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-z]+) ?|[\\[\\]\\{\\}]))");
844 bool backslashed = false;
848 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
851 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
855 if (sub.str(4) == "backslash") {
858 // 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) {
871 if (withformat && (sub.str(3) == "{"))
872 replace = accents["braceleft"];
873 else if (withformat && (sub.str(3) == "}"))
874 replace = accents["braceright"];
876 // else part should not exist
881 AccentsIterator it_ac = accents.find(sub.str(4));
882 if (it_ac == accents.end()) {
883 replace = sub.str(3);
886 replace = it_ac->second;
890 if (lastpos < (size_t) sub.position(2))
891 s += t.substr(lastpos, sub.position(2) - lastpos);
893 lastpos = sub.position(2) + sub.length(2);
897 else if (lastpos < t.length())
898 s += t.substr(lastpos, t.length() - lastpos);
902 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
903 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
904 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
905 string escape_for_regex(string s, bool withformat)
909 while (lastpos < s.size()) {
910 size_t regex_pos = s.find("\\regexp{", lastpos);
911 if (regex_pos == string::npos) {
912 regex_pos = s.size();
914 if (regex_pos > lastpos) {
915 result += string2regex(s.substr(lastpos, regex_pos-lastpos));
917 if (lastpos == s.size())
920 size_t end_pos = s.find("\\endregexp{", regex_pos + 8);
921 result += correctRegex(s.substr(regex_pos + 8, end_pos -(regex_pos + 8)), withformat);
922 lastpos = end_pos + 13;
928 /// Wrapper for lyx::regex_replace with simpler interface
929 bool regex_replace(string const & s, string & t, string const & searchstr,
930 string const & replacestr)
932 regex e(searchstr, regex_constants::ECMAScript);
934 ostream_iterator<char, char> it(oss);
935 regex_replace(it, s.begin(), s.end(), e, replacestr);
936 // tolerate t and s be references to the same variable
937 bool rv = (s != oss.str());
956 vector <string> result = vector <string>();
957 MatchResult(int len = 0): match_len(len),match_prefix(0),match2end(0), pos(0),leadsize(0),pos_len(-1),searched_size(0) {}
960 static MatchResult::range interpretMatch(MatchResult &oldres, MatchResult &newres)
962 if (newres.match2end < oldres.match2end)
963 return MatchResult::newIsTooFar;
964 if (newres.match_len < oldres.match_len)
965 return MatchResult::newIsTooFar;
967 if (newres.match_len == oldres.match_len) {
968 if (newres.match2end == oldres.match2end)
969 return MatchResult::newIsBetter;
971 return MatchResult::newIsInvalid;
974 /** The class performing a match between a position in the document and the FindAdvOptions.
977 class MatchStringAdv {
979 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt);
981 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
982 ** constructor as opt.search, under the opt.* options settings.
985 ** If set, then match is searched only against beginning of text starting at cur.
986 ** If unset, then match is searched anywhere in text starting at cur.
989 ** The length of the matching text, or zero if no match was found.
991 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
1000 /// first buffer on which search was started
1001 lyx::Buffer * const p_first_buf;
1003 FindAndReplaceOptions const & opt;
1006 /// Auxiliary find method (does not account for opt.matchword)
1007 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
1008 void CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string = "");
1010 /** Normalize a stringified or latexified LyX paragraph.
1014 ** <li>if search is not casesensitive, then lowercase the string;
1015 ** <li>remove any newline at begin or end of the string;
1016 ** <li>replace any newline in the middle of the string with a simple space;
1017 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
1020 ** @todo Normalization should also expand macros, if the corresponding
1021 ** search option was checked.
1023 string normalize(docstring const & s) const;
1024 // normalized string to search
1025 string par_as_string;
1026 // regular expression to use for searching
1027 // regexp2 is same as regexp, but prefixed with a ".*?"
1029 QRegularExpression regexp;
1030 QRegularExpression regexp2;
1035 // leading format material as string
1036 string lead_as_string;
1037 // par_as_string after removal of lead_as_string
1038 string par_as_string_nolead;
1039 // unmatched open braces in the search string/regexp
1041 // number of (.*?) subexpressions added at end of search regexp for closing
1042 // environments, math mode, styles, etc...
1043 int close_wildcards;
1045 // Are we searching with regular expressions ?
1047 static int valid_matches;
1048 static vector <string> matches;
1049 void FillResults(MatchResult &found_mr);
1052 int MatchStringAdv::valid_matches = 0;
1053 vector <string> MatchStringAdv::matches = vector <string>(10);
1055 void MatchStringAdv::FillResults(MatchResult &found_mr)
1057 if (found_mr.match_len > 0) {
1058 valid_matches = found_mr.result.size();
1059 for (size_t i = 0; i < found_mr.result.size(); i++)
1060 matches[i] = found_mr.result[i];
1066 static docstring buffer_to_latex(Buffer & buffer)
1068 //OutputParams runparams(&buffer.params().encoding());
1069 OutputParams runparams(encodings.fromLyXName("utf8"));
1070 odocstringstream ods;
1072 runparams.nice = true;
1073 runparams.flavor = Flavor::XeTeX;
1074 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
1075 // No side effect of file copying and image conversion
1076 runparams.dryrun = true;
1077 if (ignoreFormats.getDeleted())
1078 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
1080 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
1081 if (ignoreFormats.getNonContent()) {
1082 runparams.for_searchAdv |= OutputParams::SearchNonOutput;
1084 pit_type const endpit = buffer.paragraphs().size();
1085 for (pit_type pit = 0; pit != endpit; ++pit) {
1086 TeXOnePar(buffer, buffer.text(), pit, os, runparams);
1087 LYXERR(Debug::FIND, "searchString up to here: " << ods.str());
1093 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
1096 if (!opt.ignoreformat) {
1097 str = buffer_to_latex(buffer);
1099 // OutputParams runparams(&buffer.params().encoding());
1100 OutputParams runparams(encodings.fromLyXName("utf8"));
1101 runparams.nice = true;
1102 runparams.flavor = Flavor::XeTeX;
1103 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
1104 runparams.dryrun = true;
1105 int option = AS_STR_INSETS |AS_STR_PLAINTEXT;
1106 if (ignoreFormats.getDeleted()) {
1107 option |= AS_STR_SKIPDELETE;
1108 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
1111 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
1113 if (ignoreFormats.getNonContent()) {
1114 runparams.for_searchAdv |= OutputParams::SearchNonOutput;
1116 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
1117 Paragraph const & par = buffer.paragraphs().at(pit);
1118 LYXERR(Debug::FIND, "Adding to search string: '"
1119 << par.asString(pos_type(0), par.size(),
1123 str += par.asString(pos_type(0), par.size(),
1127 // Even in ignore-format we have to remove "\text{}, \lyxmathsym{}" parts
1128 string t = to_utf8(str);
1129 while (regex_replace(t, t, "\\\\(text|lyxmathsym|ensuremath)\\{([^\\}]*)\\}", "$2"));
1136 /// Return separation pos between the leading material and the rest
1137 static size_t identifyLeading(string const & s)
1140 // @TODO Support \item[text]
1141 // Kornel: Added textsl, textsf, textit, texttt and noun
1142 // + allow to search for colored text too
1143 while (regex_replace(t, t, "^\\\\(("
1144 "(author|title|subtitle|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|"
1145 "lyxaddress|lyxrightaddress|"
1146 "footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1147 "emph|noun|minisec|text(bf|md|sl|sf|it|tt))|"
1148 "((textcolor|foreignlanguage|latexenvironment)\\{[a-z]+\\*?\\})|"
1149 "(u|uu)line|(s|x)out|uwave)|((sub)?(((sub)?section)|paragraph)|part|chapter)\\*?)\\{", "")
1150 || regex_replace(t, t, "^\\$", "")
1151 || regex_replace(t, t, "^\\\\\\[", "")
1152 || regex_replace(t, t, "^ ?\\\\item\\{[a-z]+\\}", "")
1153 || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\}", ""))
1155 LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
1160 * Given a latexified string, retrieve some handled features
1161 * The features of the regex will later be compared with the features
1162 * of the searched text. If the regex features are not a
1163 * subset of the analized, then, in not format ignoring search
1164 * we can early stop the search in the relevant inset.
1166 typedef map<string, bool> Features;
1168 static Features identifyFeatures(string const & s)
1170 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
1171 static regex const valid("^("
1173 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1174 "emph|noun|text(bf|md|sl|sf|it|tt)|"
1175 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
1176 "(u|uu)line|(s|x)out|uwave|"
1177 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
1178 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
1183 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1186 if (sub.str(1).compare("regexp") == 0) {
1190 string token = sub.str(1);
1192 if (regex_match(token, sub2, valid)) {
1200 if (sub.str(1).compare("endregexp") == 0) {
1210 * defines values features of a key "\\[a-z]+{"
1215 /* Char type with content discarded
1216 * like \hspace{1cm} */
1218 /* Char, like \backslash */
1220 /* replace starting backslash with '#' */
1222 /* \part, \section*, ... */
1224 /* title, author etc */
1226 /* \foreignlanguage{ngerman}, ... */
1229 * to discard language in content */
1232 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1234 /* fonts, colors, markups, ... */
1236 /* footnotesize, ... large, ...
1237 * Ignore all of them */
1240 /* inputencoding, ...
1241 * Discard also content, because they do not help in search */
1244 * like remove, but also all arguments */
1246 /* item, listitem */
1248 /* tex, latex, ... like isChar */
1250 /* like \lettrine[lines=5]{}{} */
1252 // like isStandard, but always remove head
1254 /* End of arguments marker for lettrine,
1255 * so that they can be ignored */
1258 KeyInfo() = default;
1259 KeyInfo(KeyType type, int parcount, bool disable)
1261 parenthesiscount(parcount),
1262 disabled(disable) {}
1263 KeyType keytype = invalid;
1265 int _tokensize = -1;
1266 int _tokenstart = -1;
1267 int _dataStart = -1;
1269 int parenthesiscount = 1;
1270 bool disabled = false;
1271 bool used = false; /* by pattern */
1276 Border(int l=0, int u=0) : low(l), upper(u) {}
1281 #define MAXOPENED 30
1283 bool isPatternString_;
1285 explicit Intervall(bool isPattern, string const & p) :
1286 isPatternString_(isPattern), par(p), ignoreidx(-1), actualdeptindex(0),
1287 hasTitle(false), langcount(0)
1295 static vector<Border> borders;
1296 int depts[MAXOPENED];
1297 int closes[MAXOPENED];
1298 int actualdeptindex;
1299 int previousNotIgnored(int) const;
1300 int nextNotIgnored(int) const;
1301 void handleOpenP(int i);
1302 void handleCloseP(int i, bool closingAllowed);
1303 void resetOpenedP(int openPos);
1304 void addIntervall(int upper);
1305 void addIntervall(int low, int upper); /* if explicit */
1306 void removeAccents();
1307 void setForDefaultLang(KeyInfo const & defLang) const;
1308 int findclosing(int start, int end, char up, char down, int repeat);
1309 void handleParentheses(int lastpos, bool closingAllowed);
1311 int langcount; // Number of disabled language specs up to current position in actual interval
1312 int isOpeningPar(int pos) const;
1314 void output(ostringstream &os, int lastpos);
1315 // string show(int lastpos);
1318 vector<Border> Intervall::borders = vector<Border>(30);
1320 int Intervall::isOpeningPar(int pos) const
1322 if ((pos < 0) || (size_t(pos) >= par.size()))
1324 if (par[pos] != '{')
1326 if (size_t(pos) + 2 >= par.size())
1328 if (par[pos+2] != '}')
1330 if (par[pos+1] == '[' || par[pos+1] == ']')
1335 void Intervall::setForDefaultLang(KeyInfo const & defLang) const
1337 // Enable the use of first token again
1338 if (ignoreidx >= 0) {
1339 int value = defLang._tokenstart + defLang._tokensize;
1345 if (borders[borderidx].low < value)
1346 borders[borderidx].low = value;
1347 if (borders[borderidx].upper < value)
1348 borders[borderidx].upper = value;
1353 static void checkDepthIndex(int val)
1355 static int maxdepthidx = MAXOPENED-2;
1356 static int lastmaxdepth = 0;
