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 void setIgnoreFormat(string const & type, bool value, bool fromUser = true);
113 bool ignoreFamily_ = false;
115 bool ignoreSeries_ = false;
117 bool ignoreShape_ = false;
119 bool ignoreUnderline_ = false;
121 bool ignoreMarkUp_ = false;
123 bool ignoreStrikeOut_ = false;
125 bool ignoreSectioning_ = false;
127 bool ignoreFrontMatter_ = false;
129 bool ignoreColor_ = false;
131 bool ignoreLanguage_ = false;
132 bool userSelectedIgnoreLanguage_ = false;
134 bool ignoreDeleted_ = true;
137 void IgnoreFormats::setIgnoreFormat(string const & type, bool value, bool fromUser)
139 if (type == "color") {
140 ignoreColor_ = value;
142 else if (type == "language") {
144 userSelectedIgnoreLanguage_ = value;
145 ignoreLanguage_ = value;
148 ignoreLanguage_ = (value || userSelectedIgnoreLanguage_);
150 else if (type == "sectioning") {
151 ignoreSectioning_ = value;
152 ignoreFrontMatter_ = value;
154 else if (type == "font") {
155 ignoreSeries_ = value;
156 ignoreShape_ = value;
157 ignoreFamily_ = value;
159 else if (type == "series") {
160 ignoreSeries_ = value;
162 else if (type == "shape") {
163 ignoreShape_ = value;
165 else if (type == "family") {
166 ignoreFamily_ = value;
168 else if (type == "markup") {
169 ignoreMarkUp_ = value;
171 else if (type == "underline") {
172 ignoreUnderline_ = value;
174 else if (type == "strike") {
175 ignoreStrikeOut_ = value;
177 else if (type == "deleted") {
178 ignoreDeleted_ = value;
182 // The global variable that can be changed from outside
183 IgnoreFormats ignoreFormats;
186 void setIgnoreFormat(string const & type, bool value, bool fromUser)
188 ignoreFormats.setIgnoreFormat(type, value, fromUser);
194 bool parse_bool(docstring & howto, bool const defvalue = false)
199 howto = split(howto, var, ' ');
207 MatchString(docstring const & s, bool cs, bool mw)
208 : str(s), case_sens(cs), whole_words(mw)
211 // returns true if the specified string is at the specified position
212 // del specifies whether deleted strings in ct mode will be considered
213 int operator()(Paragraph const & par, pos_type pos, bool del = true) const
215 return par.find(str, case_sens, whole_words, pos, del);
223 // match whole words only
228 int findForward(DocIterator & cur, DocIterator const endcur,
229 MatchString const & match,
230 bool find_del = true, bool onlysel = false)
232 for (; cur; cur.forwardChar()) {
233 if (onlysel && endcur.pit() == cur.pit()
234 && endcur.idx() == cur.idx() && endcur.pos() < cur.pos())
236 if (cur.inTexted()) {
237 int len = match(cur.paragraph(), cur.pos(), find_del);
246 int findBackwards(DocIterator & cur, DocIterator const endcur,
247 MatchString const & match,
248 bool find_del = true, bool onlysel = false)
252 if (onlysel && endcur.pit() == cur.pit()
253 && endcur.idx() == cur.idx() && endcur.pos() > cur.pos())
255 if (cur.inTexted()) {
256 int len = match(cur.paragraph(), cur.pos(), find_del);
265 bool searchAllowed(docstring const & str)
268 frontend::Alert::error(_("Search error"), _("Search string is empty"));
277 bool findOne(BufferView * bv, docstring const & searchstr,
278 bool case_sens, bool whole, bool forward,
279 bool find_del, bool check_wrap, bool auto_wrap,
280 bool instant, bool onlysel)
282 // Clean up previous selections with empty searchstr on instant
283 if (searchstr.empty() && instant) {
284 if (bv->cursor().selection()) {
285 bv->setCursor(bv->cursor().selectionBegin());
286 bv->clearSelection();
291 if (!searchAllowed(searchstr))
294 DocIterator const endcur = forward ? bv->cursor().selectionEnd() : bv->cursor().selectionBegin();
296 if (onlysel && bv->cursor().selection()) {
297 docstring const matchstring = bv->cursor().selectionAsString(false);
298 docstring const lcmatchsting = support::lowercase(matchstring);
299 if (matchstring == searchstr || (!case_sens && lcmatchsting == lowercase(searchstr))) {
300 docstring q = _("The search string matches the selection, and search is limited to selection.\n"
301 "Continue search outside?");
302 int search_answer = frontend::Alert::prompt(_("Search outside selection?"),
303 q, 0, 1, _("&Yes"), _("&No"));
304 if (search_answer == 0) {
305 bv->clearSelection();
306 if (findOne(bv, searchstr, case_sens, whole, forward,
307 find_del, check_wrap, auto_wrap, false, false))
314 DocIterator cur = forward
315 ? ((instant || onlysel) ? bv->cursor().selectionBegin() : bv->cursor().selectionEnd())
316 : ((instant || onlysel) ? bv->cursor().selectionEnd() : bv->cursor().selectionBegin());
318 MatchString const match(searchstr, case_sens, whole);
320 int match_len = forward
321 ? findForward(cur, endcur, match, find_del, onlysel)
322 : findBackwards(cur, endcur, match, find_del, onlysel);
325 bv->putSelectionAt(cur, match_len, !forward);
327 docstring q = _("The search string was not found within the selection.\n"
328 "Continue search outside?");
329 int search_answer = frontend::Alert::prompt(_("Search outside selection?"),
330 q, 0, 1, _("&Yes"), _("&No"));
331 if (search_answer == 0) {
332 bv->clearSelection();
333 if (findOne(bv, searchstr, case_sens, whole, forward,
334 find_del, check_wrap, auto_wrap, false, false))
339 else if (check_wrap) {
340 DocIterator cur_orig(bv->cursor());
344 q = _("End of file reached while searching forward.\n"
345 "Continue searching from the beginning?");
347 q = _("Beginning of file reached while searching backward.\n"
348 "Continue searching from the end?");
349 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
350 q, 0, 1, _("&Yes"), _("&No"));
351 auto_wrap = wrap_answer == 0;
355 bv->cursor().clear();
356 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
358 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
359 bv->cursor().backwardPos();
361 bv->clearSelection();
362 if (findOne(bv, searchstr, case_sens, whole, forward,
363 find_del, false, false, false, false))
366 bv->cursor().setCursor(cur_orig);
370 return match_len > 0;
376 int replaceAll(BufferView * bv,
377 docstring const & searchstr, docstring const & replacestr,
378 bool case_sens, bool whole, bool onlysel)
380 Buffer & buf = bv->buffer();
382 if (!searchAllowed(searchstr) || buf.isReadonly())
385 DocIterator startcur = bv->cursor().selectionBegin();
386 DocIterator endcur = bv->cursor().selectionEnd();
387 bool const had_selection = bv->cursor().selection();
389 MatchString const match(searchstr, case_sens, whole);
392 int const rsize = replacestr.size();
393 int const ssize = searchstr.size();
396 cur.setCursor(doc_iterator_begin(&buf));
397 int match_len = findForward(cur, endcur, match, false, onlysel);
398 while (match_len > 0) {
399 // Backup current cursor position and font.
400 pos_type const pos = cur.pos();
401 Font const font = cur.paragraph().getFontSettings(buf.params(), pos);
403 int ct_deleted_text = ssize -
404 cur.paragraph().eraseChars(pos, pos + match_len,
405 buf.params().track_changes);
406 cur.paragraph().insert(pos, replacestr, font,
407 Change(buf.params().track_changes
409 : Change::UNCHANGED));
410 for (int i = 0; i < rsize + ct_deleted_text
411 && cur.pos() < cur.lastpos(); ++i)
413 if (onlysel && cur.pit() == endcur.pit() && cur.idx() == endcur.idx()) {
414 // Adjust end of selection for replace-all in selection
416 int const offset = rsize - ssize;
417 for (int i = 0; i < offset + ct_deleted_text
418 && endcur.pos() < endcur.lastpos(); ++i)
421 int const offset = ssize - rsize;
422 for (int i = 0; i < offset && endcur.pos() > 0; ++i)
423 endcur.backwardPos();
424 for (int i = 0; i < ct_deleted_text
425 && endcur.pos() < endcur.lastpos(); ++i)
430 match_len = findForward(cur, endcur, match, false, onlysel);
433 bv->putSelectionAt(doc_iterator_begin(&buf), 0, false);
435 startcur.fixIfBroken();
436 bv->setCursor(startcur);
438 // Reset selection, accounting for changes in selection
440 endcur.fixIfBroken();
441 bv->cursor().resetAnchor();
442 bv->setCursorSelectionTo(endcur);
449 // the idea here is that we are going to replace the string that
450 // is selected IF it is the search string.
451 // if there is a selection, but it is not the search string, then
452 // we basically ignore it. (FIXME We ought to replace only within
454 // if there is no selection, then:
455 // (i) if some search string has been provided, then we find it.
456 // (think of how the dialog works when you hit "replace" the
458 // (ii) if no search string has been provided, then we treat the
459 // word the cursor is in as the search string. (why? i have no
460 // idea.) but this only works in text?
462 // returns the number of replacements made (one, if any) and
463 // whether anything at all was done.
464 pair<bool, int> replaceOne(BufferView * bv, docstring searchstr,
465 docstring const & replacestr, bool case_sens,
466 bool whole, bool forward, bool findnext, bool wrap,
469 Cursor & cur = bv->cursor();
470 if (!cur.selection() || onlysel) {
471 // no selection, non-empty search string: find it
472 if (!searchstr.empty()) {
473 bool const found = findOne(bv, searchstr, case_sens, whole,
474 forward, true, findnext, wrap, false, onlysel);
475 return make_pair(found, 0);
477 // empty search string
480 return make_pair(false, 0);
481 // select current word and treat it as the search string.
482 // This causes a minor bug as undo will restore this selection,
483 // which the user did not create (#8986).
484 cur.innerText()->selectWord(cur, WHOLE_WORD);
485 searchstr = cur.selectionAsString(false, true);
488 // if we still don't have a search string, report the error
490 if (!searchAllowed(searchstr))
491 return make_pair(false, 0);
493 bool have_selection = cur.selection();
494 docstring const selected = cur.selectionAsString(false, true);
497 ? searchstr == selected
498 : compare_no_case(searchstr, selected) == 0;
500 // no selection or current selection is not search word:
501 // just find the search word
502 if (!have_selection || !match) {
503 bool const found = findOne(bv, searchstr, case_sens, whole, forward,
504 true, findnext, wrap, false, onlysel);
505 return make_pair(found, 0);
508 // we're now actually ready to replace. if the buffer is
509 // read-only, we can't, though.
510 if (bv->buffer().isReadonly())
511 return make_pair(false, 0);
513 cap::replaceSelectionWithString(cur, replacestr);
515 cur.pos() += replacestr.length();
516 LASSERT(cur.pos() <= cur.lastpos(),
517 cur.pos() = cur.lastpos());
520 findOne(bv, searchstr, case_sens, whole,
521 forward, false, findnext, wrap, false, onlysel);
523 return make_pair(true, 1);
529 docstring const find2string(docstring const & search,
530 bool casesensitive, bool matchword,
531 bool forward, bool wrap, bool instant,
536 << int(casesensitive) << ' '
537 << int(matchword) << ' '
538 << int(forward) << ' '
540 << int(instant) << ' '
546 docstring const replace2string(docstring const & replace,
547 docstring const & search,
548 bool casesensitive, bool matchword,
549 bool all, bool forward, bool findnext,
550 bool wrap, bool onlysel)
553 ss << replace << '\n'
555 << int(casesensitive) << ' '
556 << int(matchword) << ' '
558 << int(forward) << ' '
559 << int(findnext) << ' '
566 docstring const string2find(docstring const & argument,
574 // data is of the form
576 // <casesensitive> <matchword> <forward> <wrap> <onlysel>"
578 docstring howto = split(argument, search, '\n');
580 casesensitive = parse_bool(howto);
581 matchword = parse_bool(howto);
582 forward = parse_bool(howto, true);
583 wrap = parse_bool(howto);
584 instant = parse_bool(howto);
585 onlysel = parse_bool(howto);
591 bool lyxfind(BufferView * bv, FuncRequest const & ev)
593 if (!bv || ev.action() != LFUN_WORD_FIND)
596 //lyxerr << "find called, cmd: " << ev << endl;
604 docstring search = string2find(ev.argument(), casesensitive,
605 matchword, forward, wrap, instant, onlysel);
607 return findOne(bv, search, casesensitive, matchword, forward,
608 false, true, wrap, instant, onlysel);
612 bool lyxreplace(BufferView * bv, FuncRequest const & ev)
614 if (!bv || ev.action() != LFUN_WORD_REPLACE)
617 // data is of the form
620 // <casesensitive> <matchword> <all> <forward> <findnext> <wrap> <onlysel>"
623 docstring howto = split(ev.argument(), rplc, '\n');
624 howto = split(howto, search, '\n');
626 bool casesensitive = parse_bool(howto);
627 bool matchword = parse_bool(howto);
628 bool all = parse_bool(howto);
629 bool forward = parse_bool(howto, true);
630 bool findnext = parse_bool(howto, true);
631 bool wrap = parse_bool(howto);
632 bool onlysel = parse_bool(howto);
634 if (!bv->cursor().selection())
635 // only selection only makes sense with selection
640 int replace_count = 0;
642 replace_count = replaceAll(bv, search, rplc, casesensitive,
644 update = replace_count > 0;
647 replaceOne(bv, search, rplc, casesensitive, matchword,
648 forward, findnext, wrap, onlysel);
650 replace_count = rv.second;
653 Buffer const & buf = bv->buffer();
655 // emit message signal.
