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 const 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());
341 bool wrap = auto_wrap;
345 q = _("End of file reached while searching forward.\n"
346 "Continue searching from the beginning?");
348 q = _("Beginning of file reached while searching backward.\n"
349 "Continue searching from the end?");
350 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
351 q, 0, 1, _("&Yes"), _("&No"));
352 wrap = wrap_answer == 0;
356 bv->cursor().clear();
357 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
359 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
360 bv->cursor().backwardPos();
362 bv->clearSelection();
363 if (findOne(bv, searchstr, case_sens, whole, forward,
364 find_del, false, false, false, false))
367 bv->cursor().setCursor(cur_orig);
371 return match_len > 0;
377 int replaceAll(BufferView * bv,
378 docstring const & searchstr, docstring const & replacestr,
379 bool case_sens, bool whole, bool onlysel)
381 Buffer & buf = bv->buffer();
383 if (!searchAllowed(searchstr) || buf.isReadonly())
386 DocIterator startcur = bv->cursor().selectionBegin();
387 DocIterator endcur = bv->cursor().selectionEnd();
388 bool const had_selection = bv->cursor().selection();
390 MatchString const match(searchstr, case_sens, whole);
393 int const rsize = replacestr.size();
394 int const ssize = searchstr.size();
397 cur.setCursor(doc_iterator_begin(&buf));
398 int match_len = findForward(cur, endcur, match, false, onlysel);
399 while (match_len > 0) {
400 // Backup current cursor position and font.
401 pos_type const pos = cur.pos();
402 Font const font = cur.paragraph().getFontSettings(buf.params(), pos);
404 int ct_deleted_text = ssize -
405 cur.paragraph().eraseChars(pos, pos + match_len,
406 buf.params().track_changes);
407 cur.paragraph().insert(pos, replacestr, font,
408 Change(buf.params().track_changes
410 : Change::UNCHANGED));
411 for (int i = 0; i < rsize + ct_deleted_text
412 && cur.pos() < cur.lastpos(); ++i)
414 if (onlysel && cur.pit() == endcur.pit() && cur.idx() == endcur.idx()) {
415 // Adjust end of selection for replace-all in selection
417 int const offset = rsize - ssize;
418 for (int i = 0; i < offset + ct_deleted_text
419 && endcur.pos() < endcur.lastpos(); ++i)
422 int const offset = ssize - rsize;
423 for (int i = 0; i < offset && endcur.pos() > 0; ++i)
424 endcur.backwardPos();
425 for (int i = 0; i < ct_deleted_text
426 && endcur.pos() < endcur.lastpos(); ++i)
431 match_len = findForward(cur, endcur, match, false, onlysel);
434 bv->putSelectionAt(doc_iterator_begin(&buf), 0, false);
436 startcur.fixIfBroken();
437 bv->setCursor(startcur);
439 // Reset selection, accounting for changes in selection
441 endcur.fixIfBroken();
442 bv->cursor().resetAnchor();
443 bv->setCursorSelectionTo(endcur);
450 // the idea here is that we are going to replace the string that
451 // is selected IF it is the search string.
452 // if there is a selection, but it is not the search string, then
453 // we basically ignore it. (FIXME We ought to replace only within
455 // if there is no selection, then:
456 // (i) if some search string has been provided, then we find it.
457 // (think of how the dialog works when you hit "replace" the
459 // (ii) if no search string has been provided, then we treat the
460 // word the cursor is in as the search string. (why? i have no
461 // idea.) but this only works in text?
463 // returns the number of replacements made (one, if any) and
464 // whether anything at all was done.
465 pair<bool, int> replaceOne(BufferView * bv, docstring searchstr,
466 docstring const & replacestr, bool case_sens,
467 bool whole, bool forward, bool findnext, bool wrap,
470 Cursor & cur = bv->cursor();
471 if (!cur.selection() || onlysel) {
472 // no selection, non-empty search string: find it
473 if (!searchstr.empty()) {
474 bool const found = findOne(bv, searchstr, case_sens, whole,
475 forward, true, findnext, wrap, false, onlysel);
476 return make_pair(found, 0);
478 // empty search string
481 return make_pair(false, 0);
482 // select current word and treat it as the search string.
483 // This causes a minor bug as undo will restore this selection,
484 // which the user did not create (#8986).
485 cur.innerText()->selectWord(cur, WHOLE_WORD);
486 searchstr = cur.selectionAsString(false, true);
489 // if we still don't have a search string, report the error
491 if (!searchAllowed(searchstr))
492 return make_pair(false, 0);
494 bool have_selection = cur.selection();
495 docstring const selected = cur.selectionAsString(false, true);
498 ? searchstr == selected
499 : compare_no_case(searchstr, selected) == 0;
501 // no selection or current selection is not search word:
502 // just find the search word
503 if (!have_selection || !match) {
504 bool const found = findOne(bv, searchstr, case_sens, whole, forward,
505 true, findnext, wrap, false, onlysel);
506 return make_pair(found, 0);
509 // we're now actually ready to replace. if the buffer is
510 // read-only, we can't, though.