1357 if (val > lastmaxdepth) {
1358 LYXERR(Debug::INFO, "Depth reached " << val);
1361 if (val > maxdepthidx) {
1363 LYXERR(Debug::INFO, "maxdepthidx now " << val);
1368 // Not needed, because borders are now dynamically expanded
1369 static void checkIgnoreIdx(int val)
1371 static int lastmaxignore = -1;
1372 if ((lastmaxignore < val) && (size_t(val+1) >= borders.size())) {
1373 LYXERR(Debug::INFO, "IgnoreIdx reached " << val);
1374 lastmaxignore = val;
1380 * Expand the region of ignored parts of the input latex string
1381 * The region is only relevant in output()
1383 void Intervall::addIntervall(int low, int upper)
1386 if (low == upper) return;
1387 for (idx = ignoreidx+1; idx > 0; --idx) {
1388 if (low > borders[idx-1].upper) {
1392 Border br(low, upper);
1393 if (idx > ignoreidx) {
1394 if (borders.size() <= size_t(idx)) {
1395 borders.push_back(br);
1401 // checkIgnoreIdx(ignoreidx);
1405 // Expand only if one of the new bound is inside the interwall
1406 // We know here that br.low > borders[idx-1].upper
1407 if (br.upper < borders[idx].low) {
1408 // We have to insert at this pos
1409 if (size_t(ignoreidx+1) >= borders.size()) {
1410 borders.push_back(borders[ignoreidx]);
1413 borders[ignoreidx+1] = borders[ignoreidx];
1415 for (int i = ignoreidx; i > idx; --i) {
1416 borders[i] = borders[i-1];
1420 // checkIgnoreIdx(ignoreidx);
1423 // Here we know, that we are overlapping
1424 if (br.low > borders[idx].low)
1425 br.low = borders[idx].low;
1426 // check what has to be concatenated
1428 for (int i = idx; i <= ignoreidx; i++) {
1429 if (br.upper >= borders[i].low) {
1431 if (br.upper < borders[i].upper)
1432 br.upper = borders[i].upper;
1438 // count should be >= 1 here
1441 for (int i = idx + count; i <= ignoreidx; i++) {
1442 borders[i-count+1] = borders[i];
1444 ignoreidx -= count - 1;
1450 static void buildaccent(string n, string param, string values)
1454 const char delim = '|';
1455 while (getline(s, name, delim)) {
1457 for (char c : param) {
1458 string key = name + "{" + c + "}";
1459 // get the corresponding utf8-value
1460 if ((values[start] & 0xc0) != 0xc0) {
1461 // should not happen, utf8 encoding starts at least with 11xxxxxx
1462 // but value for '\dot{i}' is 'i', which is ascii
1463 if ((values[start] & 0x80) == 0) {
1465 accents[key] = values.substr(start, 1);
1466 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1471 for (int j = 1; ;j++) {
1472 if (start + j >= values.size()) {
1473 accents[key] = values.substr(start, j);
1474 start = values.size() - 1;
1477 else if ((values[start+j] & 0xc0) != 0x80) {
1478 // This is the first byte of following utf8 char
1479 accents[key] = values.substr(start, j);
1481 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1490 static string getutf8(unsigned uchar)
1493 string ret = string();
1495 if (uchar <= 0x7f) {
1496 c[maxc-1] = uchar & 0x7f;
1499 unsigned char rest = 0x40;
1500 unsigned char first = 0x80;
1502 for (int i = start; i >=0; --i) {
1504 c[i] = first + uchar;
1507 c[i] = 0x80 | (uchar & 0x3f);
1514 for (int i = 0; i < maxc; i++) {
1515 if (c[i] == 0) continue;
1521 static void addAccents(string latex_in, string unicode_out)
1523 latex_in = latex_in.substr(1);
1524 AccentsIterator it_ac = accents.find(latex_in);
1525 if (it_ac == accents.end()) {
1526 accents[latex_in] = unicode_out;
1529 LYXERR0("Key " << latex_in << " already set");
1533 void static fillMissingUnicodesymbols()
1535 addAccents("\\pounds", getutf8(0x00a3));
1536 addAccents("\\textsterling", getutf8(0x00a3));
1537 addAccents("\\textyen", getutf8(0x00a5));
1538 addAccents("\\yen", getutf8(0x00a5));
1539 addAccents("\\textsection", getutf8(0x00a7));
1540 addAccents("\\mathsection", getutf8(0x00a7));
1541 addAccents("\\textcopyright", getutf8(0x00a9));
1542 addAccents("\\copyright", getutf8(0x00a9));
1543 addAccents("\\textlnot", getutf8(0x00ac));
1544 addAccents("\\neg", getutf8(0x00ac));
1545 addAccents("\\textregistered", getutf8(0x00ae));
1546 addAccents("\\circledR", getutf8(0x00ae));
1547 addAccents("\\textpm", getutf8(0x00b1));
1548 addAccents("\\pm", getutf8(0x00b1));
1549 addAccents("\\textparagraph", getutf8(0x00b6));
1550 addAccents("\\mathparagraph", getutf8(0x00b6));
1551 addAccents("\\textperiodcentered", getutf8(0x00b7));
1552 addAccents("\\texttimes", getutf8(0x00d7));
1553 addAccents("\\times", getutf8(0x00d7));
1554 addAccents("\\O", getutf8(0x00d8));
1555 addAccents("\\dh", getutf8(0x00f0));
1556 addAccents("\\eth", getutf8(0x00f0));
1557 addAccents("\\textdiv", getutf8(0x00f7));
1558 addAccents("\\div", getutf8(0x00f7));
1559 addAccents("\\o", getutf8(0x00f8));
1560 addAccents("\\textcrlambda", getutf8(0x019b));
1561 addAccents("\\j", getutf8(0x0237));
1562 addAccents("\\textrevepsilon", getutf8(0x025c));
1563 addAccents("\\textbaru", getutf8(0x0289));
1564 addAccents("\\textquoteleft", getutf8(0x02bb));
1565 addAccents("\\textGamma", getutf8(0x0393));
1566 addAccents("\\Gamma", getutf8(0x0393));
1567 addAccents("\\textDelta", getutf8(0x0394));
1568 addAccents("\\Delta", getutf8(0x0394));
1569 addAccents("\\textTheta", getutf8(0x0398));
1570 addAccents("\\Theta", getutf8(0x0398));
1571 addAccents("\\textLambda", getutf8(0x039b));
1572 addAccents("\\Lambda", getutf8(0x039b));
1573 addAccents("\\textXi", getutf8(0x039e));
1574 addAccents("\\Xi", getutf8(0x039e));
1575 addAccents("\\textPi", getutf8(0x03a0));
1576 addAccents("\\Pi", getutf8(0x03a0));
1577 addAccents("\\textSigma", getutf8(0x03a3));
1578 addAccents("\\Sigma", getutf8(0x03a3));
1579 addAccents("\\textUpsilon", getutf8(0x03a5));
1580 addAccents("\\Upsilon", getutf8(0x03a5));
1581 addAccents("\\textPhi", getutf8(0x03a6));
1582 addAccents("\\Phi", getutf8(0x03a6));
1583 addAccents("\\textPsi", getutf8(0x03a8));
1584 addAccents("\\Psi", getutf8(0x03a8));
1585 addAccents("\\textOmega", getutf8(0x03a9));
1586 addAccents("\\Omega", getutf8(0x03a9));
1587 addAccents("\\textalpha", getutf8(0x03b1));
1588 addAccents("\\alpha", getutf8(0x03b1));
1589 addAccents("\\textbeta", getutf8(0x03b2));
1590 addAccents("\\beta", getutf8(0x03b2));
1591 addAccents("\\textgamma", getutf8(0x03b3));
1592 addAccents("\\gamma", getutf8(0x03b3));
1593 addAccents("\\textdelta", getutf8(0x03b4));
1594 addAccents("\\delta", getutf8(0x03b4));
1595 addAccents("\\textepsilon", getutf8(0x03b5));
1596 addAccents("\\varepsilon", getutf8(0x03b5));
1597 addAccents("\\textzeta", getutf8(0x03b6));
1598 addAccents("\\zeta", getutf8(0x03b6));
1599 addAccents("\\texteta", getutf8(0x03b7));
1600 addAccents("\\eta", getutf8(0x03b7));
1601 addAccents("\\texttheta", getutf8(0x03b8));
1602 addAccents("\\theta", getutf8(0x03b8));
1603 addAccents("\\textiota", getutf8(0x03b9));
1604 addAccents("\\iota", getutf8(0x03b9));
1605 addAccents("\\textkappa", getutf8(0x03ba));
1606 addAccents("\\kappa", getutf8(0x03ba));
1607 addAccents("\\textlambda", getutf8(0x03bb));
1608 addAccents("\\lambda", getutf8(0x03bb));
1609 addAccents("\\textmu", getutf8(0x03bc));
1610 addAccents("\\mu", getutf8(0x03bc));
1611 addAccents("\\textnu", getutf8(0x03bd));
1612 addAccents("\\nu", getutf8(0x03bd));
1613 addAccents("\\textxi", getutf8(0x03be));
1614 addAccents("\\xi", getutf8(0x03be));
1615 addAccents("\\textpi", getutf8(0x03c0));
1616 addAccents("\\pi", getutf8(0x03c0));
1617 addAccents("\\textrho", getutf8(0x03c1));
1618 addAccents("\\rho", getutf8(0x03c1));
1619 addAccents("\\textfinalsigma", getutf8(0x03c2));
1620 addAccents("\\varsigma", getutf8(0x03c2));
1621 addAccents("\\textsigma", getutf8(0x03c3));
1622 addAccents("\\sigma", getutf8(0x03c3));
1623 addAccents("\\texttau", getutf8(0x03c4));
1624 addAccents("\\tau", getutf8(0x03c4));
1625 addAccents("\\textupsilon", getutf8(0x03c5));
1626 addAccents("\\upsilon", getutf8(0x03c5));
1627 addAccents("\\textphi", getutf8(0x03c6));
1628 addAccents("\\varphi", getutf8(0x03c6));
1629 addAccents("\\textchi", getutf8(0x03c7));
1630 addAccents("\\chi", getutf8(0x03c7));
1631 addAccents("\\textpsi", getutf8(0x03c8));
1632 addAccents("\\psi", getutf8(0x03c8));
1633 addAccents("\\textomega", getutf8(0x03c9));
1634 addAccents("\\omega", getutf8(0x03c9));
1635 addAccents("\\textdigamma", getutf8(0x03dd));
1636 addAccents("\\digamma", getutf8(0x03dd));
1637 addAccents("\\hebalef", getutf8(0x05d0));
1638 addAccents("\\aleph", getutf8(0x05d0));
1639 addAccents("\\hebbet", getutf8(0x05d1));
1640 addAccents("\\beth", getutf8(0x05d1));
1641 addAccents("\\hebgimel", getutf8(0x05d2));
1642 addAccents("\\gimel", getutf8(0x05d2));
1643 addAccents("\\hebdalet", getutf8(0x05d3));
1644 addAccents("\\daleth", getutf8(0x05d3));
1645 addAccents("\\hebhe", getutf8(0x05d4));
1646 addAccents("\\hebvav", getutf8(0x05d5));
1647 addAccents("\\hebzayin", getutf8(0x05d6));
1648 addAccents("\\hebhet", getutf8(0x05d7));
1649 addAccents("\\hebtet", getutf8(0x05d8));
1650 addAccents("\\hebyod", getutf8(0x05d9));
1651 addAccents("\\hebfinalkaf", getutf8(0x05da));
1652 addAccents("\\hebkaf", getutf8(0x05db));
1653 addAccents("\\heblamed", getutf8(0x05dc));
1654 addAccents("\\hebfinalmem", getutf8(0x05dd));
1655 addAccents("\\hebmem", getutf8(0x05de));
1656 addAccents("\\hebfinalnun", getutf8(0x05df));
1657 addAccents("\\hebnun", getutf8(0x05e0));
1658 addAccents("\\hebsamekh", getutf8(0x05e1));
1659 addAccents("\\hebayin", getutf8(0x05e2));
1660 addAccents("\\hebfinalpe", getutf8(0x05e3));
1661 addAccents("\\hebpe", getutf8(0x05e4));
1662 addAccents("\\hebfinaltsadi", getutf8(0x05e5));
1663 addAccents("\\hebtsadi", getutf8(0x05e6));
1664 addAccents("\\hebqof", getutf8(0x05e7));
1665 addAccents("\\hebresh", getutf8(0x05e8));
1666 addAccents("\\hebshin", getutf8(0x05e9));
1667 addAccents("\\hebtav", getutf8(0x05ea));
1670 addAccents("\\thaiKoKai", getutf8(0x0e01));
1671 addAccents("\\thaiKhoKhai", getutf8(0x0e02));
1672 addAccents("\\thaiKhoKhuat", getutf8(0x0e03));
1673 addAccents("\\thaiKhoKhwai", getutf8(0x0e04));
1674 addAccents("\\thaiKhoKhon", getutf8(0x0e05));
1675 addAccents("\\thaiKhoRakhang", getutf8(0x0e06));
1676 addAccents("\\thaiNgoNgu", getutf8(0x0e07));
1677 addAccents("\\thaiChoChan", getutf8(0x0e08));
1678 addAccents("\\thaiChoChing", getutf8(0x0e09));
1679 addAccents("\\thaiChoChang", getutf8(0x0e0a));
1680 addAccents("\\thaiSoSo", getutf8(0x0e0b));
1681 addAccents("\\thaiChoChoe", getutf8(0x0e0c));
1682 addAccents("\\thaiYoYing", getutf8(0x0e0d));
1683 addAccents("\\thaiDoChada", getutf8(0x0e0e));
1684 addAccents("\\thaiToPatak", getutf8(0x0e0f));
1685 addAccents("\\thaiThoThan", getutf8(0x0e10));
1686 addAccents("\\thaiThoNangmontho", getutf8(0x0e11));
1687 addAccents("\\thaiThoPhuthao", getutf8(0x0e12));
1688 addAccents("\\thaiNoNen", getutf8(0x0e13));
1689 addAccents("\\thaiDoDek", getutf8(0x0e14));
1690 addAccents("\\thaiToTao", getutf8(0x0e15));
1691 addAccents("\\thaiThoThung", getutf8(0x0e16));
1692 addAccents("\\thaiThoThahan", getutf8(0x0e17));
1693 addAccents("\\thaiThoThong", getutf8(0x0e18));
1694 addAccents("\\thaiNoNu", getutf8(0x0e19));
1695 addAccents("\\thaiBoBaimai", getutf8(0x0e1a));
1696 addAccents("\\thaiPoPla", getutf8(0x0e1b));
1697 addAccents("\\thaiPhoPhung", getutf8(0x0e1c));
1698 addAccents("\\thaiFoFa", getutf8(0x0e1d));
1699 addAccents("\\thaiPhoPhan", getutf8(0x0e1e));
1700 addAccents("\\thaiFoFan", getutf8(0x0e1f));
1701 addAccents("\\thaiPhoSamphao", getutf8(0x0e20));
1702 addAccents("\\thaiMoMa", getutf8(0x0e21));
1703 addAccents("\\thaiYoYak", getutf8(0x0e22));
1704 addAccents("\\thaiRoRua", getutf8(0x0e23));
1705 addAccents("\\thaiRu", getutf8(0x0e24));
1706 addAccents("\\thaiLoLing", getutf8(0x0e25));
1707 addAccents("\\thaiLu", getutf8(0x0e26));
1708 addAccents("\\thaiWoWaen", getutf8(0x0e27));
1709 addAccents("\\thaiSoSala", getutf8(0x0e28));
1710 addAccents("\\thaiSoRusi", getutf8(0x0e29));
1711 addAccents("\\thaiSoSua", getutf8(0x0e2a));
1712 addAccents("\\thaiHoHip", getutf8(0x0e2b));
1713 addAccents("\\thaiLoChula", getutf8(0x0e2c));
1714 addAccents("\\thaiOAng", getutf8(0x0e2d));
1715 addAccents("\\thaiHoNokhuk", getutf8(0x0e2e));
1716 addAccents("\\thaiPaiyannoi", getutf8(0x0e2f));
1717 addAccents("\\thaiSaraA", getutf8(0x0e30));
1718 addAccents("\\thaiMaiHanakat", getutf8(0x0e31));
1719 addAccents("\\thaiSaraAa", getutf8(0x0e32));
1720 addAccents("\\thaiSaraAm", getutf8(0x0e33));
1721 addAccents("\\thaiSaraI", getutf8(0x0e34));
1722 addAccents("\\thaiSaraIi", getutf8(0x0e35));
1723 addAccents("\\thaiSaraUe", getutf8(0x0e36));
1724 addAccents("\\thaiSaraUee", getutf8(0x0e37));