657 buf.message(_("String not found in selection."));
659 buf.message(_("String not found."));
661 if (replace_count == 0) {
662 buf.message(_("String found."));
663 } else if (replace_count == 1) {
664 buf.message(_("String has been replaced."));
666 docstring const str = onlysel
667 ? bformat(_("%1$d strings have been replaced in the selection."), replace_count)
668 : bformat(_("%1$d strings have been replaced."), replace_count);
676 bool findNextChange(BufferView * bv, Cursor & cur, bool const check_wrap)
678 for (; cur; cur.forwardPos())
679 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
683 DocIterator cur_orig(bv->cursor());
684 docstring q = _("End of file reached while searching forward.\n"
685 "Continue searching from the beginning?");
686 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
687 q, 0, 1, _("&Yes"), _("&No"));
688 if (wrap_answer == 0) {
689 bv->cursor().clear();
690 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
691 bv->clearSelection();
692 cur.setCursor(bv->cursor().selectionBegin());
693 if (findNextChange(bv, cur, false))
696 bv->cursor().setCursor(cur_orig);
703 bool findPreviousChange(BufferView * bv, Cursor & cur, bool const check_wrap)
705 for (cur.backwardPos(); cur; cur.backwardPos()) {
706 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
711 DocIterator cur_orig(bv->cursor());
712 docstring q = _("Beginning of file reached while searching backward.\n"
713 "Continue searching from the end?");
714 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
715 q, 0, 1, _("&Yes"), _("&No"));
716 if (wrap_answer == 0) {
717 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
718 bv->cursor().backwardPos();
719 bv->clearSelection();
720 cur.setCursor(bv->cursor().selectionBegin());
721 if (findPreviousChange(bv, cur, false))
724 bv->cursor().setCursor(cur_orig);
731 bool selectChange(Cursor & cur, bool forward)
733 if (!cur.inTexted() || !cur.paragraph().isChanged(cur.pos()))
735 Change ch = cur.paragraph().lookupChange(cur.pos());
737 CursorSlice tip1 = cur.top();
738 for (; tip1.pit() < tip1.lastpit() || tip1.pos() < tip1.lastpos(); tip1.forwardPos()) {
739 Change ch2 = tip1.paragraph().lookupChange(tip1.pos());
740 if (!ch2.isSimilarTo(ch))
743 CursorSlice tip2 = cur.top();
744 for (; tip2.pit() > 0 || tip2.pos() > 0;) {
746 Change ch2 = tip2.paragraph().lookupChange(tip2.pos());
747 if (!ch2.isSimilarTo(ch)) {
748 // take a step forward to correctly set the selection
756 cur.bv().mouseSetCursor(cur, false);
758 cur.bv().mouseSetCursor(cur, true);
766 bool findChange(BufferView * bv, bool forward)
769 cur.setCursor(forward ? bv->cursor().selectionEnd()
770 : bv->cursor().selectionBegin());
771 forward ? findNextChange(bv, cur, true) : findPreviousChange(bv, cur, true);
772 return selectChange(cur, forward);
777 bool findNextChange(BufferView * bv)
779 return findChange(bv, true);
783 bool findPreviousChange(BufferView * bv)
785 return findChange(bv, false);
792 typedef vector<pair<string, string> > Escapes;
794 string string2regex(string in)
796 static std::regex specialChars { R"([-[\]{}()*+?.,\^$|#\s\$\\])" };
797 string temp = std::regex_replace(in, specialChars, R"(\$&)" );
802 while (fl_pos < temp.size()) {
803 fl_pos = temp.find("\\\\foreignlanguage", lastpos + offset);
804 if (fl_pos == string::npos)
807 temp2 += temp.substr(lastpos, fl_pos - lastpos);
813 if (lastpos < temp.size()) {
814 temp2 += temp.substr(lastpos, temp.size() - lastpos);
819 string correctRegex(string t, bool withformat)
821 /* Convert \backslash => \
822 * and \{, \}, \[, \] => {, }, [, ]
825 regex wordre("(\\\\)*(\\\\((backslash|mathcircumflex) ?|[\\[\\]\\{\\}]))");
828 bool backslashed = false;
829 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
832 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
836 if (sub.str(4) == "backslash") {
839 // transforms '\backslash \{' into '\{'
841 string next = t.substr(sub.position(2) + sub.str(2).length(), 2);
842 if ((next == "\\{") || (next == "\\}")) {
848 else if (sub.str(4) == "mathcircumflex")
850 else if (backslashed) {
852 if (withformat && (sub.str(3) == "{"))
853 replace = accents["braceleft"];
854 else if (withformat && (sub.str(3) == "}"))
855 replace = accents["braceright"];
857 // else part should not exist
862 replace = sub.str(3);
864 if (lastpos < (size_t) sub.position(2))
865 s += t.substr(lastpos, sub.position(2) - lastpos);
867 lastpos = sub.position(2) + sub.length(2);
871 else if (lastpos < t.length())
872 s += t.substr(lastpos, t.length() - lastpos);
876 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
877 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
878 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
879 string escape_for_regex(string s, bool withformat)
883 while (lastpos < s.size()) {
884 size_t regex_pos = s.find("\\regexp{", lastpos);
885 if (regex_pos == string::npos) {
886 regex_pos = s.size();
888 if (regex_pos > lastpos) {
889 result += string2regex(s.substr(lastpos, regex_pos-lastpos));
891 if (lastpos == s.size())
894 size_t end_pos = s.find("\\endregexp{}}", regex_pos + 8);
895 result += correctRegex(s.substr(regex_pos + 8, end_pos -(regex_pos + 8)), withformat);
896 lastpos = end_pos + 13;
902 /// Wrapper for lyx::regex_replace with simpler interface
903 bool regex_replace(string const & s, string & t, string const & searchstr,
904 string const & replacestr)
906 regex e(searchstr, regex_constants::ECMAScript);
908 ostream_iterator<char, char> it(oss);
909 regex_replace(it, s.begin(), s.end(), e, replacestr);
910 // tolerate t and s be references to the same variable
911 bool rv = (s != oss.str());
930 vector <string> result = vector <string>();
931 MatchResult(int len = 0): match_len(len),match_prefix(0),match2end(0), pos(0),leadsize(0),pos_len(-1),searched_size(0) {};
934 static MatchResult::range interpretMatch(MatchResult &oldres, MatchResult &newres)
936 if (newres.match2end < oldres.match2end)
937 return MatchResult::newIsTooFar;
938 if (newres.match_len < oldres.match_len)
939 return MatchResult::newIsTooFar;
941 if (newres.match_len == oldres.match_len) {
942 if (newres.match2end == oldres.match2end)
943 return MatchResult::newIsBetter;
945 return MatchResult::newIsInvalid;
948 /** The class performing a match between a position in the document and the FindAdvOptions.
951 class MatchStringAdv {
953 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt);
955 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
956 ** constructor as opt.search, under the opt.* options settings.
959 ** If set, then match is searched only against beginning of text starting at cur.
960 ** If unset, then match is searched anywhere in text starting at cur.
963 ** The length of the matching text, or zero if no match was found.
965 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
974 /// first buffer on which search was started
975 lyx::Buffer * const p_first_buf;
977 FindAndReplaceOptions const & opt;
980 /// Auxiliary find method (does not account for opt.matchword)
981 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
982 void CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string = "");
984 /** Normalize a stringified or latexified LyX paragraph.
988 ** <li>if search is not casesensitive, then lowercase the string;
989 ** <li>remove any newline at begin or end of the string;
990 ** <li>replace any newline in the middle of the string with a simple space;
991 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
994 ** @todo Normalization should also expand macros, if the corresponding
995 ** search option was checked.
997 string normalize(docstring const & s) const;
998 // normalized string to search
999 string par_as_string;
1000 // regular expression to use for searching
1001 // regexp2 is same as regexp, but prefixed with a ".*?"
1003 QRegularExpression regexp;
1004 QRegularExpression regexp2;
1009 // leading format material as string
1010 string lead_as_string;
1011 // par_as_string after removal of lead_as_string
1012 string par_as_string_nolead;
1013 // unmatched open braces in the search string/regexp
1015 // number of (.*?) subexpressions added at end of search regexp for closing
1016 // environments, math mode, styles, etc...
1017 int close_wildcards;
1019 // Are we searching with regular expressions ?
1021 static int valid_matches;
1022 static vector <string> matches;
1023 void FillResults(MatchResult &found_mr);
1026 int MatchStringAdv::valid_matches = 0;
1027 vector <string> MatchStringAdv::matches = vector <string>(10);
1029 void MatchStringAdv::FillResults(MatchResult &found_mr)
1031 if (found_mr.match_len > 0) {
1032 valid_matches = found_mr.result.size();
1033 for (size_t i = 0; i < found_mr.result.size(); i++)
1034 matches[i] = found_mr.result[i];
1040 static docstring buffer_to_latex(Buffer & buffer)
1042 //OutputParams runparams(&buffer.params().encoding());
1043 OutputParams runparams(encodings.fromLyXName("utf8"));
1044 odocstringstream ods;
1046 runparams.nice = true;
1047 runparams.flavor = Flavor::XeTeX;
1048 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
1049 // No side effect of file copying and image conversion
1050 runparams.dryrun = true;
1051 if (ignoreFormats.getDeleted())
1052 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
1054 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
1055 pit_type const endpit = buffer.paragraphs().size();
1056 for (pit_type pit = 0; pit != endpit; ++pit) {
1057 TeXOnePar(buffer, buffer.text(), pit, os, runparams);
1058 LYXERR(Debug::FIND, "searchString up to here: " << ods.str());
1064 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
1067 if (!opt.ignoreformat) {
1068 str = buffer_to_latex(buffer);
1070 // OutputParams runparams(&buffer.params().encoding());
1071 OutputParams runparams(encodings.fromLyXName("utf8"));
1072 runparams.nice = true;
1073 runparams.flavor = Flavor::XeTeX;
1074 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
1075 runparams.dryrun = true;
1076 int option = AS_STR_INSETS |AS_STR_PLAINTEXT;
1077 if (ignoreFormats.getDeleted()) {
1078 option |= AS_STR_SKIPDELETE;
1079 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
1082 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
1084 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
1085 Paragraph const & par = buffer.paragraphs().at(pit);
1086 LYXERR(Debug::FIND, "Adding to search string: '"
1087 << par.asString(pos_type(0), par.size(),
1091 str += par.asString(pos_type(0), par.size(),
1095 // Even in ignore-format we have to remove "\text{}, \lyxmathsym{}" parts
1096 string t = to_utf8(str);
1097 while (regex_replace(t, t, "\\\\(text|lyxmathsym|ensuremath)\\{([^\\}]*)\\}", "$2"));
1104 /// Return separation pos between the leading material and the rest
1105 static size_t identifyLeading(string const & s)
1108 // @TODO Support \item[text]
1109 // Kornel: Added textsl, textsf, textit, texttt and noun
1110 // + allow to search for colored text too
1111 while (regex_replace(t, t, "^\\\\(("
1112 "(author|title|subtitle|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|"
1113 "lyxaddress|lyxrightaddress|"
1114 "footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1115 "emph|noun|minisec|text(bf|md|sl|sf|it|tt))|"
1116 "((textcolor|foreignlanguage|latexenvironment)\\{[a-z]+\\*?\\})|"
1117 "(u|uu)line|(s|x)out|uwave)|((sub)?(((sub)?section)|paragraph)|part|chapter)\\*?)\\{", "")
1118 || regex_replace(t, t, "^\\$", "")
1119 || regex_replace(t, t, "^\\\\\\[", "")
1120 || regex_replace(t, t, "^ ?\\\\item\\{[a-z]+\\}", "")
1121 || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\}", ""))
1123 LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
1128 * Given a latexified string, retrieve some handled features
1129 * The features of the regex will later be compared with the features
1130 * of the searched text. If the regex features are not a
1131 * subset of the analized, then, in not format ignoring search
1132 * we can early stop the search in the relevant inset.
1134 typedef map<string, bool> Features;
1136 static Features identifyFeatures(string const & s)
1138 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
1139 static regex const valid("^("
1141 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1142 "emph|noun|text(bf|md|sl|sf|it|tt)|"
1143 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
1144 "(u|uu)line|(s|x)out|uwave|"
1145 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
1146 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
1151 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1154 if (sub.str(1).compare("regexp") == 0) {
1158 string token = sub.str(1);
1160 if (regex_match(token, sub2, valid)) {
1168 if (sub.str(1).compare("endregexp") == 0) {
1178 * defines values features of a key "\\[a-z]+{"
1183 /* Char type with content discarded
1184 * like \hspace{1cm} */
1186 /* Char, like \backslash */
1188 /* replace starting backslash with '#' */
1190 /* \part, \section*, ... */
1192 /* title, author etc */
1194 /* \foreignlanguage{ngerman}, ... */
1197 * to discard language in content */
1200 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1202 /* fonts, colors, markups, ... */
1204 /* footnotesize, ... large, ...
1205 * Ignore all of them */
1208 /* inputencoding, ...