511 if (bv->buffer().isReadonly())
512 return make_pair(false, 0);
514 cap::replaceSelectionWithString(cur, replacestr);
516 cur.pos() += replacestr.length();
517 LASSERT(cur.pos() <= cur.lastpos(),
518 cur.pos() = cur.lastpos());
521 findOne(bv, searchstr, case_sens, whole,
522 forward, false, findnext, wrap, false, onlysel);
524 return make_pair(true, 1);
530 docstring const find2string(docstring const & search,
531 bool casesensitive, bool matchword,
532 bool forward, bool wrap, bool instant,
537 << int(casesensitive) << ' '
538 << int(matchword) << ' '
539 << int(forward) << ' '
541 << int(instant) << ' '
547 docstring const replace2string(docstring const & replace,
548 docstring const & search,
549 bool casesensitive, bool matchword,
550 bool all, bool forward, bool findnext,
551 bool wrap, bool onlysel)
554 ss << replace << '\n'
556 << int(casesensitive) << ' '
557 << int(matchword) << ' '
559 << int(forward) << ' '
560 << int(findnext) << ' '
567 docstring const string2find(docstring const & argument,
575 // data is of the form
577 // <casesensitive> <matchword> <forward> <wrap> <onlysel>"
579 docstring howto = split(argument, search, '\n');
581 casesensitive = parse_bool(howto);
582 matchword = parse_bool(howto);
583 forward = parse_bool(howto, true);
584 wrap = parse_bool(howto);
585 instant = parse_bool(howto);
586 onlysel = parse_bool(howto);
592 bool lyxfind(BufferView * bv, FuncRequest const & ev)
594 if (!bv || ev.action() != LFUN_WORD_FIND)
597 //lyxerr << "find called, cmd: " << ev << endl;
605 docstring search = string2find(ev.argument(), casesensitive,
606 matchword, forward, wrap, instant, onlysel);
608 return findOne(bv, search, casesensitive, matchword, forward,
609 false, true, wrap, instant, onlysel);
613 bool lyxreplace(BufferView * bv, FuncRequest const & ev)
615 if (!bv || ev.action() != LFUN_WORD_REPLACE)
618 // data is of the form
621 // <casesensitive> <matchword> <all> <forward> <findnext> <wrap> <onlysel>"
624 docstring howto = split(ev.argument(), rplc, '\n');
625 howto = split(howto, search, '\n');
627 bool casesensitive = parse_bool(howto);
628 bool matchword = parse_bool(howto);
629 bool all = parse_bool(howto);
630 bool forward = parse_bool(howto, true);
631 bool findnext = parse_bool(howto, true);
632 bool wrap = parse_bool(howto);
633 bool onlysel = parse_bool(howto);
635 if (!bv->cursor().selection())
636 // only selection only makes sense with selection
641 int replace_count = 0;
643 replace_count = replaceAll(bv, search, rplc, casesensitive,
645 update = replace_count > 0;
648 replaceOne(bv, search, rplc, casesensitive, matchword,
649 forward, findnext, wrap, onlysel);
651 replace_count = rv.second;
654 Buffer const & buf = bv->buffer();
656 // emit message signal.
658 buf.message(_("String not found in selection."));
660 buf.message(_("String not found."));
662 if (replace_count == 0) {
663 buf.message(_("String found."));
664 } else if (replace_count == 1) {
665 buf.message(_("String has been replaced."));
667 docstring const str = onlysel
668 ? bformat(_("%1$d strings have been replaced in the selection."), replace_count)
669 : bformat(_("%1$d strings have been replaced."), replace_count);
677 bool findNextChange(BufferView * bv, Cursor & cur, bool const check_wrap)
679 for (; cur; cur.forwardPos())
680 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
684 DocIterator cur_orig(bv->cursor());
685 docstring q = _("End of file reached while searching forward.\n"
686 "Continue searching from the beginning?");
687 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
688 q, 0, 1, _("&Yes"), _("&No"));
689 if (wrap_answer == 0) {
690 bv->cursor().clear();
691 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
692 bv->clearSelection();
693 cur.setCursor(bv->cursor().selectionBegin());
694 if (findNextChange(bv, cur, false))
697 bv->cursor().setCursor(cur_orig);
704 bool findPreviousChange(BufferView * bv, Cursor & cur, bool const check_wrap)
706 for (cur.backwardPos(); cur; cur.backwardPos()) {
707 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
712 DocIterator cur_orig(bv->cursor());
713 docstring q = _("Beginning of file reached while searching backward.\n"
714 "Continue searching from the end?");
715 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
716 q, 0, 1, _("&Yes"), _("&No"));
717 if (wrap_answer == 0) {
718 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
719 bv->cursor().backwardPos();
720 bv->clearSelection();
721 cur.setCursor(bv->cursor().selectionBegin());
722 if (findPreviousChange(bv, cur, false))
725 bv->cursor().setCursor(cur_orig);
732 bool selectChange(Cursor & cur, bool forward)
734 if (!cur.inTexted() || !cur.paragraph().isChanged(cur.pos()))
736 Change ch = cur.paragraph().lookupChange(cur.pos());
738 CursorSlice tip1 = cur.top();
739 for (; tip1.pit() < tip1.lastpit() || tip1.pos() < tip1.lastpos(); tip1.forwardPos()) {
740 Change ch2 = tip1.paragraph().lookupChange(tip1.pos());
741 if (!ch2.isSimilarTo(ch))
744 CursorSlice tip2 = cur.top();
745 for (; tip2.pit() > 0 || tip2.pos() > 0;) {
747 Change ch2 = tip2.paragraph().lookupChange(tip2.pos());
748 if (!ch2.isSimilarTo(ch)) {
749 // take a step forward to correctly set the selection
757 cur.bv().mouseSetCursor(cur, false);
759 cur.bv().mouseSetCursor(cur, true);
767 bool findChange(BufferView * bv, bool forward)
770 cur.setCursor(forward ? bv->cursor().selectionEnd()
771 : bv->cursor().selectionBegin());
772 forward ? findNextChange(bv, cur, true) : findPreviousChange(bv, cur, true);
773 return selectChange(cur, forward);
778 bool findNextChange(BufferView * bv)
780 return findChange(bv, true);
784 bool findPreviousChange(BufferView * bv)
786 return findChange(bv, false);
793 typedef vector<pair<string, string> > Escapes;
795 string string2regex(string in)
797 static std::regex specialChars { R"([-[\]{}()*+?.,\^$|#\s\$\\])" };
798 string temp = std::regex_replace(in, specialChars, R"(\$&)" );
803 while (fl_pos < temp.size()) {
804 fl_pos = temp.find("\\\\foreignlanguage", lastpos + offset);
805 if (fl_pos == string::npos)
808 temp2 += temp.substr(lastpos, fl_pos - lastpos);
814 if (lastpos < temp.size()) {
815 temp2 += temp.substr(lastpos, temp.size() - lastpos);
820 string correctRegex(string t, bool withformat)
822 /* Convert \backslash => \
823 * and \{, \}, \[, \] => {, }, [, ]
826 regex wordre("(\\\\)*(\\\\((backslash|mathcircumflex) ?|[\\[\\]\\{\\}]))");
829 bool backslashed = false;
830 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
833 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
837 if (sub.str(4) == "backslash") {
840 // transforms '\backslash \{' into '\{'
842 string next = t.substr(sub.position(2) + sub.str(2).length(), 2);
843 if ((next == "\\{") || (next == "\\}")) {
849 else if (sub.str(4) == "mathcircumflex")
851 else if (backslashed) {
853 if (withformat && (sub.str(3) == "{"))
854 replace = accents["braceleft"];
855 else if (withformat && (sub.str(3) == "}"))
856 replace = accents["braceright"];
858 // else part should not exist
863 replace = sub.str(3);
865 if (lastpos < (size_t) sub.position(2))
866 s += t.substr(lastpos, sub.position(2) - lastpos);
868 lastpos = sub.position(2) + sub.length(2);
872 else if (lastpos < t.length())
873 s += t.substr(lastpos, t.length() - lastpos);
877 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