1725 addAccents("\\thaiSaraU", getutf8(0x0e38));
1726 addAccents("\\thaiSaraUu", getutf8(0x0e39));
1727 addAccents("\\thaiPhinthu", getutf8(0x0e3a));
1728 addAccents("\\thaiSaraE", getutf8(0x0e40));
1729 addAccents("\\thaiSaraAe", getutf8(0x0e41));
1730 addAccents("\\thaiSaraO", getutf8(0x0e42));
1731 addAccents("\\thaiSaraAiMaimuan", getutf8(0x0e43));
1732 addAccents("\\thaiSaraAiMaimalai", getutf8(0x0e44));
1733 addAccents("\\thaiLakkhangyao", getutf8(0x0e45));
1734 addAccents("\\thaiMaiyamok", getutf8(0x0e46));
1735 addAccents("\\thaiMaitaikhu", getutf8(0x0e47));
1736 addAccents("\\thaiMaiEk", getutf8(0x0e48));
1737 addAccents("\\thaiMaiTho", getutf8(0x0e49));
1738 addAccents("\\thaiMaiTri", getutf8(0x0e4a));
1739 addAccents("\\thaiMaiChattawa", getutf8(0x0e4b));
1740 addAccents("\\thaiThanthakhat", getutf8(0x0e4c));
1741 addAccents("\\thaiNikhahit", getutf8(0x0e4d));
1742 addAccents("\\thaiYamakkan", getutf8(0x0e4e));
1743 addAccents("\\thaiFongman", getutf8(0x0e4f));
1744 addAccents("\\thaizero", getutf8(0x0e50));
1745 addAccents("\\thaione", getutf8(0x0e51));
1746 addAccents("\\thaitwo", getutf8(0x0e52));
1747 addAccents("\\thaithree", getutf8(0x0e53));
1748 addAccents("\\thaifour", getutf8(0x0e54));
1749 addAccents("\\thaifive", getutf8(0x0e55));
1750 addAccents("\\thaisix", getutf8(0x0e56));
1751 addAccents("\\thaiseven", getutf8(0x0e57));
1752 addAccents("\\thaieight", getutf8(0x0e58));
1753 addAccents("\\thainine", getutf8(0x0e59));
1754 addAccents("\\thaiAngkhankhu", getutf8(0x0e5a));
1755 addAccents("\\thaiKhomut", getutf8(0x0e5b));
1756 addAccents("\\dag", getutf8(0x2020));
1757 addAccents("\\dagger", getutf8(0x2020));
1758 addAccents("\\textdagger", getutf8(0x2020));
1759 addAccents("\\ddag", getutf8(0x2021));
1760 addAccents("\\ddagger", getutf8(0x2021));
1761 addAccents("\\textdaggerdbl", getutf8(0x2021));
1762 addAccents("\\textbullet", getutf8(0x2022));
1763 addAccents("\\bullet", getutf8(0x2022));
1764 addAccents("\\dots", getutf8(0x2026));
1765 addAccents("\\ldots", getutf8(0x2026));
1766 addAccents("\\textellipsis", getutf8(0x2026));
1767 addAccents("\\textasciiacute", getutf8(0x2032));
1768 addAccents("\\prime", getutf8(0x2032));
1769 addAccents("\\textacutedbl", getutf8(0x2033));
1770 addAccents("\\dprime", getutf8(0x2033));
1771 addAccents("\\textasciigrave", getutf8(0x2035));
1772 addAccents("\\backprime", getutf8(0x2035));
1773 addAccents("\\textsubcircum{ }", getutf8(0x2038));
1774 addAccents("\\caretinsert", getutf8(0x2038));
1775 addAccents("\\textasteriskcentered", getutf8(0x204e));
1776 addAccents("\\ast", getutf8(0x204e));
1777 addAccents("\\textmho", getutf8(0x2127));
1778 addAccents("\\mho", getutf8(0x2127));
1779 addAccents("\\textleftarrow", getutf8(0x2190));
1780 addAccents("\\leftarrow", getutf8(0x2190));
1781 addAccents("\\textuparrow", getutf8(0x2191));
1782 addAccents("\\uparrow", getutf8(0x2191));
1783 addAccents("\\textrightarrow", getutf8(0x2192));
1784 addAccents("\\rightarrow", getutf8(0x2192));
1785 addAccents("\\textdownarrow", getutf8(0x2193));
1786 addAccents("\\downarrow", getutf8(0x2193));
1787 addAccents("\\textglobrise", getutf8(0x2197));
1788 addAccents("\\nearrow", getutf8(0x2197));
1789 addAccents("\\textglobfall", getutf8(0x2198));
1790 addAccents("\\searrow", getutf8(0x2198));
1791 addAccents("\\textsurd", getutf8(0x221a));
1792 addAccents("\\surd", getutf8(0x221a));
1793 addAccents("\\textbigcircle", getutf8(0x25ef));
1794 addAccents("\\bigcirc", getutf8(0x25ef));
1795 addAccents("\\FiveStar", getutf8(0x2605));
1796 addAccents("\\bigstar", getutf8(0x2605));
1797 addAccents("\\FiveStarOpen", getutf8(0x2606));
1798 addAccents("\\bigwhitestar", getutf8(0x2606));
1799 addAccents("\\Checkmark", getutf8(0x2713));
1800 addAccents("\\checkmark", getutf8(0x2713));
1801 addAccents("\\CrossMaltese", getutf8(0x2720));
1802 addAccents("\\maltese", getutf8(0x2720));
1803 addAccents("\\textlangle", getutf8(0x27e8));
1804 addAccents("\\langle", getutf8(0x27e8));
1805 addAccents("\\textrangle", getutf8(0x27e9));
1806 addAccents("\\rangle", getutf8(0x27e9));
1809 static void buildAccentsMap()
1811 accents["imath"] = "ı";
1813 accents["jmath"] = "ȷ";
1814 accents["cdot"] = "·";
1815 accents["textasciicircum"] = "^";
1816 accents["mathcircumflex"] = "^";
1817 accents["sim"] = "~";
1818 accents["guillemotright"] = "»";
1819 accents["guillemotleft"] = "«";
1820 accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
1821 accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
1822 accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
1823 accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
1824 accents["negmedspace"] = getutf8(0xf0005);
1825 accents["thickspace"] = getutf8(0xf0006);
1826 accents["negthickspace"] = getutf8(0xf0007);
1827 accents["lyx"] = getutf8(0xf0010); // Used logos
1828 accents["LyX"] = getutf8(0xf0010);
1829 accents["tex"] = getutf8(0xf0011);
1830 accents["TeX"] = getutf8(0xf0011);
1831 accents["latex"] = getutf8(0xf0012);
1832 accents["LaTeX"] = getutf8(0xf0012);
1833 accents["latexe"] = getutf8(0xf0013);
1834 accents["LaTeXe"] = getutf8(0xf0013);
1835 accents["lyxarrow"] = getutf8(0xf0020);
1836 accents["braceleft"] = getutf8(0xf0030);
1837 accents["braceright"] = getutf8(0xf0031);
1838 accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash
1839 accents["backslash LyX"] = getutf8(0xf0010);
1840 accents["backslash tex"] = getutf8(0xf0011);
1841 accents["backslash TeX"] = getutf8(0xf0011);
1842 accents["backslash latex"] = getutf8(0xf0012);
1843 accents["backslash LaTeX"] = getutf8(0xf0012);
1844 accents["backslash latexe"] = getutf8(0xf0013);
1845 accents["backslash LaTeXe"] = getutf8(0xf0013);
1846 accents["backslash lyxarrow"] = getutf8(0xf0020);
1847 accents["ddot{\\imath}"] = "ï";
1848 buildaccent("ddot", "aAeEhHiIoOtuUwWxXyY",
1849 "äÄëËḧḦïÏöÖẗüÜẅẄẍẌÿŸ"); // umlaut
1850 buildaccent("dot|.", "aAbBcCdDeEfFGghHIimMnNoOpPrRsStTwWxXyYzZ",
1851 "ȧȦḃḂċĊḋḊėĖḟḞĠġḣḢİİṁṀṅṄȯȮṗṖṙṘṡṠṫṪẇẆẋẊẏẎżŻ"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'İ'
1852 accents["acute{\\imath}"] = "í";
1853 buildaccent("acute", "aAcCeEgGkKlLmMoOnNpPrRsSuUwWyYzZiI",
1854 "áÁćĆéÉǵǴḱḰĺĹḿḾóÓńŃṕṔŕŔśŚúÚẃẂýÝźŹíÍ");
1855 buildaccent("dacute|H|h", "oOuU", "őŐűŰ"); // double acute
1856 buildaccent("mathring|r", "aAuUwy",
1858 accents["check{\\imath}"] = "ǐ";
1859 accents["check{\\jmath}"] = "ǰ";
1860 buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
1861 "čČďĎǎǍěĚǐǏǒǑǔǓǧǦǩǨȟȞľĽňŇřŘšŠŤťžŽ"); // caron
1862 accents["hat{\\imath}"] = "î";
1863 accents["hat{\\jmath}"] = "ĵ";
1864 buildaccent("hat|^", "aAcCeEgGhHiIjJoOsSuUwWyYzZ",
1865 "âÂĉĈêÊĝĜĥĤîÎĵĴôÔŝŜûÛŵŴŷŶẑẐ"); // circ
1866 accents["bar{\\imath}"] = "ī";
1867 buildaccent("bar|=", "aAeEiIoOuUyY",
1868 "āĀēĒīĪōŌūŪȳȲ"); // macron
1869 accents["tilde{\\imath}"] = "ĩ";
1870 buildaccent("tilde", "aAeEiInNoOuUvVyY",
1871 "ãÃẽẼĩĨñÑõÕũŨṽṼỹỸ"); // tilde
1872 accents["breve{\\imath}"] = "ĭ";
1873 buildaccent("breve|u", "aAeEgGiIoOuU",
1874 "ăĂĕĔğĞĭĬŏŎŭŬ"); // breve
1875 accents["grave{\\imath}"] = "ì";
1876 buildaccent("grave|`", "aAeEiIoOuUnNwWyY",
1877 "àÀèÈìÌòÒùÙǹǸẁẀỳỲ"); // grave
1878 buildaccent("subdot|d", "BbDdHhKkLlMmNnRrSsTtVvWwZzAaEeIiOoUuYy",
1879 "ḄḅḌḍḤḥḲḳḶḷṂṃṆṇṚṛṢṣṬṭṾṿẈẉẒẓẠạẸẹỊịỌọỤụỴỵ"); // dot below
1880 buildaccent("ogonek|k", "AaEeIiUuOo",
1881 "ĄąĘęĮįŲųǪǫ"); // ogonek
1882 buildaccent("cedilla|c", "CcGgKkLlNnRrSsTtEeDdHh",
1883 "ÇçĢģĶķĻļŅņŖŗŞşŢţȨȩḐḑḨḩ"); // cedilla
1884 buildaccent("subring|textsubring", "Aa",
1886 buildaccent("subhat|textsubcircum", "DdEeLlNnTtUu",
1887 "ḒḓḘḙḼḽṊṋṰṱṶṷ"); // subcircum
1888 buildaccent("subtilde|textsubtilde", "EeIiUu",
1889 "ḚḛḬḭṴṵ"); // subtilde
1890 accents["dgrave{\\imath}"] = "ȉ";
1891 accents["textdoublegrave{\\i}"] = "ȉ";
1892 buildaccent("dgrave|textdoublegrave", "AaEeIiOoRrUu",
1893 "ȀȁȄȅȈȉȌȍȐȑȔȕ"); // double grave
1894 accents["rcap{\\imath}"] = "ȋ";
1895 accents["textroundcap{\\i}"] = "ȋ";
1896 buildaccent("rcap|textroundcap", "AaEeIiOoRrUu",
1897 "ȂȃȆȇȊȋȎȏȒȓȖȗ"); // inverted breve
1898 buildaccent("slashed", "oO",
1900 fillMissingUnicodesymbols(); // Add some still not handled entries contained in 'unicodesynbols'
1901 // LYXERR0("Number of accents " << accents.size());
1905 * Created accents in math or regexp environment
1906 * are macros, but we need the utf8 equivalent
1908 void Intervall::removeAccents()
1910 if (accents.empty())
1912 static regex const accre("\\\\(([\\S]|grave|breve|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot|ogonek|"
1913 "cedilla|subring|textsubring|subhat|textsubcircum|subtilde|textsubtilde|dgrave|textdoublegrave|rcap|textroundcap|slashed)\\{[^\\{\\}]+\\}"
1914 "|((i|imath|jmath|cdot|[a-z]+space)|((backslash )?([lL]y[xX]|[tT]e[xX]|[lL]a[tT]e[xX]e?|lyxarrow))|(textquote|brace|guillemot)(left|right)|textasciicircum|mathcircumflex|sim)(?![a-zA-Z]))");
1916 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
1918 string key = sub.str(1);
1919 AccentsIterator it_ac = accents.find(key);
1920 if (it_ac != accents.end()) {
1921 string val = it_ac->second;
1922 size_t pos = sub.position(size_t(0));
1923 for (size_t i = 0; i < val.size(); i++) {
1924 par[pos+i] = val[i];
1926 // Remove possibly following space too
1927 if (par[pos+sub.str(0).size()] == ' ')
1928 addIntervall(pos+val.size(), pos + sub.str(0).size()+1);
1930 addIntervall(pos+val.size(), pos + sub.str(0).size());
1931 for (size_t i = pos+val.size(); i < pos + sub.str(0).size(); i++) {
1932 // remove traces of any remaining chars
1937 LYXERR(Debug::INFO, "Not added accent for \"" << key << "\"");
1942 void Intervall::handleOpenP(int i)
1945 depts[actualdeptindex] = i+1;
1946 closes[actualdeptindex] = -1;
1947 checkDepthIndex(actualdeptindex);
1950 void Intervall::handleCloseP(int i, bool closingAllowed)
1952 if (actualdeptindex <= 0) {
1953 if (! closingAllowed)
1954 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1955 // if we are at the very end
1956 addIntervall(i, i+1);
1959 closes[actualdeptindex] = i+1;
1964 void Intervall::resetOpenedP(int openPos)
1966 // Used as initializer for foreignlanguage entry
1967 actualdeptindex = 1;
1968 depts[1] = openPos+1;
1972 int Intervall::previousNotIgnored(int start) const
1974 int idx = 0; /* int intervalls */
1975 for (idx = ignoreidx; idx >= 0; --idx) {
1976 if (start > borders[idx].upper)
1978 if (start >= borders[idx].low)
1979 start = borders[idx].low-1;
1984 int Intervall::nextNotIgnored(int start) const
1986 int idx = 0; /* int intervalls */
1987 for (idx = 0; idx <= ignoreidx; idx++) {
1988 if (start < borders[idx].low)
1990 if (start < borders[idx].upper)
1991 start = borders[idx].upper;
1996 typedef unordered_map<string, KeyInfo> KeysMap;
1997 typedef unordered_map<string, KeyInfo>::const_iterator KeysIterator;
1998 typedef vector< KeyInfo> Entries;
1999 static KeysMap keys = unordered_map<string, KeyInfo>();
2005 Intervall interval_;
2006 void buildKeys(bool);
2007 void buildEntries(bool);
2008 void makeKey(const string &, KeyInfo, bool isPatternString);
2009 void processRegion(int start, int region_end); /* remove {} parts */
2010 void removeHead(KeyInfo const &, int count=0);
2013 LatexInfo(string const & par, bool isPatternString)
2014 : entidx_(-1), interval_(isPatternString, par)
2016 buildKeys(isPatternString);
2017 entries_ = vector<KeyInfo>();
2018 buildEntries(isPatternString);
2022 if (entries_.empty()) {
2025 if (entries_[0].keytype == KeyInfo::isTitle) {
2026 interval_.hasTitle = true;
2027 if (! entries_[0].disabled) {
2028 interval_.titleValue = entries_[0].head;
2031 interval_.titleValue = "";
2033 removeHead(entries_[0]);
2034 if (entries_.size() > 1)
2043 if (int(entries_.size()) > entidx_) {
2050 bool setNextKey(int idx) {
2051 if ((idx == entidx_) && (entidx_ >= 0)) {
2058 int find(int start, KeyInfo::KeyType keytype) const {
2062 while (tmpIdx < int(entries_.size())) {
2063 if (entries_[tmpIdx].keytype == keytype)
2069 int process(ostringstream & os, KeyInfo const & actual);
2070 int dispatch(ostringstream & os, int previousStart, KeyInfo & actual);
2071 // string show(int lastpos) { return interval.show(lastpos);}
2072 int nextNotIgnored(int start) { return interval_.nextNotIgnored(start);}
2073 KeyInfo &getKeyInfo(int keyinfo) {
2074 static KeyInfo invalidInfo = KeyInfo();
2075 if ((keyinfo < 0) || ( keyinfo >= int(entries_.size())))
2078 return entries_[keyinfo];
2080 void setForDefaultLang(KeyInfo const & defLang) {interval_.setForDefaultLang(defLang);}
2081 void addIntervall(int low, int up) { interval_.addIntervall(low, up); }
