);
915 else if (lastpos < t.length())
916 s += t.substr(lastpos, t.length() - lastpos);
917 // Handle quotes in regex
918 // substitute all '„', '“', '»', '«' with '"'
919 // and all '‚', '‘', '›', '‹' with "\'"
920 static std::regex plainquotes { R"(„|“|»|«)" };
921 static std::regex innerquotes { R"(‚|‘|›|‹)" };
922 t = std::regex_replace(s, plainquotes, R"(")");
923 s = std::regex_replace(t, innerquotes, R"(')");
924 //LYXERR0("correctRegex output '" << s << "'");
928 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
929 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
930 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
931 string escape_for_regex(string s, bool withformat)
935 while (lastpos < s.size()) {
936 size_t regex_pos = s.find("\\regexp{", lastpos);
937 if (regex_pos == string::npos) {
938 regex_pos = s.size();
940 if (regex_pos > lastpos) {
941 result += string2regex(s.substr(lastpos, regex_pos-lastpos));
943 if (lastpos == s.size())
946 size_t end_pos = s.find("\\endregexp", regex_pos + 8);
947 result += correctRegex(s.substr(regex_pos + 8, end_pos -(regex_pos + 8)), withformat);
948 lastpos = end_pos + 13;
954 /// Wrapper for lyx::regex_replace with simpler interface
955 bool regex_replace(string const & s, string & t, string const & searchstr,
956 string const & replacestr)
958 regex e(searchstr, regex_constants::ECMAScript);
960 ostream_iterator<char, char> it(oss);
961 regex_replace(it, s.begin(), s.end(), e, replacestr);
962 // tolerate t and s be references to the same variable
963 bool rv = (s != oss.str());
982 vector <string> result = vector <string>();
983 MatchResult(int len = 0): match_len(len),match_prefix(0),match2end(0), pos(0),leadsize(0),pos_len(-1),searched_size(0) {}
986 static MatchResult::range interpretMatch(MatchResult &oldres, MatchResult &newres)
988 if (newres.match2end < oldres.match2end)
989 return MatchResult::newIsTooFar;
990 if (newres.match_len < oldres.match_len)
991 return MatchResult::newIsTooFar;
993 if (newres.match_len == oldres.match_len) {
994 if (newres.match2end == oldres.match2end)
995 return MatchResult::newIsBetter;
997 return MatchResult::newIsInvalid;
1000 /** The class performing a match between a position in the document and the FindAdvOptions.
1003 class MatchStringAdv {
1005 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt);
1007 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
1008 ** constructor as opt.search, under the opt.* options settings.
1011 ** If set, then match is searched only against beginning of text starting at cur.
1012 ** If unset, then match is searched anywhere in text starting at cur.
1015 ** The length of the matching text, or zero if no match was found.
1017 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
1025 lyx::Buffer * p_buf;
1026 /// first buffer on which search was started
1027 lyx::Buffer * const p_first_buf;
1029 FindAndReplaceOptions const & opt;
1032 /// Auxiliary find method (does not account for opt.matchword)
1033 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
1034 void CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string = "");
1036 /** Normalize a stringified or latexified LyX paragraph.
1040 ** <li>if search is not casesensitive, then lowercase the string;
1041 ** <li>remove any newline at begin or end of the string;
1042 ** <li>replace any newline in the middle of the string with a simple space;
1043 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
1046 ** @todo Normalization should also expand macros, if the corresponding
1047 ** search option was checked.
1049 string normalize(docstring const & s, bool ignore_fomat) const;
1050 // normalized string to search
1051 string par_as_string;
1052 // regular expression to use for searching
1053 // regexp2 is same as regexp, but prefixed with a ".*?"
1055 QRegularExpression regexp;
1056 QRegularExpression regexp2;
1061 // leading format material as string
1062 string lead_as_string;
1063 // par_as_string after removal of lead_as_string
1064 string par_as_string_nolead;
1065 // unmatched open braces in the search string/regexp
1067 // number of (.*?) subexpressions added at end of search regexp for closing
1068 // environments, math mode, styles, etc...
1069 int close_wildcards;
1071 // Are we searching with regular expressions ?
1072 bool use_regexp = false;
1073 static int valid_matches;
1074 static vector <string> matches;
1075 void FillResults(MatchResult &found_mr);
1078 int MatchStringAdv::valid_matches = 0;
1079 vector <string> MatchStringAdv::matches = vector <string>(10);
1081 void MatchStringAdv::FillResults(MatchResult &found_mr)
1083 if (found_mr.match_len > 0) {
1084 valid_matches = found_mr.result.size();
1085 for (size_t i = 0; i < found_mr.result.size(); i++)
1086 matches[i] = found_mr.result[i];
1091 static void setFindParams(OutputParams &runparams)
1093 runparams.flavor = Flavor::XeTeX;
1094 //runparams.use_polyglossia = true;
1095 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
1096 // No side effect of file copying and image conversion
1097 runparams.dryrun = true;
1100 static docstring buffer_to_latex(Buffer & buffer)
1102 //OutputParams runparams(&buffer.params().encoding());
1103 OutputParams runparams(encodings.fromLyXName("utf8"));
1104 odocstringstream ods;
1106 runparams.nice = true;
1107 setFindParams(runparams);
1108 if (ignoreFormats.getDeleted())
1109 runparams.for_search = OutputParams::SearchWithoutDeleted;
1111 runparams.for_search = OutputParams::SearchWithDeleted;
1112 if (ignoreFormats.getNonContent()) {
1113 runparams.for_search |= OutputParams::SearchNonOutput;
1115 pit_type const endpit = buffer.paragraphs().size();
1116 for (pit_type pit = 0; pit != endpit; ++pit) {
1117 TeXOnePar(buffer, buffer.text(), pit, os, runparams, string(), -1, -1, true);
1118 LYXERR(Debug::FINDVERBOSE, "searchString up to here: " << ods.str());
1123 static string latexNamesToUtf8(docstring strIn)
1125 string addtmp = to_utf8(strIn);
1126 static regex const rmAcc("(\\\\)*("
1127 "\\\\([A-Za-z]+\\{.\\})" // e.g. "ddot{A}" == sub.str(3)
1128 "|\\\\([A-Za-z]+)( |\\{\\})?" // e.g. "LyX", "LyX{}", "LyX " == sub.str(4)
1135 if (accents.empty())
1137 for (sregex_iterator it_add(addtmp.begin(), addtmp.end(), rmAcc), end; it_add != end; ++it_add) {
1139 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
1144 if (sub.length(3) > 0)
1148 AccentsIterator it_ac = accents.find(key);
1149 if (it_ac == accents.end()) {
1150 replace = sub.str(2);
1153 replace = it_ac->second;
1156 if (lastpos < (size_t) sub.position(2))
1157 add += addtmp.substr(lastpos, sub.position(2) - lastpos);
1159 lastpos = sub.position(2) + sub.length(2);
1163 else if (addtmp.length() > lastpos)
1164 add += addtmp.substr(lastpos, addtmp.length() - lastpos);
1165 LYXERR(Debug::FINDVERBOSE, "Adding to search string: '"
1170 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
1173 if (!opt.ignoreformat) {
1174 str = buffer_to_latex(buffer);
1176 // OutputParams runparams(&buffer.params().encoding());
1177 OutputParams runparams(encodings.fromLyXName("utf8"));
1178 runparams.nice = true;
1179 setFindParams(runparams);
1180 int option = AS_STR_INSETS |AS_STR_PLAINTEXT;
1181 if (ignoreFormats.getDeleted()) {
1182 option |= AS_STR_SKIPDELETE;
1183 runparams.for_search = OutputParams::SearchWithoutDeleted;
1186 runparams.for_search = OutputParams::SearchWithDeleted;
1188 if (ignoreFormats.getNonContent()) {
1189 runparams.for_search |= OutputParams::SearchNonOutput;
1192 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
1193 Paragraph const & par = buffer.paragraphs().at(pit);
1194 string add = latexNamesToUtf8(par.asString(pos_type(0), par.size(),
1197 LYXERR(Debug::FINDVERBOSE, "Adding to search string: '"
1201 // Even in ignore-format we have to remove "\text{}, \lyxmathsym{}" parts
1202 while (regex_replace(t, t, "\\\\(text|lyxmathsym|ensuremath)\\{([^\\}]*)\\}", "$2"));
1209 /// Return separation pos between the leading material and the rest
1210 static size_t identifyLeading(string const & s)
1213 // @TODO Support \item[text]
1214 // Kornel: Added textsl, textsf, textit, texttt and noun
1215 // + allow to search for colored text too
1216 while (regex_replace(t, t, "^\\\\(("
1217 "(author|title|subtitle|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|"
1218 "lyxaddress|lyxrightaddress|"
1219 "footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1220 "emph|noun|minisec|text(bf|md|sl|sf|it|tt))|"
1221 "((textcolor|foreignlanguage|latexenvironment)\\{[a-z]+\\*?\\})|"
1222 "(u|uu)line|(s|x)out|uwave)|((sub)?(((sub)?section)|paragraph)|part|chapter)\\*?)\\{", "")
1223 || regex_replace(t, t, "^\\$", "")
1224 || regex_replace(t, t, "^\\\\\\[", "")
1225 || regex_replace(t, t, "^ ?\\\\item\\{[a-z]+\\}", "")
1226 || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\}", ""))
1228 LYXERR(Debug::FINDVERBOSE, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
1229 return s.size() - t.size();
1233 * Given a latexified string, retrieve some handled features
1234 * The features of the regex will later be compared with the features
1235 * of the searched text. If the regex features are not a
1236 * subset of the analized, then, in not format ignoring search
1237 * we can early stop the search in the relevant inset.
1239 typedef map<string, bool> Features;
1241 static Features identifyFeatures(string const & s)
1243 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
1244 static regex const valid("^("
1246 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1247 "emph|noun|text(bf|md|sl|sf|it|tt)|"
1248 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
1249 "(u|uu)line|(s|x)out|uwave|"
1250 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
1251 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
1256 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1259 if (sub.str(1).compare("regexp") == 0) {
1263 string token = sub.str(1);
1265 if (regex_match(token, sub2, valid)) {
1273 if (sub.str(1).compare("endregexp") == 0) {
1283 * defines values features of a key "\\[a-z]+{"
1288 /* Char type with content discarded
1289 * like \hspace{1cm} */
1291 /* Char, like \backslash */
1293 /* replace starting backslash with '#' */
1295 /* \part, \section*, ... */
1297 /* title, author etc */
1299 /* \foreignlanguage{ngerman}, ... */
1302 * to discard language in content */
1305 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1307 /* fonts, colors, markups, ... */
1309 /* footnotesize, ... large, ...
1310 * Ignore all of them */
1313 /* inputencoding, ...