878 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
879 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
880 string escape_for_regex(string s, bool withformat)
884 while (lastpos < s.size()) {
885 size_t regex_pos = s.find("\\regexp{", lastpos);
886 if (regex_pos == string::npos) {
887 regex_pos = s.size();
889 if (regex_pos > lastpos) {
890 result += string2regex(s.substr(lastpos, regex_pos-lastpos));
892 if (lastpos == s.size())
895 size_t end_pos = s.find("\\endregexp{}}", regex_pos + 8);
896 result += correctRegex(s.substr(regex_pos + 8, end_pos -(regex_pos + 8)), withformat);
897 lastpos = end_pos + 13;
903 /// Wrapper for lyx::regex_replace with simpler interface
904 bool regex_replace(string const & s, string & t, string const & searchstr,
905 string const & replacestr)
907 regex e(searchstr, regex_constants::ECMAScript);
909 ostream_iterator<char, char> it(oss);
910 regex_replace(it, s.begin(), s.end(), e, replacestr);
911 // tolerate t and s be references to the same variable
912 bool rv = (s != oss.str());
931 vector <string> result = vector <string>();
932 MatchResult(int len = 0): match_len(len),match_prefix(0),match2end(0), pos(0),leadsize(0),pos_len(-1),searched_size(0) {};
935 static MatchResult::range interpretMatch(MatchResult &oldres, MatchResult &newres)
937 if (newres.match2end < oldres.match2end)
938 return MatchResult::newIsTooFar;
939 if (newres.match_len < oldres.match_len)
940 return MatchResult::newIsTooFar;
942 if (newres.match_len == oldres.match_len) {
943 if (newres.match2end == oldres.match2end)
944 return MatchResult::newIsBetter;
946 return MatchResult::newIsInvalid;
949 /** The class performing a match between a position in the document and the FindAdvOptions.
952 class MatchStringAdv {
954 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt);
956 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
957 ** constructor as opt.search, under the opt.* options settings.
960 ** If set, then match is searched only against beginning of text starting at cur.
961 ** If unset, then match is searched anywhere in text starting at cur.
964 ** The length of the matching text, or zero if no match was found.
966 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
975 /// first buffer on which search was started
976 lyx::Buffer * const p_first_buf;
978 FindAndReplaceOptions const & opt;
981 /// Auxiliary find method (does not account for opt.matchword)
982 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
983 void CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string = "");
985 /** Normalize a stringified or latexified LyX paragraph.
989 ** <li>if search is not casesensitive, then lowercase the string;
990 ** <li>remove any newline at begin or end of the string;
991 ** <li>replace any newline in the middle of the string with a simple space;
992 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
995 ** @todo Normalization should also expand macros, if the corresponding
996 ** search option was checked.
998 string normalize(docstring const & s) const;
999 // normalized string to search
1000 string par_as_string;
1001 // regular expression to use for searching
1002 // regexp2 is same as regexp, but prefixed with a ".*?"
1004 QRegularExpression regexp;
1005 QRegularExpression regexp2;
1010 // leading format material as string
1011 string lead_as_string;
1012 // par_as_string after removal of lead_as_string
1013 string par_as_string_nolead;
1014 // unmatched open braces in the search string/regexp
1016 // number of (.*?) subexpressions added at end of search regexp for closing
1017 // environments, math mode, styles, etc...
1018 int close_wildcards;
1020 // Are we searching with regular expressions ?
1022 static int valid_matches;
1023 static vector <string> matches;
1024 void FillResults(MatchResult &found_mr);
1027 int MatchStringAdv::valid_matches = 0;
1028 vector <string> MatchStringAdv::matches = vector <string>(10);
1030 void MatchStringAdv::FillResults(MatchResult &found_mr)
1032 if (found_mr.match_len > 0) {
1033 valid_matches = found_mr.result.size();
1034 for (size_t i = 0; i < found_mr.result.size(); i++)
1035 matches[i] = found_mr.result[i];
1041 static docstring buffer_to_latex(Buffer & buffer)
1043 //OutputParams runparams(&buffer.params().encoding());
1044 OutputParams runparams(encodings.fromLyXName("utf8"));
1045 odocstringstream ods;
1047 runparams.nice = true;
1048 runparams.flavor = Flavor::XeTeX;
1049 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
1050 // No side effect of file copying and image conversion
1051 runparams.dryrun = true;
1052 if (ignoreFormats.getDeleted())
1053 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
1055 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
1056 pit_type const endpit = buffer.paragraphs().size();
1057 for (pit_type pit = 0; pit != endpit; ++pit) {
1058 TeXOnePar(buffer, buffer.text(), pit, os, runparams);
1059 LYXERR(Debug::FIND, "searchString up to here: " << ods.str());
1065 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
1068 if (!opt.ignoreformat) {
1069 str = buffer_to_latex(buffer);
1071 // OutputParams runparams(&buffer.params().encoding());
1072 OutputParams runparams(encodings.fromLyXName("utf8"));
1073 runparams.nice = true;
1074 runparams.flavor = Flavor::XeTeX;
1075 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
1076 runparams.dryrun = true;
1077 int option = AS_STR_INSETS |AS_STR_PLAINTEXT;
1078 if (ignoreFormats.getDeleted()) {
1079 option |= AS_STR_SKIPDELETE;
1080 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
1083 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
1085 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
1086 Paragraph const & par = buffer.paragraphs().at(pit);
1087 LYXERR(Debug::FIND, "Adding to search string: '"
1088 << par.asString(pos_type(0), par.size(),
1092 str += par.asString(pos_type(0), par.size(),
1096 // Even in ignore-format we have to remove "\text{}, \lyxmathsym{}" parts
1097 string t = to_utf8(str);
1098 while (regex_replace(t, t, "\\\\(text|lyxmathsym|ensuremath)\\{([^\\}]*)\\}", "$2"));
1105 /// Return separation pos between the leading material and the rest
1106 static size_t identifyLeading(string const & s)
1109 // @TODO Support \item[text]
1110 // Kornel: Added textsl, textsf, textit, texttt and noun
1111 // + allow to search for colored text too
1112 while (regex_replace(t, t, "^\\\\(("
1113 "(author|title|subtitle|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|"
1114 "lyxaddress|lyxrightaddress|"
1115 "footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1116 "emph|noun|minisec|text(bf|md|sl|sf|it|tt))|"
1117 "((textcolor|foreignlanguage|latexenvironment)\\{[a-z]+\\*?\\})|"
1118 "(u|uu)line|(s|x)out|uwave)|((sub)?(((sub)?section)|paragraph)|part|chapter)\\*?)\\{", "")
1119 || regex_replace(t, t, "^\\$", "")
1120 || regex_replace(t, t, "^\\\\\\[", "")
1121 || regex_replace(t, t, "^ ?\\\\item\\{[a-z]+\\}", "")
1122 || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\}", ""))
1124 LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
1129 * Given a latexified string, retrieve some handled features
1130 * The features of the regex will later be compared with the features
1131 * of the searched text. If the regex features are not a
1132 * subset of the analized, then, in not format ignoring search
1133 * we can early stop the search in the relevant inset.