2085 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
2089 for (int i = start; i < end; i += 1 + skip) {
2093 if (c == '\\') skip = 1;
2097 else if (c == down) {
2100 if ((repeat <= 0) || (par[i+1] != up))
2114 size_t mathpostfixsize;
2116 size_t mathprefixsize;
2120 vector<MathEntry> entries_;
2125 void insert(string const & wait, size_t start, size_t prefixsize, size_t end, size_t postfixsize) {
2126 MathEntry m = MathEntry();
2128 m.mathStart = start;
2129 m.mathprefixsize = prefixsize;
2130 m.mathEnd = end + postfixsize;
2131 m.mathpostfixsize = postfixsize;
2132 m.mathSize = m.mathEnd - m.mathStart;
2133 entries_.push_back(m);
2135 bool empty() const { return entries_.empty(); }
2136 size_t getEndPos() const {
2137 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2140 return entries_[actualIdx_].mathEnd;
2142 size_t getStartPos() const {
2143 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2144 return 100000; /* definitely enough? */
2146 return entries_[actualIdx_].mathStart;
2148 size_t getPrefixSize() const {
2149 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2152 return entries_[actualIdx_].mathprefixsize;
2154 size_t getPostfixSize() const {
2155 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2158 return entries_[actualIdx_].mathpostfixsize;
2160 size_t getFirstPos() {
2162 return getStartPos();
2164 size_t getSize() const {
2165 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2168 return entries_[actualIdx_].mathSize;
2170 void incrEntry() { actualIdx_++; }
2173 void LatexInfo::buildEntries(bool isPatternString)
2175 static regex const rmath("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|x?x?alignat)\\*?\\})(\\{[0-9]+\\})?)");
2176 static regex const rkeys("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?)))");
2177 static bool disableLanguageOverride = false;
2178 smatch sub, submath;
2179 bool evaluatingRegexp = false;
2181 bool evaluatingMath = false;
2182 bool evaluatingCode = false;
2184 bool evaluatingOptional = false;
2185 size_t optionalEnd = 0;
2188 bool math_end_waiting = false;
2189 size_t math_pos = 10000;
2190 size_t math_prefix_size = 1;
2192 static vector<string> usedText = vector<string>();
2193 static bool removeMathHull = false;
2195 interval_.removeAccents();
2197 for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) {
2199 if ((submath.position(2) - submath.position(0)) %2 == 1) {
2200 // prefixed by odd count of '\\'
2203 if (math_end_waiting) {
2204 size_t pos = submath.position(size_t(2));
2205 if ((math_end == "$") &&
2206 (submath.str(2) == "$")) {
2207 mi.insert("$", math_pos, 1, pos, 1);
2208 math_end_waiting = false;
2210 else if ((math_end == "\\]") &&
2211 (submath.str(2) == "\\]")) {
2212 mi.insert("\\]", math_pos, 2, pos, 2);
2213 math_end_waiting = false;
2215 else if ((submath.str(3).compare("end") == 0) &&
2216 (submath.str(5).compare(math_end) == 0)) {
2217 mi.insert(math_end, math_pos, math_prefix_size, pos, submath.str(2).length());
2218 math_end_waiting = false;
2224 if (submath.str(3).compare("begin") == 0) {
2225 math_end_waiting = true;
2226 math_end = submath.str(5);
2227 math_pos = submath.position(size_t(2));
2228 math_prefix_size = submath.str(2).length();
2230 else if (submath.str(2).compare("\\[") == 0) {
2231 math_end_waiting = true;
2233 math_pos = submath.position(size_t(2));
2235 else if (submath.str(2) == "$") {
2236 size_t pos = submath.position(size_t(2));
2237 math_end_waiting = true;
2243 // Ignore language if there is math somewhere in pattern-string
2244 if (isPatternString) {
2245 for (auto s: usedText) {
2246 // Remove entries created in previous search runs
2249 usedText = vector<string>();
2252 keys["foreignlanguage"].disabled = true;
2253 disableLanguageOverride = true;
2254 removeMathHull = false;
2257 removeMathHull = true; // used later if not isPatternString
2258 disableLanguageOverride = false;
2262 if (disableLanguageOverride) {
2263 keys["foreignlanguage"].disabled = true;
2266 math_pos = mi.getFirstPos();
2267 for (sregex_iterator it(interval_.par.begin(), interval_.par.end(), rkeys), end; it != end; ++it) {
2269 if ((sub.position(2) - sub.position(0)) %2 == 1) {
2270 // prefixed by odd count of '\\'
2273 string key = sub.str(5);
2275 if (sub.str(2)[0] == '\\')
2276 key = sub.str(2)[1];
2281 KeysIterator it_key = keys.find(key);
2282 if (it_key != keys.end()) {
2283 if (it_key->second.keytype == KeyInfo::headRemove) {
2284 KeyInfo found1 = it_key->second;
2285 found1.disabled = true;
2286 found1.head = "\\" + key + "{";
2287 found1._tokenstart = sub.position(size_t(2));
2288 found1._tokensize = found1.head.length();
2289 found1._dataStart = found1._tokenstart + found1.head.length();
2290 int endpos = interval_.findclosing(found1._dataStart, interval_.par.length(), '{', '}', 1);
2291 found1._dataEnd = endpos;
2296 if (evaluatingRegexp) {
2297 if (sub.str(3).compare("endregexp") == 0) {
2298 evaluatingRegexp = false;
2299 // found._tokenstart already set
2300 found._dataEnd = sub.position(size_t(2)) + 13;
2301 found._dataStart = found._dataEnd;
2302 found._tokensize = found._dataEnd - found._tokenstart;
2303 found.parenthesiscount = 0;
2304 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2311 if (evaluatingMath) {
2312 if (size_t(sub.position(size_t(2))) < mi.getEndPos())
2314 evaluatingMath = false;
2316 math_pos = mi.getStartPos();
2318 if (it_key == keys.end()) {
2319 found = KeyInfo(KeyInfo::isStandard, 0, true);
2320 LYXERR(Debug::INFO, "Undefined key " << key << " ==> will be used as text");
2321 found = KeyInfo(KeyInfo::isText, 0, false);
2322 if (isPatternString) {
2323 found.keytype = KeyInfo::isChar;
2324 found.disabled = false;
2328 usedText.push_back(key);
2332 if (key.compare("regexp") == 0) {
2333 evaluatingRegexp = true;
2334 found._tokenstart = sub.position(size_t(2));
2335 found._tokensize = 0;
2339 // Handle the other params of key
2340 if (found.keytype == KeyInfo::isIgnored)
2342 else if (found.keytype == KeyInfo::isMath) {
2343 if (size_t(sub.position(size_t(2))) == math_pos) {
2345 found._tokenstart = sub.position(size_t(2));
2346 found._tokensize = mi.getSize();
2347 found._dataEnd = found._tokenstart + found._tokensize;
2348 found._dataStart = found._dataEnd;
2349 found.parenthesiscount = 0;
2350 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2351 if (removeMathHull) {
2352 interval_.addIntervall(found._tokenstart, found._tokenstart + mi.getPrefixSize());
2353 interval_.addIntervall(found._dataEnd - mi.getPostfixSize(), found._dataEnd);
2356 // Treate all math constructs as simple math
2357 interval_.par[found._tokenstart] = '$';
2358 interval_.par[found._dataEnd - mi.getPostfixSize()] = '$';
2359 interval_.addIntervall(found._tokenstart + 1, found._tokenstart + mi.getPrefixSize());
2360 interval_.addIntervall(found._dataEnd - mi.getPostfixSize() + 1, found._dataEnd);
2362 evaluatingMath = true;
2365 // begin|end of unknown env, discard
2366 // First handle tables
2367 // longtable|tabular
2368 bool discardComment;
2370 found.keytype = KeyInfo::doRemove;
2371 if ((sub.str(7).compare("longtable") == 0) ||
2372 (sub.str(7).compare("tabular") == 0)) {
2373 discardComment = true; /* '%' */
2376 discardComment = false;
2377 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
2379 string token = sub.str(7);
2380 if (regex_match(token, sub2, removeArgs)) {
2381 found.keytype = KeyInfo::removeWithArg;
2384 // discard spaces before pos(2)
2385 int pos = sub.position(size_t(2));
2387 for (count = 0; pos - count > 0; count++) {
2388 char c = interval_.par[pos-count-1];
2389 if (discardComment) {
2390 if ((c != ' ') && (c != '%'))
2396 found._tokenstart = pos - count;
2397 if (sub.str(3).compare(0, 5, "begin") == 0) {
2398 size_t pos1 = pos + sub.str(2).length();
2399 if (sub.str(7).compare("cjk") == 0) {
2400 pos1 = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2401 if ((interval_.par[pos1] == '{') && (interval_.par[pos1+1] == '}'))
2403 found.keytype = KeyInfo::isMain;
2404 found._dataStart = pos1;
2405 found._dataEnd = interval_.par.length();
2406 found.disabled = keys["foreignlanguage"].disabled;
2407 found.used = keys["foreignlanguage"].used;
2408 found._tokensize = pos1 - found._tokenstart;
2409 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2412 // Swallow possible optional params
2413 while (interval_.par[pos1] == '[') {
2414 pos1 = interval_.findclosing(pos1+1, interval_.par.length(), '[', ']')+1;
2416 // Swallow also the eventual parameter
2417 if (interval_.par[pos1] == '{') {
2418 found._dataEnd = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2421 found._dataEnd = pos1;
2423 found._dataStart = found._dataEnd;
2424 found._tokensize = count + found._dataEnd - pos;
2425 found.parenthesiscount = 0;
2426 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2427 found.disabled = true;
2431 // Handle "\end{...}"
2432 found._dataStart = pos + sub.str(2).length();
2433 found._dataEnd = found._dataStart;
2434 found._tokensize = count + found._dataEnd - pos;
2435 found.parenthesiscount = 0;
2436 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2437 found.disabled = true;
2441 else if (found.keytype != KeyInfo::isRegex) {
2442 found._tokenstart = sub.position(size_t(2));
2443 if (found.parenthesiscount == 0) {
2444 // Probably to be discarded
2445 size_t following_pos = sub.position(size_t(2)) + sub.str(5).length() + 1;
2446 char following = interval_.par[following_pos];
2447 if (following == ' ')
2448 found.head = "\\" + sub.str(5) + " ";
2449 else if (following == '=') {
2450 // like \uldepth=1000pt
2451 found.head = sub.str(2);
2454 found.head = "\\" + key;
2455 found._tokensize = found.head.length();
2456 found._dataEnd = found._tokenstart + found._tokensize;
2457 found._dataStart = found._dataEnd;
2460 int params = found._tokenstart + key.length() + 1;
2461 if (evaluatingOptional) {
2462 if (size_t(found._tokenstart) > optionalEnd) {
2463 evaluatingOptional = false;
2466 found.disabled = true;
2469 int optend = params;
2470 while (interval_.par[optend] == '[') {
2471 // discard optional parameters
2472 optend = interval_.findclosing(optend+1, interval_.par.length(), '[', ']') + 1;
2474 if (optend > params) {
2475 key += interval_.par.substr(params, optend-params);
2476 evaluatingOptional = true;
2477 optionalEnd = optend;
2478 if (found.keytype == KeyInfo::isSectioning) {
2479 // Remove optional values (but still keep in header)
2480 interval_.addIntervall(params, optend);
2483 string token = sub.str(7);
2485 if (interval_.par[optend] != '{') {
2487 found.parenthesiscount = 0;
2488 found.head = "\\" + key;
2491 closings = found.parenthesiscount;
2492 if (found.parenthesiscount == 1) {
2493 found.head = "\\" + key + "{";
2495 else if (found.parenthesiscount > 1) {
2497 found.head = sub.str(2) + "{";
2498 closings = found.parenthesiscount - 1;
2501 found.head = "\\" + key + "{";
2504 found._tokensize = found.head.length();
2505 found._dataStart = found._tokenstart + found.head.length();
2506 if (found.keytype == KeyInfo::doRemove) {
2508 size_t endpar = 2 + interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2509 if (endpar >= interval_.par.length())
2510 found._dataStart = interval_.par.length();
2512 found._dataStart = endpar;
2513 found._tokensize = found._dataStart - found._tokenstart;
2516 found._dataStart = found._tokenstart + found._tokensize;
2520 if (interval_.par.substr(found._dataStart, 15).compare("\\endarguments{}") == 0) {
2521 found._dataStart += 15;
2525 endpos = found._dataStart - 1;
2527 endpos = interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2528 if (found.keytype == KeyInfo::isList) {
2529 // Check if it really is list env
2530 static regex const listre("^([a-z]+)$");
2532 if (!regex_match(token, sub2, listre)) {
2533 // Change the key of this entry. It is not in a list/item environment
2534 found.keytype = KeyInfo::endArguments;
2537 if (found.keytype == KeyInfo::noMain) {
2538 evaluatingCode = true;
2540 codeStart = found._dataStart;
2542 else if (evaluatingCode) {
2543 if (size_t(found._dataStart) > codeEnd)
2544 evaluatingCode = false;
2545 else if (found.keytype == KeyInfo::isMain) {
2546 // Disable this key, treate it as standard
2547 found.keytype = KeyInfo::isStandard;
2548 found.disabled = true;
2549 if ((codeEnd +1 >= interval_.par.length()) &&
2550 (found._tokenstart == codeStart)) {
2551 // trickery, because the code inset starts
2552 // with \selectlanguage ...