1314 * Discard also content, because they do not help in search */
1317 * like remove, but also all arguments */
1319 /* item, listitem */
1321 /* tex, latex, ... like isChar */
1323 /* like \lettrine[lines=5]{}{} */
1325 // like isStandard, but always remove head
1327 /* End of arguments marker for lettrine,
1328 * so that they can be ignored */
1331 KeyInfo() = default;
1332 KeyInfo(KeyType type, int parcount, bool disable)
1334 parenthesiscount(parcount),
1335 disabled(disable) {}
1336 KeyType keytype = invalid;
1338 int _tokensize = -1;
1339 int _tokenstart = -1;
1340 int _dataStart = -1;
1342 int parenthesiscount = 1;
1343 bool disabled = false;
1344 bool used = false; /* by pattern */
1349 Border(int l=0, int u=0) : low(l), upper(u) {}
1354 #define MAXOPENED 30
1356 bool isPatternString_;
1358 explicit Intervall(bool isPattern, string const & p)
1359 : isPatternString_(isPattern), par(p), ignoreidx(-1),
1360 actualdeptindex(0), hasTitle(false), langcount(0)
1368 static vector<Border> borders;
1369 int depts[MAXOPENED];
1370 int closes[MAXOPENED];
1371 int actualdeptindex;
1372 int previousNotIgnored(int) const;
1373 int nextNotIgnored(int) const;
1374 void handleOpenP(int i);
1375 void handleCloseP(int i, bool closingAllowed);
1376 void resetOpenedP(int openPos);
1377 void addIntervall(int upper);
1378 void addIntervall(int low, int upper); /* if explicit */
1379 void removeAccents();
1380 void setForDefaultLang(KeyInfo const & defLang) const;
1381 int findclosing(int start, int end, char up, char down, int repeat);
1382 void handleParentheses(int lastpos, bool closingAllowed);
1384 // Number of disabled language specs up
1385 // to current position in actual interval
1387 int isOpeningPar(int pos) const;
1389 void output(ostringstream &os, int lastpos);
1390 // string show(int lastpos);
1393 vector<Border> Intervall::borders = vector<Border>(30);
1395 int Intervall::isOpeningPar(int pos) const
1397 if ((pos < 0) || (size_t(pos) >= par.size()))
1399 if (par[pos] != '{')
1401 if (size_t(pos) + 2 >= par.size())
1403 if (par[pos+2] != '}')
1405 if (par[pos+1] == '[' || par[pos+1] == ']')
1410 void Intervall::setForDefaultLang(KeyInfo const & defLang) const
1412 // Enable the use of first token again
1413 if (ignoreidx >= 0) {
1414 int value = defLang._tokenstart + defLang._tokensize;
1419 if (borders[borderidx].low < value)
1420 borders[borderidx].low = value;
1421 if (borders[borderidx].upper < value)
1422 borders[borderidx].upper = value;
1427 static void checkDepthIndex(int val)
1429 static int maxdepthidx = MAXOPENED-2;
1430 static int lastmaxdepth = 0;
1431 if (val > lastmaxdepth) {
1432 LYXERR(Debug::INFO, "Depth reached " << val);
1435 if (val > maxdepthidx) {
1437 LYXERR(Debug::INFO, "maxdepthidx now " << val);
1442 // Not needed, because borders are now dynamically expanded
1443 static void checkIgnoreIdx(int val)
1445 static int lastmaxignore = -1;
1446 if ((lastmaxignore < val) && (size_t(val+1) >= borders.size())) {
1447 LYXERR(Debug::INFO, "IgnoreIdx reached " << val);
1448 lastmaxignore = val;
1454 * Expand the region of ignored parts of the input latex string
1455 * The region is only relevant in output()
1457 void Intervall::addIntervall(int low, int upper)
1460 if (low == upper) return;
1461 for (idx = ignoreidx+1; idx > 0; --idx) {
1462 if (low > borders[idx-1].upper) {
1466 Border br(low, upper);
1467 if (idx > ignoreidx) {
1468 if (borders.size() <= size_t(idx)) {
1469 borders.push_back(br);
1475 // checkIgnoreIdx(ignoreidx);
1479 // Expand only if one of the new bound is inside the interwall
1480 // We know here that br.low > borders[idx-1].upper
1481 if (br.upper < borders[idx].low) {
1482 // We have to insert at this pos
1483 if (size_t(ignoreidx+1) >= borders.size()) {
1484 borders.push_back(borders[ignoreidx]);
1487 borders[ignoreidx+1] = borders[ignoreidx];
1489 for (int i = ignoreidx; i > idx; --i) {
1490 borders[i] = borders[i-1];
1494 // checkIgnoreIdx(ignoreidx);
1497 // Here we know, that we are overlapping
1498 if (br.low > borders[idx].low)
1499 br.low = borders[idx].low;
1500 // check what has to be concatenated
1502 for (int i = idx; i <= ignoreidx; i++) {
1503 if (br.upper >= borders[i].low) {
1505 if (br.upper < borders[i].upper)
1506 br.upper = borders[i].upper;
1512 // count should be >= 1 here
1515 for (int i = idx + count; i <= ignoreidx; i++) {
1516 borders[i-count+1] = borders[i];
1518 ignoreidx -= count - 1;
1524 static void buildaccent(string n, string param, string values)
1528 const char delim = '|';
1529 while (getline(s, name, delim)) {
1531 for (char c : param) {
1532 string key = name + "{" + c + "}";
1533 // get the corresponding utf8-value
1534 if ((values[start] & 0xc0) != 0xc0) {
1535 // should not happen, utf8 encoding starts at least with 11xxxxxx
1536 // but value for '\dot{i}' is 'i', which is ascii
1537 if ((values[start] & 0x80) == 0) {
1539 accents[key] = values.substr(start, 1);
1540 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1545 for (int j = 1; ;j++) {
1546 if (start + j >= values.size()) {
1547 accents[key] = values.substr(start, j);
1548 start = values.size() - 1;
1551 else if ((values[start+j] & 0xc0) != 0x80) {
1552 // This is the first byte of following utf8 char
1553 accents[key] = values.substr(start, j);
1555 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1564 static string getutf8(unsigned uchar)
1567 string ret = string();
1569 if (uchar <= 0x7f) {
1570 c[maxc-1] = uchar & 0x7f;
1573 unsigned char rest = 0x40;
1574 unsigned char first = 0x80;
1576 for (int i = start; i >=0; --i) {
1578 c[i] = first + uchar;
1581 c[i] = 0x80 | (uchar & 0x3f);
1588 for (int i = 0; i < maxc; i++) {
1589 if (c[i] == 0) continue;
1595 static void addAccents(string latex_in, string unicode_out)
1597 latex_in = latex_in.substr(1);
1598 AccentsIterator it_ac = accents.find(latex_in);
1599 if (it_ac == accents.end()) {
1600 accents[latex_in] = unicode_out;
1603 LYXERR0("Key " << latex_in << " already set");
1607 void static fillMissingUnicodesymbols()
1609 addAccents("\\pounds", getutf8(0x00a3));
1610 addAccents("\\textsterling", getutf8(0x00a3));
1611 addAccents("\\textyen", getutf8(0x00a5));
1612 addAccents("\\yen", getutf8(0x00a5));
1613 addAccents("\\textsection", getutf8(0x00a7));
1614 addAccents("\\mathsection", getutf8(0x00a7));
1615 addAccents("\\textcopyright", getutf8(0x00a9));
1616 addAccents("\\copyright", getutf8(0x00a9));
1617 addAccents("\\textlnot", getutf8(0x00ac));
1618 addAccents("\\neg", getutf8(0x00ac));
1619 addAccents("\\textregistered", getutf8(0x00ae));
1620 addAccents("\\circledR", getutf8(0x00ae));
1621 addAccents("\\textpm", getutf8(0x00b1));
1622 addAccents("\\pm", getutf8(0x00b1));
1623 addAccents("\\textparagraph", getutf8(0x00b6));
1624 addAccents("\\mathparagraph", getutf8(0x00b6));
1625 addAccents("\\textperiodcentered", getutf8(0x00b7));
1626 addAccents("\\texttimes", getutf8(0x00d7));
1627 addAccents("\\times", getutf8(0x00d7));
1628 addAccents("\\O", getutf8(0x00d8));
1629 addAccents("\\dh", getutf8(0x00f0));
1630 addAccents("\\eth", getutf8(0x00f0));
1631 addAccents("\\textdiv", getutf8(0x00f7));
1632 addAccents("\\div", getutf8(0x00f7));
1633 addAccents("\\o", getutf8(0x00f8));
1634 addAccents("\\textcrlambda", getutf8(0x019b));
1635 addAccents("\\j", getutf8(0x0237));
1636 addAccents("\\textrevepsilon", getutf8(0x025c));
1637 addAccents("\\textbaru", getutf8(0x0289));
1638 addAccents("\\textquoteleft", getutf8(0x02bb));
1639 addAccents("\\textGamma", getutf8(0x0393));
1640 addAccents("\\Gamma", getutf8(0x0393));
1641 addAccents("\\textDelta", getutf8(0x0394));
1642 addAccents("\\Delta", getutf8(0x0394));
1643 addAccents("\\textTheta", getutf8(0x0398));
1644 addAccents("\\Theta", getutf8(0x0398));
1645 addAccents("\\textLambda", getutf8(0x039b));
1646 addAccents("\\Lambda", getutf8(0x039b));
1647 addAccents("\\textXi", getutf8(0x039e));
1648 addAccents("\\Xi", getutf8(0x039e));
1649 addAccents("\\textPi", getutf8(0x03a0));
1650 addAccents("\\Pi", getutf8(0x03a0));
1651 addAccents("\\textSigma", getutf8(0x03a3));
1652 addAccents("\\Sigma", getutf8(0x03a3));
1653 addAccents("\\textUpsilon", getutf8(0x03a5));
1654 addAccents("\\Upsilon", getutf8(0x03a5));
1655 addAccents("\\textPhi", getutf8(0x03a6));
1656 addAccents("\\Phi", getutf8(0x03a6));
1657 addAccents("\\textPsi", getutf8(0x03a8));
1658 addAccents("\\Psi", getutf8(0x03a8));
1659 addAccents("\\textOmega", getutf8(0x03a9));
1660 addAccents("\\Omega", getutf8(0x03a9));
1661 addAccents("\\textalpha", getutf8(0x03b1));
1662 addAccents("\\alpha", getutf8(0x03b1));
1663 addAccents("\\textbeta", getutf8(0x03b2));
1664 addAccents("\\beta", getutf8(0x03b2));
1665 addAccents("\\textgamma", getutf8(0x03b3));
1666 addAccents("\\gamma", getutf8(0x03b3));
1667 addAccents("\\textdelta", getutf8(0x03b4));
1668 addAccents("\\delta", getutf8(0x03b4));
1669 addAccents("\\textepsilon", getutf8(0x03b5));
1670 addAccents("\\varepsilon", getutf8(0x03b5));
1671 addAccents("\\textzeta", getutf8(0x03b6));
1672 addAccents("\\zeta", getutf8(0x03b6));
1673 addAccents("\\texteta", getutf8(0x03b7));
1674 addAccents("\\eta", getutf8(0x03b7));
1675 addAccents("\\texttheta", getutf8(0x03b8));
1676 addAccents("\\theta", getutf8(0x03b8));
1677 addAccents("\\textiota", getutf8(0x03b9));
1678 addAccents("\\iota", getutf8(0x03b9));
1679 addAccents("\\textkappa", getutf8(0x03ba));
1680 addAccents("\\kappa", getutf8(0x03ba));
1681 addAccents("\\textlambda", getutf8(0x03bb));
1682 addAccents("\\lambda", getutf8(0x03bb));
1683 addAccents("\\textmu", getutf8(0x03bc));
1684 addAccents("\\mu", getutf8(0x03bc));
1685 addAccents("\\textnu", getutf8(0x03bd));
1686 addAccents("\\nu", getutf8(0x03bd));
1687 addAccents("\\textxi", getutf8(0x03be));
1688 addAccents("\\xi", getutf8(0x03be));
1689 addAccents("\\textpi", getutf8(0x03c0));
1690 addAccents("\\pi", getutf8(0x03c0));
1691 addAccents("\\textrho", getutf8(0x03c1));
1692 addAccents("\\rho", getutf8(0x03c1));
1693 addAccents("\\textfinalsigma", getutf8(0x03c2));
1694 addAccents("\\varsigma", getutf8(0x03c2));
1695 addAccents("\\textsigma", getutf8(0x03c3));
1696 addAccents("\\sigma", getutf8(0x03c3));
1697 addAccents("\\texttau", getutf8(0x03c4));
1698 addAccents("\\tau", getutf8(0x03c4));
1699 addAccents("\\textupsilon", getutf8(0x03c5));
1700 addAccents("\\upsilon", getutf8(0x03c5));
1701 addAccents("\\textphi", getutf8(0x03c6));
1702 addAccents("\\varphi", getutf8(0x03c6));
1703 addAccents("\\textchi", getutf8(0x03c7));
1704 addAccents("\\chi", getutf8(0x03c7));
1705 addAccents("\\textpsi", getutf8(0x03c8));
1706 addAccents("\\psi", getutf8(0x03c8));
1707 addAccents("\\textomega", getutf8(0x03c9));
1708 addAccents("\\omega", getutf8(0x03c9));
1709 addAccents("\\textdigamma", getutf8(0x03dd));
1710 addAccents("\\digamma", getutf8(0x03dd));
1711 addAccents("\\hebalef", getutf8(0x05d0));
1712 addAccents("\\aleph", getutf8(0x05d0));
1713 addAccents("\\hebbet", getutf8(0x05d1));
1714 addAccents("\\beth", getutf8(0x05d1));
1715 addAccents("\\hebgimel", getutf8(0x05d2));
1716 addAccents("\\gimel", getutf8(0x05d2));
1717 addAccents("\\hebdalet", getutf8(0x05d3));
1718 addAccents("\\daleth", getutf8(0x05d3));
1719 addAccents("\\hebhe", getutf8(0x05d4));
1720 addAccents("\\hebvav", getutf8(0x05d5));
1721 addAccents("\\hebzayin", getutf8(0x05d6));
1722 addAccents("\\hebhet", getutf8(0x05d7));
1723 addAccents("\\hebtet", getutf8(0x05d8));
1724 addAccents("\\hebyod", getutf8(0x05d9));
1725 addAccents("\\hebfinalkaf", getutf8(0x05da));
1726 addAccents("\\hebkaf", getutf8(0x05db));
1727 addAccents("\\heblamed", getutf8(0x05dc));
1728 addAccents("\\hebfinalmem", getutf8(0x05dd));
1729 addAccents("\\hebmem", getutf8(0x05de));
1730 addAccents("\\hebfinalnun", getutf8(0x05df));
1731 addAccents("\\hebnun", getutf8(0x05e0));
1732 addAccents("\\hebsamekh", getutf8(0x05e1));
1733 addAccents("\\hebayin", getutf8(0x05e2));
1734 addAccents("\\hebfinalpe", getutf8(0x05e3));
1735 addAccents("\\hebpe", getutf8(0x05e4));
1736 addAccents("\\hebfinaltsadi", getutf8(0x05e5));
1737 addAccents("\\hebtsadi", getutf8(0x05e6));
1738 addAccents("\\hebqof", getutf8(0x05e7));
1739 addAccents("\\hebresh", getutf8(0x05e8));
1740 addAccents("\\hebshin", getutf8(0x05e9));
1741 addAccents("\\hebtav", getutf8(0x05ea));
1744 addAccents("\\thaiKoKai", getutf8(0x0e01));
1745 addAccents("\\thaiKhoKhai", getutf8(0x0e02));
1746 addAccents("\\thaiKhoKhuat", getutf8(0x0e03));
1747 addAccents("\\thaiKhoKhwai", getutf8(0x0e04));
1748 addAccents("\\thaiKhoKhon", getutf8(0x0e05));
1749 addAccents("\\thaiKhoRakhang", getutf8(0x0e06));
1750 addAccents("\\thaiNgoNgu", getutf8(0x0e07));
1751 addAccents("\\thaiChoChan", getutf8(0x0e08));
1752 addAccents("\\thaiChoChing", getutf8(0x0e09));
1753 addAccents("\\thaiChoChang", getutf8(0x0e0a));
1754 addAccents("\\thaiSoSo", getutf8(0x0e0b));
1755 addAccents("\\thaiChoChoe", getutf8(0x0e0c));
1756 addAccents("\\thaiYoYing", getutf8(0x0e0d));
1757 addAccents("\\thaiDoChada", getutf8(0x0e0e));
1758 addAccents("\\thaiToPatak", getutf8(0x0e0f));
1759 addAccents("\\thaiThoThan", getutf8(0x0e10));
1760 addAccents("\\thaiThoNangmontho", getutf8(0x0e11));
1761 addAccents("\\thaiThoPhuthao", getutf8(0x0e12));