1135 typedef map<string, bool> Features;
1137 static Features identifyFeatures(string const & s)
1139 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
1140 static regex const valid("^("
1142 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1143 "emph|noun|text(bf|md|sl|sf|it|tt)|"
1144 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
1145 "(u|uu)line|(s|x)out|uwave|"
1146 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
1147 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
1152 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1155 if (sub.str(1).compare("regexp") == 0) {
1159 string token = sub.str(1);
1161 if (regex_match(token, sub2, valid)) {
1169 if (sub.str(1).compare("endregexp") == 0) {
1179 * defines values features of a key "\\[a-z]+{"
1184 /* Char type with content discarded
1185 * like \hspace{1cm} */
1187 /* Char, like \backslash */
1189 /* replace starting backslash with '#' */
1191 /* \part, \section*, ... */
1193 /* title, author etc */
1195 /* \foreignlanguage{ngerman}, ... */
1198 * to discard language in content */
1201 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1203 /* fonts, colors, markups, ... */
1205 /* footnotesize, ... large, ...
1206 * Ignore all of them */
1209 /* inputencoding, ...
1210 * Discard also content, because they do not help in search */
1213 * like remove, but also all arguments */
1215 /* item, listitem */
1217 /* tex, latex, ... like isChar */
1219 /* like \lettrine[lines=5]{}{} */
1221 // like isStandard, but always remove head
1223 /* End of arguments marker for lettrine,
1224 * so that they can be ignored */
1227 KeyInfo() = default;
1228 KeyInfo(KeyType type, int parcount, bool disable)
1230 parenthesiscount(parcount),
1231 disabled(disable) {}
1232 KeyType keytype = invalid;
1234 int _tokensize = -1;
1235 int _tokenstart = -1;
1236 int _dataStart = -1;
1238 int parenthesiscount = 1;
1239 bool disabled = false;
1240 bool used = false; /* by pattern */
1245 Border(int l=0, int u=0) : low(l), upper(u) {};
1250 #define MAXOPENED 30
1252 bool isPatternString_;
1254 explicit Intervall(bool isPattern, string const & p) :
1255 isPatternString_(isPattern), par(p), ignoreidx(-1), actualdeptindex(0),
1256 hasTitle(false), langcount(0)
1264 static vector<Border> borders;
1265 int depts[MAXOPENED];
1266 int closes[MAXOPENED];
1267 int actualdeptindex;
1268 int previousNotIgnored(int) const;
1269 int nextNotIgnored(int) const;
1270 void handleOpenP(int i);
1271 void handleCloseP(int i, bool closingAllowed);
1272 void resetOpenedP(int openPos);
1273 void addIntervall(int upper);
1274 void addIntervall(int low, int upper); /* if explicit */
1275 void removeAccents();
1276 void setForDefaultLang(KeyInfo const & defLang) const;
1277 int findclosing(int start, int end, char up, char down, int repeat);
1278 void handleParentheses(int lastpos, bool closingAllowed);
1280 int langcount; // Number of disabled language specs up to current position in actual interval
1281 int isOpeningPar(int pos) const;
1283 void output(ostringstream &os, int lastpos);
1284 // string show(int lastpos);
1287 vector<Border> Intervall::borders = vector<Border>(30);
1289 int Intervall::isOpeningPar(int pos) const
1291 if ((pos < 0) || (size_t(pos) >= par.size()))
1293 if (par[pos] != '{')
1295 if (size_t(pos) + 2 >= par.size())
1297 if (par[pos+2] != '}')
1299 if (par[pos+1] == '[' || par[pos+1] == ']')
1304 void Intervall::setForDefaultLang(KeyInfo const & defLang) const
1306 // Enable the use of first token again
1307 if (ignoreidx >= 0) {
1308 int value = defLang._tokenstart + defLang._tokensize;
1314 if (borders[borderidx].low < value)
1315 borders[borderidx].low = value;
1316 if (borders[borderidx].upper < value)
1317 borders[borderidx].upper = value;
1322 static void checkDepthIndex(int val)
1324 static int maxdepthidx = MAXOPENED-2;
1325 static int lastmaxdepth = 0;
1326 if (val > lastmaxdepth) {
1327 LYXERR(Debug::INFO, "Depth reached " << val);
1330 if (val > maxdepthidx) {
1332 LYXERR(Debug::INFO, "maxdepthidx now " << val);
1337 // Not needed, because borders are now dynamically expanded
1338 static void checkIgnoreIdx(int val)
1340 static int lastmaxignore = -1;
1341 if ((lastmaxignore < val) && (size_t(val+1) >= borders.size())) {
1342 LYXERR(Debug::INFO, "IgnoreIdx reached " << val);
1343 lastmaxignore = val;
1349 * Expand the region of ignored parts of the input latex string
1350 * The region is only relevant in output()
1352 void Intervall::addIntervall(int low, int upper)
1355 if (low == upper) return;
1356 for (idx = ignoreidx+1; idx > 0; --idx) {
1357 if (low > borders[idx-1].upper) {
1361 Border br(low, upper);
1362 if (idx > ignoreidx) {