1209 * Discard also content, because they do not help in search */
1212 * like remove, but also all arguments */
1214 /* item, listitem */
1216 /* tex, latex, ... like isChar */
1218 /* like \lettrine[lines=5]{}{} */
1220 // like isStandard, but always remove head
1222 /* End of arguments marker for lettrine,
1223 * so that they can be ignored */
1226 KeyInfo() = default;
1227 KeyInfo(KeyType type, int parcount, bool disable)
1229 parenthesiscount(parcount),
1230 disabled(disable) {}
1231 KeyType keytype = invalid;
1233 int _tokensize = -1;
1234 int _tokenstart = -1;
1235 int _dataStart = -1;
1237 int parenthesiscount = 1;
1238 bool disabled = false;
1239 bool used = false; /* by pattern */
1244 Border(int l=0, int u=0) : low(l), upper(u) {};
1249 #define MAXOPENED 30
1251 bool isPatternString_;
1253 explicit Intervall(bool isPattern, string const & p) :
1254 isPatternString_(isPattern), par(p), ignoreidx(-1), actualdeptindex(0),
1255 hasTitle(false), langcount(0)
1263 static vector<Border> borders;
1264 int depts[MAXOPENED];
1265 int closes[MAXOPENED];
1266 int actualdeptindex;
1267 int previousNotIgnored(int) const;
1268 int nextNotIgnored(int) const;
1269 void handleOpenP(int i);
1270 void handleCloseP(int i, bool closingAllowed);
1271 void resetOpenedP(int openPos);
1272 void addIntervall(int upper);
1273 void addIntervall(int low, int upper); /* if explicit */
1274 void removeAccents();
1275 void setForDefaultLang(KeyInfo const & defLang) const;
1276 int findclosing(int start, int end, char up, char down, int repeat);
1277 void handleParentheses(int lastpos, bool closingAllowed);
1279 int langcount; // Number of disabled language specs up to current position in actual interval
1280 int isOpeningPar(int pos) const;
1282 void output(ostringstream &os, int lastpos);
1283 // string show(int lastpos);
1286 vector<Border> Intervall::borders = vector<Border>(30);
1288 int Intervall::isOpeningPar(int pos) const
1290 if ((pos < 0) || (size_t(pos) >= par.size()))
1292 if (par[pos] != '{')
1294 if (size_t(pos) + 2 >= par.size())
1296 if (par[pos+2] != '}')
1298 if (par[pos+1] == '[' || par[pos+1] == ']')
1303 void Intervall::setForDefaultLang(KeyInfo const & defLang) const
1305 // Enable the use of first token again
1306 if (ignoreidx >= 0) {
1307 int value = defLang._tokenstart + defLang._tokensize;
1313 if (borders[borderidx].low < value)
1314 borders[borderidx].low = value;
1315 if (borders[borderidx].upper < value)
1316 borders[borderidx].upper = value;
1321 static void checkDepthIndex(int val)
1323 static int maxdepthidx = MAXOPENED-2;
1324 static int lastmaxdepth = 0;
1325 if (val > lastmaxdepth) {
1326 LYXERR(Debug::INFO, "Depth reached " << val);
1329 if (val > maxdepthidx) {
1331 LYXERR(Debug::INFO, "maxdepthidx now " << val);
1336 // Not needed, because borders are now dynamically expanded
1337 static void checkIgnoreIdx(int val)
1339 static int lastmaxignore = -1;
1340 if ((lastmaxignore < val) && (size_t(val+1) >= borders.size())) {
1341 LYXERR(Debug::INFO, "IgnoreIdx reached " << val);
1342 lastmaxignore = val;
1348 * Expand the region of ignored parts of the input latex string
1349 * The region is only relevant in output()
1351 void Intervall::addIntervall(int low, int upper)
1354 if (low == upper) return;
1355 for (idx = ignoreidx+1; idx > 0; --idx) {
1356 if (low > borders[idx-1].upper) {
1360 Border br(low, upper);
1361 if (idx > ignoreidx) {
1362 if (borders.size() <= size_t(idx)) {
1363 borders.push_back(br);
1369 // checkIgnoreIdx(ignoreidx);
1373 // Expand only if one of the new bound is inside the interwall
1374 // We know here that br.low > borders[idx-1].upper
1375 if (br.upper < borders[idx].low) {
1376 // We have to insert at this pos
1377 if (size_t(ignoreidx+1) >= borders.size()) {
1378 borders.push_back(borders[ignoreidx]);
1381 borders[ignoreidx+1] = borders[ignoreidx];
1383 for (int i = ignoreidx; i > idx; --i) {
1384 borders[i] = borders[i-1];
1388 // checkIgnoreIdx(ignoreidx);
1391 // Here we know, that we are overlapping
1392 if (br.low > borders[idx].low)
1393 br.low = borders[idx].low;
1394 // check what has to be concatenated
1396 for (int i = idx; i <= ignoreidx; i++) {
1397 if (br.upper >= borders[i].low) {
1399 if (br.upper < borders[i].upper)
1400 br.upper = borders[i].upper;
1406 // count should be >= 1 here
1409 for (int i = idx + count; i <= ignoreidx; i++) {
1410 borders[i-count+1] = borders[i];
1412 ignoreidx -= count - 1;
1418 static void buildaccent(string n, string param, string values)
1422 const char delim = '|';
1423 while (getline(s, name, delim)) {
1425 for (char c : param) {
1426 string key = name + "{" + c + "}";
1427 // get the corresponding utf8-value
1428 if ((values[start] & 0xc0) != 0xc0) {
1429 // should not happen, utf8 encoding starts at least with 11xxxxxx
1430 // but value for '\dot{i}' is 'i', which is ascii
1431 if ((values[start] & 0x80) == 0) {
1433 accents[key] = values.substr(start, 1);
1434 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1439 for (int j = 1; ;j++) {
1440 if (start + j >= values.size()) {
1441 accents[key] = values.substr(start, j);
1442 start = values.size() - 1;
1445 else if ((values[start+j] & 0xc0) != 0x80) {
1446 // This is the first byte of following utf8 char
1447 accents[key] = values.substr(start, j);
1449 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1458 static string getutf8(unsigned uchar)
1461 string ret = string();
1463 if (uchar <= 0x7f) {
1464 c[maxc-1] = uchar & 0x7f;
1467 unsigned char rest = 0x40;
1468 unsigned char first = 0x80;
1470 for (int i = start; i >=0; --i) {
1472 c[i] = first + uchar;
1475 c[i] = 0x80 | (uchar & 0x3f);
1482 for (int i = 0; i < maxc; i++) {
1483 if (c[i] == 0) continue;
1489 static void addAccents(string latex_in, string unicode_out)
1491 latex_in = latex_in.substr(1);
1492 AccentsIterator it_ac = accents.find(latex_in);
1493 if (it_ac == accents.end()) {
1494 accents[latex_in] = unicode_out;
1497 LYXERR0("Key " << latex_in << " already set");
1501 void static fillMissingUnicodesymbols()
1503 addAccents("\\textyen", getutf8(0x00a5));
1504 addAccents("\\yen", getutf8(0x00a5));
1505 addAccents("\\textsection", getutf8(0x00a7));
1506 addAccents("\\mathsection", getutf8(0x00a7));
1507 addAccents("\\textlnot", getutf8(0x00ac));
1508 addAccents("\\neg", getutf8(0x00ac));
1509 addAccents("\\textpm", getutf8(0x00b1));
1510 addAccents("\\pm", getutf8(0x00b1));
1511 addAccents("\\textparagraph", getutf8(0x00b6));
1512 addAccents("\\mathparagraph", getutf8(0x00b6));
1513 addAccents("\\textperiodcentered", getutf8(0x00b7));
1514 addAccents("\\texttimes", getutf8(0x00d7));
1515 addAccents("\\times", getutf8(0x00d7));
1516 addAccents("\\dh", getutf8(0x00f0));
1517 addAccents("\\eth", getutf8(0x00f0));
1518 addAccents("\\textdiv", getutf8(0x00f7));
1519 addAccents("\\div", getutf8(0x00f7));
1520 addAccents("\\o", getutf8(0x00f8));
1521 addAccents("\\j", getutf8(0x0237));
1522 addAccents("\\textalpha", getutf8(0x03b1));
1523 addAccents("\\alpha", getutf8(0x03b1));
1524 addAccents("\\textbeta", getutf8(0x03b2));
1525 addAccents("\\beta", getutf8(0x03b2));
1526 addAccents("\\textgamma", getutf8(0x03b3));
1527 addAccents("\\gamma", getutf8(0x03b3));
1528 addAccents("\\textdelta", getutf8(0x03b4));
1529 addAccents("\\delta", getutf8(0x03b4));
1530 addAccents("\\textepsilon", getutf8(0x03b5));
1531 addAccents("\\varepsilon", getutf8(0x03b5));
1532 addAccents("\\textzeta", getutf8(0x03b6));
1533 addAccents("\\zeta", getutf8(0x03b6));
1534 addAccents("\\texteta", getutf8(0x03b7));
1535 addAccents("\\eta", getutf8(0x03b7));
1536 addAccents("\\texttheta", getutf8(0x03b8));
1537 addAccents("\\theta", getutf8(0x03b8));
1538 addAccents("\\textiota", getutf8(0x03b9));
1539 addAccents("\\iota", getutf8(0x03b9));
1540 addAccents("\\textkappa", getutf8(0x03ba));
1541 addAccents("\\kappa", getutf8(0x03ba));
1542 addAccents("\\textlambda", getutf8(0x03bb));
1543 addAccents("\\lambda", getutf8(0x03bb));
1544 addAccents("\\textmu", getutf8(0x03bc));
1545 addAccents("\\mu", getutf8(0x03bc));
1546 addAccents("\\textnu", getutf8(0x03bd));
1547 addAccents("\\nu", getutf8(0x03bd));
1548 addAccents("\\textxi", getutf8(0x03be));
1549 addAccents("\\xi", getutf8(0x03be));
1550 addAccents("\\textpi", getutf8(0x03c0));
1551 addAccents("\\pi", getutf8(0x03c0));
1552 addAccents("\\textrho", getutf8(0x03c1));
1553 addAccents("\\rho", getutf8(0x03c1));
1554 addAccents("\\textfinalsigma", getutf8(0x03c2));
1555 addAccents("\\varsigma", getutf8(0x03c2));
1556 addAccents("\\textsigma", getutf8(0x03c3));
1557 addAccents("\\sigma", getutf8(0x03c3));
1558 addAccents("\\texttau", getutf8(0x03c4));
1559 addAccents("\\tau", getutf8(0x03c4));
1560 addAccents("\\textupsilon", getutf8(0x03c5));
1561 addAccents("\\upsilon", getutf8(0x03c5));
1562 addAccents("\\textphi", getutf8(0x03c6));
1563 addAccents("\\varphi", getutf8(0x03c6));
1564 addAccents("\\textchi", getutf8(0x03c7));
1565 addAccents("\\chi", getutf8(0x03c7));
1566 addAccents("\\textpsi", getutf8(0x03c8));
1567 addAccents("\\psi", getutf8(0x03c8));
1568 addAccents("\\textomega", getutf8(0x03c9));
1569 addAccents("\\omega", getutf8(0x03c9));
1570 addAccents("\\textdigamma", getutf8(0x03dd));
1571 addAccents("\\digamma", getutf8(0x03dd));
1572 addAccents("\\hebalef", getutf8(0x05d0));
1573 addAccents("\\aleph", getutf8(0x05d0));
1574 addAccents("\\hebbet", getutf8(0x05d1));
1575 addAccents("\\beth", getutf8(0x05d1));
1576 addAccents("\\hebgimel", getutf8(0x05d2));
1577 addAccents("\\gimel", getutf8(0x05d2));
1578 addAccents("\\hebdalet", getutf8(0x05d3));
1579 addAccents("\\daleth", getutf8(0x05d3));
1580 addAccents("\\dag", getutf8(0x2020));
1581 addAccents("\\dagger", getutf8(0x2020));
1582 addAccents("\\ddag", getutf8(0x2021));
1583 addAccents("\\ddagger", getutf8(0x2021));
1584 addAccents("\\textbullet", getutf8(0x2022));
1585 addAccents("\\bullet", getutf8(0x2022));
1586 addAccents("\\dots", getutf8(0x2026));
1587 addAccents("\\ldots", getutf8(0x2026));
1588 addAccents("\\textasciiacute", getutf8(0x2032));
1589 addAccents("\\prime", getutf8(0x2032));
1590 addAccents("\\textasciigrave", getutf8(0x2035));
1591 addAccents("\\backprime", getutf8(0x2035));
1592 addAccents("\\textasteriskcentered", getutf8(0x204e));
1593 addAccents("\\ast", getutf8(0x204e));
1594 addAccents("\\textmho", getutf8(0x2127));
1595 addAccents("\\mho", getutf8(0x2127));
1596 addAccents("\\textleftarrow", getutf8(0x2190));
1597 addAccents("\\leftarrow", getutf8(0x2190));
1598 addAccents("\\textuparrow", getutf8(0x2191));
1599 addAccents("\\uparrow", getutf8(0x2191));
1600 addAccents("\\textrightarrow", getutf8(0x2192));
1601 addAccents("\\rightarrow", getutf8(0x2192));
1602 addAccents("\\textdownarrow", getutf8(0x2193));
1603 addAccents("\\downarrow", getutf8(0x2193));
1604 addAccents("\\textglobrise", getutf8(0x2197));
1605 addAccents("\\nearrow", getutf8(0x2197));
1606 addAccents("\\textglobfall", getutf8(0x2198));
1607 addAccents("\\searrow", getutf8(0x2198));
1608 addAccents("\\textsurd", getutf8(0x221a));
1609 addAccents("\\surd", getutf8(0x221a));
1610 addAccents("\\textbigcircle", getutf8(0x25ef));
1611 addAccents("\\bigcirc", getutf8(0x25ef));
1612 addAccents("\\textlangle", getutf8(0x27e8));
1613 addAccents("\\langle", getutf8(0x27e8));
1614 addAccents("\\textrangle", getutf8(0x27e9));
1615 addAccents("\\rangle", getutf8(0x27e9));
1618 static void buildAccentsMap()
1620 accents["imath"] = "ı";
1622 accents["jmath"] = "ȷ";
1623 accents["cdot"] = "·";
1624 accents["textasciicircum"] = "^";
1625 accents["mathcircumflex"] = "^";
1626 accents["sim"] = "~";
1627 accents["guillemotright"] = "»";
1628 accents["guillemotleft"] = "«";
1629 accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
1630 accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
1631 accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
1632 accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
1633 accents["negmedspace"] = getutf8(0xf0005);
1634 accents["thickspace"] = getutf8(0xf0006);
1635 accents["negthickspace"] = getutf8(0xf0007);
1636 accents["lyx"] = getutf8(0xf0010); // Used logos
1637 accents["LyX"] = getutf8(0xf0010);
1638 accents["tex"] = getutf8(0xf0011);
1639 accents["TeX"] = getutf8(0xf0011);
1640 accents["latex"] = getutf8(0xf0012);
1641 accents["LaTeX"] = getutf8(0xf0012);
1642 accents["latexe"] = getutf8(0xf0013);
1643 accents["LaTeXe"] = getutf8(0xf0013);
1644 accents["lyxarrow"] = getutf8(0xf0020);
1645 accents["braceleft"] = getutf8(0xf0030);
1646 accents["braceright"] = getutf8(0xf0031);
1647 accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash
1648 accents["backslash LyX"] = getutf8(0xf0010);
1649 accents["backslash tex"] = getutf8(0xf0011);
1650 accents["backslash TeX"] = getutf8(0xf0011);
1651 accents["backslash latex"] = getutf8(0xf0012);
1652 accents["backslash LaTeX"] = getutf8(0xf0012);
1653 accents["backslash latexe"] = getutf8(0xf0013);
1654 accents["backslash LaTeXe"] = getutf8(0xf0013);
1655 accents["backslash lyxarrow"] = getutf8(0xf0020);
1656 accents["ddot{\\imath}"] = "ï";
1657 buildaccent("ddot", "aAeEhHiIioOtuUwWxXyY",
1658 "äÄëËḧḦïÏïöÖẗüÜẅẄẍẌÿŸ"); // umlaut
1659 buildaccent("dot|.", "aAbBcCdDeEfFGghHIimMnNoOpPrRsStTwWxXyYzZ",
1660 "ȧȦḃḂċĊḋḊėĖḟḞĠġḣḢİİṁṀṅṄȯȮṗṖṙṘṡṠṫṪẇẆẋẊẏẎżŻ"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'İ'
1661 accents["acute{\\imath}"] = "í";
1662 buildaccent("acute", "aAcCeEgGkKlLmMoOnNpPrRsSuUwWyYzZiI",
1663 "áÁćĆéÉǵǴḱḰĺĹḿḾóÓńŃṕṔŕŔśŚúÚẃẂýÝźŹíÍ");
1664 buildaccent("dacute|H|h", "oOuU", "őŐűŰ"); // double acute
1665 buildaccent("mathring|r", "aAuUwy",
1667 accents["check{\\imath}"] = "ǐ";
1668 accents["check{\\jmath}"] = "ǰ";
1669 buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
1670 "čČďĎǎǍěĚǐǏǒǑǔǓǧǦǩǨȟȞľĽňŇřŘšŠŤťžŽ"); // caron
1671 accents["hat{\\imath}"] = "î";
1672 accents["hat{\\jmath}"] = "ĵ";
1673 buildaccent("hat|^", "aAcCeEgGhHiIjJoOsSuUwWyYzZ",
1674 "âÂĉĈêÊĝĜĥĤîÎĵĴôÔŝŜûÛŵŴŷŶẑẐ"); // circ
1675 accents["bar{\\imath}"] = "ī";
1676 buildaccent("bar|=", "aAeEiIoOuUyY",
1677 "āĀēĒīĪōŌūŪȳȲ"); // macron
1678 accents["tilde{\\imath}"] = "ĩ";
1679 buildaccent("tilde", "aAeEiInNoOuUvVyY",
1680 "ãÃẽẼĩĨñÑõÕũŨṽṼỹỸ"); // tilde
1681 accents["breve{\\imath}"] = "ĭ";
1682 buildaccent("breve|u", "aAeEgGiIoOuU",
1683 "ăĂĕĔğĞĭĬŏŎŭŬ"); // breve
1684 accents["grave{\\imath}"] = "ì";
1685 buildaccent("grave|`", "aAeEiIoOuUnNwWyY",
1686 "àÀèÈìÌòÒùÙǹǸẁẀỳỲ"); // grave
1687 buildaccent("subdot|d", "BbDdHhKkLlMmNnRrSsTtVvWwZzAaEeIiOoUuYy",
1688 "ḄḅḌḍḤḥḲḳḶḷṂṃṆṇṚṛṢṣṬṭṾṿẈẉẒẓẠạẸẹỊịỌọỤụỴỵ"); // dot below
1689 buildaccent("ogonek|k", "AaEeIiUuOo",
1690 "ĄąĘęĮįŲųǪǫ"); // ogonek
1691 buildaccent("cedilla|c", "CcGgKkLlNnRrSsTtEeDdHh",
1692 "ÇçĢģĶķĻļŅņŖŗŞşŢţȨȩḐḑḨḩ"); // cedilla
1693 buildaccent("subring|textsubring", "Aa",
1695 buildaccent("subhat|textsubcircum", "DdEeLlNnTtUu",
1696 "ḒḓḘḙḼḽṊṋṰṱṶṷ"); // subcircum
1697 buildaccent("subtilde|textsubtilde", "EeIiUu",
1698 "ḚḛḬḭṴṵ"); // subtilde
1699 accents["dgrave{\\imath}"] = "ȉ";
1700 accents["textdoublegrave{\\i}"] = "ȉ";
1701 buildaccent("dgrave|textdoublegrave", "AaEeIiOoRrUu",