1762 addAccents("\\thaiNoNen", getutf8(0x0e13));
1763 addAccents("\\thaiDoDek", getutf8(0x0e14));
1764 addAccents("\\thaiToTao", getutf8(0x0e15));
1765 addAccents("\\thaiThoThung", getutf8(0x0e16));
1766 addAccents("\\thaiThoThahan", getutf8(0x0e17));
1767 addAccents("\\thaiThoThong", getutf8(0x0e18));
1768 addAccents("\\thaiNoNu", getutf8(0x0e19));
1769 addAccents("\\thaiBoBaimai", getutf8(0x0e1a));
1770 addAccents("\\thaiPoPla", getutf8(0x0e1b));
1771 addAccents("\\thaiPhoPhung", getutf8(0x0e1c));
1772 addAccents("\\thaiFoFa", getutf8(0x0e1d));
1773 addAccents("\\thaiPhoPhan", getutf8(0x0e1e));
1774 addAccents("\\thaiFoFan", getutf8(0x0e1f));
1775 addAccents("\\thaiPhoSamphao", getutf8(0x0e20));
1776 addAccents("\\thaiMoMa", getutf8(0x0e21));
1777 addAccents("\\thaiYoYak", getutf8(0x0e22));
1778 addAccents("\\thaiRoRua", getutf8(0x0e23));
1779 addAccents("\\thaiRu", getutf8(0x0e24));
1780 addAccents("\\thaiLoLing", getutf8(0x0e25));
1781 addAccents("\\thaiLu", getutf8(0x0e26));
1782 addAccents("\\thaiWoWaen", getutf8(0x0e27));
1783 addAccents("\\thaiSoSala", getutf8(0x0e28));
1784 addAccents("\\thaiSoRusi", getutf8(0x0e29));
1785 addAccents("\\thaiSoSua", getutf8(0x0e2a));
1786 addAccents("\\thaiHoHip", getutf8(0x0e2b));
1787 addAccents("\\thaiLoChula", getutf8(0x0e2c));
1788 addAccents("\\thaiOAng", getutf8(0x0e2d));
1789 addAccents("\\thaiHoNokhuk", getutf8(0x0e2e));
1790 addAccents("\\thaiPaiyannoi", getutf8(0x0e2f));
1791 addAccents("\\thaiSaraA", getutf8(0x0e30));
1792 addAccents("\\thaiMaiHanakat", getutf8(0x0e31));
1793 addAccents("\\thaiSaraAa", getutf8(0x0e32));
1794 addAccents("\\thaiSaraAm", getutf8(0x0e33));
1795 addAccents("\\thaiSaraI", getutf8(0x0e34));
1796 addAccents("\\thaiSaraIi", getutf8(0x0e35));
1797 addAccents("\\thaiSaraUe", getutf8(0x0e36));
1798 addAccents("\\thaiSaraUee", getutf8(0x0e37));
1799 addAccents("\\thaiSaraU", getutf8(0x0e38));
1800 addAccents("\\thaiSaraUu", getutf8(0x0e39));
1801 addAccents("\\thaiPhinthu", getutf8(0x0e3a));
1802 addAccents("\\thaiSaraE", getutf8(0x0e40));
1803 addAccents("\\thaiSaraAe", getutf8(0x0e41));
1804 addAccents("\\thaiSaraO", getutf8(0x0e42));
1805 addAccents("\\thaiSaraAiMaimuan", getutf8(0x0e43));
1806 addAccents("\\thaiSaraAiMaimalai", getutf8(0x0e44));
1807 addAccents("\\thaiLakkhangyao", getutf8(0x0e45));
1808 addAccents("\\thaiMaiyamok", getutf8(0x0e46));
1809 addAccents("\\thaiMaitaikhu", getutf8(0x0e47));
1810 addAccents("\\thaiMaiEk", getutf8(0x0e48));
1811 addAccents("\\thaiMaiTho", getutf8(0x0e49));
1812 addAccents("\\thaiMaiTri", getutf8(0x0e4a));
1813 addAccents("\\thaiMaiChattawa", getutf8(0x0e4b));
1814 addAccents("\\thaiThanthakhat", getutf8(0x0e4c));
1815 addAccents("\\thaiNikhahit", getutf8(0x0e4d));
1816 addAccents("\\thaiYamakkan", getutf8(0x0e4e));
1817 addAccents("\\thaiFongman", getutf8(0x0e4f));
1818 addAccents("\\thaizero", getutf8(0x0e50));
1819 addAccents("\\thaione", getutf8(0x0e51));
1820 addAccents("\\thaitwo", getutf8(0x0e52));
1821 addAccents("\\thaithree", getutf8(0x0e53));
1822 addAccents("\\thaifour", getutf8(0x0e54));
1823 addAccents("\\thaifive", getutf8(0x0e55));
1824 addAccents("\\thaisix", getutf8(0x0e56));
1825 addAccents("\\thaiseven", getutf8(0x0e57));
1826 addAccents("\\thaieight", getutf8(0x0e58));
1827 addAccents("\\thainine", getutf8(0x0e59));
1828 addAccents("\\thaiAngkhankhu", getutf8(0x0e5a));
1829 addAccents("\\thaiKhomut", getutf8(0x0e5b));
1830 addAccents("\\dag", getutf8(0x2020));
1831 addAccents("\\dagger", getutf8(0x2020));
1832 addAccents("\\textdagger", getutf8(0x2020));
1833 addAccents("\\ddag", getutf8(0x2021));
1834 addAccents("\\ddagger", getutf8(0x2021));
1835 addAccents("\\textdaggerdbl", getutf8(0x2021));
1836 addAccents("\\textbullet", getutf8(0x2022));
1837 addAccents("\\bullet", getutf8(0x2022));
1838 addAccents("\\dots", getutf8(0x2026));
1839 addAccents("\\ldots", getutf8(0x2026));
1840 addAccents("\\textellipsis", getutf8(0x2026));
1841 addAccents("\\textasciiacute", getutf8(0x2032));
1842 addAccents("\\prime", getutf8(0x2032));
1843 addAccents("\\textacutedbl", getutf8(0x2033));
1844 addAccents("\\dprime", getutf8(0x2033));
1845 addAccents("\\textasciigrave", getutf8(0x2035));
1846 addAccents("\\backprime", getutf8(0x2035));
1847 addAccents("\\textsubcircum{ }", getutf8(0x2038));
1848 addAccents("\\caretinsert", getutf8(0x2038));
1849 addAccents("\\textasteriskcentered", getutf8(0x204e));
1850 addAccents("\\ast", getutf8(0x204e));
1851 addAccents("\\textmho", getutf8(0x2127));
1852 addAccents("\\mho", getutf8(0x2127));
1853 addAccents("\\textleftarrow", getutf8(0x2190));
1854 addAccents("\\leftarrow", getutf8(0x2190));
1855 addAccents("\\textuparrow", getutf8(0x2191));
1856 addAccents("\\uparrow", getutf8(0x2191));
1857 addAccents("\\textrightarrow", getutf8(0x2192));
1858 addAccents("\\rightarrow", getutf8(0x2192));
1859 addAccents("\\textdownarrow", getutf8(0x2193));
1860 addAccents("\\downarrow", getutf8(0x2193));
1861 addAccents("\\textglobrise", getutf8(0x2197));
1862 addAccents("\\nearrow", getutf8(0x2197));
1863 addAccents("\\textglobfall", getutf8(0x2198));
1864 addAccents("\\searrow", getutf8(0x2198));
1865 addAccents("\\textsurd", getutf8(0x221a));
1866 addAccents("\\surd", getutf8(0x221a));
1867 addAccents("\\textbigcircle", getutf8(0x25ef));
1868 addAccents("\\bigcirc", getutf8(0x25ef));
1869 addAccents("\\FiveStar", getutf8(0x2605));
1870 addAccents("\\bigstar", getutf8(0x2605));
1871 addAccents("\\FiveStarOpen", getutf8(0x2606));
1872 addAccents("\\bigwhitestar", getutf8(0x2606));
1873 addAccents("\\Checkmark", getutf8(0x2713));
1874 addAccents("\\checkmark", getutf8(0x2713));
1875 addAccents("\\CrossMaltese", getutf8(0x2720));
1876 addAccents("\\maltese", getutf8(0x2720));
1877 addAccents("\\textlangle", getutf8(0x27e8));
1878 addAccents("\\langle", getutf8(0x27e8));
1879 addAccents("\\textrangle", getutf8(0x27e9));
1880 addAccents("\\rangle", getutf8(0x27e9));
1883 static void buildAccentsMap()
1885 accents["imath"] = "ı";
1887 accents["jmath"] = "ȷ";
1888 accents["cdot"] = "·";
1889 accents["textasciicircum"] = "^";
1890 accents["mathcircumflex"] = "^";
1891 accents["sim"] = "~";
1892 accents["guillemotright"] = "»";
1893 accents["guillemotleft"] = "«";
1894 accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
1895 accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
1896 accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
1897 accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
1898 accents["negmedspace"] = getutf8(0xf0005);
1899 accents["thickspace"] = getutf8(0xf0006);
1900 accents["negthickspace"] = getutf8(0xf0007);
1901 accents["lyx"] = getutf8(0xf0010); // Used logos
1902 accents["LyX"] = getutf8(0xf0010);
1903 accents["tex"] = getutf8(0xf0011);
1904 accents["TeX"] = getutf8(0xf0011);
1905 accents["latex"] = getutf8(0xf0012);
1906 accents["LaTeX"] = getutf8(0xf0012);
1907 accents["latexe"] = getutf8(0xf0013);
1908 accents["LaTeXe"] = getutf8(0xf0013);
1909 accents["lyxarrow"] = getutf8(0xf0020);
1910 accents["braceleft"] = getutf8(0xf0030);
1911 accents["braceright"] = getutf8(0xf0031);
1912 accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash
1913 accents["backslash LyX"] = getutf8(0xf0010);
1914 accents["backslash tex"] = getutf8(0xf0011);
1915 accents["backslash TeX"] = getutf8(0xf0011);
1916 accents["backslash latex"] = getutf8(0xf0012);
1917 accents["backslash LaTeX"] = getutf8(0xf0012);
1918 accents["backslash latexe"] = getutf8(0xf0013);
1919 accents["backslash LaTeXe"] = getutf8(0xf0013);
1920 accents["backslash lyxarrow"] = getutf8(0xf0020);
1921 accents["ddot{\\imath}"] = "ï";
1922 buildaccent("ddot", "aAeEhHiIoOtuUwWxXyY",
1923 "äÄëËḧḦïÏöÖẗüÜẅẄẍẌÿŸ"); // umlaut
1924 buildaccent("dot|.", "aAbBcCdDeEfFGghHIimMnNoOpPrRsStTwWxXyYzZ",
1925 "ȧȦḃḂċĊḋḊėĖḟḞĠġḣḢİİṁṀṅṄȯȮṗṖṙṘṡṠṫṪẇẆẋẊẏẎżŻ"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'İ'
1926 accents["acute{\\imath}"] = "í";
1927 buildaccent("acute", "aAcCeEgGkKlLmMoOnNpPrRsSuUwWyYzZiI",
1928 "áÁćĆéÉǵǴḱḰĺĹḿḾóÓńŃṕṔŕŔśŚúÚẃẂýÝźŹíÍ");
1929 buildaccent("dacute|H|h", "oOuU", "őŐűŰ"); // double acute
1930 buildaccent("mathring|r", "aAuUwy",
1932 accents["check{\\imath}"] = "ǐ";
1933 accents["check{\\jmath}"] = "ǰ";
1934 buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
1935 "čČďĎǎǍěĚǐǏǒǑǔǓǧǦǩǨȟȞľĽňŇřŘšŠŤťžŽ"); // caron
1936 accents["hat{\\imath}"] = "î";
1937 accents["hat{\\jmath}"] = "ĵ";
1938 buildaccent("hat|^", "aAcCeEgGhHiIjJoOsSuUwWyYzZ",
1939 "âÂĉĈêÊĝĜĥĤîÎĵĴôÔŝŜûÛŵŴŷŶẑẐ"); // circ
1940 accents["bar{\\imath}"] = "ī";
1941 buildaccent("bar|=", "aAeEiIoOuUyY",
1942 "āĀēĒīĪōŌūŪȳȲ"); // macron
1943 accents["tilde{\\imath}"] = "ĩ";
1944 buildaccent("tilde", "aAeEiInNoOuUvVyY",
1945 "ãÃẽẼĩĨñÑõÕũŨṽṼỹỸ"); // tilde
1946 accents["breve{\\imath}"] = "ĭ";
1947 buildaccent("breve|u", "aAeEgGiIoOuU",
1948 "ăĂĕĔğĞĭĬŏŎŭŬ"); // breve
1949 accents["grave{\\imath}"] = "ì";
1950 buildaccent("grave|`", "aAeEiIoOuUnNwWyY",
1951 "àÀèÈìÌòÒùÙǹǸẁẀỳỲ"); // grave
1952 buildaccent("subdot|d", "BbDdHhKkLlMmNnRrSsTtVvWwZzAaEeIiOoUuYy",
1953 "ḄḅḌḍḤḥḲḳḶḷṂṃṆṇṚṛṢṣṬṭṾṿẈẉẒẓẠạẸẹỊịỌọỤụỴỵ"); // dot below
1954 buildaccent("ogonek|k", "AaEeIiUuOo",
1955 "ĄąĘęĮįŲųǪǫ"); // ogonek
1956 buildaccent("cedilla|c", "CcGgKkLlNnRrSsTtEeDdHh",
1957 "ÇçĢģĶķĻļŅņŖŗŞşŢţȨȩḐḑḨḩ"); // cedilla
1958 buildaccent("subring|textsubring", "Aa",
1960 buildaccent("subhat|textsubcircum", "DdEeLlNnTtUu",
1961 "ḒḓḘḙḼḽṊṋṰṱṶṷ"); // subcircum
1962 buildaccent("subtilde|textsubtilde", "EeIiUu",
1963 "ḚḛḬḭṴṵ"); // subtilde
1964 accents["dgrave{\\imath}"] = "ȉ";
1965 accents["textdoublegrave{\\i}"] = "ȉ";
1966 buildaccent("dgrave|textdoublegrave", "AaEeIiOoRrUu",
1967 "ȀȁȄȅȈȉȌȍȐȑȔȕ"); // double grave
1968 accents["rcap{\\imath}"] = "ȋ";
1969 accents["textroundcap{\\i}"] = "ȋ";
1970 buildaccent("rcap|textroundcap", "AaEeIiOoRrUu",
1971 "ȂȃȆȇȊȋȎȏȒȓȖȗ"); // inverted breve
1972 buildaccent("slashed", "oO",
1974 fillMissingUnicodesymbols(); // Add some still not handled entries contained in 'unicodesynbols'
1975 // LYXERR0("Number of accents " << accents.size());
1979 * Created accents in math or regexp environment
1980 * are macros, but we need the utf8 equivalent
1982 void Intervall::removeAccents()
1984 if (accents.empty())
1986 static regex const accre("\\\\("
1987 "([\\S]|[A-Za-z]+)\\{[^\\{\\}]+\\}"
1989 "(backslash ([lL]y[xX]|[tT]e[xX]|[lL]a[tT]e[xX]e?|lyxarrow))"
1994 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
1996 string key = sub.str(1);
1997 AccentsIterator it_ac = accents.find(key);
1998 if (it_ac != accents.end()) {
1999 string val = it_ac->second;
2000 size_t pos = sub.position(size_t(0));
2001 for (size_t i = 0; i < val.size(); i++) {
2002 par[pos+i] = val[i];
2004 // Remove possibly following space too
2005 if (par[pos+sub.str(0).size()] == ' ')
2006 addIntervall(pos+val.size(), pos + sub.str(0).size()+1);
2008 addIntervall(pos+val.size(), pos + sub.str(0).size());
2009 for (size_t i = pos+val.size(); i < pos + sub.str(0).size(); i++) {
2010 // remove traces of any remaining chars
2015 LYXERR(Debug::INFO, "Not added accent for \"" << key << "\"");
2020 void Intervall::handleOpenP(int i)
2023 depts[actualdeptindex] = i+1;
2024 closes[actualdeptindex] = -1;
2025 checkDepthIndex(actualdeptindex);
2028 void Intervall::handleCloseP(int i, bool closingAllowed)
2030 if (actualdeptindex <= 0) {
2031 if (! closingAllowed)
2032 LYXERR(Debug::FINDVERBOSE, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
2033 // if we are at the very end
2034 addIntervall(i, i+1);
2037 closes[actualdeptindex] = i+1;
2042 void Intervall::resetOpenedP(int openPos)
2044 // Used as initializer for foreignlanguage entry
2045 actualdeptindex = 1;
2046 depts[1] = openPos+1;
2050 int Intervall::previousNotIgnored(int start) const
2052 int idx = 0; /* int intervalls */
2053 for (idx = ignoreidx; idx >= 0; --idx) {
2054 if (start > borders[idx].upper)
2056 if (start >= borders[idx].low)
2057 start = borders[idx].low-1;
2062 int Intervall::nextNotIgnored(int start) const
2064 int idx = 0; /* int intervalls */
2065 for (idx = 0; idx <= ignoreidx; idx++) {
2066 if (start < borders[idx].low)
2068 if (start < borders[idx].upper)
2069 start = borders[idx].upper;
2074 typedef unordered_map<string, KeyInfo> KeysMap;
2075 typedef unordered_map<string, KeyInfo>::const_iterator KeysIterator;
2076 typedef vector< KeyInfo> Entries;
2077 static KeysMap keys = unordered_map<string, KeyInfo>();
2083 Intervall interval_;
2084 void buildKeys(bool);
2085 void buildEntries(bool);
2086 void makeKey(const string &, KeyInfo, bool isPatternString);
2087 void processRegion(int start, int region_end); /* remove {} parts */
2088 void removeHead(KeyInfo const &, int count=0);
2091 LatexInfo(string const & par, bool isPatternString)
2092 : entidx_(-1), interval_(isPatternString, par)
2094 buildKeys(isPatternString);
2095 entries_ = vector<KeyInfo>();
2096 buildEntries(isPatternString);
2100 if (entries_.empty()) {
2103 if (entries_[0].keytype == KeyInfo::isTitle) {
2104 interval_.hasTitle = true;
2105 if (! entries_[0].disabled) {
2106 interval_.titleValue = entries_[0].head;
2109 interval_.titleValue = "";
2111 removeHead(entries_[0]);
2112 if (entries_.size() > 1)
2121 if (int(entries_.size()) > entidx_) {
2128 bool setNextKey(int idx) {
2129 if ((idx == entidx_) && (entidx_ >= 0)) {