1363 if (borders.size() <= size_t(idx)) {
1364 borders.push_back(br);
1370 // checkIgnoreIdx(ignoreidx);
1374 // Expand only if one of the new bound is inside the interwall
1375 // We know here that br.low > borders[idx-1].upper
1376 if (br.upper < borders[idx].low) {
1377 // We have to insert at this pos
1378 if (size_t(ignoreidx+1) >= borders.size()) {
1379 borders.push_back(borders[ignoreidx]);
1382 borders[ignoreidx+1] = borders[ignoreidx];
1384 for (int i = ignoreidx; i > idx; --i) {
1385 borders[i] = borders[i-1];
1389 // checkIgnoreIdx(ignoreidx);
1392 // Here we know, that we are overlapping
1393 if (br.low > borders[idx].low)
1394 br.low = borders[idx].low;
1395 // check what has to be concatenated
1397 for (int i = idx; i <= ignoreidx; i++) {
1398 if (br.upper >= borders[i].low) {
1400 if (br.upper < borders[i].upper)
1401 br.upper = borders[i].upper;
1407 // count should be >= 1 here
1410 for (int i = idx + count; i <= ignoreidx; i++) {
1411 borders[i-count+1] = borders[i];
1413 ignoreidx -= count - 1;
1419 static void buildaccent(string n, string param, string values)
1423 const char delim = '|';
1424 while (getline(s, name, delim)) {
1426 for (char c : param) {
1427 string key = name + "{" + c + "}";
1428 // get the corresponding utf8-value
1429 if ((values[start] & 0xc0) != 0xc0) {
1430 // should not happen, utf8 encoding starts at least with 11xxxxxx
1431 // but value for '\dot{i}' is 'i', which is ascii
1432 if ((values[start] & 0x80) == 0) {
1434 accents[key] = values.substr(start, 1);
1435 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1440 for (int j = 1; ;j++) {
1441 if (start + j >= values.size()) {
1442 accents[key] = values.substr(start, j);
1443 start = values.size() - 1;
1446 else if ((values[start+j] & 0xc0) != 0x80) {
1447 // This is the first byte of following utf8 char
1448 accents[key] = values.substr(start, j);
1450 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1459 static string getutf8(unsigned uchar)
1462 string ret = string();
1464 if (uchar <= 0x7f) {
1465 c[maxc-1] = uchar & 0x7f;
1468 unsigned char rest = 0x40;
1469 unsigned char first = 0x80;
1471 for (int i = start; i >=0; --i) {
1473 c[i] = first + uchar;
1476 c[i] = 0x80 | (uchar & 0x3f);
1483 for (int i = 0; i < maxc; i++) {
1484 if (c[i] == 0) continue;
1490 static void addAccents(string latex_in, string unicode_out)
1492 latex_in = latex_in.substr(1);
1493 AccentsIterator it_ac = accents.find(latex_in);
1494 if (it_ac == accents.end()) {
1495 accents[latex_in] = unicode_out;
1498 LYXERR0("Key " << latex_in << " already set");
1502 void static fillMissingUnicodesymbols()
1504 addAccents("\\textyen", getutf8(0x00a5));
1505 addAccents("\\yen", getutf8(0x00a5));
1506 addAccents("\\textsection", getutf8(0x00a7));
1507 addAccents("\\mathsection", getutf8(0x00a7));
1508 addAccents("\\textlnot", getutf8(0x00ac));
1509 addAccents("\\neg", getutf8(0x00ac));
1510 addAccents("\\textpm", getutf8(0x00b1));
1511 addAccents("\\pm", getutf8(0x00b1));
1512 addAccents("\\textparagraph", getutf8(0x00b6));
1513 addAccents("\\mathparagraph", getutf8(0x00b6));
1514 addAccents("\\textperiodcentered", getutf8(0x00b7));
1515 addAccents("\\texttimes", getutf8(0x00d7));
1516 addAccents("\\times", getutf8(0x00d7));
1517 addAccents("\\dh", getutf8(0x00f0));
1518 addAccents("\\eth", getutf8(0x00f0));
1519 addAccents("\\textdiv", getutf8(0x00f7));
1520 addAccents("\\div", getutf8(0x00f7));
1521 addAccents("\\o", getutf8(0x00f8));
1522 addAccents("\\j", getutf8(0x0237));
1523 addAccents("\\textalpha", getutf8(0x03b1));
1524 addAccents("\\alpha", getutf8(0x03b1));
1525 addAccents("\\textbeta", getutf8(0x03b2));
1526 addAccents("\\beta", getutf8(0x03b2));
1527 addAccents("\\textgamma", getutf8(0x03b3));
1528 addAccents("\\gamma", getutf8(0x03b3));
1529 addAccents("\\textdelta", getutf8(0x03b4));
1530 addAccents("\\delta", getutf8(0x03b4));
1531 addAccents("\\textepsilon", getutf8(0x03b5));
1532 addAccents("\\varepsilon", getutf8(0x03b5));
1533 addAccents("\\textzeta", getutf8(0x03b6));
1534 addAccents("\\zeta", getutf8(0x03b6));
1535 addAccents("\\texteta", getutf8(0x03b7));
1536 addAccents("\\eta", getutf8(0x03b7));
1537 addAccents("\\texttheta", getutf8(0x03b8));
1538 addAccents("\\theta", getutf8(0x03b8));
1539 addAccents("\\textiota", getutf8(0x03b9));
1540 addAccents("\\iota", getutf8(0x03b9));
1541 addAccents("\\textkappa", getutf8(0x03ba));
1542 addAccents("\\kappa", getutf8(0x03ba));
1543 addAccents("\\textlambda", getutf8(0x03bb));
1544 addAccents("\\lambda", getutf8(0x03bb));
1545 addAccents("\\textmu", getutf8(0x03bc));
1546 addAccents("\\mu", getutf8(0x03bc));
1547 addAccents("\\textnu", getutf8(0x03bd));
1548 addAccents("\\nu", getutf8(0x03bd));