1702 "ȀȁȄȅȈȉȌȍȐȑȔȕ"); // double grave
1703 accents["rcap{\\imath}"] = "ȉ";
1704 accents["textroundcap{\\i}"] = "ȉ";
1705 buildaccent("rcap|textroundcap", "AaEeIiOoRrUu",
1706 "ȂȃȆȇȊȋȎȏȒȓȖȗ"); // inverted breve
1707 buildaccent("slashed", "oO",
1709 fillMissingUnicodesymbols(); // Add some still not handled entries contained in 'unicodesynbols'
1710 // LYXERR0("Number of accents " << accents.size());
1714 * Created accents in math or regexp environment
1715 * are macros, but we need the utf8 equivalent
1717 void Intervall::removeAccents()
1719 if (accents.empty())
1721 static regex const accre("\\\\(([\\S]|grave|breve|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot|ogonek|"
1722 "cedilla|subring|textsubring|subhat|textsubcircum|subtilde|textsubtilde|dgrave|textdoublegrave|rcap|textroundcap|slashed)\\{[^\\{\\}]+\\}"
1723 "|((i|imath|jmath|cdot|[a-z]+space)|((backslash )?([lL]y[xX]|[tT]e[xX]|[lL]a[tT]e[xX]e?|lyxarrow))|(brace|guillemot)(left|right)|textasciicircum|mathcircumflex|sim)(?![a-zA-Z]))");
1725 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
1727 string key = sub.str(1);
1728 AccentsIterator it_ac = accents.find(key);
1729 if (it_ac != accents.end()) {
1730 string val = it_ac->second;
1731 size_t pos = sub.position(size_t(0));
1732 for (size_t i = 0; i < val.size(); i++) {
1733 par[pos+i] = val[i];
1735 // Remove possibly following space too
1736 if (par[pos+sub.str(0).size()] == ' ')
1737 addIntervall(pos+val.size(), pos + sub.str(0).size()+1);
1739 addIntervall(pos+val.size(), pos + sub.str(0).size());
1740 for (size_t i = pos+val.size(); i < pos + sub.str(0).size(); i++) {
1741 // remove traces of any remaining chars
1746 LYXERR(Debug::INFO, "Not added accent for \"" << key << "\"");
1751 void Intervall::handleOpenP(int i)
1754 depts[actualdeptindex] = i+1;
1755 closes[actualdeptindex] = -1;
1756 checkDepthIndex(actualdeptindex);
1759 void Intervall::handleCloseP(int i, bool closingAllowed)
1761 if (actualdeptindex <= 0) {
1762 if (! closingAllowed)
1763 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1764 // if we are at the very end
1765 addIntervall(i, i+1);
1768 closes[actualdeptindex] = i+1;
1773 void Intervall::resetOpenedP(int openPos)
1775 // Used as initializer for foreignlanguage entry
1776 actualdeptindex = 1;
1777 depts[1] = openPos+1;
1781 int Intervall::previousNotIgnored(int start) const
1783 int idx = 0; /* int intervalls */
1784 for (idx = ignoreidx; idx >= 0; --idx) {
1785 if (start > borders[idx].upper)
1787 if (start >= borders[idx].low)
1788 start = borders[idx].low-1;
1793 int Intervall::nextNotIgnored(int start) const
1795 int idx = 0; /* int intervalls */
1796 for (idx = 0; idx <= ignoreidx; idx++) {
1797 if (start < borders[idx].low)
1799 if (start < borders[idx].upper)
1800 start = borders[idx].upper;
1805 typedef unordered_map<string, KeyInfo> KeysMap;
1806 typedef unordered_map<string, KeyInfo>::const_iterator KeysIterator;
1807 typedef vector< KeyInfo> Entries;
1808 static KeysMap keys = unordered_map<string, KeyInfo>();
1814 Intervall interval_;
1815 void buildKeys(bool);
1816 void buildEntries(bool);
1817 void makeKey(const string &, KeyInfo, bool isPatternString);
1818 void processRegion(int start, int region_end); /* remove {} parts */
1819 void removeHead(KeyInfo const &, int count=0);
1822 LatexInfo(string const & par, bool isPatternString)
1823 : entidx_(-1), interval_(isPatternString, par)
1825 buildKeys(isPatternString);
1826 entries_ = vector<KeyInfo>();
1827 buildEntries(isPatternString);
1831 if (entries_.empty()) {
1834 if (entries_[0].keytype == KeyInfo::isTitle) {
1835 interval_.hasTitle = true;
1836 if (! entries_[0].disabled) {
1837 interval_.titleValue = entries_[0].head;
1840 interval_.titleValue = "";
1842 removeHead(entries_[0]);
1843 if (entries_.size() > 1)
1852 if (int(entries_.size()) > entidx_) {
1859 bool setNextKey(int idx) {
1860 if ((idx == entidx_) && (entidx_ >= 0)) {
1867 int find(int start, KeyInfo::KeyType keytype) const {
1871 while (tmpIdx < int(entries_.size())) {
1872 if (entries_[tmpIdx].keytype == keytype)
1878 int process(ostringstream & os, KeyInfo const & actual);
1879 int dispatch(ostringstream & os, int previousStart, KeyInfo & actual);
1880 // string show(int lastpos) { return interval.show(lastpos);};
1881 int nextNotIgnored(int start) { return interval_.nextNotIgnored(start);};
1882 KeyInfo &getKeyInfo(int keyinfo) {
1883 static KeyInfo invalidInfo = KeyInfo();
1884 if ((keyinfo < 0) || ( keyinfo >= int(entries_.size())))
1887 return entries_[keyinfo];
1889 void setForDefaultLang(KeyInfo const & defLang) {interval_.setForDefaultLang(defLang);};
1890 void addIntervall(int low, int up) { interval_.addIntervall(low, up); };
1894 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
1898 for (int i = start; i < end; i += 1 + skip) {
1902 if (c == '\\') skip = 1;
1906 else if (c == down) {
1909 if ((repeat <= 0) || (par[i+1] != up))
1923 size_t mathpostfixsize;
1925 size_t mathprefixsize;
1929 vector<MathEntry> entries_;
1934 void insert(string const & wait, size_t start, size_t prefixsize, size_t end, size_t postfixsize) {
1935 MathEntry m = MathEntry();
1937 m.mathStart = start;
1938 m.mathprefixsize = prefixsize;
1939 m.mathEnd = end + postfixsize;
1940 m.mathpostfixsize = postfixsize;
1941 m.mathSize = m.mathEnd - m.mathStart;
1942 entries_.push_back(m);
1944 bool empty() const { return entries_.empty(); };
1945 size_t getEndPos() const {
1946 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1949 return entries_[actualIdx_].mathEnd;
1951 size_t getStartPos() const {
1952 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1953 return 100000; /* definitely enough? */
1955 return entries_[actualIdx_].mathStart;
1957 size_t getPrefixSize() const {
1958 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1961 return entries_[actualIdx_].mathprefixsize;
1963 size_t getPostfixSize() const {
1964 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1967 return entries_[actualIdx_].mathpostfixsize;
1969 size_t getFirstPos() {
1971 return getStartPos();
1973 size_t getSize() const {
1974 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1977 return entries_[actualIdx_].mathSize;
1979 void incrEntry() { actualIdx_++; };
1982 void LatexInfo::buildEntries(bool isPatternString)
1984 static regex const rmath("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|x?x?alignat)\\*?\\})(\\{[0-9]+\\})?)");
1985 static regex const rkeys("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?)))");
1986 static bool disableLanguageOverride = false;
1987 smatch sub, submath;
1988 bool evaluatingRegexp = false;
1990 bool evaluatingMath = false;
1991 bool evaluatingCode = false;
1993 bool evaluatingOptional = false;
1994 size_t optionalEnd = 0;
1997 bool math_end_waiting = false;
1998 size_t math_pos = 10000;
1999 size_t math_prefix_size = 1;
2001 static vector<string> usedText = vector<string>();
2002 static bool removeMathHull = false;
2004 interval_.removeAccents();
2006 for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) {
2008 if ((submath.position(2) - submath.position(0)) %2 == 1) {
2009 // prefixed by odd count of '\\'
2012 if (math_end_waiting) {
2013 size_t pos = submath.position(size_t(2));
2014 if ((math_end == "$") &&
2015 (submath.str(2) == "$")) {
2016 mi.insert("$", math_pos, 1, pos, 1);
2017 math_end_waiting = false;
2019 else if ((math_end == "\\]") &&
2020 (submath.str(2) == "\\]")) {
2021 mi.insert("\\]", math_pos, 2, pos, 2);
2022 math_end_waiting = false;
2024 else if ((submath.str(3).compare("end") == 0) &&
2025 (submath.str(5).compare(math_end) == 0)) {
2026 mi.insert(math_end, math_pos, math_prefix_size, pos, submath.str(2).length());
2027 math_end_waiting = false;
2033 if (submath.str(3).compare("begin") == 0) {
2034 math_end_waiting = true;
2035 math_end = submath.str(5);
2036 math_pos = submath.position(size_t(2));
2037 math_prefix_size = submath.str(2).length();
2039 else if (submath.str(2).compare("\\[") == 0) {
2040 math_end_waiting = true;
2042 math_pos = submath.position(size_t(2));
2044 else if (submath.str(2) == "$") {
2045 size_t pos = submath.position(size_t(2));
2046 math_end_waiting = true;
2052 // Ignore language if there is math somewhere in pattern-string
2053 if (isPatternString) {
2054 for (auto s: usedText) {
2055 // Remove entries created in previous search runs
2058 usedText = vector<string>();
2061 keys["foreignlanguage"].disabled = true;
2062 disableLanguageOverride = true;
2063 removeMathHull = false;
2066 removeMathHull = true; // used later if not isPatternString
2067 disableLanguageOverride = false;
2071 if (disableLanguageOverride) {
2072 keys["foreignlanguage"].disabled = true;
2075 math_pos = mi.getFirstPos();
2076 for (sregex_iterator it(interval_.par.begin(), interval_.par.end(), rkeys), end; it != end; ++it) {
2078 if ((sub.position(2) - sub.position(0)) %2 == 1) {
2079 // prefixed by odd count of '\\'
2082 string key = sub.str(5);
2084 if (sub.str(2)[0] == '\\')
2085 key = sub.str(2)[1];
2090 KeysIterator it_key = keys.find(key);
2091 if (it_key != keys.end()) {
2092 if (it_key->second.keytype == KeyInfo::headRemove) {
2093 KeyInfo found1 = it_key->second;
2094 found1.disabled = true;
2095 found1.head = "\\" + key + "{";
2096 found1._tokenstart = sub.position(size_t(2));
2097 found1._tokensize = found1.head.length();
2098 found1._dataStart = found1._tokenstart + found1.head.length();
2099 int endpos = interval_.findclosing(found1._dataStart, interval_.par.length(), '{', '}', 1);
2100 found1._dataEnd = endpos;
2105 if (evaluatingRegexp) {
2106 if (sub.str(3).compare("endregexp") == 0) {
2107 evaluatingRegexp = false;
2108 // found._tokenstart already set
2109 found._dataEnd = sub.position(size_t(2)) + 13;
2110 found._dataStart = found._dataEnd;
2111 found._tokensize = found._dataEnd - found._tokenstart;
2112 found.parenthesiscount = 0;
2113 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2120 if (evaluatingMath) {
2121 if (size_t(sub.position(size_t(2))) < mi.getEndPos())
2123 evaluatingMath = false;
2125 math_pos = mi.getStartPos();
2127 if (it_key == keys.end()) {
2128 found = KeyInfo(KeyInfo::isStandard, 0, true);
2129 LYXERR(Debug::INFO, "Undefined key " << key << " ==> will be used as text");
2130 found = KeyInfo(KeyInfo::isText, 0, false);
2131 if (isPatternString) {
2132 found.keytype = KeyInfo::isChar;
2133 found.disabled = false;
2137 usedText.push_back(key);
2141 if (key.compare("regexp") == 0) {
2142 evaluatingRegexp = true;
2143 found._tokenstart = sub.position(size_t(2));
2144 found._tokensize = 0;
2148 // Handle the other params of key
2149 if (found.keytype == KeyInfo::isIgnored)
2151 else if (found.keytype == KeyInfo::isMath) {
2152 if (size_t(sub.position(size_t(2))) == math_pos) {
2154 found._tokenstart = sub.position(size_t(2));
2155 found._tokensize = mi.getSize();
2156 found._dataEnd = found._tokenstart + found._tokensize;
2157 found._dataStart = found._dataEnd;
2158 found.parenthesiscount = 0;
2159 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2160 if (removeMathHull) {
2161 interval_.addIntervall(found._tokenstart, found._tokenstart + mi.getPrefixSize());
2162 interval_.addIntervall(found._dataEnd - mi.getPostfixSize(), found._dataEnd);
2165 // Treate all math constructs as simple math
2166 interval_.par[found._tokenstart] = '$';
2167 interval_.par[found._dataEnd - mi.getPostfixSize()] = '$';
2168 interval_.addIntervall(found._tokenstart + 1, found._tokenstart + mi.getPrefixSize());
2169 interval_.addIntervall(found._dataEnd - mi.getPostfixSize() + 1, found._dataEnd);
2171 evaluatingMath = true;
2174 // begin|end of unknown env, discard
2175 // First handle tables
2176 // longtable|tabular
2177 bool discardComment;
2179 found.keytype = KeyInfo::doRemove;
2180 if ((sub.str(7).compare("longtable") == 0) ||
2181 (sub.str(7).compare("tabular") == 0)) {
2182 discardComment = true; /* '%' */
2185 discardComment = false;
2186 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
2188 string token = sub.str(7);
2189 if (regex_match(token, sub2, removeArgs)) {
2190 found.keytype = KeyInfo::removeWithArg;
2193 // discard spaces before pos(2)
2194 int pos = sub.position(size_t(2));
2196 for (count = 0; pos - count > 0; count++) {
2197 char c = interval_.par[pos-count-1];
2198 if (discardComment) {
2199 if ((c != ' ') && (c != '%'))
2205 found._tokenstart = pos - count;
2206 if (sub.str(3).compare(0, 5, "begin") == 0) {
2207 size_t pos1 = pos + sub.str(2).length();
2208 if (sub.str(7).compare("cjk") == 0) {
2209 pos1 = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2210 if ((interval_.par[pos1] == '{') && (interval_.par[pos1+1] == '}'))
2212 found.keytype = KeyInfo::isMain;
2213 found._dataStart = pos1;
2214 found._dataEnd = interval_.par.length();
2215 found.disabled = keys["foreignlanguage"].disabled;
2216 found.used = keys["foreignlanguage"].used;
2217 found._tokensize = pos1 - found._tokenstart;
2218 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2221 // Swallow possible optional params
2222 while (interval_.par[pos1] == '[') {
2223 pos1 = interval_.findclosing(pos1+1, interval_.par.length(), '[', ']')+1;
2225 // Swallow also the eventual parameter
2226 if (interval_.par[pos1] == '{') {
2227 found._dataEnd = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2230 found._dataEnd = pos1;
2232 found._dataStart = found._dataEnd;
2233 found._tokensize = count + found._dataEnd - pos;
2234 found.parenthesiscount = 0;
2235 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2236 found.disabled = true;
2240 // Handle "\end{...}"
2241 found._dataStart = pos + sub.str(2).length();
2242 found._dataEnd = found._dataStart;
2243 found._tokensize = count + found._dataEnd - pos;
2244 found.parenthesiscount = 0;
2245 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2246 found.disabled = true;
2250 else if (found.keytype != KeyInfo::isRegex) {
2251 found._tokenstart = sub.position(size_t(2));
2252 if (found.parenthesiscount == 0) {
2253 // Probably to be discarded
2254 size_t following_pos = sub.position(size_t(2)) + sub.str(5).length() + 1;
2255 char following = interval_.par[following_pos];
2256 if (following == ' ')
2257 found.head = "\\" + sub.str(5) + " ";
2258 else if (following == '=') {
2259 // like \uldepth=1000pt
2260 found.head = sub.str(2);
2263 found.head = "\\" + key;
2264 found._tokensize = found.head.length();
2265 found._dataEnd = found._tokenstart + found._tokensize;
2266 found._dataStart = found._dataEnd;
2269 int params = found._tokenstart + key.length() + 1;
2270 if (evaluatingOptional) {
2271 if (size_t(found._tokenstart) > optionalEnd) {
2272 evaluatingOptional = false;
2275 found.disabled = true;
2278 int optend = params;
2279 while (interval_.par[optend] == '[') {
2280 // discard optional parameters
2281 optend = interval_.findclosing(optend+1, interval_.par.length(), '[', ']') + 1;
2283 if (optend > params) {
2284 key += interval_.par.substr(params, optend-params);
2285 evaluatingOptional = true;
2286 optionalEnd = optend;