2136 int find(int start, KeyInfo::KeyType keytype) const {
2140 while (tmpIdx < int(entries_.size())) {
2141 if (entries_[tmpIdx].keytype == keytype)
2147 int process(ostringstream & os, KeyInfo const & actual);
2148 int dispatch(ostringstream & os, int previousStart, KeyInfo & actual);
2149 // string show(int lastpos) { return interval.show(lastpos);}
2150 int nextNotIgnored(int start) { return interval_.nextNotIgnored(start);}
2151 KeyInfo &getKeyInfo(int keyinfo) {
2152 static KeyInfo invalidInfo = KeyInfo();
2153 if ((keyinfo < 0) || ( keyinfo >= int(entries_.size())))
2156 return entries_[keyinfo];
2158 void setForDefaultLang(KeyInfo const & defLang) {interval_.setForDefaultLang(defLang);}
2159 void addIntervall(int low, int up) { interval_.addIntervall(low, up); }
2163 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
2167 for (int i = start; i < end; i += 1 + skip) {
2171 if (c == '\\') skip = 1;
2175 else if (c == down) {
2178 if ((repeat <= 0) || (par[i+1] != up))
2192 size_t mathpostfixsize;
2194 size_t mathprefixsize;
2198 vector<MathEntry> entries_;
2203 void insert(string const & wait, size_t start, size_t prefixsize, size_t end, size_t postfixsize) {
2204 MathEntry m = MathEntry();
2206 m.mathStart = start;
2207 m.mathprefixsize = prefixsize;
2208 m.mathEnd = end + postfixsize;
2209 m.mathpostfixsize = postfixsize;
2210 m.mathSize = m.mathEnd - m.mathStart;
2211 entries_.push_back(m);
2213 bool empty() const { return entries_.empty(); }
2214 size_t getEndPos() const {
2215 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2218 return entries_[actualIdx_].mathEnd;
2220 size_t getStartPos() const {
2221 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2222 return 100000; /* definitely enough? */
2224 return entries_[actualIdx_].mathStart;
2226 size_t getPrefixSize() const {
2227 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2230 return entries_[actualIdx_].mathprefixsize;
2232 size_t getPostfixSize() const {
2233 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2236 return entries_[actualIdx_].mathpostfixsize;
2238 size_t getFirstPos() {
2240 return getStartPos();
2242 size_t getSize() const {
2243 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2246 return entries_[actualIdx_].mathSize;
2248 void incrEntry() { actualIdx_++; }
2251 void LatexInfo::buildEntries(bool isPatternString)
2253 static regex const rmath("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|x?x?alignat)\\*?\\})(\\{[0-9]+\\})?)");
2254 static regex const rkeys("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?)))");
2255 static bool disableLanguageOverride = false;
2256 smatch sub, submath;
2257 bool evaluatingRegexp = false;
2259 bool evaluatingMath = false;
2260 bool evaluatingCode = false;
2262 bool evaluatingOptional = false;
2263 size_t optionalEnd = 0;
2266 bool math_end_waiting = false;
2267 size_t math_pos = 10000;
2268 size_t math_prefix_size = 1;
2270 static vector<string> usedText = vector<string>();
2271 static bool removeMathHull = false;
2273 interval_.removeAccents();
2275 for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) {
2277 if ((submath.position(2) - submath.position(0)) %2 == 1) {
2278 // prefixed by odd count of '\\'
2281 if (math_end_waiting) {
2282 size_t pos = submath.position(size_t(2));
2283 if ((math_end == "$") &&
2284 (submath.str(2) == "$")) {
2285 mi.insert("$", math_pos, 1, pos, 1);
2286 math_end_waiting = false;
2288 else if ((math_end == "\\]") &&
2289 (submath.str(2) == "\\]")) {
2290 mi.insert("\\]", math_pos, 2, pos, 2);
2291 math_end_waiting = false;
2293 else if ((submath.str(3).compare("end") == 0) &&
2294 (submath.str(5).compare(math_end) == 0)) {
2295 mi.insert(math_end, math_pos, math_prefix_size, pos, submath.str(2).length());
2296 math_end_waiting = false;
2302 if (submath.str(3).compare("begin") == 0) {
2303 math_end_waiting = true;
2304 math_end = submath.str(5);
2305 math_pos = submath.position(size_t(2));
2306 math_prefix_size = submath.str(2).length();
2308 else if (submath.str(2).compare("\\[") == 0) {
2309 math_end_waiting = true;
2311 math_pos = submath.position(size_t(2));
2313 else if (submath.str(2) == "$") {
2314 size_t pos = submath.position(size_t(2));
2315 math_end_waiting = true;
2321 // Ignore language if there is math somewhere in pattern-string
2322 if (isPatternString) {
2323 for (auto s: usedText) {
2324 // Remove entries created in previous search runs
2327 usedText = vector<string>();
2330 keys["foreignlanguage"].disabled = true;
2331 disableLanguageOverride = true;
2332 removeMathHull = false;
2335 removeMathHull = true; // used later if not isPatternString
2336 disableLanguageOverride = false;
2340 if (disableLanguageOverride) {
2341 keys["foreignlanguage"].disabled = true;
2344 math_pos = mi.getFirstPos();
2345 for (sregex_iterator it(interval_.par.begin(), interval_.par.end(), rkeys), end; it != end; ++it) {
2347 if ((sub.position(2) - sub.position(0)) %2 == 1) {
2348 // prefixed by odd count of '\\'
2351 string key = sub.str(5);
2353 if (sub.str(2)[0] == '\\')
2354 key = sub.str(2)[1];
2359 KeysIterator it_key = keys.find(key);
2360 if (it_key != keys.end()) {
2361 if (it_key->second.keytype == KeyInfo::headRemove) {
2362 KeyInfo found1 = it_key->second;
2363 found1.disabled = true;
2364 found1.head = "\\" + key + "{";
2365 found1._tokenstart = sub.position(size_t(2));
2366 found1._tokensize = found1.head.length();
2367 found1._dataStart = found1._tokenstart + found1.head.length();
2368 int endpos = interval_.findclosing(found1._dataStart, interval_.par.length(), '{', '}', 1);
2369 found1._dataEnd = endpos;
2374 if (evaluatingRegexp) {
2375 if (sub.str(3).compare("endregexp") == 0) {
2376 evaluatingRegexp = false;
2377 // found._tokenstart already set
2378 found._dataEnd = sub.position(size_t(2)) + 13;
2379 found._dataStart = found._dataEnd;
2380 found._tokensize = found._dataEnd - found._tokenstart;
2381 found.parenthesiscount = 0;
2382 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2389 if (evaluatingMath) {
2390 if (size_t(sub.position(size_t(2))) < mi.getEndPos())
2392 evaluatingMath = false;
2394 math_pos = mi.getStartPos();
2396 if (it_key == keys.end()) {
2397 found = KeyInfo(KeyInfo::isStandard, 0, true);
2398 LYXERR(Debug::INFO, "Undefined key " << key << " ==> will be used as text");
2399 found = KeyInfo(KeyInfo::isText, 0, false);
2400 if (isPatternString) {
2401 found.keytype = KeyInfo::isChar;
2402 found.disabled = false;
2406 usedText.push_back(key);
2410 if (key.compare("regexp") == 0) {
2411 evaluatingRegexp = true;
2412 found._tokenstart = sub.position(size_t(2));
2413 found._tokensize = 0;
2417 // Handle the other params of key
2418 if (found.keytype == KeyInfo::isIgnored)
2420 else if (found.keytype == KeyInfo::isMath) {
2421 if (size_t(sub.position(size_t(2))) == math_pos) {
2423 found._tokenstart = sub.position(size_t(2));
2424 found._tokensize = mi.getSize();
2425 found._dataEnd = found._tokenstart + found._tokensize;
2426 found._dataStart = found._dataEnd;
2427 found.parenthesiscount = 0;
2428 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2429 if (removeMathHull) {
2430 interval_.addIntervall(found._tokenstart, found._tokenstart + mi.getPrefixSize());
2431 interval_.addIntervall(found._dataEnd - mi.getPostfixSize(), found._dataEnd);
2434 // Treate all math constructs as simple math
2435 interval_.par[found._tokenstart] = '$';
2436 interval_.par[found._dataEnd - mi.getPostfixSize()] = '$';
2437 interval_.addIntervall(found._tokenstart + 1, found._tokenstart + mi.getPrefixSize());
2438 interval_.addIntervall(found._dataEnd - mi.getPostfixSize() + 1, found._dataEnd);
2440 evaluatingMath = true;
2443 // begin|end of unknown env, discard
2444 // First handle tables
2445 // longtable|tabular
2446 bool discardComment;
2448 found.keytype = KeyInfo::doRemove;
2449 if ((sub.str(7).compare("longtable") == 0) ||
2450 (sub.str(7).compare("tabular") == 0)) {
2451 discardComment = true; /* '%' */
2454 discardComment = false;
2455 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
2457 string token = sub.str(7);
2458 if (regex_match(token, sub2, removeArgs)) {
2459 found.keytype = KeyInfo::removeWithArg;
2462 // discard spaces before pos(2)
2463 int pos = sub.position(size_t(2));
2465 for (count = 0; pos - count > 0; count++) {
2466 char c = interval_.par[pos-count-1];
2467 if (discardComment) {
2468 if ((c != ' ') && (c != '%'))
2474 found._tokenstart = pos - count;
2475 if (sub.str(3).compare(0, 5, "begin") == 0) {
2476 size_t pos1 = pos + sub.str(2).length();
2477 if (sub.str(7).compare("cjk") == 0) {
2478 pos1 = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2479 if ((interval_.par[pos1] == '{') && (interval_.par[pos1+1] == '}'))
2481 found.keytype = KeyInfo::isMain;
2482 found._dataStart = pos1;
2483 found._dataEnd = interval_.par.length();
2484 found.disabled = keys["foreignlanguage"].disabled;
2485 found.used = keys["foreignlanguage"].used;
2486 found._tokensize = pos1 - found._tokenstart;
2487 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2490 // Swallow possible optional params
2491 while (interval_.par[pos1] == '[') {
2492 pos1 = interval_.findclosing(pos1+1, interval_.par.length(), '[', ']')+1;
2494 // Swallow also the eventual parameter
2495 if (interval_.par[pos1] == '{') {
2496 found._dataEnd = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2499 found._dataEnd = pos1;
2501 found._dataStart = found._dataEnd;
2502 found._tokensize = count + found._dataEnd - pos;
2503 found.parenthesiscount = 0;
2504 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2505 found.disabled = true;
2509 // Handle "\end{...}"
2510 found._dataStart = pos + sub.str(2).length();
2511 found._dataEnd = found._dataStart;
2512 found._tokensize = count + found._dataEnd - pos;
2513 found.parenthesiscount = 0;
2514 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2515 found.disabled = true;
2519 else if (found.keytype != KeyInfo::isRegex) {
2520 found._tokenstart = sub.position(size_t(2));
2521 if (found.parenthesiscount == 0) {
2522 // Probably to be discarded
2523 size_t following_pos = sub.position(size_t(2)) + sub.str(5).length() + 1;
2524 char following = interval_.par[following_pos];
2525 if (following == ' ')
2526 found.head = "\\" + sub.str(5) + " ";
2527 else if (following == '=') {
2528 // like \uldepth=1000pt
2529 found.head = sub.str(2);
2532 found.head = "\\" + key;
2533 found._tokensize = found.head.length();
2534 found._dataEnd = found._tokenstart + found._tokensize;
2535 found._dataStart = found._dataEnd;
2538 int params = found._tokenstart + key.length() + 1;
2539 if (evaluatingOptional) {
2540 if (size_t(found._tokenstart) > optionalEnd) {
2541 evaluatingOptional = false;
2544 found.disabled = true;
2547 int optend = params;
2548 while (interval_.par[optend] == '[') {
2549 // discard optional parameters
2550 optend = interval_.findclosing(optend+1, interval_.par.length(), '[', ']') + 1;
2552 if (optend > params) {
2553 key += interval_.par.substr(params, optend-params);
2554 evaluatingOptional = true;
2555 optionalEnd = optend;
2556 if (found.keytype == KeyInfo::isSectioning) {
2557 // Remove optional values (but still keep in header)
2558 interval_.addIntervall(params, optend);
2561 string token = sub.str(7);
2563 if (interval_.par[optend] != '{') {
2565 found.parenthesiscount = 0;
2566 found.head = "\\" + key;
2569 closings = found.parenthesiscount;
2570 if (found.parenthesiscount == 1) {
2571 found.head = "\\" + key + "{";
2573 else if (found.parenthesiscount > 1) {
2575 found.head = sub.str(2) + "{";
2576 closings = found.parenthesiscount - 1;
2579 found.head = "\\" + key + "{";
2582 found._tokensize = found.head.length();
2583 found._dataStart = found._tokenstart + found.head.length();
2584 if (found.keytype == KeyInfo::doRemove) {
2586 size_t endpar = 2 + interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2587 if (endpar >= interval_.par.length())
2588 found._dataStart = interval_.par.length();
2590 found._dataStart = endpar;
2591 found._tokensize = found._dataStart - found._tokenstart;
2594 found._dataStart = found._tokenstart + found._tokensize;
2598 if (interval_.par.substr(found._dataStart, 15).compare("\\endarguments{}") == 0) {
2599 found._dataStart += 15;
2603 endpos = found._dataStart - 1;
2605 endpos = interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2606 if (found.keytype == KeyInfo::isList) {
2607 // Check if it really is list env
2608 static regex const listre("^([a-z]+)$");
2610 if (!regex_match(token, sub2, listre)) {
2611 // Change the key of this entry. It is not in a list/item environment
2612 found.keytype = KeyInfo::endArguments;
2615 if (found.keytype == KeyInfo::noMain) {
2616 evaluatingCode = true;
2618 codeStart = found._dataStart;
2620 else if (evaluatingCode) {
2621 if (size_t(found._dataStart) > codeEnd)
2622 evaluatingCode = false;
2623 else if (found.keytype == KeyInfo::isMain) {
2624 // Disable this key, treate it as standard
2625 found.keytype = KeyInfo::isStandard;
2626 found.disabled = true;
2627 if ((codeEnd +1 >= interval_.par.length()) &&
2628 (found._tokenstart == codeStart)) {
2629 // trickery, because the code inset starts
2630 // with \selectlanguage ...