1549 addAccents("\\textxi", getutf8(0x03be));
1550 addAccents("\\xi", getutf8(0x03be));
1551 addAccents("\\textpi", getutf8(0x03c0));
1552 addAccents("\\pi", getutf8(0x03c0));
1553 addAccents("\\textrho", getutf8(0x03c1));
1554 addAccents("\\rho", getutf8(0x03c1));
1555 addAccents("\\textfinalsigma", getutf8(0x03c2));
1556 addAccents("\\varsigma", getutf8(0x03c2));
1557 addAccents("\\textsigma", getutf8(0x03c3));
1558 addAccents("\\sigma", getutf8(0x03c3));
1559 addAccents("\\texttau", getutf8(0x03c4));
1560 addAccents("\\tau", getutf8(0x03c4));
1561 addAccents("\\textupsilon", getutf8(0x03c5));
1562 addAccents("\\upsilon", getutf8(0x03c5));
1563 addAccents("\\textphi", getutf8(0x03c6));
1564 addAccents("\\varphi", getutf8(0x03c6));
1565 addAccents("\\textchi", getutf8(0x03c7));
1566 addAccents("\\chi", getutf8(0x03c7));
1567 addAccents("\\textpsi", getutf8(0x03c8));
1568 addAccents("\\psi", getutf8(0x03c8));
1569 addAccents("\\textomega", getutf8(0x03c9));
1570 addAccents("\\omega", getutf8(0x03c9));
1571 addAccents("\\textdigamma", getutf8(0x03dd));
1572 addAccents("\\digamma", getutf8(0x03dd));
1573 addAccents("\\hebalef", getutf8(0x05d0));
1574 addAccents("\\aleph", getutf8(0x05d0));
1575 addAccents("\\hebbet", getutf8(0x05d1));
1576 addAccents("\\beth", getutf8(0x05d1));
1577 addAccents("\\hebgimel", getutf8(0x05d2));
1578 addAccents("\\gimel", getutf8(0x05d2));
1579 addAccents("\\hebdalet", getutf8(0x05d3));
1580 addAccents("\\daleth", getutf8(0x05d3));
1582 addAccents("\\thaiKoKai", getutf8(0x0e01));
1583 addAccents("\\thaiKhoKhai", getutf8(0x0e02));
1584 addAccents("\\thaiKhoKhuat", getutf8(0x0e03));
1585 addAccents("\\thaiKhoKhwai", getutf8(0x0e04));
1586 addAccents("\\thaiKhoKhon", getutf8(0x0e05));
1587 addAccents("\\thaiKhoRakhang", getutf8(0x0e06));
1588 addAccents("\\thaiNgoNgu", getutf8(0x0e07));
1589 addAccents("\\thaiChoChan", getutf8(0x0e08));
1590 addAccents("\\thaiChoChing", getutf8(0x0e09));
1591 addAccents("\\thaiChoChang", getutf8(0x0e0a));
1592 addAccents("\\thaiSoSo", getutf8(0x0e0b));
1593 addAccents("\\thaiChoChoe", getutf8(0x0e0c));
1594 addAccents("\\thaiYoYing", getutf8(0x0e0d));
1595 addAccents("\\thaiDoChada", getutf8(0x0e0e));
1596 addAccents("\\thaiToPatak", getutf8(0x0e0f));
1597 addAccents("\\thaiThoThan", getutf8(0x0e10));
1598 addAccents("\\thaiThoNangmontho", getutf8(0x0e11));
1599 addAccents("\\thaiThoPhuthao", getutf8(0x0e12));
1600 addAccents("\\thaiNoNen", getutf8(0x0e13));
1601 addAccents("\\thaiDoDek", getutf8(0x0e14));
1602 addAccents("\\thaiToTao", getutf8(0x0e15));
1603 addAccents("\\thaiThoThung", getutf8(0x0e16));
1604 addAccents("\\thaiThoThahan", getutf8(0x0e17));
1605 addAccents("\\thaiThoThong", getutf8(0x0e18));
1606 addAccents("\\thaiNoNu", getutf8(0x0e19));
1607 addAccents("\\thaiBoBaimai", getutf8(0x0e1a));
1608 addAccents("\\thaiPoPla", getutf8(0x0e1b));
1609 addAccents("\\thaiPhoPhung", getutf8(0x0e1c));
1610 addAccents("\\thaiFoFa", getutf8(0x0e1d));
1611 addAccents("\\thaiPhoPhan", getutf8(0x0e1e));
1612 addAccents("\\thaiFoFan", getutf8(0x0e1f));
1613 addAccents("\\thaiPhoSamphao", getutf8(0x0e20));
1614 addAccents("\\thaiMoMa", getutf8(0x0e21));
1615 addAccents("\\thaiYoYak", getutf8(0x0e22));
1616 addAccents("\\thaiRoRua", getutf8(0x0e23));
1617 addAccents("\\thaiRu", getutf8(0x0e24));
1618 addAccents("\\thaiLoLing", getutf8(0x0e25));
1619 addAccents("\\thaiLu", getutf8(0x0e26));
1620 addAccents("\\thaiWoWaen", getutf8(0x0e27));
1621 addAccents("\\thaiSoSala", getutf8(0x0e28));
1622 addAccents("\\thaiSoRusi", getutf8(0x0e29));
1623 addAccents("\\thaiSoSua", getutf8(0x0e2a));
1624 addAccents("\\thaiHoHip", getutf8(0x0e2b));
1625 addAccents("\\thaiLoChula", getutf8(0x0e2c));
1626 addAccents("\\thaiOAng", getutf8(0x0e2d));
1627 addAccents("\\thaiHoNokhuk", getutf8(0x0e2e));
1628 addAccents("\\thaiPaiyannoi", getutf8(0x0e2f));
1629 addAccents("\\thaiSaraA", getutf8(0x0e30));
1630 addAccents("\\thaiMaiHanakat", getutf8(0x0e31));
1631 addAccents("\\thaiSaraAa", getutf8(0x0e32));
1632 addAccents("\\thaiSaraAm", getutf8(0x0e33));
1633 addAccents("\\thaiSaraI", getutf8(0x0e34));
1634 addAccents("\\thaiSaraIi", getutf8(0x0e35));
1635 addAccents("\\thaiSaraUe", getutf8(0x0e36));
1636 addAccents("\\thaiSaraUee", getutf8(0x0e37));
1637 addAccents("\\thaiSaraU", getutf8(0x0e38));
1638 addAccents("\\thaiSaraUu", getutf8(0x0e39));
1639 addAccents("\\thaiPhinthu", getutf8(0x0e3a));
1640 addAccents("\\thaiSaraE", getutf8(0x0e40));
1641 addAccents("\\thaiSaraAe", getutf8(0x0e41));
1642 addAccents("\\thaiSaraO", getutf8(0x0e42));
1643 addAccents("\\thaiSaraAiMaimuan", getutf8(0x0e43));
1644 addAccents("\\thaiSaraAiMaimalai", getutf8(0x0e44));