2287 if (found.keytype == KeyInfo::isSectioning) {
2288 // Remove optional values (but still keep in header)
2289 interval_.addIntervall(params, optend);
2292 string token = sub.str(7);
2294 if (interval_.par[optend] != '{') {
2296 found.parenthesiscount = 0;
2297 found.head = "\\" + key;
2300 closings = found.parenthesiscount;
2301 if (found.parenthesiscount == 1) {
2302 found.head = "\\" + key + "{";
2304 else if (found.parenthesiscount > 1) {
2306 found.head = sub.str(2) + "{";
2307 closings = found.parenthesiscount - 1;
2310 found.head = "\\" + key + "{";
2313 found._tokensize = found.head.length();
2314 found._dataStart = found._tokenstart + found.head.length();
2315 if (found.keytype == KeyInfo::doRemove) {
2317 size_t endpar = 2 + interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2318 if (endpar >= interval_.par.length())
2319 found._dataStart = interval_.par.length();
2321 found._dataStart = endpar;
2322 found._tokensize = found._dataStart - found._tokenstart;
2325 found._dataStart = found._tokenstart + found._tokensize;
2329 if (interval_.par.substr(found._dataStart, 15).compare("\\endarguments{}") == 0) {
2330 found._dataStart += 15;
2334 endpos = found._dataStart - 1;
2336 endpos = interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2337 if (found.keytype == KeyInfo::isList) {
2338 // Check if it really is list env
2339 static regex const listre("^([a-z]+)$");
2341 if (!regex_match(token, sub2, listre)) {
2342 // Change the key of this entry. It is not in a list/item environment
2343 found.keytype = KeyInfo::endArguments;
2346 if (found.keytype == KeyInfo::noMain) {
2347 evaluatingCode = true;
2349 codeStart = found._dataStart;
2351 else if (evaluatingCode) {
2352 if (size_t(found._dataStart) > codeEnd)
2353 evaluatingCode = false;
2354 else if (found.keytype == KeyInfo::isMain) {
2355 // Disable this key, treate it as standard
2356 found.keytype = KeyInfo::isStandard;
2357 found.disabled = true;
2358 if ((codeEnd +1 >= interval_.par.length()) &&
2359 (found._tokenstart == codeStart)) {
2360 // trickery, because the code inset starts
2361 // with \selectlanguage ...
2363 if (entries_.size() > 1) {
2364 entries_[entries_.size()-1]._dataEnd = codeEnd;
2369 if ((endpos == interval_.par.length()) &&
2370 (found.keytype == KeyInfo::doRemove)) {
2371 // Missing closing => error in latex-input?
2372 // therefore do not delete remaining data
2373 found._dataStart -= 1;
2374 found._dataEnd = found._dataStart;
2377 found._dataEnd = endpos;
2379 if (isPatternString) {
2380 keys[key].used = true;
2383 entries_.push_back(found);
2387 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
2389 stringstream s(keysstring);
2391 const char delim = '|';
2392 while (getline(s, key, delim)) {
2393 KeyInfo keyII(keyI);
2394 if (isPatternString) {
2397 else if ( !keys[key].used)
2398 keyII.disabled = true;
2403 void LatexInfo::buildKeys(bool isPatternString)
2406 static bool keysBuilt = false;
2407 if (keysBuilt && !isPatternString) return;
2409 // Keys to ignore in any case
2410 makeKey("text|textcyrillic|lyxmathsym|ensuremath", KeyInfo(KeyInfo::headRemove, 1, true), true);
2411 // Known standard keys with 1 parameter.
2412 // Split is done, if not at start of region
2413 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
2414 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
2415 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
2416 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
2417 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
2418 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
2420 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
2421 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2422 makeKey("section*|subsection*|subsubsection*|paragraph*",
2423 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2424 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2425 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
2427 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
2429 // Split is done, if not at start of region
2430 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
2431 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
2433 // Split is done always.
2434 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
2438 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2439 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2440 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2441 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2443 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2444 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2445 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2446 makeKey("thickspace|medspace|thinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2448 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2449 // Custom space/skip, remove the content (== length value)
2450 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
2451 // Found in fr/UserGuide.lyx
2452 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2454 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2455 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2456 // Known macros to remove (including their parameter)
2458 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
2459 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
2460 // handle like standard keys with 1 parameter.
2461 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
2463 // Ignore deleted text
2464 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString);
2465 // but preserve added text
2466 makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2468 // Macros to remove, but let the parameter survive
2470 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2472 // Remove language spec from content of these insets
2473 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
2475 // Same effect as previous, parameter will survive (because there is no one anyway)
2477 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2478 // Remove table decorations
2479 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
2480 // Discard shape-header.
2481 // For footnote or shortcut too, because of lang settings
2482 // and wrong handling if used 'KeyInfo::noMain'
2483 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2484 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2485 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2486 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2487 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2488 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
2489 // like ('tiny{}' or '\tiny ' ... )
2490 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
2492 // Survives, like known character
2493 // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2494 makeKey("tableofcontents", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2495 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
2497 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2498 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2499 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2501 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2502 // Remove RTL/LTR marker
2503 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2504 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
2505 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
2506 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
2507 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
2508 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
2509 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2510 if (isPatternString) {
2511 // Allow the first searched string to rebuild the keys too
2515 // no need to rebuild again
2521 * Keep the list of actual opened parentheses actual
2522 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2524 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2527 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2531 if (c == '\\') skip = 1;
2532 else if (c == '{') {
2535 else if (c == '}') {
2536 handleCloseP(i, closingAllowed);
2542 string Intervall::show(int lastpos)
2544 int idx = 0; /* int intervalls */
2547 for (idx = 0; idx <= ignoreidx; idx++) {
2548 while (i < lastpos) {
2550 if (i <= borders[idx].low) {
2551 if (borders[idx].low > lastpos)
2552 printsize = lastpos - i;
2554 printsize = borders[idx].low - i;
2555 s += par.substr(i, printsize);
2557 if (i >= borders[idx].low)
2558 i = borders[idx].upper;
2561 i = borders[idx].upper;
2567 s += par.substr(i, lastpos-i);
2573 void Intervall::output(ostringstream &os, int lastpos)
2575 // get number of chars to output
2576 int idx = 0; /* int intervalls */
2579 string startTitle = titleValue;
2580 for (idx = 0; idx <= ignoreidx; idx++) {
2582 if (i <= borders[idx].low) {
2584 if (borders[idx].low > lastpos)
2585 printsize = lastpos - i;
2587 printsize = borders[idx].low - i;
2588 if (printsize > 0) {
2589 os << startTitle << par.substr(i, printsize);
2591 printed += printsize;
2594 handleParentheses(i, false);
2595 if (i >= borders[idx].low)
2596 i = borders[idx].upper;
2599 i = borders[idx].upper;
2606 os << startTitle << par.substr(i, lastpos-i);
2607 printed += lastpos-i;
2609 handleParentheses(lastpos, false);
2611 if (keys["foreignlanguage"].disabled)
2612 startindex = actualdeptindex-langcount;
2614 startindex = actualdeptindex;
2615 for (int i = startindex; i > 0; --i) {
2618 if (hasTitle && (printed > 0))
2620 if (! isPatternString_)
2622 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2625 void LatexInfo::processRegion(int start, int region_end)
2627 while (start < region_end) { /* Let {[} and {]} survive */
2628 int cnt = interval_.isOpeningPar(start);
2630 // Closing is allowed past the region
2631 int closing = interval_.findclosing(start+1, interval_.par.length());
2632 interval_.addIntervall(start, start+1);
2633 interval_.addIntervall(closing, closing+1);
2637 start = interval_.nextNotIgnored(start+1);
2641 void LatexInfo::removeHead(KeyInfo const & actual, int count)
2643 if (actual.parenthesiscount == 0) {
2644 // "{\tiny{} ...}" ==> "{{} ...}"
2645 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2648 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2649 interval_.addIntervall(actual._tokenstart - count, actual._dataStart);
2650 interval_.addIntervall(actual._dataEnd, actual._dataEnd+1);
2654 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2657 switch (actual.keytype)
2659 case KeyInfo::isTitle: {
2661 nextKeyIdx = getNextKey();
2664 case KeyInfo::cleanToStart: {
2665 actual._dataEnd = actual._dataStart;
2666 nextKeyIdx = getNextKey();
2667 // Search for end of arguments
2668 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2670 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2671 entries_[i].disabled = true;
2673 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2675 while (interval_.par[actual._dataEnd] == ' ')
2677 interval_.addIntervall(0, actual._dataEnd+1);
2678 interval_.actualdeptindex = 0;
2679 interval_.depts[0] = actual._dataEnd+1;
2680 interval_.closes[0] = -1;
2683 case KeyInfo::isText:
2684 interval_.par[actual._tokenstart] = '#';
2685 //interval_.addIntervall(actual._tokenstart, actual._tokenstart+1);
2686 nextKeyIdx = getNextKey();
2688 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2689 if (actual.disabled)
2690 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2692 interval_.addIntervall(actual._dataStart, actual._dataEnd);
2695 case KeyInfo::isChar: {
2696 nextKeyIdx = getNextKey();
2699 case KeyInfo::isSize: {
2700 if (actual.disabled || (interval_.par[actual._dataStart] != '{') || (interval_.par[actual._dataStart-1] == ' ')) {
2701 if (actual.parenthesiscount == 0)
2702 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2704 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2706 nextKeyIdx = getNextKey();
2708 // Here _dataStart points to '{', so correct it
2709 actual._dataStart += 1;
2710 actual._tokensize += 1;
2711 actual.parenthesiscount = 1;
2712 if (interval_.par[actual._dataStart] == '}') {
2713 // Determine the end if used like '{\tiny{}...}'
2714 actual._dataEnd = interval_.findclosing(actual._dataStart+1, interval_.par.length()) + 1;
2715 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2718 // Determine the end if used like '\tiny{...}'
2719 actual._dataEnd = interval_.findclosing(actual._dataStart, interval_.par.length()) + 1;
2721 // Split on this key if not at start
2722 int start = interval_.nextNotIgnored(previousStart);
2723 if (start < actual._tokenstart) {
2724 interval_.output(os, actual._tokenstart);
2725 interval_.addIntervall(start, actual._tokenstart);
2727 // discard entry if at end of actual
2728 nextKeyIdx = process(os, actual);
2732 case KeyInfo::endArguments: {
2733 // Remove trailing '{}' too
2734 actual._dataStart += 1;
2735 actual._dataEnd += 1;
2736 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2737 nextKeyIdx = getNextKey();
2740 case KeyInfo::noMain:
2742 case KeyInfo::isStandard: {
2743 if (actual.disabled) {
2745 processRegion(actual._dataStart, actual._dataStart+1);
2746 nextKeyIdx = getNextKey();
2748 // Split on this key if not at datastart of calling entry
2749 int start = interval_.nextNotIgnored(previousStart);
2750 if (start < actual._tokenstart) {
2751 interval_.output(os, actual._tokenstart);
2752 interval_.addIntervall(start, actual._tokenstart);
2754 // discard entry if at end of actual
2755 nextKeyIdx = process(os, actual);
2759 case KeyInfo::removeWithArg: {
2760 nextKeyIdx = getNextKey();
2761 // Search for end of arguments
2762 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2764 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2765 entries_[i].disabled = true;
2767 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2769 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2772 case KeyInfo::doRemove: {
2773 // Remove the key with all parameters and following spaces
2776 if (interval_.par[actual._dataEnd-1] == ' ')
2777 start = actual._dataEnd;
2779 start = actual._dataEnd+1;
2780 for (pos = start; pos < interval_.par.length(); pos++) {
2781 if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
2784 // Remove also enclosing parentheses [] and {}
2787 while (actual._tokenstart > numpars) {
2788 if (pos+numpars >= interval_.par.size())
2790 else if (interval_.par[pos+numpars] == ']' && interval_.par[actual._tokenstart-numpars-1] == '[')
2792 else if (interval_.par[pos+numpars] == '}' && interval_.par[actual._tokenstart-numpars-1] == '{')
2798 if (interval_.par[pos+numpars] == ' ')
2802 interval_.addIntervall(actual._tokenstart-numpars, pos+numpars+spaces);
2803 nextKeyIdx = getNextKey();
2806 case KeyInfo::isList: {
2807 // Discard space before _tokenstart
2809 for (count = 0; count < actual._tokenstart; count++) {
2810 if (interval_.par[actual._tokenstart-count-1] != ' ')
2813 nextKeyIdx = getNextKey();
2814 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2816 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2818 // How else can we catch this one?