2632 if (entries_.size() > 1) {
2633 entries_[entries_.size()-1]._dataEnd = codeEnd;
2638 if ((endpos == interval_.par.length()) &&
2639 (found.keytype == KeyInfo::doRemove)) {
2640 // Missing closing => error in latex-input?
2641 // therefore do not delete remaining data
2642 found._dataStart -= 1;
2643 found._dataEnd = found._dataStart;
2646 found._dataEnd = endpos;
2648 if (isPatternString) {
2649 keys[key].used = true;
2652 entries_.push_back(found);
2656 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
2658 stringstream s(keysstring);
2660 const char delim = '|';
2661 while (getline(s, key, delim)) {
2662 KeyInfo keyII(keyI);
2663 if (isPatternString) {
2666 else if ( !keys[key].used)
2667 keyII.disabled = true;
2672 void LatexInfo::buildKeys(bool isPatternString)
2675 static bool keysBuilt = false;
2676 if (keysBuilt && !isPatternString) return;
2678 // Keys to ignore in any case
2679 makeKey("text|textcyrillic|lyxmathsym|ensuremath", KeyInfo(KeyInfo::headRemove, 1, true), true);
2680 // Known standard keys with 1 parameter.
2681 // Split is done, if not at start of region
2682 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
2683 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
2684 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
2685 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
2686 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
2687 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
2689 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
2690 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2691 makeKey("section*|subsection*|subsubsection*|paragraph*",
2692 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2693 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2694 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
2696 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
2698 // Split is done, if not at start of region
2699 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
2700 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
2702 // Split is done always.
2703 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
2707 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2708 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2709 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2710 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2712 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2713 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2714 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2715 makeKey("thickspace|medspace|thinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2717 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2718 // Custom space/skip, remove the content (== length value)
2719 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
2720 // Found in fr/UserGuide.lyx
2721 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2723 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2724 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2725 // Known macros to remove (including their parameter)
2727 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
2728 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
2729 // handle like standard keys with 1 parameter.
2730 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
2732 // Ignore deleted text
2733 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString);
2734 // but preserve added text
2735 makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2737 // Macros to remove, but let the parameter survive
2739 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2741 // Remove language spec from content of these insets
2742 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
2744 // Same effect as previous, parameter will survive (because there is no one anyway)
2746 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2747 // Remove table decorations
2748 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
2749 // Discard shape-header.
2750 // For footnote or shortcut too, because of lang settings
2751 // and wrong handling if used 'KeyInfo::noMain'
2752 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2753 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2754 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2755 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2756 makeKey("hphantom|vphantom|note|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2757 makeKey("textgreek|textcyrillic", KeyInfo(KeyInfo::isStandard, 1, true), false);
2758 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
2759 // like ('tiny{}' or '\tiny ' ... )
2760 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
2762 // Survives, like known character
2763 // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2764 makeKey("tableofcontents", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2765 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
2767 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2768 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2769 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2771 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2772 // Remove RTL/LTR marker
2773 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2774 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
2775 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
2776 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
2777 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
2778 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
2779 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2780 if (isPatternString) {
2781 // Allow the first searched string to rebuild the keys too
2785 // no need to rebuild again
2791 * Keep the list of actual opened parentheses actual
2792 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2794 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2797 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2801 if (c == '\\') skip = 1;
2802 else if (c == '{') {
2805 else if (c == '}') {
2806 handleCloseP(i, closingAllowed);
2812 string Intervall::show(int lastpos)
2814 int idx = 0; /* int intervalls */
2817 for (idx = 0; idx <= ignoreidx; idx++) {
2818 while (i < lastpos) {
2820 if (i <= borders[idx].low) {
2821 if (borders[idx].low > lastpos)
2822 printsize = lastpos - i;
2824 printsize = borders[idx].low - i;
2825 s += par.substr(i, printsize);
2827 if (i >= borders[idx].low)
2828 i = borders[idx].upper;
2831 i = borders[idx].upper;
2837 s += par.substr(i, lastpos-i);
2843 void Intervall::output(ostringstream &os, int lastpos)
2845 // get number of chars to output
2846 int idx = 0; /* int intervalls */
2849 string startTitle = titleValue;
2850 for (idx = 0; idx <= ignoreidx; idx++) {
2852 if (i <= borders[idx].low) {
2854 if (borders[idx].low > lastpos)
2855 printsize = lastpos - i;
2857 printsize = borders[idx].low - i;
2858 if (printsize > 0) {
2859 os << startTitle << par.substr(i, printsize);
2861 printed += printsize;
2864 handleParentheses(i, false);
2865 if (i >= borders[idx].low)
2866 i = borders[idx].upper;
2869 i = borders[idx].upper;
2876 os << startTitle << par.substr(i, lastpos-i);
2877 printed += lastpos-i;
2879 handleParentheses(lastpos, false);
2881 if (keys["foreignlanguage"].disabled)
2882 startindex = actualdeptindex-langcount;
2884 startindex = actualdeptindex;
2885 for (int i = startindex; i > 0; --i) {
2888 if (hasTitle && (printed > 0))
2890 if (! isPatternString_)
2892 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2895 void LatexInfo::processRegion(int start, int region_end)
2897 while (start < region_end) { /* Let {[} and {]} survive */
2898 int cnt = interval_.isOpeningPar(start);
2900 // Closing is allowed past the region
2901 int closing = interval_.findclosing(start+1, interval_.par.length());
2902 interval_.addIntervall(start, start+1);
2903 interval_.addIntervall(closing, closing+1);
2907 start = interval_.nextNotIgnored(start+1);
2911 void LatexInfo::removeHead(KeyInfo const & actual, int count)
2913 if (actual.parenthesiscount == 0) {
2914 // "{\tiny{} ...}" ==> "{{} ...}"
2915 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2918 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2919 interval_.addIntervall(actual._tokenstart - count, actual._dataStart);
2920 interval_.addIntervall(actual._dataEnd, actual._dataEnd+1);
2924 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2927 switch (actual.keytype)
2929 case KeyInfo::isTitle: {
2931 nextKeyIdx = getNextKey();
2934 case KeyInfo::cleanToStart: {
2935 actual._dataEnd = actual._dataStart;
2936 nextKeyIdx = getNextKey();
2937 // Search for end of arguments
2938 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2940 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2941 entries_[i].disabled = true;
2943 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2945 while (interval_.par[actual._dataEnd] == ' ')
2947 interval_.addIntervall(0, actual._dataEnd+1);
2948 interval_.actualdeptindex = 0;
2949 interval_.depts[0] = actual._dataEnd+1;
2950 interval_.closes[0] = -1;
2953 case KeyInfo::isText:
2954 interval_.par[actual._tokenstart] = '#';
2955 //interval_.addIntervall(actual._tokenstart, actual._tokenstart+1);
2956 nextKeyIdx = getNextKey();
2958 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2959 if (actual.disabled)
2960 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2962 interval_.addIntervall(actual._dataStart, actual._dataEnd);
2965 case KeyInfo::isChar: {
2966 nextKeyIdx = getNextKey();
2969 case KeyInfo::isSize: {
2970 if (actual.disabled || (interval_.par[actual._dataStart] != '{') || (interval_.par[actual._dataStart-1] == ' ')) {
2971 if (actual.parenthesiscount == 0)
2972 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2974 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2976 nextKeyIdx = getNextKey();
2978 // Here _dataStart points to '{', so correct it
2979 actual._dataStart += 1;
2980 actual._tokensize += 1;
2981 actual.parenthesiscount = 1;
2982 if (interval_.par[actual._dataStart] == '}') {
2983 // Determine the end if used like '{\tiny{}...}'
2984 actual._dataEnd = interval_.findclosing(actual._dataStart+1, interval_.par.length()) + 1;
2985 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2988 // Determine the end if used like '\tiny{...}'
2989 actual._dataEnd = interval_.findclosing(actual._dataStart, interval_.par.length()) + 1;
2991 // Split on this key if not at start
2992 int start = interval_.nextNotIgnored(previousStart);
2993 if (start < actual._tokenstart) {
2994 interval_.output(os, actual._tokenstart);
2995 interval_.addIntervall(start, actual._tokenstart);
2997 // discard entry if at end of actual
2998 nextKeyIdx = process(os, actual);
3002 case KeyInfo::endArguments: {
3003 // Remove trailing '{}' too
3004 actual._dataStart += 1;
3005 actual._dataEnd += 1;
3006 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
3007 nextKeyIdx = getNextKey();
3010 case KeyInfo::noMain:
3012 case KeyInfo::isStandard: {
3013 if (actual.disabled) {
3015 processRegion(actual._dataStart, actual._dataStart+1);
3016 nextKeyIdx = getNextKey();
3018 // Split on this key if not at datastart of calling entry
3019 int start = interval_.nextNotIgnored(previousStart);
3020 if (start < actual._tokenstart) {
3021 interval_.output(os, actual._tokenstart);
3022 interval_.addIntervall(start, actual._tokenstart);
3024 // discard entry if at end of actual
3025 nextKeyIdx = process(os, actual);
3029 case KeyInfo::removeWithArg: {
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 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
3042 case KeyInfo::doRemove: {
3043 // Remove the key with all parameters and following spaces
3046 if (interval_.par[actual._dataEnd-1] == ' ' || interval_.par[actual._dataEnd-1] == '}')
3047 start = actual._dataEnd;
3049 start = actual._dataEnd+1;
3050 for (pos = start; pos < interval_.par.length(); pos++) {
3051 if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
3054 // Remove also enclosing parentheses [] and {}
3057 while (actual._tokenstart > numpars) {
3058 if (pos+numpars >= interval_.par.size())
3060 else if (interval_.par[pos+numpars] == ']' && interval_.par[actual._tokenstart-numpars-1] == '[')
3062 else if (interval_.par[pos+numpars] == '}' && interval_.par[actual._tokenstart-numpars-1] == '{')
3068 if (interval_.par[pos+numpars] == ' ')
3072 interval_.addIntervall(actual._tokenstart-numpars, pos+numpars+spaces);
3073 nextKeyIdx = getNextKey();
3076 case KeyInfo::isList: {
3077 // Discard space before _tokenstart
3079 for (count = 0; count < actual._tokenstart; count++) {
3080 if (interval_.par[actual._tokenstart-count-1] != ' ')
3083 nextKeyIdx = getNextKey();
3084 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
3086 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
3088 // How else can we catch this one?
3089 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
3090 entries_[i].disabled = true;
3092 actual._dataEnd = entries_[tmpIdx]._dataEnd;
3094 else if (nextKeyIdx > 0) {
3095 // Ignore any lang entries inside data region
3096 for (int i = nextKeyIdx; i < int(entries_.size()) && entries_[i]._tokenstart < actual._dataEnd; i++) {
3097 if (entries_[i].keytype == KeyInfo::isMain)
3098 entries_[i].disabled = true;
3101 if (actual.disabled) {
3102 interval_.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
3105 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
3107 if (interval_.par[actual._dataEnd+1] == '[') {
3108 int posdown = interval_.findclosing(actual._dataEnd+2, interval_.par.length(), '[', ']');
3109 if ((interval_.par[actual._dataEnd+2] == '{') &&
3110 (interval_.par[posdown-1] == '}')) {
3111 interval_.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
3112 interval_.addIntervall(posdown-1, posdown+1);
3115 interval_.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
3116 interval_.addIntervall(posdown, posdown+1);
3118 int blk = interval_.nextNotIgnored(actual._dataEnd+1);
3119 if (blk > posdown) {
3120 // Discard at most 1 space after empty item
3122 for (count = 0; count < 1; count++) {
3123 if (interval_.par[blk+count] != ' ')
3127 interval_.addIntervall(blk, blk+count);
3132 case KeyInfo::isSectioning: {
3133 // Discard spaces before _tokenstart
3135 int val = actual._tokenstart;
3136 for (count = 0; count < actual._tokenstart;) {
3137 val = interval_.previousNotIgnored(val-1);
3138 if (val < 0 || interval_.par[val] != ' ')
3141 count = actual._tokenstart - val;
3144 if (actual.disabled) {
3145 removeHead(actual, count);
3146 nextKeyIdx = getNextKey();
3148 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
3149 nextKeyIdx = process(os, actual);
3153 case KeyInfo::isMath: {
3154 // Same as regex, use the content unchanged
3155 nextKeyIdx = getNextKey();
3158 case KeyInfo::isRegex: {
3159 // DO NOT SPLIT ON REGEX
3161 nextKeyIdx = getNextKey();
3164 case KeyInfo::isIgnored: {
3165 // Treat like a character for now
3166 nextKeyIdx = getNextKey();
3169 case KeyInfo::isMain: {
3170 if (interval_.par.substr(actual._dataStart, 2) == "% ")
3171 interval_.addIntervall(actual._dataStart, actual._dataStart+2);
3172 if (actual._tokenstart > 0) {
3173 int prev = interval_.previousNotIgnored(actual._tokenstart - 1);
3174 if ((prev >= 0) && interval_.par[prev] == '%')
3175 interval_.addIntervall(prev, prev+1);
3177 if (actual.disabled) {
3179 interval_.langcount++;
3180 if ((interval_.par.substr(actual._dataStart, 3) == " \\[") ||
3181 (interval_.par.substr(actual._dataStart, 8) == " \\begin{")) {
3182 // Discard also the space before math-equation
3183 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
3185 nextKeyIdx = getNextKey();
3186 // interval.resetOpenedP(actual._dataStart-1);
3189 if (actual._tokenstart < 26) {
3190 // for the first (and maybe dummy) language
3191 interval_.setForDefaultLang(actual);
3193 interval_.resetOpenedP(actual._dataStart-1);
3197 case KeyInfo::invalid:
3198 case KeyInfo::headRemove:
3199 // These two cases cannot happen, already handled
3202 // LYXERR(Debug::INFO, "Unhandled keytype");
3203 nextKeyIdx = getNextKey();
3210 int LatexInfo::process(ostringstream & os, KeyInfo const & actual )
3212 int end = interval_.nextNotIgnored(actual._dataEnd);
3213 int oldStart = actual._dataStart;
3214 int nextKeyIdx = getNextKey();
3216 if ((nextKeyIdx < 0) ||
3217 (entries_[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
3218 (entries_[nextKeyIdx].keytype == KeyInfo::invalid)) {
3219 if (oldStart <= end) {
3220 processRegion(oldStart, end);
3225 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
3227 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
3228 (void) dispatch(os, actual._dataStart, nextKey);
3229 end = nextKey._tokenstart;
3232 processRegion(oldStart, nextKey._tokenstart);
3233 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
3235 oldStart = nextKey._dataEnd+1;
3237 // now nextKey is either invalid or is outside of actual._dataEnd
3238 // output the remaining and discard myself
3239 if (oldStart <= end) {
3240 processRegion(oldStart, end);
3242 if (interval_.par.size() > (size_t) end && interval_.par[end] == '}') {
3244 // This is the normal case.