1645 addAccents("\\thaiLakkhangyao", getutf8(0x0e45));
1646 addAccents("\\thaiMaiyamok", getutf8(0x0e46));
1647 addAccents("\\thaiMaitaikhu", getutf8(0x0e47));
1648 addAccents("\\thaiMaiEk", getutf8(0x0e48));
1649 addAccents("\\thaiMaiTho", getutf8(0x0e49));
1650 addAccents("\\thaiMaiTri", getutf8(0x0e4a));
1651 addAccents("\\thaiMaiChattawa", getutf8(0x0e4b));
1652 addAccents("\\thaiThanthakhat", getutf8(0x0e4c));
1653 addAccents("\\thaiNikhahit", getutf8(0x0e4d));
1654 addAccents("\\thaiYamakkan", getutf8(0x0e4e));
1655 addAccents("\\thaiFongman", getutf8(0x0e4f));
1656 addAccents("\\thaizero", getutf8(0x0e50));
1657 addAccents("\\thaione", getutf8(0x0e51));
1658 addAccents("\\thaitwo", getutf8(0x0e52));
1659 addAccents("\\thaithree", getutf8(0x0e53));
1660 addAccents("\\thaifour", getutf8(0x0e54));
1661 addAccents("\\thaifive", getutf8(0x0e55));
1662 addAccents("\\thaisix", getutf8(0x0e56));
1663 addAccents("\\thaiseven", getutf8(0x0e57));
1664 addAccents("\\thaieight", getutf8(0x0e58));
1665 addAccents("\\thainine", getutf8(0x0e59));
1666 addAccents("\\thaiAngkhankhu", getutf8(0x0e5a));
1667 addAccents("\\thaiKhomut", getutf8(0x0e5b));
1669 addAccents("\\dag", getutf8(0x2020));
1670 addAccents("\\dagger", getutf8(0x2020));
1671 addAccents("\\ddag", getutf8(0x2021));
1672 addAccents("\\ddagger", getutf8(0x2021));
1673 addAccents("\\textbullet", getutf8(0x2022));
1674 addAccents("\\bullet", getutf8(0x2022));
1675 addAccents("\\dots", getutf8(0x2026));
1676 addAccents("\\ldots", getutf8(0x2026));
1677 addAccents("\\textasciiacute", getutf8(0x2032));
1678 addAccents("\\prime", getutf8(0x2032));
1679 addAccents("\\textasciigrave", getutf8(0x2035));
1680 addAccents("\\backprime", getutf8(0x2035));
1681 addAccents("\\textasteriskcentered", getutf8(0x204e));
1682 addAccents("\\ast", getutf8(0x204e));
1683 addAccents("\\textmho", getutf8(0x2127));
1684 addAccents("\\mho", getutf8(0x2127));
1685 addAccents("\\textleftarrow", getutf8(0x2190));
1686 addAccents("\\leftarrow", getutf8(0x2190));
1687 addAccents("\\textuparrow", getutf8(0x2191));
1688 addAccents("\\uparrow", getutf8(0x2191));
1689 addAccents("\\textrightarrow", getutf8(0x2192));
1690 addAccents("\\rightarrow", getutf8(0x2192));
1691 addAccents("\\textdownarrow", getutf8(0x2193));
1692 addAccents("\\downarrow", getutf8(0x2193));
1693 addAccents("\\textglobrise", getutf8(0x2197));
1694 addAccents("\\nearrow", getutf8(0x2197));
1695 addAccents("\\textglobfall", getutf8(0x2198));
1696 addAccents("\\searrow", getutf8(0x2198));
1697 addAccents("\\textsurd", getutf8(0x221a));
1698 addAccents("\\surd", getutf8(0x221a));
1699 addAccents("\\textbigcircle", getutf8(0x25ef));
1700 addAccents("\\bigcirc", getutf8(0x25ef));
1701 addAccents("\\textlangle", getutf8(0x27e8));
1702 addAccents("\\langle", getutf8(0x27e8));
1703 addAccents("\\textrangle", getutf8(0x27e9));
1704 addAccents("\\rangle", getutf8(0x27e9));
1707 static void buildAccentsMap()
1709 accents["imath"] = "ı";
1711 accents["jmath"] = "ȷ";
1712 accents["cdot"] = "·";
1713 accents["textasciicircum"] = "^";
1714 accents["mathcircumflex"] = "^";
1715 accents["sim"] = "~";
1716 accents["guillemotright"] = "»";
1717 accents["guillemotleft"] = "«";
1718 accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
1719 accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
1720 accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
1721 accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
1722 accents["negmedspace"] = getutf8(0xf0005);
1723 accents["thickspace"] = getutf8(0xf0006);
1724 accents["negthickspace"] = getutf8(0xf0007);
1725 accents["lyx"] = getutf8(0xf0010); // Used logos
1726 accents["LyX"] = getutf8(0xf0010);
1727 accents["tex"] = getutf8(0xf0011);
1728 accents["TeX"] = getutf8(0xf0011);
1729 accents["latex"] = getutf8(0xf0012);
1730 accents["LaTeX"] = getutf8(0xf0012);
1731 accents["latexe"] = getutf8(0xf0013);
1732 accents["LaTeXe"] = getutf8(0xf0013);
1733 accents["lyxarrow"] = getutf8(0xf0020);
1734 accents["braceleft"] = getutf8(0xf0030);
1735 accents["braceright"] = getutf8(0xf0031);
1736 accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash
1737 accents["backslash LyX"] = getutf8(0xf0010);
1738 accents["backslash tex"] = getutf8(0xf0011);
1739 accents["backslash TeX"] = getutf8(0xf0011);
1740 accents["backslash latex"] = getutf8(0xf0012);
1741 accents["backslash LaTeX"] = getutf8(0xf0012);
1742 accents["backslash latexe"] = getutf8(0xf0013);
1743 accents["backslash LaTeXe"] = getutf8(0xf0013);
1744 accents["backslash lyxarrow"] = getutf8(0xf0020);
1745 accents["ddot{\\imath}"] = "ï";
1746 buildaccent("ddot", "aAeEhHiIioOtuUwWxXyY",
1747 "äÄëËḧḦïÏïöÖẗüÜẅẄẍẌÿŸ"); // umlaut