2819 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2820 entries_[i].disabled = true;
2822 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2824 else if (nextKeyIdx > 0) {
2825 // Ignore any lang entries inside data region
2826 for (int i = nextKeyIdx; i < int(entries_.size()) && entries_[i]._tokenstart < actual._dataEnd; i++) {
2827 if (entries_[i].keytype == KeyInfo::isMain)
2828 entries_[i].disabled = true;
2831 if (actual.disabled) {
2832 interval_.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2835 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2837 if (interval_.par[actual._dataEnd+1] == '[') {
2838 int posdown = interval_.findclosing(actual._dataEnd+2, interval_.par.length(), '[', ']');
2839 if ((interval_.par[actual._dataEnd+2] == '{') &&
2840 (interval_.par[posdown-1] == '}')) {
2841 interval_.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
2842 interval_.addIntervall(posdown-1, posdown+1);
2845 interval_.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
2846 interval_.addIntervall(posdown, posdown+1);
2848 int blk = interval_.nextNotIgnored(actual._dataEnd+1);
2849 if (blk > posdown) {
2850 // Discard at most 1 space after empty item
2852 for (count = 0; count < 1; count++) {
2853 if (interval_.par[blk+count] != ' ')
2857 interval_.addIntervall(blk, blk+count);
2862 case KeyInfo::isSectioning: {
2863 // Discard spaces before _tokenstart
2865 int val = actual._tokenstart;
2866 for (count = 0; count < actual._tokenstart;) {
2867 val = interval_.previousNotIgnored(val-1);
2868 if (val < 0 || interval_.par[val] != ' ')
2871 count = actual._tokenstart - val;
2874 if (actual.disabled) {
2875 removeHead(actual, count);
2876 nextKeyIdx = getNextKey();
2878 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2879 nextKeyIdx = process(os, actual);
2883 case KeyInfo::isMath: {
2884 // Same as regex, use the content unchanged
2885 nextKeyIdx = getNextKey();
2888 case KeyInfo::isRegex: {
2889 // DO NOT SPLIT ON REGEX
2891 nextKeyIdx = getNextKey();
2894 case KeyInfo::isIgnored: {
2895 // Treat like a character for now
2896 nextKeyIdx = getNextKey();
2899 case KeyInfo::isMain: {
2900 if (interval_.par.substr(actual._dataStart, 2) == "% ")
2901 interval_.addIntervall(actual._dataStart, actual._dataStart+2);
2902 if (actual._tokenstart > 0) {
2903 int prev = interval_.previousNotIgnored(actual._tokenstart - 1);
2904 if ((prev >= 0) && interval_.par[prev] == '%')
2905 interval_.addIntervall(prev, prev+1);
2907 if (actual.disabled) {
2909 interval_.langcount++;
2910 if ((interval_.par.substr(actual._dataStart, 3) == " \\[") ||
2911 (interval_.par.substr(actual._dataStart, 8) == " \\begin{")) {
2912 // Discard also the space before math-equation
2913 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2915 nextKeyIdx = getNextKey();
2916 // interval.resetOpenedP(actual._dataStart-1);
2919 if (actual._tokenstart < 26) {
2920 // for the first (and maybe dummy) language
2921 interval_.setForDefaultLang(actual);
2923 interval_.resetOpenedP(actual._dataStart-1);
2927 case KeyInfo::invalid:
2928 case KeyInfo::headRemove:
2929 // These two cases cannot happen, already handled
2932 // LYXERR(Debug::INFO, "Unhandled keytype");
2933 nextKeyIdx = getNextKey();
2940 int LatexInfo::process(ostringstream & os, KeyInfo const & actual )
2942 int end = interval_.nextNotIgnored(actual._dataEnd);
2943 int oldStart = actual._dataStart;
2944 int nextKeyIdx = getNextKey();
2946 if ((nextKeyIdx < 0) ||
2947 (entries_[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
2948 (entries_[nextKeyIdx].keytype == KeyInfo::invalid)) {
2949 if (oldStart <= end) {
2950 processRegion(oldStart, end);
2955 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
2957 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
2958 (void) dispatch(os, actual._dataStart, nextKey);
2959 end = nextKey._tokenstart;
2962 processRegion(oldStart, nextKey._tokenstart);
2963 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
2965 oldStart = nextKey._dataEnd+1;
2967 // now nextKey is either invalid or is outside of actual._dataEnd
2968 // output the remaining and discard myself
2969 if (oldStart <= end) {
2970 processRegion(oldStart, end);
2972 if (interval_.par.size() > (size_t) end && interval_.par[end] == '}') {
2974 // This is the normal case.
2975 // But if using the firstlanguage, the closing may be missing
2977 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
2979 if (actual._dataEnd < end)
2980 output_end = interval_.nextNotIgnored(actual._dataEnd);
2981 else if (interval_.par.size() > (size_t) end)
2982 output_end = interval_.nextNotIgnored(end);
2984 output_end = interval_.par.size();
2985 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
2986 interval_.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
2988 // Remove possible empty data
2989 int dstart = interval_.nextNotIgnored(actual._dataStart);
2990 while (interval_.isOpeningPar(dstart) == 1) {
2991 interval_.addIntervall(dstart, dstart+1);
2992 int dend = interval_.findclosing(dstart+1, output_end);
2993 interval_.addIntervall(dend, dend+1);
2994 dstart = interval_.nextNotIgnored(dstart+1);
2996 if (dstart < output_end)
2997 interval_.output(os, output_end);
2999 interval_.addIntervall(0, end);
3001 interval_.addIntervall(actual._tokenstart, end);
3005 string splitOnKnownMacros(string par, bool isPatternString)
3008 LatexInfo li(par, isPatternString);
3009 // LYXERR(Debug::INFO, "Berfore split: " << par);
3010 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
3012 DummyKey._tokensize = 0;
3013 DummyKey._dataStart = 0;
3014 DummyKey._dataEnd = par.length();
3015 DummyKey.disabled = true;
3016 int firstkeyIdx = li.getFirstKey();
3018 if (firstkeyIdx >= 0) {
3019 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
3020 DummyKey._tokenstart = firstKey._tokenstart;
3022 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
3023 // Use dummy firstKey
3024 firstKey = DummyKey;
3025 (void) li.setNextKey(firstkeyIdx);
3028 if (par.substr(firstKey._dataStart, 2) == "% ")
3029 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
3031 nextkeyIdx = li.process(os, firstKey);
3032 while (nextkeyIdx >= 0) {
3033 // Check for a possible gap between the last
3034 // entry and this one
3035 int datastart = li.nextNotIgnored(firstKey._dataStart);
3036 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
3037 if ((nextKey._tokenstart > datastart)) {
3039 firstKey._dataStart = datastart;
3040 firstKey._dataEnd = par.length();
3041 (void) li.setNextKey(nextkeyIdx);
3042 // Fake the last opened parenthesis
3043 li.setForDefaultLang(firstKey);
3044 nextkeyIdx = li.process(os, firstKey);
3047 if (nextKey.keytype != KeyInfo::isMain) {
3048 firstKey._dataStart = datastart;
3049 firstKey._dataEnd = nextKey._dataEnd+1;
3050 (void) li.setNextKey(nextkeyIdx);
3051 li.setForDefaultLang(firstKey);
3052 nextkeyIdx = li.process(os, firstKey);
3055 nextkeyIdx = li.process(os, nextKey);
3059 // Handle the remaining
3060 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
3061 firstKey._dataEnd = par.length();
3063 if ((firstKey._dataStart < firstKey._dataEnd) &&
3064 (par[firstKey._dataStart] != '}')) {
3065 li.setForDefaultLang(firstKey);
3066 (void) li.process(os, firstKey);
3070 // return string definitelly impossible to match
3071 s = "\\foreignlanguage{ignore}{ }";
3075 s = par; /* no known macros found */
3076 // LYXERR(Debug::INFO, "After split: " << s);
3081 * Try to unify the language specs in the latexified text.
3082 * Resulting modified string is set to "", if
3083 * the searched tex does not contain all the features in the search pattern
3085 static string correctlanguagesetting(string par, bool isPatternString, bool withformat, lyx::Buffer *pbuf = nullptr)
3087 static Features regex_f;
3088 static int missed = 0;
3089 static bool regex_with_format = false;
3091 int parlen = par.length();
3093 while ((parlen > 0) && (par[parlen-1] == '\n')) {
3096 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
3097 // Happens to be there in case of description or labeling environment
3102 // Split the latex input into pieces which
3103 // can be digested by our search engine
3104 LYXERR(Debug::FIND, "input: \"" << par << "\"");
3105 if (isPatternString && (pbuf != nullptr)) { // Check if we should disable/enable test for language
3106 // We check for polyglossia, because in runparams.flavor we use Flavor::XeTeX
3107 string doclang = pbuf->params().language->polyglossia();
3108 static regex langre("\\\\(foreignlanguage)\\{([^\\}]+)\\}");
3110 bool toIgnoreLang = true;
3111 for (sregex_iterator it(par.begin(), par.end(), langre), end; it != end; ++it) {
3113 if (sub.str(2) != doclang) {
3114 toIgnoreLang = false;
3118 setIgnoreFormat("language", toIgnoreLang, false);
3121 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
3122 LYXERR(Debug::FIND, "After splitOnKnownMacros:\n\"" << result << "\"");
3125 result = par.substr(0, parlen);
3126 if (isPatternString) {
3129 regex_f = identifyFeatures(result);
3130 string features = "";
3131 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3132 string a = it->first;
3133 regex_with_format = true;
3134 features += " " + a;
3135 // LYXERR(Debug::INFO, "Identified regex format:" << a);
3137 LYXERR(Debug::FIND, "Identified Features" << features);
3140 } else if (regex_with_format) {
3141 Features info = identifyFeatures(result);
3142 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3143 string a = it->first;
3144 bool b = it->second;
3145 if (b && ! info[a]) {
3147 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
3154 // LYXERR(Debug::INFO, "No regex formats");
3160 // Remove trailing closure of math, macros and environments, so to catch parts of them.
3161 static int identifyClosing(string & t)
3163 int open_braces = 0;
3165 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
3166 if (regex_replace(t, t, "(.*[^\\\\])\\$$", "$1"))
3168 if (regex_replace(t, t, "(.*[^\\\\])\\\\\\]$", "$1"))
3170 if (regex_replace(t, t, "(.*[^\\\\])\\\\end\\{[a-zA-Z_]*\\*?\\}$", "$1"))
3172 if (regex_replace(t, t, "(.*[^\\\\])\\}$", "$1")) {
3181 static int num_replaced = 0;
3182 static bool previous_single_replace = true;
3184 void MatchStringAdv::CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string)
3187 // Handle \w properly
3188 QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
3189 if (! opt.casesensitive) {
3190 popts |= QRegularExpression::CaseInsensitiveOption;
3192 regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
3193 regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
3195 if (regexp.isValid() && regexp2.isValid()) {
3196 regexIsValid = true;
3197 // Check '{', '}' pairs inside the regex
3200 for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
3201 char c = par_as_string[i];
3208 else if (c == '}') {
3215 if (balanced != 0) {
3216 regexIsValid = false;
3217 regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
3221 regexIsValid = false;
3222 if (!regexp.isValid())
3223 regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
3225 regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
3228 if (opt.casesensitive) {
3229 regexp = regex(regexp_str);
3230 regexp2 = regex(regexp2_str);
3233 regexp = regex(regexp_str, std::regex_constants::icase);
3234 regexp2 = regex(regexp2_str, std::regex_constants::icase);
3239 static void modifyRegexForMatchWord(string &t)
3242 regex wordre("(\\\\)*((\\.|\\\\b))");
3245 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
3247 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
3250 else if (sub.str(2) == "\\\\b")
3252 if (lastpos < (size_t) sub.position(2))
3253 s += t.substr(lastpos, sub.position(2) - lastpos);
3255 lastpos = sub.position(2) + sub.length(2);
3258 s = "\\b" + t + "\\b";
3262 else if (lastpos < t.length())
3263 s += t.substr(lastpos, t.length() - lastpos);
3264 t = "\\b" + s + "\\b";
3267 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt)
3268 : p_buf(&buf), p_first_buf(&buf), opt(opt)
3270 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
3271 docstring const & ds = stringifySearchBuffer(find_buf, opt);
3272 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
3273 if (opt.replace_all && previous_single_replace) {
3274 previous_single_replace = false;
3277 else if (!opt.replace_all) {
3278 num_replaced = 0; // count number of replaced strings
3279 previous_single_replace = true;
3281 // When using regexp, braces are hacked already by escape_for_regex()
3282 par_as_string = normalize(ds);
3284 close_wildcards = 0;
3286 size_t lead_size = 0;
3287 // correct the language settings
3288 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat, &buf);
3289 opt.matchAtStart = false;
3291 identifyClosing(par_as_string); // Removes math closings ($, ], ...) at end of string
3292 if (opt.ignoreformat) {
3296 lead_size = identifyLeading(par_as_string);
3298 lead_as_string = par_as_string.substr(0, lead_size);
3299 string lead_as_regex_string = string2regex(lead_as_string);
3300 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3301 string par_as_regex_string_nolead = string2regex(par_as_string_nolead);
3302 /* Handle whole words too in this case
3304 if (opt.matchword) {
3305 par_as_regex_string_nolead = "\\b" + par_as_regex_string_nolead + "\\b";
3306 opt.matchword = false;
3308 string regexp_str = "(" + lead_as_regex_string + ")()" + par_as_regex_string_nolead;
3309 string regexp2_str = "(" + lead_as_regex_string + ")(.*?)" + par_as_regex_string_nolead;
3310 CreateRegexp(opt, regexp_str, regexp2_str);
3312 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3313 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3317 if (!opt.ignoreformat) {
3318 lead_size = identifyLeading(par_as_string);
3319 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
3320 lead_as_string = par_as_string.substr(0, lead_size);
3321 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3324 // Here we are using regexp
3325 LASSERT(use_regexp, /**/);
3327 string lead_as_regexp;
3328 if (lead_size > 0) {
3329 lead_as_regexp = string2regex(par_as_string.substr(0, lead_size));
3330 regex_replace(par_as_string_nolead, par_as_string_nolead, "}$", "");
3331 par_as_string = par_as_string_nolead;
3332 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
3333 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3335 // LYXERR(Debug::FIND, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
3336 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
3337 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
3338 // LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3340 size_t lng = par_as_string.size();
3341 if (!opt.ignoreformat) {
3342 // Remove extra '\}' at end if not part of \{\.\}
3344 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
3346 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
3355 if (lng < par_as_string.size())
3356 par_as_string = par_as_string.substr(0,lng);
3358 LYXERR(Debug::FIND, "par_as_string after correctRegex is '" << par_as_string << "'");
3359 if ((lng > 0) && (par_as_string[0] == '^')) {
3360 par_as_string = par_as_string.substr(1);
3362 opt.matchAtStart = true;
3364 // LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3365 // LYXERR(Debug::FIND, "Open braces: " << open_braces);
3366 // LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
3368 // If entered regexp must match at begin of searched string buffer
3369 // Kornel: Added parentheses to use $1 for size of the leading string
3373 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
3374 // Unfortunately is '\1', '\2', etc not working for strings with extra format
3375 // so the convert has no effect in that case
3376 for (int i = 7; i > 0; --i) {
3377 string orig = "\\\\" + std::to_string(i);
3378 string dest = "\\" + std::to_string(i+2);
3379 while (regex_replace(par_as_string, par_as_string, orig, dest));
3381 if (opt.matchword) {
3382 modifyRegexForMatchWord(par_as_string);
3383 opt.matchword = false;
3385 regexp_str = "(" + lead_as_regexp + ")()" + par_as_string;
3386 regexp2_str = "(" + lead_as_regexp + ")(.*?)" + par_as_string;
3388 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3389 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3390 CreateRegexp(opt, regexp_str, regexp2_str, par_as_string);
3394 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
3398 mres.searched_size = len;
3400 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
3403 docstring docstr = stringifyFromForSearch(opt, cur, len);
3405 str = normalize(docstr);
3406 if (!opt.ignoreformat) {
3407 str = correctlanguagesetting(str, false, !opt.ignoreformat);
3408 // remove closing '}' and '\n' to allow for use of '$' in regex
3409 size_t lng = str.size();
3410 while ((lng > 1) && ((str[lng -1] == '}') || (str[lng -1] == '\n')))
3412 if (lng != str.size()) {
3413 str = str.substr(0, lng);
3417 mres.match_len = -1;
3420 LYXERR(Debug::FIND, "After normalization: Matching against:\n'" << str << "'");
3422 LASSERT(use_regexp, /**/);
3424 // use_regexp always true
3425 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
3427 QString qstr = QString::fromStdString(str);
3428 QRegularExpression const *p_regexp;
3429 QRegularExpression::MatchType flags = QRegularExpression::NormalMatch;
3433 p_regexp = ®exp2;
3435 QRegularExpressionMatch match = p_regexp->match(qstr, 0, flags);
3436 if (!match.hasMatch())
3439 regex const *p_regexp;
3440 regex_constants::match_flag_type flags;
3442 flags = regex_constants::match_continuous;
3445 flags = regex_constants::match_default;
3446 p_regexp = ®exp2;
3448 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
3449 if (re_it == sregex_iterator())
3451 match_results<string::const_iterator> const & m = *re_it;
3453 // Whole found string, including the leading
3454 // std: m[0].second - m[0].first
3455 // Qt: match.capturedEnd(0) - match.capturedStart(0)
3457 // Size of the leading string
3458 // std: m[1].second - m[1].first
3459 // Qt: match.capturedEnd(1) - match.capturedStart(1)
3460 int leadingsize = 0;
3462 if (match.lastCapturedIndex() > 0) {
3463 leadingsize = match.capturedEnd(1) - match.capturedStart(1);
3468 leadingsize = m[1].second - m[1].first;
3472 mres.match_prefix = match.capturedEnd(2) - match.capturedStart(2);
3473 mres.match_len = match.capturedEnd(0) - match.capturedEnd(2);
3474 // because of different number of closing at end of string
3475 // we have to 'unify' the length of the post-match.