2554 if (entries_.size() > 1) {
2555 entries_[entries_.size()-1]._dataEnd = codeEnd;
2560 if ((endpos == interval_.par.length()) &&
2561 (found.keytype == KeyInfo::doRemove)) {
2562 // Missing closing => error in latex-input?
2563 // therefore do not delete remaining data
2564 found._dataStart -= 1;
2565 found._dataEnd = found._dataStart;
2568 found._dataEnd = endpos;
2570 if (isPatternString) {
2571 keys[key].used = true;
2574 entries_.push_back(found);
2578 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
2580 stringstream s(keysstring);
2582 const char delim = '|';
2583 while (getline(s, key, delim)) {
2584 KeyInfo keyII(keyI);
2585 if (isPatternString) {
2588 else if ( !keys[key].used)
2589 keyII.disabled = true;
2594 void LatexInfo::buildKeys(bool isPatternString)
2597 static bool keysBuilt = false;
2598 if (keysBuilt && !isPatternString) return;
2600 // Keys to ignore in any case
2601 makeKey("text|textcyrillic|lyxmathsym|ensuremath", KeyInfo(KeyInfo::headRemove, 1, true), true);
2602 // Known standard keys with 1 parameter.
2603 // Split is done, if not at start of region
2604 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
2605 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
2606 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
2607 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
2608 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
2609 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
2611 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
2612 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2613 makeKey("section*|subsection*|subsubsection*|paragraph*",
2614 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2615 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2616 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
2618 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
2620 // Split is done, if not at start of region
2621 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
2622 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
2624 // Split is done always.
2625 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
2629 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2630 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2631 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2632 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2634 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2635 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2636 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2637 makeKey("thickspace|medspace|thinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2639 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2640 // Custom space/skip, remove the content (== length value)
2641 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
2642 // Found in fr/UserGuide.lyx
2643 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2645 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2646 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2647 // Known macros to remove (including their parameter)
2649 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
2650 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
2651 // handle like standard keys with 1 parameter.
2652 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
2654 // Ignore deleted text
2655 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString);
2656 // but preserve added text
2657 makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2659 // Macros to remove, but let the parameter survive
2661 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2663 // Remove language spec from content of these insets
2664 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
2666 // Same effect as previous, parameter will survive (because there is no one anyway)
2668 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2669 // Remove table decorations
2670 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
2671 // Discard shape-header.
2672 // For footnote or shortcut too, because of lang settings
2673 // and wrong handling if used 'KeyInfo::noMain'
2674 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2675 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2676 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2677 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2678 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2679 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
2680 // like ('tiny{}' or '\tiny ' ... )
2681 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
2683 // Survives, like known character
2684 // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2685 makeKey("tableofcontents", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2686 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
2688 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2689 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2690 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2692 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2693 // Remove RTL/LTR marker
2694 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2695 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
2696 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
2697 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
2698 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
2699 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
2700 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2701 if (isPatternString) {
2702 // Allow the first searched string to rebuild the keys too
2706 // no need to rebuild again
2712 * Keep the list of actual opened parentheses actual
2713 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2715 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2718 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2722 if (c == '\\') skip = 1;
2723 else if (c == '{') {
2726 else if (c == '}') {
2727 handleCloseP(i, closingAllowed);
2733 string Intervall::show(int lastpos)
2735 int idx = 0; /* int intervalls */
2738 for (idx = 0; idx <= ignoreidx; idx++) {
2739 while (i < lastpos) {
2741 if (i <= borders[idx].low) {
2742 if (borders[idx].low > lastpos)
2743 printsize = lastpos - i;
2745 printsize = borders[idx].low - i;
2746 s += par.substr(i, printsize);
2748 if (i >= borders[idx].low)
2749 i = borders[idx].upper;
2752 i = borders[idx].upper;
2758 s += par.substr(i, lastpos-i);
2764 void Intervall::output(ostringstream &os, int lastpos)
2766 // get number of chars to output
2767 int idx = 0; /* int intervalls */
2770 string startTitle = titleValue;
2771 for (idx = 0; idx <= ignoreidx; idx++) {
2773 if (i <= borders[idx].low) {
2775 if (borders[idx].low > lastpos)
2776 printsize = lastpos - i;
2778 printsize = borders[idx].low - i;
2779 if (printsize > 0) {
2780 os << startTitle << par.substr(i, printsize);
2782 printed += printsize;
2785 handleParentheses(i, false);
2786 if (i >= borders[idx].low)
2787 i = borders[idx].upper;
2790 i = borders[idx].upper;
2797 os << startTitle << par.substr(i, lastpos-i);
2798 printed += lastpos-i;
2800 handleParentheses(lastpos, false);
2802 if (keys["foreignlanguage"].disabled)
2803 startindex = actualdeptindex-langcount;
2805 startindex = actualdeptindex;
2806 for (int i = startindex; i > 0; --i) {
2809 if (hasTitle && (printed > 0))
2811 if (! isPatternString_)
2813 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2816 void LatexInfo::processRegion(int start, int region_end)
2818 while (start < region_end) { /* Let {[} and {]} survive */
2819 int cnt = interval_.isOpeningPar(start);
2821 // Closing is allowed past the region
2822 int closing = interval_.findclosing(start+1, interval_.par.length());
2823 interval_.addIntervall(start, start+1);
2824 interval_.addIntervall(closing, closing+1);
2828 start = interval_.nextNotIgnored(start+1);
2832 void LatexInfo::removeHead(KeyInfo const & actual, int count)
2834 if (actual.parenthesiscount == 0) {
2835 // "{\tiny{} ...}" ==> "{{} ...}"
2836 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2839 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2840 interval_.addIntervall(actual._tokenstart - count, actual._dataStart);
2841 interval_.addIntervall(actual._dataEnd, actual._dataEnd+1);
2845 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2848 switch (actual.keytype)
2850 case KeyInfo::isTitle: {
2852 nextKeyIdx = getNextKey();
2855 case KeyInfo::cleanToStart: {
2856 actual._dataEnd = actual._dataStart;
2857 nextKeyIdx = getNextKey();
2858 // Search for end of arguments
2859 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2861 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2862 entries_[i].disabled = true;
2864 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2866 while (interval_.par[actual._dataEnd] == ' ')
2868 interval_.addIntervall(0, actual._dataEnd+1);
2869 interval_.actualdeptindex = 0;
2870 interval_.depts[0] = actual._dataEnd+1;
2871 interval_.closes[0] = -1;
2874 case KeyInfo::isText:
2875 interval_.par[actual._tokenstart] = '#';
2876 //interval_.addIntervall(actual._tokenstart, actual._tokenstart+1);
2877 nextKeyIdx = getNextKey();
2879 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2880 if (actual.disabled)
2881 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2883 interval_.addIntervall(actual._dataStart, actual._dataEnd);
2886 case KeyInfo::isChar: {
2887 nextKeyIdx = getNextKey();
2890 case KeyInfo::isSize: {
2891 if (actual.disabled || (interval_.par[actual._dataStart] != '{') || (interval_.par[actual._dataStart-1] == ' ')) {
2892 if (actual.parenthesiscount == 0)
2893 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2895 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2897 nextKeyIdx = getNextKey();
2899 // Here _dataStart points to '{', so correct it
2900 actual._dataStart += 1;
2901 actual._tokensize += 1;
2902 actual.parenthesiscount = 1;
2903 if (interval_.par[actual._dataStart] == '}') {
2904 // Determine the end if used like '{\tiny{}...}'
2905 actual._dataEnd = interval_.findclosing(actual._dataStart+1, interval_.par.length()) + 1;
2906 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2909 // Determine the end if used like '\tiny{...}'
2910 actual._dataEnd = interval_.findclosing(actual._dataStart, interval_.par.length()) + 1;
2912 // Split on this key if not at start
2913 int start = interval_.nextNotIgnored(previousStart);
2914 if (start < actual._tokenstart) {
2915 interval_.output(os, actual._tokenstart);
2916 interval_.addIntervall(start, actual._tokenstart);
2918 // discard entry if at end of actual
2919 nextKeyIdx = process(os, actual);
2923 case KeyInfo::endArguments: {
2924 // Remove trailing '{}' too
2925 actual._dataStart += 1;
2926 actual._dataEnd += 1;
2927 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2928 nextKeyIdx = getNextKey();
2931 case KeyInfo::noMain:
2933 case KeyInfo::isStandard: {
2934 if (actual.disabled) {
2936 processRegion(actual._dataStart, actual._dataStart+1);
2937 nextKeyIdx = getNextKey();
2939 // Split on this key if not at datastart of calling entry
2940 int start = interval_.nextNotIgnored(previousStart);
2941 if (start < actual._tokenstart) {
2942 interval_.output(os, actual._tokenstart);
2943 interval_.addIntervall(start, actual._tokenstart);
2945 // discard entry if at end of actual
2946 nextKeyIdx = process(os, actual);
2950 case KeyInfo::removeWithArg: {
2951 nextKeyIdx = getNextKey();
2952 // Search for end of arguments
2953 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2955 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2956 entries_[i].disabled = true;
2958 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2960 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2963 case KeyInfo::doRemove: {
2964 // Remove the key with all parameters and following spaces
2967 if (interval_.par[actual._dataEnd-1] == ' ')
2968 start = actual._dataEnd;
2970 start = actual._dataEnd+1;
2971 for (pos = start; pos < interval_.par.length(); pos++) {
2972 if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
2975 // Remove also enclosing parentheses [] and {}
2978 while (actual._tokenstart > numpars) {
2979 if (pos+numpars >= interval_.par.size())
2981 else if (interval_.par[pos+numpars] == ']' && interval_.par[actual._tokenstart-numpars-1] == '[')
2983 else if (interval_.par[pos+numpars] == '}' && interval_.par[actual._tokenstart-numpars-1] == '{')
2989 if (interval_.par[pos+numpars] == ' ')
2993 interval_.addIntervall(actual._tokenstart-numpars, pos+numpars+spaces);
2994 nextKeyIdx = getNextKey();
2997 case KeyInfo::isList: {
2998 // Discard space before _tokenstart
3000 for (count = 0; count < actual._tokenstart; count++) {
3001 if (interval_.par[actual._tokenstart-count-1] != ' ')
3004 nextKeyIdx = getNextKey();
3005 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
3007 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
3009 // How else can we catch this one?