3245 // But if using the firstlanguage, the closing may be missing
3247 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
3249 if (actual._dataEnd < end)
3250 output_end = interval_.nextNotIgnored(actual._dataEnd);
3251 else if (interval_.par.size() > (size_t) end)
3252 output_end = interval_.nextNotIgnored(end);
3254 output_end = interval_.par.size();
3255 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
3256 interval_.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
3258 // Remove possible empty data
3259 int dstart = interval_.nextNotIgnored(actual._dataStart);
3260 while (interval_.isOpeningPar(dstart) == 1) {
3261 interval_.addIntervall(dstart, dstart+1);
3262 int dend = interval_.findclosing(dstart+1, output_end);
3263 interval_.addIntervall(dend, dend+1);
3264 dstart = interval_.nextNotIgnored(dstart+1);
3266 if (dstart < output_end)
3267 interval_.output(os, output_end);
3269 interval_.addIntervall(0, end);
3271 interval_.addIntervall(actual._tokenstart, end);
3275 string splitOnKnownMacros(string par, bool isPatternString)
3278 LatexInfo li(par, isPatternString);
3279 // LYXERR(Debug::INFO, "Berfore split: " << par);
3280 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
3282 DummyKey._tokensize = 0;
3283 DummyKey._dataStart = 0;
3284 DummyKey._dataEnd = par.length();
3285 DummyKey.disabled = true;
3286 int firstkeyIdx = li.getFirstKey();
3288 if (firstkeyIdx >= 0) {
3289 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
3290 DummyKey._tokenstart = firstKey._tokenstart;
3292 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
3293 // Use dummy firstKey
3294 firstKey = DummyKey;
3295 (void) li.setNextKey(firstkeyIdx);
3298 if (par.substr(firstKey._dataStart, 2) == "% ")
3299 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
3301 nextkeyIdx = li.process(os, firstKey);
3302 while (nextkeyIdx >= 0) {
3303 // Check for a possible gap between the last
3304 // entry and this one
3305 int datastart = li.nextNotIgnored(firstKey._dataStart);
3306 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
3307 if ((nextKey._tokenstart > datastart)) {
3309 firstKey._dataStart = datastart;
3310 firstKey._dataEnd = par.length();
3311 (void) li.setNextKey(nextkeyIdx);
3312 // Fake the last opened parenthesis
3313 li.setForDefaultLang(firstKey);
3314 nextkeyIdx = li.process(os, firstKey);
3317 if (nextKey.keytype != KeyInfo::isMain) {
3318 firstKey._dataStart = datastart;
3319 firstKey._dataEnd = nextKey._dataEnd+1;
3320 (void) li.setNextKey(nextkeyIdx);
3321 li.setForDefaultLang(firstKey);
3322 nextkeyIdx = li.process(os, firstKey);
3325 nextkeyIdx = li.process(os, nextKey);
3329 // Handle the remaining
3330 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
3331 firstKey._dataEnd = par.length();
3333 if ((firstKey._dataStart < firstKey._dataEnd) &&
3334 (par[firstKey._dataStart] != '}')) {
3335 li.setForDefaultLang(firstKey);
3336 (void) li.process(os, firstKey);
3339 // return string definitelly impossible to match, but should be known
3342 s = par; /* no known macros found */
3343 // LYXERR(Debug::INFO, "After split: " << s);
3348 * Try to unify the language specs in the latexified text.
3349 * Resulting modified string is set to "", if
3350 * the searched tex does not contain all the features in the search pattern
3352 static string correctlanguagesetting(string par, bool isPatternString, bool withformat, lyx::Buffer *pbuf = nullptr)
3354 static Features regex_f;
3355 static int missed = 0;
3356 static bool regex_with_format = false;
3358 int parlen = par.length();
3360 while ((parlen > 0) && (par[parlen-1] == '\n')) {
3363 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
3364 // Happens to be there in case of description or labeling environment
3369 // Split the latex input into pieces which
3370 // can be digested by our search engine
3371 LYXERR(Debug::FINDVERBOSE, "input: \"" << par << "\"");
3372 if (isPatternString && (pbuf != nullptr)) { // Check if we should disable/enable test for language
3373 // We check for polyglossia, because in runparams.flavor we use Flavor::XeTeX
3374 string doclang = pbuf->params().language->polyglossia();
3375 static regex langre("\\\\(foreignlanguage)\\{([^\\}]+)\\}");
3377 bool toIgnoreLang = true;
3378 for (sregex_iterator it(par.begin(), par.end(), langre), end; it != end; ++it) {
3380 if (sub.str(2) != doclang) {
3381 toIgnoreLang = false;
3385 setIgnoreFormat("language", toIgnoreLang, false);
3388 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
3389 LYXERR(Debug::FINDVERBOSE, "After splitOnKnownMacros:\n\"" << result << "\"");
3392 result = par.substr(0, parlen);
3393 if (isPatternString) {
3396 regex_f = identifyFeatures(result);
3397 string features = "";
3398 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3399 string a = it->first;
3400 regex_with_format = true;
3401 features += " " + a;
3402 // LYXERR(Debug::INFO, "Identified regex format:" << a);
3404 LYXERR(Debug::FINDVERBOSE, "Identified Features" << features);
3407 } else if (regex_with_format) {
3408 Features info = identifyFeatures(result);
3409 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3410 string a = it->first;
3411 bool b = it->second;
3412 if (b && ! info[a]) {
3414 LYXERR(Debug::FINDVERBOSE, "Missed(" << missed << " " << a <<", srclen = " << parlen );
3421 // LYXERR(Debug::INFO, "No regex formats");
3427 // Remove trailing closure of math, macros and environments, so to catch parts of them.
3428 static void identifyClosing(string & t, bool ignoreformat)
3431 LYXERR(Debug::FINDVERBOSE, "identifyClosing(): t now is '" << t << "'");
3432 if (regex_replace(t, t, "(.*[^\\\\])\\$$", "$1"))
3434 if (regex_replace(t, t, "(.*[^\\\\])\\\\\\]$", "$1"))
3436 if (regex_replace(t, t, "(.*[^\\\\])\\\\end\\{[a-zA-Z_]+\\*?\\}$", "$1"))
3438 if (! ignoreformat) {
3439 if (regex_replace(t, t, "(.*[^\\\\])\\}$", "$1"))
3447 static int num_replaced = 0;
3448 static bool previous_single_replace = true;
3450 void MatchStringAdv::CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string)
3453 if (regexp_str.empty() || regexp2_str.empty()) {
3454 regexIsValid = false;
3455 regexError = "Invalid empty regex";
3458 // Handle \w properly
3459 QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
3460 if (! opt.casesensitive) {
3461 popts |= QRegularExpression::CaseInsensitiveOption;
3463 regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
3464 regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
3466 if (regexp.isValid() && regexp2.isValid()) {
3467 regexIsValid = true;
3468 // Check '{', '}' pairs inside the regex
3471 for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
3472 char c = par_as_string[i];
3479 else if (c == '}') {
3486 if (balanced != 0) {
3487 regexIsValid = false;
3488 regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
3492 regexIsValid = false;
3493 if (!regexp.isValid())
3494 regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
3496 regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
3499 (void)par_as_string;
3500 if (opt.casesensitive) {
3501 regexp = regex(regexp_str);
3502 regexp2 = regex(regexp2_str);
3505 regexp = regex(regexp_str, std::regex_constants::icase);
3506 regexp2 = regex(regexp2_str, std::regex_constants::icase);
3511 static void modifyRegexForMatchWord(string &t)
3514 regex wordre("(\\\\)*((\\.|\\\\b))");
3517 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
3519 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
3522 else if (sub.str(2) == "\\\\b")
3524 if (lastpos < (size_t) sub.position(2))
3525 s += t.substr(lastpos, sub.position(2) - lastpos);
3527 lastpos = sub.position(2) + sub.length(2);
3530 s = "\\b" + t + "\\b";
3534 else if (lastpos < t.length())
3535 s += t.substr(lastpos, t.length() - lastpos);
3536 t = "\\b" + s + "\\b";
3539 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt)
3540 : p_buf(&buf), p_first_buf(&buf), opt(opt)
3542 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
3543 docstring const & ds = stringifySearchBuffer(find_buf, opt);
3545 CreateRegexp(opt, "", "", "");
3548 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
3549 if (opt.replace_all && previous_single_replace) {
3550 previous_single_replace = false;
3553 else if (!opt.replace_all) {
3554 num_replaced = 0; // count number of replaced strings
3555 previous_single_replace = true;
3557 // When using regexp, braces are hacked already by escape_for_regex()
3558 par_as_string = normalize(ds, opt.ignoreformat);
3560 close_wildcards = 0;
3562 size_t lead_size = 0;
3563 // correct the language settings
3564 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat, &buf);
3565 if (par_as_string.empty()) {
3566 CreateRegexp(opt, "", "", "");
3569 opt.matchAtStart = false;
3571 identifyClosing(par_as_string, opt.ignoreformat); // Removes math closings ($, ], ...) at end of string
3572 if (opt.ignoreformat) {
3576 lead_size = identifyLeading(par_as_string);
3578 lead_as_string = par_as_string.substr(0, lead_size);
3579 string lead_as_regex_string = string2regex(lead_as_string);
3580 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3581 string par_as_regex_string_nolead = string2regex(par_as_string_nolead);
3582 /* Handle whole words too in this case
3584 if (opt.matchword) {
3585 par_as_regex_string_nolead = "\\b" + par_as_regex_string_nolead + "\\b";
3586 opt.matchword = false;
3588 string regexp_str = "(" + lead_as_regex_string + ")()" + par_as_regex_string_nolead;
3589 string regexp2_str = "(" + lead_as_regex_string + ")(.*?)" + par_as_regex_string_nolead;
3590 CreateRegexp(opt, regexp_str, regexp2_str);
3592 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3593 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3597 if (!opt.ignoreformat) {
3598 lead_size = identifyLeading(par_as_string);
3599 LYXERR(Debug::FINDVERBOSE, "Lead_size: " << lead_size);
3600 lead_as_string = par_as_string.substr(0, lead_size);
3601 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3604 // Here we are using regexp
3605 LASSERT(use_regexp, /**/);
3607 string lead_as_regexp;
3608 if (lead_size > 0) {
3609 lead_as_regexp = string2regex(par_as_string.substr(0, lead_size));
3610 (void)regex_replace(par_as_string_nolead, par_as_string_nolead, "\\$$", "");
3611 (void)regex_replace(par_as_string_nolead, par_as_string_nolead, "}$", "");
3612 par_as_string = par_as_string_nolead;
3613 LYXERR(Debug::FINDVERBOSE, "lead_as_regexp is '" << lead_as_regexp << "'");
3614 LYXERR(Debug::FINDVERBOSE, "par_as_string now is '" << par_as_string << "'");
3616 // LYXERR(Debug::FINDVERBOSE, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
3617 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
3618 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
3619 // LYXERR(Debug::FINDVERBOSE, "par_as_string now is '" << par_as_string << "'");
3621 size_t lng = par_as_string.size();
3622 if (!opt.ignoreformat) {
3623 // Remove extra '\}' at end if not part of \{\.\}
3625 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
3627 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
3636 if (lng < par_as_string.size())
3637 par_as_string = par_as_string.substr(0,lng);
3639 LYXERR(Debug::FINDVERBOSE, "par_as_string after correctRegex is '" << par_as_string << "'");
3640 if ((lng > 0) && (par_as_string[0] == '^')) {
3641 par_as_string = par_as_string.substr(1);
3643 opt.matchAtStart = true;
3645 // LYXERR(Debug::FINDVERBOSE, "par_as_string now is '" << par_as_string << "'");
3646 // LYXERR(Debug::FINDVERBOSE, "Open braces: " << open_braces);
3647 // LYXERR(Debug::FINDVERBOSE, "Replaced text (to be used as regex): " << par_as_string);
3649 // If entered regexp must match at begin of searched string buffer
3650 // Kornel: Added parentheses to use $1 for size of the leading string
3654 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
3655 // Unfortunately is '\1', '\2', etc not working for strings with extra format
3656 // so the convert has no effect in that case
3657 for (int i = 7; i > 0; --i) {
3658 string orig = "\\\\" + std::to_string(i);
3659 string dest = "\\" + std::to_string(i+2);
3660 while (regex_replace(par_as_string, par_as_string, orig, dest));
3662 if (opt.matchword) {
3663 modifyRegexForMatchWord(par_as_string);
3664 opt.matchword = false;
3666 regexp_str = "(" + lead_as_regexp + ")()" + par_as_string;
3667 regexp2_str = "(" + lead_as_regexp + ")(.*?)" + par_as_string;
3669 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3670 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3671 CreateRegexp(opt, regexp_str, regexp2_str, par_as_string);
3675 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
3679 mres.searched_size = len;
3681 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
3684 docstring docstr = stringifyFromForSearch(opt, cur, len);
3686 str = normalize(docstr, opt.ignoreformat);
3687 if (!opt.ignoreformat) {
3688 str = correctlanguagesetting(str, false, !opt.ignoreformat);
3689 // remove closing '}' and '\n' to allow for use of '$' in regex
3690 size_t lng = str.size();
3691 while ((lng > 1) && ((str[lng -1] == '}') || (str[lng -1] == '\n')))
3693 if (lng != str.size()) {
3694 str = str.substr(0, lng);
3698 mres.match_len = -1;
3701 LYXERR(Debug::FINDVERBOSE, "After normalization: Matching against:\n'" << str << "'");
3703 LASSERT(use_regexp, /**/);
3705 // use_regexp always true
3706 LYXERR(Debug::FINDVERBOSE, "Searching in regexp mode: at_begin=" << at_begin);
3708 QString qstr = QString::fromStdString(str);
3709 QRegularExpression const *p_regexp;
3710 QRegularExpression::MatchType flags = QRegularExpression::NormalMatch;
3714 p_regexp = ®exp2;
3716 QRegularExpressionMatch match = p_regexp->match(qstr, 0, flags);
3717 if (!match.hasMatch())
3720 regex const *p_regexp;
3721 regex_constants::match_flag_type flags;
3723 flags = regex_constants::match_continuous;
3726 flags = regex_constants::match_default;
3727 p_regexp = ®exp2;
3729 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
3730 if (re_it == sregex_iterator())
3732 match_results<string::const_iterator> const & m = *re_it;
3734 // Whole found string, including the leading
3735 // std: m[0].second - m[0].first
3736 // Qt: match.capturedEnd(0) - match.capturedStart(0)
3738 // Size of the leading string
3739 // std: m[1].second - m[1].first
3740 // Qt: match.capturedEnd(1) - match.capturedStart(1)
3741 int leadingsize = 0;
3743 if (match.lastCapturedIndex() > 0) {
3744 leadingsize = match.capturedEnd(1) - match.capturedStart(1);
3749 leadingsize = m[1].second - m[1].first;
3753 mres.match_prefix = match.capturedEnd(2) - match.capturedStart(2);
3754 mres.match_len = match.capturedEnd(0) - match.capturedEnd(2);
3755 // because of different number of closing at end of string
3756 // we have to 'unify' the length of the post-match.