3476 // Done by ignoring closing parenthesis and linefeeds at string end
3477 int matchend = match.capturedEnd(0);
3478 size_t strsize = qstr.size();
3479 if (!opt.ignoreformat) {
3480 while (mres.match_len > 0) {
3481 QChar c = qstr.at(matchend - 1);
3482 if ((c == '\n') || (c == '}') || (c == '{')) {
3489 while (strsize > (size_t) match.capturedEnd(0)) {
3490 QChar c = qstr.at(strsize-1);
3491 if ((c == '\n') || (c == '}')) {
3498 // LYXERR0(qstr.toStdString());
3499 mres.match2end = strsize - matchend;
3500 mres.pos = match.capturedStart(2);
3502 mres.match_prefix = m[2].second - m[2].first;
3503 mres.match_len = m[0].second - m[2].second;
3504 // ignore closing parenthesis and linefeeds at string end
3505 size_t strend = m[0].second - m[0].first;
3506 int matchend = strend;
3507 size_t strsize = str.size();
3508 if (!opt.ignoreformat) {
3509 while (mres.match_len > 0) {
3510 char c = str.at(matchend - 1);
3511 if ((c == '\n') || (c == '}') || (c == '{')) {
3518 while (strsize > strend) {
3519 if ((str.at(strsize-1) == '}') || (str.at(strsize-1) == '\n')) {
3527 mres.match2end = strsize - matchend;
3528 mres.pos = m[2].first - m[0].first;;
3530 if (mres.match2end < 0)
3532 mres.leadsize = leadingsize;
3534 if (mres.match_len > 0) {
3535 string a0 = match.captured(0).mid(mres.pos + mres.match_prefix, mres.match_len).toStdString();
3536 mres.result.push_back(a0);
3537 for (int i = 3; i <= match.lastCapturedIndex(); i++) {
3538 mres.result.push_back(match.captured(i).toStdString());
3542 if (mres.match_len > 0) {
3543 string a0 = m[0].str().substr(mres.pos + mres.match_prefix, mres.match_len);
3544 mres.result.push_back(a0);
3545 for (size_t i = 3; i < m.size(); i++) {
3546 mres.result.push_back(m[i]);
3555 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
3557 MatchResult mres = findAux(cur, len, at_begin);
3558 int res = mres.match_len;
3560 "res=" << res << ", at_begin=" << at_begin
3561 << ", matchAtStart=" << opt.matchAtStart
3562 << ", inTexted=" << cur.inTexted());
3563 if (opt.matchAtStart) {
3566 else if (mres.match_prefix > 0)
3575 static bool simple_replace(string &t, string from, string to)
3577 regex repl("(\\\\)*(" + from + ")");
3581 for (sregex_iterator it(t.begin(), t.end(), repl), end; it != end; ++it) {
3583 if ((sub.position(2) - sub.position(0)) % 2 == 1)
3585 if (lastpos < (size_t) sub.position(2))
3586 s += t.substr(lastpos, sub.position(2) - lastpos);
3588 lastpos = sub.position(2) + sub.length(2);
3592 else if (lastpos < t.length())
3593 s += t.substr(lastpos, t.length() - lastpos);
3599 string MatchStringAdv::normalize(docstring const & s) const
3602 t = lyx::to_utf8(s);
3603 // Remove \n at begin
3604 while (!t.empty() && t[0] == '\n')
3607 while (!t.empty() && t[t.size() - 1] == '\n')
3608 t = t.substr(0, t.size() - 1);
3610 // Handle all other '\n'
3611 while ((pos = t.find("\n")) != string::npos) {
3612 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
3614 if (isAlnumASCII(t[pos+1])) {
3615 t.replace(pos-2, 3, " ");
3618 t.replace(pos-2, 3, "");
3621 else if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
3622 // '\n' adjacent to non-alpha-numerics, discard
3623 t.replace(pos, 1, "");
3626 // Replace all other \n with spaces
3627 t.replace(pos, 1, " ");
3630 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
3631 // Kornel: Added textsl, textsf, textit, texttt and noun
3632 // + allow to seach for colored text too
3633 LYXERR(Debug::FIND, "Removing stale empty macros from: " << t);
3634 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
3635 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3636 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
3637 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3638 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
3644 docstring stringifyFromCursor(DocIterator const & cur, int len)
3646 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
3647 if (cur.inTexted()) {
3648 Paragraph const & par = cur.paragraph();
3649 // TODO what about searching beyond/across paragraph breaks ?
3650 // TODO Try adding a AS_STR_INSERTS as last arg
3651 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
3652 int(par.size()) : cur.pos() + len;
3653 // OutputParams runparams(&cur.buffer()->params().encoding());
3654 OutputParams runparams(encodings.fromLyXName("utf8"));
3655 runparams.nice = true;
3656 runparams.flavor = Flavor::XeTeX;
3657 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
3658 // No side effect of file copying and image conversion
3659 runparams.dryrun = true;
3660 int option = AS_STR_INSETS | AS_STR_PLAINTEXT;
3661 if (ignoreFormats.getDeleted()) {
3662 option |= AS_STR_SKIPDELETE;
3663 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
3666 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
3668 LYXERR(Debug::FIND, "Stringifying with cur: "
3669 << cur << ", from pos: " << cur.pos() << ", end: " << end);
3670 return par.asString(cur.pos(), end,
3673 } else if (cur.inMathed()) {
3674 CursorSlice cs = cur.top();
3675 MathData md = cs.cell();
3676 MathData::const_iterator it_end =
3677 (( len == -1 || cs.pos() + len > int(md.size()))
3679 : md.begin() + cs.pos() + len );
3681 for (MathData::const_iterator it = md.begin() + cs.pos();
3684 docstring s = asString(md2);
3685 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
3688 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3693 /** Computes the LaTeX export of buf starting from cur and ending len positions
3694 * after cur, if len is positive, or at the paragraph or innermost inset end
3697 docstring latexifyFromCursor(DocIterator const & cur, int len)
3700 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
3701 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
3702 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
3704 Buffer const & buf = *cur.buffer();
3706 odocstringstream ods;
3708 //OutputParams runparams(&buf.params().encoding());
3709 OutputParams runparams(encodings.fromLyXName("utf8"));
3710 runparams.nice = false;
3711 runparams.flavor = Flavor::XeTeX;
3712 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3713 // No side effect of file copying and image conversion
3714 runparams.dryrun = true;
3715 if (ignoreFormats.getDeleted()) {
3716 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
3719 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
3722 if (cur.inTexted()) {
3723 // @TODO what about searching beyond/across paragraph breaks ?
3724 pos_type endpos = cur.paragraph().size();
3725 if (len != -1 && endpos > cur.pos() + len)
3726 endpos = cur.pos() + len;
3727 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
3728 string(), cur.pos(), endpos);
3729 string s = lyx::to_utf8(ods.str());
3730 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
3731 return(lyx::from_utf8(s));
3732 } else if (cur.inMathed()) {
3733 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
3734 for (int s = cur.depth() - 1; s >= 0; --s) {
3735 CursorSlice const & cs = cur[s];
3736 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
3737 TeXMathStream ws(os);
3738 cs.asInsetMath()->asHullInset()->header_write(ws);
3743 CursorSlice const & cs = cur.top();
3744 MathData md = cs.cell();
3745 MathData::const_iterator it_end =
3746 ((len == -1 || cs.pos() + len > int(md.size()))
3748 : md.begin() + cs.pos() + len);
3750 for (MathData::const_iterator it = md.begin() + cs.pos();
3754 ods << asString(md2);
3755 // Retrieve the math environment type, and add '$' or '$]'
3756 // or others (\end{equation}) accordingly
3757 for (int s = cur.depth() - 1; s >= 0; --s) {
3758 CursorSlice const & cs2 = cur[s];
3759 InsetMath * inset = cs2.asInsetMath();
3760 if (inset && inset->asHullInset()) {
3761 TeXMathStream ws(os);
3762 inset->asHullInset()->footer_write(ws);
3766 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3768 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3773 #if defined(ResultsDebug)
3775 static void displayMResult(MatchResult &mres, string from, DocIterator & cur)
3777 LYXERR0( "from:\t\t\t" << from);
3779 if (mres.pos_len > 0) {
3781 status = "FINALSEARCH";
3784 if (mres.match_len > 0) {
3785 if ((mres.match_prefix == 0) && (mres.pos == mres.leadsize))
3786 status = "Good Match";
3788 status = "Matched in";
3791 status = "MissedSearch";
3794 LYXERR0( status << "(" << cur.pos() << " ... " << mres.searched_size + cur.pos() << ") cur.lastpos(" << cur.lastpos() << ")");
3795 if ((mres.leadsize > 0) || (mres.match_len > 0) || (mres.match2end > 0))
3796 LYXERR0( "leadsize(" << mres.leadsize << ") match_len(" << mres.match_len << ") match2end(" << mres.match2end << ")");
3797 if ((mres.pos > 0) || (mres.match_prefix > 0))
3798 LYXERR0( "pos(" << mres.pos << ") match_prefix(" << mres.match_prefix << ")");
3799 for (size_t i = 0; i < mres.result.size(); i++)
3800 LYXERR0( "Match " << i << " = \"" << mres.result[i] << "\"");
3802 #define displayMres(s, txt, cur) displayMResult(s, txt, cur);
3804 #define displayMres(s, txt, cur)
3807 /** Finalize an advanced find operation, advancing the cursor to the innermost
3808 ** position that matches, plus computing the length of the matching text to
3810 ** Return the cur.pos() difference between start and end of found match
3812 MatchResult findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, MatchResult const & expected = MatchResult(-1))
3814 // Search the foremost position that matches (avoids find of entire math
3815 // inset when match at start of it)
3816 DocIterator old_cur(cur.buffer());
3818 static MatchResult fail = MatchResult();
3819 MatchResult max_match;
3820 // If (prefix_len > 0) means that forwarding 1 position will remove the complete entry
3821 // Happens with e.g. hyperlinks
3822 // either one sees "http://www.bla.bla" or nothing
3823 // so the search for "www" gives prefix_len = 7 (== sizeof("http://")
3824 // and although we search for only 3 chars, we find the whole hyperlink inset
3825 bool at_begin = (expected.match_prefix == 0);
3826 if (!match.opt.forward && match.opt.ignoreformat) {
3827 if (expected.pos > 0)
3830 LASSERT(at_begin, /**/);
3831 if (expected.match_len > 0 && at_begin) {
3832 // Search for deepest match
3834 max_match = expected;
3836 size_t d = cur.depth();
3840 if (cur.depth() < d)
3842 if (cur.depth() == d)
3845 while (cur && cur.depth() > lastd) {
3846 lastd = cur.depth();
3847 mres = match(cur, -1, at_begin);
3848 displayMres(mres, "Checking innermost", cur);
3849 if (mres.match_len > 0)
3854 if (mres.match_len < expected.match_len)
3862 // (expected.match_len <= 0)
3863 mres = match(cur); /* match valid only if not searching whole words */
3864 displayMres(mres, "Start with negative match", cur);
3867 if (max_match.match_len <= 0) return fail;
3868 LYXERR(Debug::FIND, "Ok");
3870 // Compute the match length
3872 if (cur.pos() + len > cur.lastpos())
3875 LASSERT(match.use_regexp, /**/);
3878 int maxl = cur.lastpos() - cur.pos();
3879 // Greedy behaviour while matching regexps
3880 while (maxl > minl) {
3882 mres2 = match(cur, len, at_begin);
3883 displayMres(mres2, "Finalize loop", cur);
3884 int actual_match_len = mres2.match_len;
3885 if (actual_match_len >= max_match.match_len) {
3886 // actual_match_len > max_match _can_ happen,
3887 // if the search area splits
3888 // some following word so that the regex
3889 // (e.g. 'r.*r\b' matches 'r' from the middle of the
3891 // This means, the len value is too big
3892 actual_match_len = max_match.match_len;
3894 max_match.match_len = actual_match_len;
3896 if (maxl - minl < 4)
3897 len = (int)((maxl + minl)/2);
3899 len = (int)(minl + (maxl - minl + 3)/4);
3902 // (actual_match_len < max_match.match_len)
3904 len = (int)((maxl + minl)/2);
3909 // Search for real start of matched characters
3911 MatchResult actual_match;
3914 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
3915 if (cur.depth() < old_cur.depth()) {
3917 LYXERR(Debug::INFO, "cur.depth() < old_cur.depth(), this should never happen");
3920 if (cur.pos() != old_cur.pos()) {
3921 // OK, forwarded 1 pos in actual inset
3922 actual_match = match(cur, len-1, at_begin);
3923 if (actual_match.match_len == max_match.match_len) {
3924 // Ha, got it! The shorter selection has the same match length
3927 max_match = actual_match;
3930 // OK, the shorter selection matches less chars, revert to previous value
3936 LYXERR(Debug::INFO, "cur.pos() == old_cur.pos(), this should never happen");
3937 actual_match = match(cur, len, at_begin);
3938 if (actual_match.match_len == max_match.match_len) {
3940 max_match = actual_match;
3947 max_match.pos_len = len;
3948 displayMres(max_match, "SEARCH RESULT", cur)
3955 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
3959 bool repeat = false;
3960 DocIterator orig_cur; // to be used if repeat not successful
3961 MatchResult orig_mres;
3962 while (!theApp()->longOperationCancelled() && cur) {
3963 //(void) findAdvForwardInnermost(cur);
3964 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
3965 MatchResult mres = match(cur, -1, false);
3966 string msg = "Starting";
3969 displayMres(mres, msg + " findForwardAdv", cur)
3970 int match_len = mres.match_len;
3971 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
3972 LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
3975 if (match_len <= 0) {
3976 // This should exit nested insets, if any, or otherwise undefine the currsor.