3010 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
3011 entries_[i].disabled = true;
3013 actual._dataEnd = entries_[tmpIdx]._dataEnd;
3015 else if (nextKeyIdx > 0) {
3016 // Ignore any lang entries inside data region
3017 for (int i = nextKeyIdx; i < int(entries_.size()) && entries_[i]._tokenstart < actual._dataEnd; i++) {
3018 if (entries_[i].keytype == KeyInfo::isMain)
3019 entries_[i].disabled = true;
3022 if (actual.disabled) {
3023 interval_.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
3026 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
3028 if (interval_.par[actual._dataEnd+1] == '[') {
3029 int posdown = interval_.findclosing(actual._dataEnd+2, interval_.par.length(), '[', ']');
3030 if ((interval_.par[actual._dataEnd+2] == '{') &&
3031 (interval_.par[posdown-1] == '}')) {
3032 interval_.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
3033 interval_.addIntervall(posdown-1, posdown+1);
3036 interval_.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
3037 interval_.addIntervall(posdown, posdown+1);
3039 int blk = interval_.nextNotIgnored(actual._dataEnd+1);
3040 if (blk > posdown) {
3041 // Discard at most 1 space after empty item
3043 for (count = 0; count < 1; count++) {
3044 if (interval_.par[blk+count] != ' ')
3048 interval_.addIntervall(blk, blk+count);
3053 case KeyInfo::isSectioning: {
3054 // Discard spaces before _tokenstart
3056 int val = actual._tokenstart;
3057 for (count = 0; count < actual._tokenstart;) {
3058 val = interval_.previousNotIgnored(val-1);
3059 if (val < 0 || interval_.par[val] != ' ')
3062 count = actual._tokenstart - val;
3065 if (actual.disabled) {
3066 removeHead(actual, count);
3067 nextKeyIdx = getNextKey();
3069 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
3070 nextKeyIdx = process(os, actual);
3074 case KeyInfo::isMath: {
3075 // Same as regex, use the content unchanged
3076 nextKeyIdx = getNextKey();
3079 case KeyInfo::isRegex: {
3080 // DO NOT SPLIT ON REGEX
3082 nextKeyIdx = getNextKey();
3085 case KeyInfo::isIgnored: {
3086 // Treat like a character for now
3087 nextKeyIdx = getNextKey();
3090 case KeyInfo::isMain: {
3091 if (interval_.par.substr(actual._dataStart, 2) == "% ")
3092 interval_.addIntervall(actual._dataStart, actual._dataStart+2);
3093 if (actual._tokenstart > 0) {
3094 int prev = interval_.previousNotIgnored(actual._tokenstart - 1);
3095 if ((prev >= 0) && interval_.par[prev] == '%')
3096 interval_.addIntervall(prev, prev+1);
3098 if (actual.disabled) {
3100 interval_.langcount++;
3101 if ((interval_.par.substr(actual._dataStart, 3) == " \\[") ||
3102 (interval_.par.substr(actual._dataStart, 8) == " \\begin{")) {
3103 // Discard also the space before math-equation
3104 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
3106 nextKeyIdx = getNextKey();
3107 // interval.resetOpenedP(actual._dataStart-1);
3110 if (actual._tokenstart < 26) {
3111 // for the first (and maybe dummy) language
3112 interval_.setForDefaultLang(actual);
3114 interval_.resetOpenedP(actual._dataStart-1);
3118 case KeyInfo::invalid:
3119 case KeyInfo::headRemove:
3120 // These two cases cannot happen, already handled
3123 // LYXERR(Debug::INFO, "Unhandled keytype");
3124 nextKeyIdx = getNextKey();
3131 int LatexInfo::process(ostringstream & os, KeyInfo const & actual )
3133 int end = interval_.nextNotIgnored(actual._dataEnd);
3134 int oldStart = actual._dataStart;
3135 int nextKeyIdx = getNextKey();
3137 if ((nextKeyIdx < 0) ||
3138 (entries_[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
3139 (entries_[nextKeyIdx].keytype == KeyInfo::invalid)) {
3140 if (oldStart <= end) {
3141 processRegion(oldStart, end);
3146 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
3148 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
3149 (void) dispatch(os, actual._dataStart, nextKey);
3150 end = nextKey._tokenstart;
3153 processRegion(oldStart, nextKey._tokenstart);
3154 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
3156 oldStart = nextKey._dataEnd+1;
3158 // now nextKey is either invalid or is outside of actual._dataEnd
3159 // output the remaining and discard myself
3160 if (oldStart <= end) {
3161 processRegion(oldStart, end);
3163 if (interval_.par.size() > (size_t) end && interval_.par[end] == '}') {
3165 // This is the normal case.
3166 // But if using the firstlanguage, the closing may be missing
3168 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
3170 if (actual._dataEnd < end)
3171 output_end = interval_.nextNotIgnored(actual._dataEnd);
3172 else if (interval_.par.size() > (size_t) end)
3173 output_end = interval_.nextNotIgnored(end);
3175 output_end = interval_.par.size();
3176 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
3177 interval_.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
3179 // Remove possible empty data
3180 int dstart = interval_.nextNotIgnored(actual._dataStart);
3181 while (interval_.isOpeningPar(dstart) == 1) {
3182 interval_.addIntervall(dstart, dstart+1);
3183 int dend = interval_.findclosing(dstart+1, output_end);
3184 interval_.addIntervall(dend, dend+1);
3185 dstart = interval_.nextNotIgnored(dstart+1);
3187 if (dstart < output_end)
3188 interval_.output(os, output_end);
3190 interval_.addIntervall(0, end);
3192 interval_.addIntervall(actual._tokenstart, end);
3196 string splitOnKnownMacros(string par, bool isPatternString)
3199 LatexInfo li(par, isPatternString);
3200 // LYXERR(Debug::INFO, "Berfore split: " << par);
3201 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
3203 DummyKey._tokensize = 0;
3204 DummyKey._dataStart = 0;
3205 DummyKey._dataEnd = par.length();
3206 DummyKey.disabled = true;
3207 int firstkeyIdx = li.getFirstKey();
3209 if (firstkeyIdx >= 0) {
3210 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
3211 DummyKey._tokenstart = firstKey._tokenstart;
3213 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
3214 // Use dummy firstKey
3215 firstKey = DummyKey;
3216 (void) li.setNextKey(firstkeyIdx);
3219 if (par.substr(firstKey._dataStart, 2) == "% ")
3220 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
3222 nextkeyIdx = li.process(os, firstKey);
3223 while (nextkeyIdx >= 0) {
3224 // Check for a possible gap between the last
3225 // entry and this one
3226 int datastart = li.nextNotIgnored(firstKey._dataStart);
3227 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
3228 if ((nextKey._tokenstart > datastart)) {
3230 firstKey._dataStart = datastart;
3231 firstKey._dataEnd = par.length();
3232 (void) li.setNextKey(nextkeyIdx);
3233 // Fake the last opened parenthesis
3234 li.setForDefaultLang(firstKey);
3235 nextkeyIdx = li.process(os, firstKey);
3238 if (nextKey.keytype != KeyInfo::isMain) {
3239 firstKey._dataStart = datastart;
3240 firstKey._dataEnd = nextKey._dataEnd+1;
3241 (void) li.setNextKey(nextkeyIdx);
3242 li.setForDefaultLang(firstKey);
3243 nextkeyIdx = li.process(os, firstKey);
3246 nextkeyIdx = li.process(os, nextKey);
3250 // Handle the remaining
3251 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
3252 firstKey._dataEnd = par.length();
3254 if ((firstKey._dataStart < firstKey._dataEnd) &&
3255 (par[firstKey._dataStart] != '}')) {
3256 li.setForDefaultLang(firstKey);
3257 (void) li.process(os, firstKey);
3261 // return string definitelly impossible to match
3262 s = "\\foreignlanguage{ignore}{ }";
3266 s = par; /* no known macros found */
3267 // LYXERR(Debug::INFO, "After split: " << s);
3272 * Try to unify the language specs in the latexified text.
3273 * Resulting modified string is set to "", if
3274 * the searched tex does not contain all the features in the search pattern
3276 static string correctlanguagesetting(string par, bool isPatternString, bool withformat, lyx::Buffer *pbuf = nullptr)
3278 static Features regex_f;
3279 static int missed = 0;
3280 static bool regex_with_format = false;
3282 int parlen = par.length();
3284 while ((parlen > 0) && (par[parlen-1] == '\n')) {
3287 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
3288 // Happens to be there in case of description or labeling environment
3293 // Split the latex input into pieces which
3294 // can be digested by our search engine
3295 LYXERR(Debug::FIND, "input: \"" << par << "\"");
3296 if (isPatternString && (pbuf != nullptr)) { // Check if we should disable/enable test for language
3297 // We check for polyglossia, because in runparams.flavor we use Flavor::XeTeX
3298 string doclang = pbuf->params().language->polyglossia();
3299 static regex langre("\\\\(foreignlanguage)\\{([^\\}]+)\\}");
3301 bool toIgnoreLang = true;
3302 for (sregex_iterator it(par.begin(), par.end(), langre), end; it != end; ++it) {
3304 if (sub.str(2) != doclang) {
3305 toIgnoreLang = false;
3309 setIgnoreFormat("language", toIgnoreLang, false);
3312 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
3313 LYXERR(Debug::FIND, "After splitOnKnownMacros:\n\"" << result << "\"");
3316 result = par.substr(0, parlen);
3317 if (isPatternString) {
3320 regex_f = identifyFeatures(result);
3321 string features = "";
3322 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3323 string a = it->first;
3324 regex_with_format = true;
3325 features += " " + a;
3326 // LYXERR(Debug::INFO, "Identified regex format:" << a);
3328 LYXERR(Debug::FIND, "Identified Features" << features);
3331 } else if (regex_with_format) {
3332 Features info = identifyFeatures(result);
3333 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3334 string a = it->first;
3335 bool b = it->second;
3336 if (b && ! info[a]) {
3338 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
3345 // LYXERR(Debug::INFO, "No regex formats");
3351 // Remove trailing closure of math, macros and environments, so to catch parts of them.
3352 static int identifyClosing(string & t)
3354 int open_braces = 0;
3356 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
3357 if (regex_replace(t, t, "(.*[^\\\\])\\$$", "$1"))
3359 if (regex_replace(t, t, "(.*[^\\\\])\\\\\\]$", "$1"))
3361 if (regex_replace(t, t, "(.*[^\\\\])\\\\end\\{[a-zA-Z_]*\\*?\\}$", "$1"))
3363 if (regex_replace(t, t, "(.*[^\\\\])\\}$", "$1")) {
3372 static int num_replaced = 0;
3373 static bool previous_single_replace = true;
3375 void MatchStringAdv::CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string)
3378 // Handle \w properly
3379 QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
3380 if (! opt.casesensitive) {
3381 popts |= QRegularExpression::CaseInsensitiveOption;
3383 regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
3384 regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
3386 if (regexp.isValid() && regexp2.isValid()) {
3387 regexIsValid = true;
3388 // Check '{', '}' pairs inside the regex
3391 for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
3392 char c = par_as_string[i];
3399 else if (c == '}') {
3406 if (balanced != 0) {
3407 regexIsValid = false;
3408 regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
3412 regexIsValid = false;
3413 if (!regexp.isValid())
3414 regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
3416 regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
3419 (void)par_as_string;
3420 if (opt.casesensitive) {
3421 regexp = regex(regexp_str);
3422 regexp2 = regex(regexp2_str);
3425 regexp = regex(regexp_str, std::regex_constants::icase);
3426 regexp2 = regex(regexp2_str, std::regex_constants::icase);
3431 static void modifyRegexForMatchWord(string &t)
3434 regex wordre("(\\\\)*((\\.|\\\\b))");
3437 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
3439 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
3442 else if (sub.str(2) == "\\\\b")
3444 if (lastpos < (size_t) sub.position(2))
3445 s += t.substr(lastpos, sub.position(2) - lastpos);
3447 lastpos = sub.position(2) + sub.length(2);
3450 s = "\\b" + t + "\\b";
3454 else if (lastpos < t.length())
3455 s += t.substr(lastpos, t.length() - lastpos);
3456 t = "\\b" + s + "\\b";
3459 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt)
3460 : p_buf(&buf), p_first_buf(&buf), opt(opt)
3462 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
3463 docstring const & ds = stringifySearchBuffer(find_buf, opt);
3464 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
3465 if (opt.replace_all && previous_single_replace) {
3466 previous_single_replace = false;
3469 else if (!opt.replace_all) {
3470 num_replaced = 0; // count number of replaced strings
3471 previous_single_replace = true;
3473 // When using regexp, braces are hacked already by escape_for_regex()
3474 par_as_string = normalize(ds);
3476 close_wildcards = 0;
3478 size_t lead_size = 0;
3479 // correct the language settings
3480 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat, &buf);
3481 opt.matchAtStart = false;
3483 identifyClosing(par_as_string); // Removes math closings ($, ], ...) at end of string
3484 if (opt.ignoreformat) {
3488 lead_size = identifyLeading(par_as_string);
3490 lead_as_string = par_as_string.substr(0, lead_size);
3491 string lead_as_regex_string = string2regex(lead_as_string);
3492 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3493 string par_as_regex_string_nolead = string2regex(par_as_string_nolead);
3494 /* Handle whole words too in this case
3496 if (opt.matchword) {
3497 par_as_regex_string_nolead = "\\b" + par_as_regex_string_nolead + "\\b";
3498 opt.matchword = false;
3500 string regexp_str = "(" + lead_as_regex_string + ")()" + par_as_regex_string_nolead;
3501 string regexp2_str = "(" + lead_as_regex_string + ")(.*?)" + par_as_regex_string_nolead;
3502 CreateRegexp(opt, regexp_str, regexp2_str);
3504 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3505 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3509 if (!opt.ignoreformat) {
3510 lead_size = identifyLeading(par_as_string);
3511 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
3512 lead_as_string = par_as_string.substr(0, lead_size);
3513 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3516 // Here we are using regexp
3517 LASSERT(use_regexp, /**/);
3519 string lead_as_regexp;
3520 if (lead_size > 0) {
3521 lead_as_regexp = string2regex(par_as_string.substr(0, lead_size));
3522 (void)regex_replace(par_as_string_nolead, par_as_string_nolead, "\\$$", "");
3523 (void)regex_replace(par_as_string_nolead, par_as_string_nolead, "}$", "");
3524 par_as_string = par_as_string_nolead;
3525 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
3526 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3528 // LYXERR(Debug::FIND, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
3529 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
3530 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
3531 // LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3533 size_t lng = par_as_string.size();
3534 if (!opt.ignoreformat) {
3535 // Remove extra '\}' at end if not part of \{\.\}
3537 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
3539 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
3548 if (lng < par_as_string.size())
3549 par_as_string = par_as_string.substr(0,lng);
3551 LYXERR(Debug::FIND, "par_as_string after correctRegex is '" << par_as_string << "'");
3552 if ((lng > 0) && (par_as_string[0] == '^')) {
3553 par_as_string = par_as_string.substr(1);
3555 opt.matchAtStart = true;
3557 // LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3558 // LYXERR(Debug::FIND, "Open braces: " << open_braces);
3559 // LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
3561 // If entered regexp must match at begin of searched string buffer
3562 // Kornel: Added parentheses to use $1 for size of the leading string
3566 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
3567 // Unfortunately is '\1', '\2', etc not working for strings with extra format
3568 // so the convert has no effect in that case
3569 for (int i = 7; i > 0; --i) {
3570 string orig = "\\\\" + std::to_string(i);
3571 string dest = "\\" + std::to_string(i+2);
3572 while (regex_replace(par_as_string, par_as_string, orig, dest));
3574 if (opt.matchword) {
3575 modifyRegexForMatchWord(par_as_string);
3576 opt.matchword = false;
3578 regexp_str = "(" + lead_as_regexp + ")()" + par_as_string;
3579 regexp2_str = "(" + lead_as_regexp + ")(.*?)" + par_as_string;
3581 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3582 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3583 CreateRegexp(opt, regexp_str, regexp2_str, par_as_string);
3587 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
3591 mres.searched_size = len;
3593 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
3596 docstring docstr = stringifyFromForSearch(opt, cur, len);
3598 str = normalize(docstr);
3599 if (!opt.ignoreformat) {
3600 str = correctlanguagesetting(str, false, !opt.ignoreformat);
3601 // remove closing '}' and '\n' to allow for use of '$' in regex
3602 size_t lng = str.size();
3603 while ((lng > 1) && ((str[lng -1] == '}') || (str[lng -1] == '\n')))
3605 if (lng != str.size()) {
3606 str = str.substr(0, lng);
3610 mres.match_len = -1;
3613 LYXERR(Debug::FIND, "After normalization: Matching against:\n'" << str << "'");
3615 LASSERT(use_regexp, /**/);
3617 // use_regexp always true
3618 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
3620 QString qstr = QString::fromStdString(str);
3621 QRegularExpression const *p_regexp;
3622 QRegularExpression::MatchType flags = QRegularExpression::NormalMatch;
3626 p_regexp = ®exp2;
3628 QRegularExpressionMatch match = p_regexp->match(qstr, 0, flags);
3629 if (!match.hasMatch())
3632 regex const *p_regexp;
3633 regex_constants::match_flag_type flags;
3635 flags = regex_constants::match_continuous;
3638 flags = regex_constants::match_default;
3639 p_regexp = ®exp2;
3641 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
3642 if (re_it == sregex_iterator())
3644 match_results<string::const_iterator> const & m = *re_it;
3646 // Whole found string, including the leading
3647 // std: m[0].second - m[0].first
3648 // Qt: match.capturedEnd(0) - match.capturedStart(0)
3650 // Size of the leading string
3651 // std: m[1].second - m[1].first
3652 // Qt: match.capturedEnd(1) - match.capturedStart(1)
3653 int leadingsize = 0;
3655 if (match.lastCapturedIndex() > 0) {
3656 leadingsize = match.capturedEnd(1) - match.capturedStart(1);
3661 leadingsize = m[1].second - m[1].first;
3665 mres.match_prefix = match.capturedEnd(2) - match.capturedStart(2);
3666 mres.match_len = match.capturedEnd(0) - match.capturedEnd(2);
3667 // because of different number of closing at end of string
3668 // we have to 'unify' the length of the post-match.