3757 // Done by ignoring closing parenthesis and linefeeds at string end
3758 int matchend = match.capturedEnd(0);
3759 size_t strsize = qstr.size();
3760 if (!opt.ignoreformat) {
3761 while (mres.match_len > 0) {
3762 QChar c = qstr.at(matchend - 1);
3763 if ((c == '\n') || (c == '}') || (c == '{')) {
3770 while (strsize > (size_t) match.capturedEnd(0)) {
3771 QChar c = qstr.at(strsize-1);
3772 if ((c == '\n') || (c == '}')) {
3779 // LYXERR0(qstr.toStdString());
3780 mres.match2end = strsize - matchend;
3781 mres.pos = match.capturedStart(2);
3783 mres.match_prefix = m[2].second - m[2].first;
3784 mres.match_len = m[0].second - m[2].second;
3785 // ignore closing parenthesis and linefeeds at string end
3786 size_t strend = m[0].second - m[0].first;
3787 int matchend = strend;
3788 size_t strsize = str.size();
3789 if (!opt.ignoreformat) {
3790 while (mres.match_len > 0) {
3791 char c = str.at(matchend - 1);
3792 if ((c == '\n') || (c == '}') || (c == '{')) {
3799 while (strsize > strend) {
3800 if ((str.at(strsize-1) == '}') || (str.at(strsize-1) == '\n')) {
3808 mres.match2end = strsize - matchend;
3809 mres.pos = m[2].first - m[0].first;;
3811 if (mres.match2end < 0)
3813 mres.leadsize = leadingsize;
3815 if (mres.match_len > 0) {
3816 string a0 = match.captured(0).mid(mres.pos + mres.match_prefix, mres.match_len).toStdString();
3817 mres.result.push_back(a0);
3818 for (int i = 3; i <= match.lastCapturedIndex(); i++) {
3819 mres.result.push_back(match.captured(i).toStdString());
3823 if (mres.match_len > 0) {
3824 string a0 = m[0].str().substr(mres.pos + mres.match_prefix, mres.match_len);
3825 mres.result.push_back(a0);
3826 for (size_t i = 3; i < m.size(); i++) {
3827 mres.result.push_back(m[i]);
3836 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
3838 MatchResult mres = findAux(cur, len, at_begin);
3839 int res = mres.match_len;
3840 LYXERR(Debug::FINDVERBOSE,
3841 "res=" << res << ", at_begin=" << at_begin
3842 << ", matchAtStart=" << opt.matchAtStart
3843 << ", inTexted=" << cur.inTexted());
3844 if (opt.matchAtStart) {
3847 else if (mres.match_prefix > 0)
3856 static bool simple_replace(string &t, string from, string to)
3858 regex repl("(\\\\)*(" + from + ")");
3862 for (sregex_iterator it(t.begin(), t.end(), repl), end; it != end; ++it) {
3864 if ((sub.position(2) - sub.position(0)) % 2 == 1)
3866 if (lastpos < (size_t) sub.position(2))
3867 s += t.substr(lastpos, sub.position(2) - lastpos);
3869 lastpos = sub.position(2) + sub.length(2);
3873 else if (lastpos < t.length())
3874 s += t.substr(lastpos, t.length() - lastpos);
3880 string MatchStringAdv::normalize(docstring const & s, bool ignore_format) const
3883 t = lyx::to_utf8(s);
3884 // Remove \n at begin
3885 while (!t.empty() && t[0] == '\n')
3887 // Remove \n* at end
3888 while (!t.empty() && t[t.size() - 1] == '\n') {
3889 t = t.substr(0, t.size() - 1);
3892 // Handle all other '\n'
3893 while ((pos = t.find("\n")) != string::npos) {
3894 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
3896 if (isAlnumASCII(t[pos+1])) {
3897 t.replace(pos-2, 3, " ");
3900 t.replace(pos-2, 3, "");
3904 if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
3905 // '\n' adjacent to non-alpha-numerics, discard
3906 t.replace(pos, 1, "");
3909 // Replace all other \n with spaces
3910 t.replace(pos, 1, " ");
3912 if (!ignore_format) {
3914 while ((pos > count + 1) && (t[pos - 1 -count] == '%')) {
3918 t.replace(pos - count, count, "");
3923 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
3924 // Kornel: Added textsl, textsf, textit, texttt and noun
3925 // + allow to seach for colored text too
3926 LYXERR(Debug::FINDVERBOSE, "Removing stale empty macros from: " << t);
3927 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
3928 LYXERR(Debug::FINDVERBOSE, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3929 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
3930 LYXERR(Debug::FINDVERBOSE, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3931 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
3937 docstring stringifyFromCursor(DocIterator const & cur, int len)
3939 LYXERR(Debug::FINDVERBOSE, "Stringifying with len=" << len << " from cursor at pos: " << cur);
3940 if (cur.inTexted()) {
3941 Paragraph const & par = cur.paragraph();
3942 // TODO what about searching beyond/across paragraph breaks ?
3943 // TODO Try adding a AS_STR_INSERTS as last arg
3944 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
3945 int(par.size()) : cur.pos() + len;
3946 // OutputParams runparams(&cur.buffer()->params().encoding());
3947 OutputParams runparams(encodings.fromLyXName("utf8"));
3948 runparams.nice = true;
3949 setFindParams(runparams);
3950 int option = AS_STR_INSETS | AS_STR_PLAINTEXT;
3951 if (ignoreFormats.getDeleted()) {
3952 option |= AS_STR_SKIPDELETE;
3953 runparams.for_search = OutputParams::SearchWithoutDeleted;
3956 runparams.for_search = OutputParams::SearchWithDeleted;
3958 if (ignoreFormats.getNonContent()) {
3959 runparams.for_search |= OutputParams::SearchNonOutput;
3961 LYXERR(Debug::FINDVERBOSE, "Stringifying with cur: "
3962 << cur << ", from pos: " << cur.pos() << ", end: " << end);
3963 docstring res = from_utf8(latexNamesToUtf8(par.asString(cur.pos(), end,
3966 LYXERR(Debug::FIND, "Stringified text from pos(" << cur.pos() << ") len(" << len << "): " << res);
3968 } else if (cur.inMathed()) {
3969 CursorSlice cs = cur.top();
3970 MathData md = cs.cell();
3971 MathData::const_iterator it_end =
3972 (( len == -1 || cs.pos() + len > int(md.size()))
3974 : md.begin() + cs.pos() + len );
3976 for (MathData::const_iterator it = md.begin() + cs.pos(); it != it_end; ++it)
3978 docstring res = from_utf8(latexNamesToUtf8(asString(md2)));
3979 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Stringified math from pos(" << cur.pos() << ") len(" << len << "): " << res);
3982 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Don't know how to stringify from here: " << cur);
3986 /** Computes the LaTeX export of buf starting from cur and ending len positions
3987 * after cur, if len is positive, or at the paragraph or innermost inset end
3990 docstring latexifyFromCursor(DocIterator const & cur, int len)
3993 LYXERR(Debug::FINDVERBOSE, "Latexifying with len=" << len << " from cursor at pos: " << cur);
3994 LYXERR(Debug::FINDVERBOSE, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
3995 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
3997 Buffer const & buf = *cur.buffer();
3999 odocstringstream ods;
4001 //OutputParams runparams(&buf.params().encoding());
4002 OutputParams runparams(encodings.fromLyXName("utf8"));
4003 runparams.nice = false;
4004 setFindParams(runparams);
4005 if (ignoreFormats.getDeleted()) {
4006 runparams.for_search = OutputParams::SearchWithoutDeleted;
4009 runparams.for_search = OutputParams::SearchWithDeleted;
4011 if (ignoreFormats.getNonContent()) {
4012 runparams.for_search |= OutputParams::SearchNonOutput;
4015 if (cur.inTexted()) {
4016 // @TODO what about searching beyond/across paragraph breaks ?
4017 pos_type endpos = cur.paragraph().size();
4018 if (len != -1 && endpos > cur.pos() + len)
4019 endpos = cur.pos() + len;
4020 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
4021 string(), cur.pos(), endpos, true);
4022 LYXERR(Debug::FIND, "Latexified text from pos(" << cur.pos() << ") len(" << len << "): " << ods.str());
4024 } else if (cur.inMathed()) {
4025 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
4026 for (int s = cur.depth() - 1; s >= 0; --s) {
4027 CursorSlice const & cs = cur[s];
4028 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
4029 TeXMathStream ws(os);
4030 cs.asInsetMath()->asHullInset()->header_write(ws);
4035 CursorSlice const & cs = cur.top();
4036 MathData md = cs.cell();
4037 MathData::const_iterator it_end =
4038 ((len == -1 || cs.pos() + len > int(md.size()))
4040 : md.begin() + cs.pos() + len);
4042 for (MathData::const_iterator it = md.begin() + cs.pos();
4046 ods << asString(md2);
4047 // Retrieve the math environment type, and add '$' or '$]'
4048 // or others (\end{equation}) accordingly
4049 for (int s = cur.depth() - 1; s >= 0; --s) {
4050 CursorSlice const & cs2 = cur[s];
4051 InsetMath * inset = cs2.asInsetMath();
4052 if (inset && inset->asHullInset()) {
4053 TeXMathStream ws(os);
4054 inset->asHullInset()->footer_write(ws);
4058 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Latexified math from pos(" << cur.pos() << ") len(" << len << "): " << ods.str());
4060 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Don't know how to stringify from here: " << cur);
4065 #if defined(ResultsDebug)
4067 static void displayMResult(MatchResult &mres, string from, DocIterator & cur)
4069 LYXERR0( "from:\t\t\t" << from);
4071 if (mres.pos_len > 0) {
4073 status = "FINALSEARCH";
4076 if (mres.match_len > 0) {
4077 if ((mres.match_prefix == 0) && (mres.pos == mres.leadsize))
4078 status = "Good Match";
4080 status = "Matched in";
4083 status = "MissedSearch";
4086 LYXERR0( status << "(" << cur.pos() << " ... " << mres.searched_size + cur.pos() << ") cur.lastpos(" << cur.lastpos() << ")");
4087 if ((mres.leadsize > 0) || (mres.match_len > 0) || (mres.match2end > 0))
4088 LYXERR0( "leadsize(" << mres.leadsize << ") match_len(" << mres.match_len << ") match2end(" << mres.match2end << ")");
4089 if ((mres.pos > 0) || (mres.match_prefix > 0))
4090 LYXERR0( "pos(" << mres.pos << ") match_prefix(" << mres.match_prefix << ")");
4091 for (size_t i = 0; i < mres.result.size(); i++)
4092 LYXERR0( "Match " << i << " = \"" << mres.result[i] << "\"");
4094 #define displayMres(s, txt, cur) displayMResult(s, txt, cur);
4096 #define displayMres(s, txt, cur)
4099 /** Finalize an advanced find operation, advancing the cursor to the innermost
4100 ** position that matches, plus computing the length of the matching text to
4102 ** Return the cur.pos() difference between start and end of found match
4104 MatchResult findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, MatchResult const & expected = MatchResult(-1))
4106 // Search the foremost position that matches (avoids find of entire math
4107 // inset when match at start of it)
4108 DocIterator old_cur(cur.buffer());
4110 static MatchResult fail = MatchResult();
4111 MatchResult max_match;
4112 // If (prefix_len > 0) means that forwarding 1 position will remove the complete entry
4113 // Happens with e.g. hyperlinks
4114 // either one sees "http://www.bla.bla" or nothing
4115 // so the search for "www" gives prefix_len = 7 (== sizeof("http://")
4116 // and although we search for only 3 chars, we find the whole hyperlink inset
4117 bool at_begin = (expected.match_prefix == 0);
4118 if (!match.opt.forward && match.opt.ignoreformat) {
4119 if (expected.pos > 0)
4122 LASSERT(at_begin, /**/);
4123 if (expected.match_len > 0 && at_begin) {
4124 // Search for deepest match
4126 max_match = expected;
4128 size_t d = cur.depth();
4132 if (cur.depth() < d)
4134 if (cur.depth() == d)
4137 while (cur && cur.depth() > lastd) {
4138 lastd = cur.depth();
4139 mres = match(cur, -1, at_begin);
4140 displayMres(mres, "Checking innermost", cur);
4141 if (mres.match_len > 0)
4146 if (mres.match_len < expected.match_len)
4154 // (expected.match_len <= 0)
4155 mres = match(cur); /* match valid only if not searching whole words */
4156 displayMres(mres, "Start with negative match", cur);
4159 if (max_match.match_len <= 0) return fail;
4160 LYXERR(Debug::FINDVERBOSE, "Ok");
4162 // Compute the match length
4164 if (cur.pos() + len > cur.lastpos())
4167 LASSERT(match.use_regexp, /**/);
4170 int maxl = cur.lastpos() - cur.pos();
4171 // Greedy behaviour while matching regexps
4172 while (maxl > minl) {
4174 mres2 = match(cur, len, at_begin);
4175 displayMres(mres2, "Finalize loop", cur);
4176 int actual_match_len = mres2.match_len;
4177 if (actual_match_len >= max_match.match_len) {
4178 // actual_match_len > max_match _can_ happen,
4179 // if the search area splits
4180 // some following word so that the regex
4181 // (e.g. 'r.*r\b' matches 'r' from the middle of the
4183 // This means, the len value is too big
4184 actual_match_len = max_match.match_len;
4186 max_match.match_len = actual_match_len;
4188 if (maxl - minl < 4)
4189 len = (maxl + minl)/2;
4191 len = minl + (maxl - minl + 3)/4;
4194 // (actual_match_len < max_match.match_len)
4196 len = (maxl + minl)/2;
4201 // Search for real start of matched characters
4203 MatchResult actual_match;
4206 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
4207 if (cur.depth() < old_cur.depth()) {
4209 LYXERR(Debug::INFO, "cur.depth() < old_cur.depth(), this should never happen");
4212 if (cur.pos() != old_cur.pos()) {
4213 // OK, forwarded 1 pos in actual inset
4214 actual_match = match(cur, len-1, at_begin);
4215 if (actual_match.match_len == max_match.match_len) {
4216 // Ha, got it! The shorter selection has the same match length
4219 max_match = actual_match;
4222 // OK, the shorter selection matches less chars, revert to previous value
4228 LYXERR(Debug::INFO, "cur.pos() == old_cur.pos(), this should never happen");
4229 actual_match = match(cur, len, at_begin);
4230 if (actual_match.match_len == max_match.match_len) {
4232 max_match = actual_match;
4239 max_match.pos_len = len;
4240 displayMres(max_match, "SEARCH RESULT", cur)
4247 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
4251 bool repeat = false;
4252 DocIterator orig_cur; // to be used if repeat not successful
4253 MatchResult orig_mres;
4254 while (!theApp()->longOperationCancelled() && cur) {
4255 //(void) findAdvForwardInnermost(cur);
4256 LYXERR(Debug::FINDVERBOSE, "findForwardAdv() cur: " << cur);
4257 MatchResult mres = match(cur, -1, false);
4258 string msg = "Starting";
4261 displayMres(mres, msg + " findForwardAdv", cur)
4262 int match_len = mres.match_len;
4263 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
4264 LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
4267 if (match_len <= 0) {
4268 // This should exit nested insets, if any, or otherwise undefine the currsor.