3977 cur.pos() = cur.lastpos();
3978 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
3981 else { // match_len > 0
3982 // Try to find the begin of searched string
3984 int firstInvalid = cur.lastpos() - cur.pos();
3986 int incrmatch = (mres.match_prefix + mres.pos - mres.leadsize + 1)*3/4;
3987 int incrcur = (firstInvalid + 1 )*3/4;
3988 if (incrcur < incrmatch)
3989 increment = incrcur;
3991 increment = incrmatch;
3995 LYXERR(Debug::FIND, "Set increment to " << increment);
3996 while (increment > 0) {
3997 DocIterator old_cur = cur;
3998 if (cur.pos() + increment >= cur.lastpos()) {
4002 cur.pos() = cur.pos() + increment;
4003 MatchResult mres2 = match(cur, -1, false);
4004 displayMres(mres2, "findForwardAdv loop", cur)
4005 switch (interpretMatch(mres, mres2)) {
4006 case MatchResult::newIsTooFar:
4007 // behind the expected match
4008 firstInvalid = increment;
4012 case MatchResult::newIsBetter:
4013 // not reached yet, but cur.pos()+increment is bettert
4015 firstInvalid -= increment;
4016 if (increment > firstInvalid*3/4)
4017 increment = firstInvalid*3/4;
4018 if ((mres2.pos == mres2.leadsize) && (increment >= mres2.match_prefix)) {
4019 if (increment >= mres2.match_prefix)
4020 increment = (mres2.match_prefix+1)*3/4;
4025 // Handle not like MatchResult::newIsTooFar
4026 LYXERR0( "Probably too far: Increment = " << increment << " match_prefix = " << mres.match_prefix);
4028 increment = increment*3/4;
4033 if (mres.match_len > 0) {
4034 if (mres.match_prefix + mres.pos - mres.leadsize > 0) {
4035 // The match seems to indicate some deeper level
4044 // should never be reached.
4048 // LYXERR0("Leaving first loop");
4049 LYXERR(Debug::FIND, "Finalizing 1");
4050 MatchResult found_match = findAdvFinalize(cur, match, mres);
4051 if (found_match.match_len > 0) {
4052 LASSERT(found_match.pos_len > 0, /**/);
4053 match.FillResults(found_match);
4054 return found_match.pos_len;
4057 // try next possible match
4068 /// Find the most backward consecutive match within same paragraph while searching backwards.
4069 MatchResult findMostBackwards(DocIterator & cur, MatchStringAdv const & match, MatchResult &expected)
4071 DocIterator cur_begin = cur;
4072 cur_begin.pos() = 0;
4073 DocIterator tmp_cur = cur;
4074 MatchResult mr = findAdvFinalize(tmp_cur, match, expected);
4075 Inset & inset = cur.inset();
4076 for (; cur != cur_begin; cur.backwardPos()) {
4077 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
4078 DocIterator new_cur = cur;
4079 new_cur.backwardPos();
4080 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
4082 MatchResult new_mr = findAdvFinalize(new_cur, match, expected);
4083 if (new_mr.match_len == mr.match_len)
4087 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
4093 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
4097 // Backup of original position
4098 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
4099 if (cur == cur_begin)
4102 DocIterator cur_orig(cur);
4103 bool pit_changed = false;
4106 MatchResult found_match = match(cur, -1, false);
4108 if (found_match.match_len > 0) {
4110 cur.pos() = cur.lastpos();
4112 cur.pos() = cur_orig.pos();
4113 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
4114 DocIterator cur_prev_iter;
4116 found_match = match(cur);
4117 LYXERR(Debug::FIND, "findBackAdv3: found_match="
4118 << (found_match.match_len > 0) << ", cur: " << cur);
4119 if (found_match.match_len > 0) {
4120 MatchResult found_mr = findMostBackwards(cur, match, found_match);
4121 if (found_mr.pos_len > 0) {
4122 match.FillResults(found_mr);
4123 return found_mr.pos_len;
4127 // Stop if begin of document reached
4128 if (cur == cur_begin)
4130 cur_prev_iter = cur;
4134 if (cur == cur_begin)
4141 } while (!theApp()->longOperationCancelled());
4149 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
4150 DocIterator const & cur, int len)
4152 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
4154 if (!opt.ignoreformat)
4155 return latexifyFromCursor(cur, len);
4157 return stringifyFromCursor(cur, len);
4161 FindAndReplaceOptions::FindAndReplaceOptions(
4162 docstring const & _find_buf_name, bool _casesensitive,
4163 bool _matchword, bool _forward, bool _expandmacros, bool _ignoreformat,
4164 docstring const & _repl_buf_name, bool _keep_case,
4165 SearchScope _scope, SearchRestriction _restr, bool _replace_all)
4166 : find_buf_name(_find_buf_name), casesensitive(_casesensitive), matchword(_matchword),
4167 forward(_forward), expandmacros(_expandmacros), ignoreformat(_ignoreformat),
4168 repl_buf_name(_repl_buf_name), keep_case(_keep_case), scope(_scope), restr(_restr), replace_all(_replace_all)
4176 /** Check if 'len' letters following cursor are all non-lowercase */
4177 static bool allNonLowercase(Cursor const & cur, int len)
4179 pos_type beg_pos = cur.selectionBegin().pos();
4180 pos_type end_pos = cur.selectionBegin().pos() + len;
4181 if (len > cur.lastpos() + 1 - beg_pos) {
4182 LYXERR(Debug::FIND, "This should not happen, more debug needed");
4183 len = cur.lastpos() + 1 - beg_pos;
4184 end_pos = beg_pos + len;
4186 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
4187 if (isLowerCase(cur.paragraph().getChar(pos)))
4193 /** Check if first letter is upper case and second one is lower case */
4194 static bool firstUppercase(Cursor const & cur)
4197 pos_type pos = cur.selectionBegin().pos();
4198 if (pos >= cur.lastpos() - 1) {
4199 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
4202 ch1 = cur.paragraph().getChar(pos);
4203 ch2 = cur.paragraph().getChar(pos + 1);
4204 bool result = isUpperCase(ch1) && isLowerCase(ch2);
4205 LYXERR(Debug::FIND, "firstUppercase(): "
4206 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
4207 << ch2 << "(" << char(ch2) << ")"
4208 << ", result=" << result << ", cur=" << cur);
4213 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
4215 ** \fixme What to do with possible further paragraphs in replace buffer ?
4217 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
4219 ParagraphList::iterator pit = buffer.paragraphs().begin();
4220 LASSERT(!pit->empty(), /**/);
4221 pos_type right = pos_type(1);
4222 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
4223 right = pit->size();
4224 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
4228 static bool replaceMatches(string &t, int maxmatchnum, vector <string> const & replacements)
4230 // Should replace the string "$" + std::to_string(matchnum) with replacement
4231 // if the char '$' is not prefixed with odd number of char '\\'
4232 static regex const rematch("(\\\\)*(\\$\\$([0-9]))");
4236 for (sregex_iterator it(t.begin(), t.end(), rematch), end; it != end; ++it) {
4238 if ((sub.position(2) - sub.position(0)) % 2 == 1)
4240 int num = stoi(sub.str(3), nullptr, 10);
4241 if (num >= maxmatchnum)
4243 if (lastpos < (size_t) sub.position(2))
4244 s += t.substr(lastpos, sub.position(2) - lastpos);
4245 s += replacements[num];
4246 lastpos = sub.position(2) + sub.length(2);
4250 else if (lastpos < t.length())
4251 s += t.substr(lastpos, t.length() - lastpos);
4257 static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
4259 Cursor & cur = bv->cursor();
4260 if (opt.repl_buf_name.empty()
4261 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
4262 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4265 DocIterator sel_beg = cur.selectionBegin();
4266 DocIterator sel_end = cur.selectionEnd();
4267 if (&sel_beg.inset() != &sel_end.inset()
4268 || sel_beg.pit() != sel_end.pit()
4269 || sel_beg.idx() != sel_end.idx())
4271 int sel_len = sel_end.pos() - sel_beg.pos();
4272 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
4273 << ", sel_len: " << sel_len << endl);
4276 LASSERT(sel_len > 0, return 0);
4278 if (!matchAdv(sel_beg, sel_len).match_len)
4281 // Build a copy of the replace buffer, adapted to the KeepCase option
4282 Buffer const & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
4284 repl_buffer_orig.write(oss);
4285 string lyx = oss.str();
4286 if (matchAdv.valid_matches > 0) {
4287 replaceMatches(lyx, matchAdv.valid_matches, matchAdv.matches);
4289 Buffer repl_buffer("", false);
4290 repl_buffer.setUnnamed(true);
4291 LASSERT(repl_buffer.readString(lyx), return 0);
4292 if (opt.keep_case && sel_len >= 2) {
4293 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
4294 if (cur.inTexted()) {
4295 if (firstUppercase(cur))
4296 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
4297 else if (allNonLowercase(cur, sel_len))
4298 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
4301 cap::cutSelection(cur, false);
4302 if (cur.inTexted()) {
4303 repl_buffer.changeLanguage(
4304 repl_buffer.language(),
4305 cur.getFont().language());
4306 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
4307 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
4308 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
4309 repl_buffer.params().documentClassPtr(),
4310 repl_buffer.params().authors(),
4311 bv->buffer().errorList("Paste"));
4312 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
4313 sel_len = repl_buffer.paragraphs().begin()->size();
4314 } else if (cur.inMathed()) {
4315 odocstringstream ods;
4317 // OutputParams runparams(&repl_buffer.params().encoding());
4318 OutputParams runparams(encodings.fromLyXName("utf8"));
4319 runparams.nice = false;
4320 runparams.flavor = Flavor::XeTeX;
4321 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
4322 runparams.dryrun = true;
4323 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
4324 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
4325 docstring repl_latex = ods.str();
4326 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
4328 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
4329 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
4330 repl_latex = from_utf8(s);
4331 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
4332 MathData ar(cur.buffer());
4333 asArray(repl_latex, ar, Parse::NORMAL);
4335 sel_len = ar.size();
4336 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4338 if (cur.pos() >= sel_len)
4339 cur.pos() -= sel_len;
4342 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4343 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
4344 bv->processUpdateFlags(Update::Force);
4349 /// Perform a FindAdv operation.
4350 bool findAdv(BufferView * bv, FindAndReplaceOptions & opt)
4355 // e.g., when invoking word-findadv from mini-buffer wither with
4356 // wrong options syntax or before ever opening advanced F&R pane
4357 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4361 MatchStringAdv matchAdv(bv->buffer(), opt);
4363 if (!matchAdv.regexIsValid) {
4364 bv->message(lyx::from_utf8(matchAdv.regexError));
4368 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
4370 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
4371 num_replaced += findAdvReplace(bv, opt, matchAdv);
4374 pos_len = findForwardAdv(cur, matchAdv);
4376 pos_len = findBackwardsAdv(cur, matchAdv);
4377 } catch (exception & ex) {
4378 bv->message(from_utf8(ex.what()));
4383 if (num_replaced > 0) {
4384 switch (num_replaced)
4387 bv->message(_("One match has been replaced."));
4390 bv->message(_("Two matches have been replaced."));
4393 bv->message(bformat(_("%1$d matches have been replaced."), num_replaced));
4399 bv->message(_("Match not found."));
4404 if (num_replaced > 0)
4405 bv->message(_("Match has been replaced."));
4407 bv->message(_("Match found."));
4409 if (cur.pos() + pos_len > cur.lastpos()) {
4410 // Prevent crash in bv->putSelectionAt()
4411 // Should never happen, maybe LASSERT() here?
4412 pos_len = cur.lastpos() - cur.pos();
4414 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << pos_len);
4415 bv->putSelectionAt(cur, pos_len, !opt.forward);
4421 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
4423 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
4424 << opt.casesensitive << ' '
4425 << opt.matchword << ' '
4426 << opt.forward << ' '
4427 << opt.expandmacros << ' '
4428 << opt.ignoreformat << ' '
4429 << opt.replace_all << ' '
4430 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
4431 << opt.keep_case << ' '
4432 << int(opt.scope) << ' '
4435 LYXERR(Debug::FIND, "built: " << os.str());
4441 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
4443 // LYXERR(Debug::FIND, "parsing");
4447 while (line != "EOSS") {
4451 if (is.eof()) // Tolerate malformed request
4455 // LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
4456 opt.find_buf_name = from_utf8(s);
4457 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.replace_all;
4458 is.get(); // Waste space before replace string
4461 while (line != "EOSS") {
4465 if (is.eof()) // Tolerate malformed request
4469 // LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
4470 opt.repl_buf_name = from_utf8(s);
4471 is >> opt.keep_case;
4474 opt.scope = FindAndReplaceOptions::SearchScope(i);
4476 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
4479 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
4480 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
4481 << opt.scope << ' ' << opt.restr);