3669 // Done by ignoring closing parenthesis and linefeeds at string end
3670 int matchend = match.capturedEnd(0);
3671 size_t strsize = qstr.size();
3672 if (!opt.ignoreformat) {
3673 while (mres.match_len > 0) {
3674 QChar c = qstr.at(matchend - 1);
3675 if ((c == '\n') || (c == '}') || (c == '{')) {
3682 while (strsize > (size_t) match.capturedEnd(0)) {
3683 QChar c = qstr.at(strsize-1);
3684 if ((c == '\n') || (c == '}')) {
3691 // LYXERR0(qstr.toStdString());
3692 mres.match2end = strsize - matchend;
3693 mres.pos = match.capturedStart(2);
3695 mres.match_prefix = m[2].second - m[2].first;
3696 mres.match_len = m[0].second - m[2].second;
3697 // ignore closing parenthesis and linefeeds at string end
3698 size_t strend = m[0].second - m[0].first;
3699 int matchend = strend;
3700 size_t strsize = str.size();
3701 if (!opt.ignoreformat) {
3702 while (mres.match_len > 0) {
3703 char c = str.at(matchend - 1);
3704 if ((c == '\n') || (c == '}') || (c == '{')) {
3711 while (strsize > strend) {
3712 if ((str.at(strsize-1) == '}') || (str.at(strsize-1) == '\n')) {
3720 mres.match2end = strsize - matchend;
3721 mres.pos = m[2].first - m[0].first;;
3723 if (mres.match2end < 0)
3725 mres.leadsize = leadingsize;
3727 if (mres.match_len > 0) {
3728 string a0 = match.captured(0).mid(mres.pos + mres.match_prefix, mres.match_len).toStdString();
3729 mres.result.push_back(a0);
3730 for (int i = 3; i <= match.lastCapturedIndex(); i++) {
3731 mres.result.push_back(match.captured(i).toStdString());
3735 if (mres.match_len > 0) {
3736 string a0 = m[0].str().substr(mres.pos + mres.match_prefix, mres.match_len);
3737 mres.result.push_back(a0);
3738 for (size_t i = 3; i < m.size(); i++) {
3739 mres.result.push_back(m[i]);
3748 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
3750 MatchResult mres = findAux(cur, len, at_begin);
3751 int res = mres.match_len;
3753 "res=" << res << ", at_begin=" << at_begin
3754 << ", matchAtStart=" << opt.matchAtStart
3755 << ", inTexted=" << cur.inTexted());
3756 if (opt.matchAtStart) {
3759 else if (mres.match_prefix > 0)
3768 static bool simple_replace(string &t, string from, string to)
3770 regex repl("(\\\\)*(" + from + ")");
3774 for (sregex_iterator it(t.begin(), t.end(), repl), end; it != end; ++it) {
3776 if ((sub.position(2) - sub.position(0)) % 2 == 1)
3778 if (lastpos < (size_t) sub.position(2))
3779 s += t.substr(lastpos, sub.position(2) - lastpos);
3781 lastpos = sub.position(2) + sub.length(2);
3785 else if (lastpos < t.length())
3786 s += t.substr(lastpos, t.length() - lastpos);
3792 string MatchStringAdv::normalize(docstring const & s) const
3795 t = lyx::to_utf8(s);
3796 // Remove \n at begin
3797 while (!t.empty() && t[0] == '\n')
3800 while (!t.empty() && t[t.size() - 1] == '\n')
3801 t = t.substr(0, t.size() - 1);
3803 // Handle all other '\n'
3804 while ((pos = t.find("\n")) != string::npos) {
3805 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
3807 if (isAlnumASCII(t[pos+1])) {
3808 t.replace(pos-2, 3, " ");
3811 t.replace(pos-2, 3, "");
3814 else if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
3815 // '\n' adjacent to non-alpha-numerics, discard
3816 t.replace(pos, 1, "");
3819 // Replace all other \n with spaces
3820 t.replace(pos, 1, " ");
3823 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
3824 // Kornel: Added textsl, textsf, textit, texttt and noun
3825 // + allow to seach for colored text too
3826 LYXERR(Debug::FIND, "Removing stale empty macros from: " << t);
3827 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
3828 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3829 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
3830 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3831 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
3837 docstring stringifyFromCursor(DocIterator const & cur, int len)
3839 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
3840 if (cur.inTexted()) {
3841 Paragraph const & par = cur.paragraph();
3842 // TODO what about searching beyond/across paragraph breaks ?
3843 // TODO Try adding a AS_STR_INSERTS as last arg
3844 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
3845 int(par.size()) : cur.pos() + len;
3846 // OutputParams runparams(&cur.buffer()->params().encoding());
3847 OutputParams runparams(encodings.fromLyXName("utf8"));
3848 runparams.nice = true;
3849 runparams.flavor = Flavor::XeTeX;
3850 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
3851 // No side effect of file copying and image conversion
3852 runparams.dryrun = true;
3853 int option = AS_STR_INSETS | AS_STR_PLAINTEXT;
3854 if (ignoreFormats.getDeleted()) {
3855 option |= AS_STR_SKIPDELETE;
3856 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
3859 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
3861 if (ignoreFormats.getNonContent()) {
3862 runparams.for_searchAdv |= OutputParams::SearchNonOutput;
3864 LYXERR(Debug::FIND, "Stringifying with cur: "
3865 << cur << ", from pos: " << cur.pos() << ", end: " << end);
3866 return par.asString(cur.pos(), end,
3869 } else if (cur.inMathed()) {
3870 CursorSlice cs = cur.top();
3871 MathData md = cs.cell();
3872 MathData::const_iterator it_end =
3873 (( len == -1 || cs.pos() + len > int(md.size()))
3875 : md.begin() + cs.pos() + len );
3877 for (MathData::const_iterator it = md.begin() + cs.pos();
3880 docstring s = asString(md2);
3881 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
3884 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3888 /** Computes the LaTeX export of buf starting from cur and ending len positions
3889 * after cur, if len is positive, or at the paragraph or innermost inset end
3892 docstring latexifyFromCursor(DocIterator const & cur, int len)
3895 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
3896 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
3897 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
3899 Buffer const & buf = *cur.buffer();
3901 odocstringstream ods;
3903 //OutputParams runparams(&buf.params().encoding());
3904 OutputParams runparams(encodings.fromLyXName("utf8"));
3905 runparams.nice = false;
3906 runparams.flavor = Flavor::XeTeX;
3907 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3908 // No side effect of file copying and image conversion
3909 runparams.dryrun = true;
3910 if (ignoreFormats.getDeleted()) {
3911 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
3914 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
3916 if (ignoreFormats.getNonContent()) {
3917 runparams.for_searchAdv |= OutputParams::SearchNonOutput;
3920 if (cur.inTexted()) {
3921 // @TODO what about searching beyond/across paragraph breaks ?
3922 pos_type endpos = cur.paragraph().size();
3923 if (len != -1 && endpos > cur.pos() + len)
3924 endpos = cur.pos() + len;
3925 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
3926 string(), cur.pos(), endpos);
3927 string s = lyx::to_utf8(ods.str());
3928 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
3929 return(lyx::from_utf8(s));
3930 } else if (cur.inMathed()) {
3931 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
3932 for (int s = cur.depth() - 1; s >= 0; --s) {
3933 CursorSlice const & cs = cur[s];
3934 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
3935 TeXMathStream ws(os);
3936 cs.asInsetMath()->asHullInset()->header_write(ws);
3941 CursorSlice const & cs = cur.top();
3942 MathData md = cs.cell();
3943 MathData::const_iterator it_end =
3944 ((len == -1 || cs.pos() + len > int(md.size()))
3946 : md.begin() + cs.pos() + len);
3948 for (MathData::const_iterator it = md.begin() + cs.pos();
3952 ods << asString(md2);
3953 // Retrieve the math environment type, and add '$' or '$]'
3954 // or others (\end{equation}) accordingly
3955 for (int s = cur.depth() - 1; s >= 0; --s) {
3956 CursorSlice const & cs2 = cur[s];
3957 InsetMath * inset = cs2.asInsetMath();
3958 if (inset && inset->asHullInset()) {
3959 TeXMathStream ws(os);
3960 inset->asHullInset()->footer_write(ws);
3964 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3966 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3971 #if defined(ResultsDebug)
3973 static void displayMResult(MatchResult &mres, string from, DocIterator & cur)
3975 LYXERR0( "from:\t\t\t" << from);
3977 if (mres.pos_len > 0) {
3979 status = "FINALSEARCH";
3982 if (mres.match_len > 0) {
3983 if ((mres.match_prefix == 0) && (mres.pos == mres.leadsize))
3984 status = "Good Match";
3986 status = "Matched in";
3989 status = "MissedSearch";
3992 LYXERR0( status << "(" << cur.pos() << " ... " << mres.searched_size + cur.pos() << ") cur.lastpos(" << cur.lastpos() << ")");
3993 if ((mres.leadsize > 0) || (mres.match_len > 0) || (mres.match2end > 0))
3994 LYXERR0( "leadsize(" << mres.leadsize << ") match_len(" << mres.match_len << ") match2end(" << mres.match2end << ")");
3995 if ((mres.pos > 0) || (mres.match_prefix > 0))
3996 LYXERR0( "pos(" << mres.pos << ") match_prefix(" << mres.match_prefix << ")");
3997 for (size_t i = 0; i < mres.result.size(); i++)
3998 LYXERR0( "Match " << i << " = \"" << mres.result[i] << "\"");
4000 #define displayMres(s, txt, cur) displayMResult(s, txt, cur);
4002 #define displayMres(s, txt, cur)
4005 /** Finalize an advanced find operation, advancing the cursor to the innermost
4006 ** position that matches, plus computing the length of the matching text to
4008 ** Return the cur.pos() difference between start and end of found match
4010 MatchResult findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, MatchResult const & expected = MatchResult(-1))
4012 // Search the foremost position that matches (avoids find of entire math
4013 // inset when match at start of it)
4014 DocIterator old_cur(cur.buffer());
4016 static MatchResult fail = MatchResult();
4017 MatchResult max_match;
4018 // If (prefix_len > 0) means that forwarding 1 position will remove the complete entry
4019 // Happens with e.g. hyperlinks
4020 // either one sees "http://www.bla.bla" or nothing
4021 // so the search for "www" gives prefix_len = 7 (== sizeof("http://")
4022 // and although we search for only 3 chars, we find the whole hyperlink inset
4023 bool at_begin = (expected.match_prefix == 0);
4024 if (!match.opt.forward && match.opt.ignoreformat) {
4025 if (expected.pos > 0)
4028 LASSERT(at_begin, /**/);
4029 if (expected.match_len > 0 && at_begin) {
4030 // Search for deepest match
4032 max_match = expected;
4034 size_t d = cur.depth();
4038 if (cur.depth() < d)
4040 if (cur.depth() == d)
4043 while (cur && cur.depth() > lastd) {
4044 lastd = cur.depth();
4045 mres = match(cur, -1, at_begin);
4046 displayMres(mres, "Checking innermost", cur);
4047 if (mres.match_len > 0)
4052 if (mres.match_len < expected.match_len)
4060 // (expected.match_len <= 0)
4061 mres = match(cur); /* match valid only if not searching whole words */
4062 displayMres(mres, "Start with negative match", cur);
4065 if (max_match.match_len <= 0) return fail;
4066 LYXERR(Debug::FIND, "Ok");
4068 // Compute the match length
4070 if (cur.pos() + len > cur.lastpos())
4073 LASSERT(match.use_regexp, /**/);
4076 int maxl = cur.lastpos() - cur.pos();
4077 // Greedy behaviour while matching regexps
4078 while (maxl > minl) {
4080 mres2 = match(cur, len, at_begin);
4081 displayMres(mres2, "Finalize loop", cur);
4082 int actual_match_len = mres2.match_len;
4083 if (actual_match_len >= max_match.match_len) {
4084 // actual_match_len > max_match _can_ happen,
4085 // if the search area splits
4086 // some following word so that the regex
4087 // (e.g. 'r.*r\b' matches 'r' from the middle of the
4089 // This means, the len value is too big
4090 actual_match_len = max_match.match_len;
4092 max_match.match_len = actual_match_len;
4094 if (maxl - minl < 4)
4095 len = (maxl + minl)/2;
4097 len = minl + (maxl - minl + 3)/4;
4100 // (actual_match_len < max_match.match_len)
4102 len = (maxl + minl)/2;
4107 // Search for real start of matched characters
4109 MatchResult actual_match;
4112 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
4113 if (cur.depth() < old_cur.depth()) {
4115 LYXERR(Debug::INFO, "cur.depth() < old_cur.depth(), this should never happen");
4118 if (cur.pos() != old_cur.pos()) {
4119 // OK, forwarded 1 pos in actual inset
4120 actual_match = match(cur, len-1, at_begin);
4121 if (actual_match.match_len == max_match.match_len) {
4122 // Ha, got it! The shorter selection has the same match length
4125 max_match = actual_match;
4128 // OK, the shorter selection matches less chars, revert to previous value
4134 LYXERR(Debug::INFO, "cur.pos() == old_cur.pos(), this should never happen");
4135 actual_match = match(cur, len, at_begin);
4136 if (actual_match.match_len == max_match.match_len) {
4138 max_match = actual_match;
4145 max_match.pos_len = len;
4146 displayMres(max_match, "SEARCH RESULT", cur)
4153 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
4157 bool repeat = false;
4158 DocIterator orig_cur; // to be used if repeat not successful
4159 MatchResult orig_mres;
4160 while (!theApp()->longOperationCancelled() && cur) {
4161 //(void) findAdvForwardInnermost(cur);
4162 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
4163 MatchResult mres = match(cur, -1, false);
4164 string msg = "Starting";
4167 displayMres(mres, msg + " findForwardAdv", cur)
4168 int match_len = mres.match_len;
4169 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
4170 LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
4173 if (match_len <= 0) {
4174 // This should exit nested insets, if any, or otherwise undefine the currsor.