4269 cur.pos() = cur.lastpos();
4270 LYXERR(Debug::FINDVERBOSE, "Advancing pos: cur=" << cur);
4273 else { // match_len > 0
4274 // Try to find the begin of searched string
4276 int firstInvalid = cur.lastpos() - cur.pos();
4278 int incrmatch = (mres.match_prefix + mres.pos - mres.leadsize + 1)*3/4;
4279 int incrcur = (firstInvalid + 1 )*3/4;
4280 if (incrcur < incrmatch)
4281 increment = incrcur;
4283 increment = incrmatch;
4287 LYXERR(Debug::FINDVERBOSE, "Set increment to " << increment);
4288 while (increment > 0) {
4289 DocIterator old_cur = cur;
4290 if (cur.pos() + increment >= cur.lastpos()) {
4294 cur.pos() = cur.pos() + increment;
4295 MatchResult mres2 = match(cur, -1, false);
4296 displayMres(mres2, "findForwardAdv loop", cur)
4297 switch (interpretMatch(mres, mres2)) {
4298 case MatchResult::newIsTooFar:
4299 // behind the expected match
4300 firstInvalid = increment;
4304 case MatchResult::newIsBetter:
4305 // not reached yet, but cur.pos()+increment is bettert
4307 firstInvalid -= increment;
4308 if (increment > firstInvalid*3/4)
4309 increment = firstInvalid*3/4;
4310 if ((mres2.pos == mres2.leadsize) && (increment >= mres2.match_prefix)) {
4311 if (increment >= mres2.match_prefix)
4312 increment = (mres2.match_prefix+1)*3/4;
4317 // Handle not like MatchResult::newIsTooFar
4318 LYXERR0( "Probably too far: Increment = " << increment << " match_prefix = " << mres.match_prefix);
4320 increment = increment*3/4;
4325 if (mres.match_len > 0) {
4326 if (mres.match_prefix + mres.pos - mres.leadsize > 0) {
4327 // The match seems to indicate some deeper level
4336 // should never be reached.
4340 // LYXERR0("Leaving first loop");
4341 LYXERR(Debug::FINDVERBOSE, "Finalizing 1");
4342 MatchResult found_match = findAdvFinalize(cur, match, mres);
4343 if (found_match.match_len > 0) {
4344 match.FillResults(found_match);
4345 return found_match.pos_len;
4348 // try next possible match
4359 /// Find the most backward consecutive match within same paragraph while searching backwards.
4360 MatchResult findMostBackwards(DocIterator & cur, MatchStringAdv const & match, MatchResult &expected)
4362 DocIterator cur_begin = cur;
4363 cur_begin.pos() = 0;
4364 DocIterator tmp_cur = cur;
4365 MatchResult mr = findAdvFinalize(tmp_cur, match, expected);
4366 Inset & inset = cur.inset();
4367 for (; cur != cur_begin; cur.backwardPos()) {
4368 LYXERR(Debug::FINDVERBOSE, "findMostBackwards(): cur=" << cur);
4369 DocIterator new_cur = cur;
4370 new_cur.backwardPos();
4371 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
4373 MatchResult new_mr = findAdvFinalize(new_cur, match, expected);
4374 if (new_mr.match_len == mr.match_len)
4378 LYXERR(Debug::FINDVERBOSE, "findMostBackwards(): exiting with cur=" << cur);
4384 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
4388 // Backup of original position
4389 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
4390 if (cur == cur_begin)
4393 DocIterator cur_orig(cur);
4394 bool pit_changed = false;
4397 MatchResult found_match = match(cur, -1, false);
4399 if (found_match.match_len > 0) {
4401 cur.pos() = cur.lastpos();
4403 cur.pos() = cur_orig.pos();
4404 LYXERR(Debug::FINDVERBOSE, "findBackAdv2: cur: " << cur);
4405 DocIterator cur_prev_iter;
4407 found_match = match(cur);
4408 LYXERR(Debug::FINDVERBOSE, "findBackAdv3: found_match="
4409 << (found_match.match_len > 0) << ", cur: " << cur);
4410 if (found_match.match_len > 0) {
4411 MatchResult found_mr = findMostBackwards(cur, match, found_match);
4412 if (found_mr.pos_len > 0) {
4413 match.FillResults(found_mr);
4414 return found_mr.pos_len;
4418 // Stop if begin of document reached
4419 if (cur == cur_begin)
4421 cur_prev_iter = cur;
4425 if (cur == cur_begin)
4432 } while (!theApp()->longOperationCancelled());
4440 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
4441 DocIterator const & cur, int len)
4443 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
4445 if (!opt.ignoreformat)
4446 return latexifyFromCursor(cur, len);
4448 return stringifyFromCursor(cur, len);
4452 FindAndReplaceOptions::FindAndReplaceOptions(
4453 docstring const & _find_buf_name, bool _casesensitive,
4454 bool _matchword, bool _forward, bool _expandmacros, bool _ignoreformat,
4455 docstring const & _repl_buf_name, bool _keep_case,
4456 SearchScope _scope, SearchRestriction _restr, bool _replace_all)
4457 : find_buf_name(_find_buf_name), casesensitive(_casesensitive), matchword(_matchword),
4458 forward(_forward), expandmacros(_expandmacros), ignoreformat(_ignoreformat),
4459 repl_buf_name(_repl_buf_name), keep_case(_keep_case), scope(_scope), restr(_restr), replace_all(_replace_all)
4467 /** Check if 'len' letters following cursor are all non-lowercase */
4468 static bool allNonLowercase(Cursor const & cur, int len)
4470 pos_type beg_pos = cur.selectionBegin().pos();
4471 pos_type end_pos = cur.selectionBegin().pos() + len;
4472 if (len > cur.lastpos() + 1 - beg_pos) {
4473 LYXERR(Debug::FINDVERBOSE, "This should not happen, more debug needed");
4474 len = cur.lastpos() + 1 - beg_pos;
4475 end_pos = beg_pos + len;
4477 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
4478 if (isLowerCase(cur.paragraph().getChar(pos)))
4484 /** Check if first letter is upper case and second one is lower case */
4485 static bool firstUppercase(Cursor const & cur)
4488 pos_type pos = cur.selectionBegin().pos();
4489 if (pos >= cur.lastpos() - 1) {
4490 LYXERR(Debug::FINDVERBOSE, "No upper-case at cur: " << cur);
4493 ch1 = cur.paragraph().getChar(pos);
4494 ch2 = cur.paragraph().getChar(pos + 1);
4495 bool result = isUpperCase(ch1) && isLowerCase(ch2);
4496 LYXERR(Debug::FINDVERBOSE, "firstUppercase(): "
4497 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
4498 << ch2 << "(" << char(ch2) << ")"
4499 << ", result=" << result << ", cur=" << cur);
4504 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
4506 ** \fixme What to do with possible further paragraphs in replace buffer ?
4508 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
4510 ParagraphList::iterator pit = buffer.paragraphs().begin();
4511 LASSERT(!pit->empty(), /**/);
4512 pos_type right = pos_type(1);
4513 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
4514 right = pit->size();
4515 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
4519 static bool replaceMatches(string &t, int maxmatchnum, vector <string> const & replacements)
4521 // Should replace the string "$" + std::to_string(matchnum) with replacement
4522 // if the char '$' is not prefixed with odd number of char '\\'
4523 static regex const rematch("(\\\\)*(\\$\\$([0-9]))");
4527 for (sregex_iterator it(t.begin(), t.end(), rematch), end; it != end; ++it) {
4529 if ((sub.position(2) - sub.position(0)) % 2 == 1)
4531 int num = stoi(sub.str(3), nullptr, 10);
4532 if (num >= maxmatchnum)
4534 if (lastpos < (size_t) sub.position(2))
4535 s += t.substr(lastpos, sub.position(2) - lastpos);
4536 s += replacements[num];
4537 lastpos = sub.position(2) + sub.length(2);
4541 else if (lastpos < t.length())
4542 s += t.substr(lastpos, t.length() - lastpos);
4548 static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
4550 Cursor & cur = bv->cursor();
4551 if (opt.repl_buf_name.empty()
4552 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
4553 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4556 DocIterator sel_beg = cur.selectionBegin();
4557 DocIterator sel_end = cur.selectionEnd();
4558 if (&sel_beg.inset() != &sel_end.inset()
4559 || sel_beg.pit() != sel_end.pit()
4560 || sel_beg.idx() != sel_end.idx())
4562 int sel_len = sel_end.pos() - sel_beg.pos();
4563 LYXERR(Debug::FINDVERBOSE, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
4564 << ", sel_len: " << sel_len << endl);
4567 LASSERT(sel_len > 0, return 0);
4569 if (!matchAdv(sel_beg, sel_len).match_len)
4572 // Build a copy of the replace buffer, adapted to the KeepCase option
4573 Buffer const & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
4575 repl_buffer_orig.write(oss);
4576 string lyx = oss.str();
4577 if (matchAdv.valid_matches > 0)
4578 replaceMatches(lyx, matchAdv.valid_matches, matchAdv.matches);
4579 Buffer repl_buffer(string(), false);
4580 repl_buffer.setInternal(true);
4581 repl_buffer.setUnnamed(true);
4582 LASSERT(repl_buffer.readString(lyx), return 0);
4583 if (opt.keep_case && sel_len >= 2) {
4584 LYXERR(Debug::FINDVERBOSE, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
4585 if (cur.inTexted()) {
4586 if (firstUppercase(cur))
4587 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
4588 else if (allNonLowercase(cur, sel_len))
4589 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
4592 cap::cutSelection(cur, false);
4593 if (cur.inTexted()) {
4594 repl_buffer.changeLanguage(
4595 repl_buffer.language(),
4596 cur.getFont().language());
4597 LYXERR(Debug::FINDVERBOSE, "Replacing by pasteParagraphList()ing repl_buffer");
4598 LYXERR(Debug::FINDVERBOSE, "Before pasteParagraphList() cur=" << cur << endl);
4599 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
4600 repl_buffer.params().documentClassPtr(),
4601 repl_buffer.params().authors(),
4602 bv->buffer().errorList("Paste"));
4603 LYXERR(Debug::FINDVERBOSE, "After pasteParagraphList() cur=" << cur << endl);
4604 sel_len = repl_buffer.paragraphs().begin()->size();
4605 } else if (cur.inMathed()) {
4606 odocstringstream ods;
4608 // OutputParams runparams(&repl_buffer.params().encoding());
4609 OutputParams runparams(encodings.fromLyXName("utf8"));
4610 runparams.nice = false;
4611 setFindParams(runparams);
4612 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams, string(), -1, -1, true);
4613 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
4614 docstring repl_latex = ods.str();
4615 LYXERR(Debug::FINDVERBOSE, "Latexified replace_buffer: '" << repl_latex << "'");
4617 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
4618 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
4619 repl_latex = from_utf8(s);
4620 LYXERR(Debug::FINDVERBOSE, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
4621 MathData ar(cur.buffer());
4622 asArray(repl_latex, ar, Parse::NORMAL);
4624 sel_len = ar.size();
4625 LYXERR(Debug::FINDVERBOSE, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4627 if (cur.pos() >= sel_len)
4628 cur.pos() -= sel_len;
4631 LYXERR(Debug::FINDVERBOSE, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4632 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
4633 bv->processUpdateFlags(Update::Force);
4638 /// Perform a FindAdv operation.
4639 bool findAdv(BufferView * bv, FindAndReplaceOptions & opt)
4644 // e.g., when invoking word-findadv from mini-buffer wither with
4645 // wrong options syntax or before ever opening advanced F&R pane
4646 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4650 MatchStringAdv matchAdv(bv->buffer(), opt);
4652 if (!matchAdv.regexIsValid) {
4653 bv->message(lyx::from_utf8(matchAdv.regexError));
4657 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
4659 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
4660 num_replaced += findAdvReplace(bv, opt, matchAdv);
4663 pos_len = findForwardAdv(cur, matchAdv);
4665 pos_len = findBackwardsAdv(cur, matchAdv);
4666 } catch (exception & ex) {
4667 bv->message(from_utf8(ex.what()));
4672 if (num_replaced > 0) {
4673 switch (num_replaced)
4676 bv->message(_("One match has been replaced."));
4679 bv->message(_("Two matches have been replaced."));
4682 bv->message(bformat(_("%1$d matches have been replaced."), num_replaced));
4688 bv->message(_("Match not found."));
4693 if (num_replaced > 0)
4694 bv->message(_("Match has been replaced."));
4696 bv->message(_("Match found."));
4698 if (cur.pos() + pos_len > cur.lastpos()) {
4699 // Prevent crash in bv->putSelectionAt()
4700 // Should never happen, maybe LASSERT() here?
4701 pos_len = cur.lastpos() - cur.pos();
4703 LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Putting selection at cur=" << cur << " with len: " << pos_len);
4704 bv->putSelectionAt(cur, pos_len, !opt.forward);
4710 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
4712 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
4713 << opt.casesensitive << ' '
4714 << opt.matchword << ' '
4715 << opt.forward << ' '
4716 << opt.expandmacros << ' '
4717 << opt.ignoreformat << ' '
4718 << opt.replace_all << ' '
4719 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
4720 << opt.keep_case << ' '
4721 << int(opt.scope) << ' '
4724 LYXERR(Debug::FINDVERBOSE, "built: " << os.str());
4730 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
4732 // LYXERR(Debug::FINDVERBOSE, "parsing");
4736 while (line != "EOSS") {
4740 if (is.eof()) // Tolerate malformed request
4744 // LYXERR(Debug::FINDVERBOSE, "file_buf_name: '" << s << "'");
4745 opt.find_buf_name = from_utf8(s);
4746 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.replace_all;
4747 is.get(); // Waste space before replace string
4750 while (line != "EOSS") {
4754 if (is.eof()) // Tolerate malformed request
4758 // LYXERR(Debug::FINDVERBOSE, "repl_buf_name: '" << s << "'");
4759 opt.repl_buf_name = from_utf8(s);
4760 is >> opt.keep_case;
4763 opt.scope = FindAndReplaceOptions::SearchScope(i);
4765 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
4768 LYXERR(Debug::FINDVERBOSE, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
4769 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
4770 << opt.scope << ' ' << opt.restr);