4175 cur.pos() = cur.lastpos();
4176 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
4179 else { // match_len > 0
4180 // Try to find the begin of searched string
4182 int firstInvalid = cur.lastpos() - cur.pos();
4184 int incrmatch = (mres.match_prefix + mres.pos - mres.leadsize + 1)*3/4;
4185 int incrcur = (firstInvalid + 1 )*3/4;
4186 if (incrcur < incrmatch)
4187 increment = incrcur;
4189 increment = incrmatch;
4193 LYXERR(Debug::FIND, "Set increment to " << increment);
4194 while (increment > 0) {
4195 DocIterator old_cur = cur;
4196 if (cur.pos() + increment >= cur.lastpos()) {
4200 cur.pos() = cur.pos() + increment;
4201 MatchResult mres2 = match(cur, -1, false);
4202 displayMres(mres2, "findForwardAdv loop", cur)
4203 switch (interpretMatch(mres, mres2)) {
4204 case MatchResult::newIsTooFar:
4205 // behind the expected match
4206 firstInvalid = increment;
4210 case MatchResult::newIsBetter:
4211 // not reached yet, but cur.pos()+increment is bettert
4213 firstInvalid -= increment;
4214 if (increment > firstInvalid*3/4)
4215 increment = firstInvalid*3/4;
4216 if ((mres2.pos == mres2.leadsize) && (increment >= mres2.match_prefix)) {
4217 if (increment >= mres2.match_prefix)
4218 increment = (mres2.match_prefix+1)*3/4;
4223 // Handle not like MatchResult::newIsTooFar
4224 LYXERR0( "Probably too far: Increment = " << increment << " match_prefix = " << mres.match_prefix);
4226 increment = increment*3/4;
4231 if (mres.match_len > 0) {
4232 if (mres.match_prefix + mres.pos - mres.leadsize > 0) {
4233 // The match seems to indicate some deeper level
4242 // should never be reached.
4246 // LYXERR0("Leaving first loop");
4247 LYXERR(Debug::FIND, "Finalizing 1");
4248 MatchResult found_match = findAdvFinalize(cur, match, mres);
4249 if (found_match.match_len > 0) {
4250 LASSERT(found_match.pos_len > 0, /**/);
4251 match.FillResults(found_match);
4252 return found_match.pos_len;
4255 // try next possible match
4266 /// Find the most backward consecutive match within same paragraph while searching backwards.
4267 MatchResult findMostBackwards(DocIterator & cur, MatchStringAdv const & match, MatchResult &expected)
4269 DocIterator cur_begin = cur;
4270 cur_begin.pos() = 0;
4271 DocIterator tmp_cur = cur;
4272 MatchResult mr = findAdvFinalize(tmp_cur, match, expected);
4273 Inset & inset = cur.inset();
4274 for (; cur != cur_begin; cur.backwardPos()) {
4275 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
4276 DocIterator new_cur = cur;
4277 new_cur.backwardPos();
4278 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
4280 MatchResult new_mr = findAdvFinalize(new_cur, match, expected);
4281 if (new_mr.match_len == mr.match_len)
4285 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
4291 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
4295 // Backup of original position
4296 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
4297 if (cur == cur_begin)
4300 DocIterator cur_orig(cur);
4301 bool pit_changed = false;
4304 MatchResult found_match = match(cur, -1, false);
4306 if (found_match.match_len > 0) {
4308 cur.pos() = cur.lastpos();
4310 cur.pos() = cur_orig.pos();
4311 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
4312 DocIterator cur_prev_iter;
4314 found_match = match(cur);
4315 LYXERR(Debug::FIND, "findBackAdv3: found_match="
4316 << (found_match.match_len > 0) << ", cur: " << cur);
4317 if (found_match.match_len > 0) {
4318 MatchResult found_mr = findMostBackwards(cur, match, found_match);
4319 if (found_mr.pos_len > 0) {
4320 match.FillResults(found_mr);
4321 return found_mr.pos_len;
4325 // Stop if begin of document reached
4326 if (cur == cur_begin)
4328 cur_prev_iter = cur;
4332 if (cur == cur_begin)
4339 } while (!theApp()->longOperationCancelled());
4347 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
4348 DocIterator const & cur, int len)
4350 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
4352 if (!opt.ignoreformat)
4353 return latexifyFromCursor(cur, len);
4355 return stringifyFromCursor(cur, len);
4359 FindAndReplaceOptions::FindAndReplaceOptions(
4360 docstring const & _find_buf_name, bool _casesensitive,
4361 bool _matchword, bool _forward, bool _expandmacros, bool _ignoreformat,
4362 docstring const & _repl_buf_name, bool _keep_case,
4363 SearchScope _scope, SearchRestriction _restr, bool _replace_all)
4364 : find_buf_name(_find_buf_name), casesensitive(_casesensitive), matchword(_matchword),
4365 forward(_forward), expandmacros(_expandmacros), ignoreformat(_ignoreformat),
4366 repl_buf_name(_repl_buf_name), keep_case(_keep_case), scope(_scope), restr(_restr), replace_all(_replace_all)
4374 /** Check if 'len' letters following cursor are all non-lowercase */
4375 static bool allNonLowercase(Cursor const & cur, int len)
4377 pos_type beg_pos = cur.selectionBegin().pos();
4378 pos_type end_pos = cur.selectionBegin().pos() + len;
4379 if (len > cur.lastpos() + 1 - beg_pos) {
4380 LYXERR(Debug::FIND, "This should not happen, more debug needed");
4381 len = cur.lastpos() + 1 - beg_pos;
4382 end_pos = beg_pos + len;
4384 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
4385 if (isLowerCase(cur.paragraph().getChar(pos)))
4391 /** Check if first letter is upper case and second one is lower case */
4392 static bool firstUppercase(Cursor const & cur)
4395 pos_type pos = cur.selectionBegin().pos();
4396 if (pos >= cur.lastpos() - 1) {
4397 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
4400 ch1 = cur.paragraph().getChar(pos);
4401 ch2 = cur.paragraph().getChar(pos + 1);
4402 bool result = isUpperCase(ch1) && isLowerCase(ch2);
4403 LYXERR(Debug::FIND, "firstUppercase(): "
4404 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
4405 << ch2 << "(" << char(ch2) << ")"
4406 << ", result=" << result << ", cur=" << cur);
4411 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
4413 ** \fixme What to do with possible further paragraphs in replace buffer ?
4415 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
4417 ParagraphList::iterator pit = buffer.paragraphs().begin();
4418 LASSERT(!pit->empty(), /**/);
4419 pos_type right = pos_type(1);
4420 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
4421 right = pit->size();
4422 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
4426 static bool replaceMatches(string &t, int maxmatchnum, vector <string> const & replacements)
4428 // Should replace the string "$" + std::to_string(matchnum) with replacement
4429 // if the char '$' is not prefixed with odd number of char '\\'
4430 static regex const rematch("(\\\\)*(\\$\\$([0-9]))");
4434 for (sregex_iterator it(t.begin(), t.end(), rematch), end; it != end; ++it) {
4436 if ((sub.position(2) - sub.position(0)) % 2 == 1)
4438 int num = stoi(sub.str(3), nullptr, 10);
4439 if (num >= maxmatchnum)
4441 if (lastpos < (size_t) sub.position(2))
4442 s += t.substr(lastpos, sub.position(2) - lastpos);
4443 s += replacements[num];
4444 lastpos = sub.position(2) + sub.length(2);
4448 else if (lastpos < t.length())
4449 s += t.substr(lastpos, t.length() - lastpos);
4455 static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
4457 Cursor & cur = bv->cursor();
4458 if (opt.repl_buf_name.empty()
4459 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
4460 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4463 DocIterator sel_beg = cur.selectionBegin();
4464 DocIterator sel_end = cur.selectionEnd();
4465 if (&sel_beg.inset() != &sel_end.inset()
4466 || sel_beg.pit() != sel_end.pit()
4467 || sel_beg.idx() != sel_end.idx())
4469 int sel_len = sel_end.pos() - sel_beg.pos();
4470 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
4471 << ", sel_len: " << sel_len << endl);
4474 LASSERT(sel_len > 0, return 0);
4476 if (!matchAdv(sel_beg, sel_len).match_len)
4479 // Build a copy of the replace buffer, adapted to the KeepCase option
4480 Buffer const & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
4482 repl_buffer_orig.write(oss);
4483 string lyx = oss.str();
4484 if (matchAdv.valid_matches > 0)
4485 replaceMatches(lyx, matchAdv.valid_matches, matchAdv.matches);
4486 Buffer repl_buffer(string(), false);
4487 repl_buffer.setInternal(true);
4488 repl_buffer.setUnnamed(true);
4489 LASSERT(repl_buffer.readString(lyx), return 0);
4490 if (opt.keep_case && sel_len >= 2) {
4491 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
4492 if (cur.inTexted()) {
4493 if (firstUppercase(cur))
4494 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
4495 else if (allNonLowercase(cur, sel_len))
4496 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
4499 cap::cutSelection(cur, false);
4500 if (cur.inTexted()) {
4501 repl_buffer.changeLanguage(
4502 repl_buffer.language(),
4503 cur.getFont().language());
4504 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
4505 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
4506 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
4507 repl_buffer.params().documentClassPtr(),
4508 repl_buffer.params().authors(),
4509 bv->buffer().errorList("Paste"));
4510 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
4511 sel_len = repl_buffer.paragraphs().begin()->size();
4512 } else if (cur.inMathed()) {
4513 odocstringstream ods;
4515 // OutputParams runparams(&repl_buffer.params().encoding());
4516 OutputParams runparams(encodings.fromLyXName("utf8"));
4517 runparams.nice = false;
4518 runparams.flavor = Flavor::XeTeX;
4519 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
4520 runparams.dryrun = true;
4521 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
4522 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
4523 docstring repl_latex = ods.str();
4524 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
4526 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
4527 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
4528 repl_latex = from_utf8(s);
4529 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
4530 MathData ar(cur.buffer());
4531 asArray(repl_latex, ar, Parse::NORMAL);
4533 sel_len = ar.size();
4534 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4536 if (cur.pos() >= sel_len)
4537 cur.pos() -= sel_len;
4540 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4541 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
4542 bv->processUpdateFlags(Update::Force);
4547 /// Perform a FindAdv operation.
4548 bool findAdv(BufferView * bv, FindAndReplaceOptions & opt)
4553 // e.g., when invoking word-findadv from mini-buffer wither with
4554 // wrong options syntax or before ever opening advanced F&R pane
4555 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4559 MatchStringAdv matchAdv(bv->buffer(), opt);
4561 if (!matchAdv.regexIsValid) {
4562 bv->message(lyx::from_utf8(matchAdv.regexError));
4566 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
4568 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
4569 num_replaced += findAdvReplace(bv, opt, matchAdv);
4572 pos_len = findForwardAdv(cur, matchAdv);
4574 pos_len = findBackwardsAdv(cur, matchAdv);
4575 } catch (exception & ex) {
4576 bv->message(from_utf8(ex.what()));
4581 if (num_replaced > 0) {
4582 switch (num_replaced)
4585 bv->message(_("One match has been replaced."));
4588 bv->message(_("Two matches have been replaced."));
4591 bv->message(bformat(_("%1$d matches have been replaced."), num_replaced));
4597 bv->message(_("Match not found."));
4602 if (num_replaced > 0)
4603 bv->message(_("Match has been replaced."));
4605 bv->message(_("Match found."));
4607 if (cur.pos() + pos_len > cur.lastpos()) {
4608 // Prevent crash in bv->putSelectionAt()
4609 // Should never happen, maybe LASSERT() here?
4610 pos_len = cur.lastpos() - cur.pos();
4612 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << pos_len);
4613 bv->putSelectionAt(cur, pos_len, !opt.forward);
4619 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
4621 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
4622 << opt.casesensitive << ' '
4623 << opt.matchword << ' '
4624 << opt.forward << ' '
4625 << opt.expandmacros << ' '
4626 << opt.ignoreformat << ' '
4627 << opt.replace_all << ' '
4628 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
4629 << opt.keep_case << ' '
4630 << int(opt.scope) << ' '
4633 LYXERR(Debug::FIND, "built: " << os.str());
4639 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
4641 // LYXERR(Debug::FIND, "parsing");
4645 while (line != "EOSS") {
4649 if (is.eof()) // Tolerate malformed request
4653 // LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
4654 opt.find_buf_name = from_utf8(s);
4655 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.replace_all;
4656 is.get(); // Waste space before replace string
4659 while (line != "EOSS") {
4663 if (is.eof()) // Tolerate malformed request
4667 // LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
4668 opt.repl_buf_name = from_utf8(s);
4669 is >> opt.keep_case;
4672 opt.scope = FindAndReplaceOptions::SearchScope(i);
4674 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
4677 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
4678 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
4679 << opt.scope << ' ' << opt.restr);