1748 buildaccent("dot|.", "aAbBcCdDeEfFGghHIimMnNoOpPrRsStTwWxXyYzZ",
1749 "ȧȦḃḂċĊḋḊėĖḟḞĠġḣḢİİṁṀṅṄȯȮṗṖṙṘṡṠṫṪẇẆẋẊẏẎżŻ"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'İ'
1750 accents["acute{\\imath}"] = "í";
1751 buildaccent("acute", "aAcCeEgGkKlLmMoOnNpPrRsSuUwWyYzZiI",
1752 "áÁćĆéÉǵǴḱḰĺĹḿḾóÓńŃṕṔŕŔśŚúÚẃẂýÝźŹíÍ");
1753 buildaccent("dacute|H|h", "oOuU", "őŐűŰ"); // double acute
1754 buildaccent("mathring|r", "aAuUwy",
1756 accents["check{\\imath}"] = "ǐ";
1757 accents["check{\\jmath}"] = "ǰ";
1758 buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
1759 "čČďĎǎǍěĚǐǏǒǑǔǓǧǦǩǨȟȞľĽňŇřŘšŠŤťžŽ"); // caron
1760 accents["hat{\\imath}"] = "î";
1761 accents["hat{\\jmath}"] = "ĵ";
1762 buildaccent("hat|^", "aAcCeEgGhHiIjJoOsSuUwWyYzZ",
1763 "âÂĉĈêÊĝĜĥĤîÎĵĴôÔŝŜûÛŵŴŷŶẑẐ"); // circ
1764 accents["bar{\\imath}"] = "ī";
1765 buildaccent("bar|=", "aAeEiIoOuUyY",
1766 "āĀēĒīĪōŌūŪȳȲ"); // macron
1767 accents["tilde{\\imath}"] = "ĩ";
1768 buildaccent("tilde", "aAeEiInNoOuUvVyY",
1769 "ãÃẽẼĩĨñÑõÕũŨṽṼỹỸ"); // tilde
1770 accents["breve{\\imath}"] = "ĭ";
1771 buildaccent("breve|u", "aAeEgGiIoOuU",
1772 "ăĂĕĔğĞĭĬŏŎŭŬ"); // breve
1773 accents["grave{\\imath}"] = "ì";
1774 buildaccent("grave|`", "aAeEiIoOuUnNwWyY",
1775 "àÀèÈìÌòÒùÙǹǸẁẀỳỲ"); // grave
1776 buildaccent("subdot|d", "BbDdHhKkLlMmNnRrSsTtVvWwZzAaEeIiOoUuYy",
1777 "ḄḅḌḍḤḥḲḳḶḷṂṃṆṇṚṛṢṣṬṭṾṿẈẉẒẓẠạẸẹỊịỌọỤụỴỵ"); // dot below
1778 buildaccent("ogonek|k", "AaEeIiUuOo",
1779 "ĄąĘęĮįŲųǪǫ"); // ogonek
1780 buildaccent("cedilla|c", "CcGgKkLlNnRrSsTtEeDdHh",
1781 "ÇçĢģĶķĻļŅņŖŗŞşŢţȨȩḐḑḨḩ"); // cedilla
1782 buildaccent("subring|textsubring", "Aa",
1784 buildaccent("subhat|textsubcircum", "DdEeLlNnTtUu",
1785 "ḒḓḘḙḼḽṊṋṰṱṶṷ"); // subcircum
1786 buildaccent("subtilde|textsubtilde", "EeIiUu",
1787 "ḚḛḬḭṴṵ"); // subtilde
1788 accents["dgrave{\\imath}"] = "ȉ";
1789 accents["textdoublegrave{\\i}"] = "ȉ";
1790 buildaccent("dgrave|textdoublegrave", "AaEeIiOoRrUu",
1791 "ȀȁȄȅȈȉȌȍȐȑȔȕ"); // double grave
1792 accents["rcap{\\imath}"] = "ȉ";
1793 accents["textroundcap{\\i}"] = "ȉ";
1794 buildaccent("rcap|textroundcap", "AaEeIiOoRrUu",
1795 "ȂȃȆȇȊȋȎȏȒȓȖȗ"); // inverted breve
1796 buildaccent("slashed", "oO",
1798 fillMissingUnicodesymbols(); // Add some still not handled entries contained in 'unicodesynbols'
1799 // LYXERR0("Number of accents " << accents.size());
1803 * Created accents in math or regexp environment
1804 * are macros, but we need the utf8 equivalent
1806 void Intervall::removeAccents()
1808 if (accents.empty())
1810 static regex const accre("\\\\(([\\S]|grave|breve|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot|ogonek|"
1811 "cedilla|subring|textsubring|subhat|textsubcircum|subtilde|textsubtilde|dgrave|textdoublegrave|rcap|textroundcap|slashed)\\{[^\\{\\}]+\\}"
1812 "|((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]))");
1814 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
1816 string key = sub.str(1);
1817 AccentsIterator it_ac = accents.find(key);
1818 if (it_ac != accents.end()) {
1819 string val = it_ac->second;
1820 size_t pos = sub.position(size_t(0));
1821 for (size_t i = 0; i < val.size(); i++) {
1822 par[pos+i] = val[i];
1824 // Remove possibly following space too
1825 if (par[pos+sub.str(0).size()] == ' ')
1826 addIntervall(pos+val.size(), pos + sub.str(0).size()+1);
1828 addIntervall(pos+val.size(), pos + sub.str(0).size());
1829 for (size_t i = pos+val.size(); i < pos + sub.str(0).size(); i++) {
1830 // remove traces of any remaining chars
1835 LYXERR(Debug::INFO, "Not added accent for \"" << key << "\"");
1840 void Intervall::handleOpenP(int i)
1843 depts[actualdeptindex] = i+1;
1844 closes[actualdeptindex] = -1;
1845 checkDepthIndex(actualdeptindex);
1848 void Intervall::handleCloseP(int i, bool closingAllowed)
1850 if (actualdeptindex <= 0) {
1851 if (! closingAllowed)
1852 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1853 // if we are at the very end
1854 addIntervall(i, i+1);
1857 closes[actualdeptindex] = i+1;
1862 void Intervall::resetOpenedP(int openPos)
1864 // Used as initializer for foreignlanguage entry
1865 actualdeptindex = 1;
1866 depts[1] = openPos+1;
1870 int Intervall::previousNotIgnored(int start) const
1872 int idx = 0; /* int intervalls */
1873 for (idx = ignoreidx; idx >= 0; --idx) {
1874 if (start > borders[idx].upper)
1876 if (start >= borders[idx].low)
1877 start = borders[idx].low-1;
1882 int Intervall::nextNotIgnored(int start) const
1884 int idx = 0; /* int intervalls */
1885 for (idx = 0; idx <= ignoreidx; idx++) {
1886 if (start < borders[idx].low)
1888 if (start < borders[idx].upper)
1889 start = borders[idx].upper;
1894 typedef unordered_map<string, KeyInfo> KeysMap;
1895 typedef unordered_map<string, KeyInfo>::const_iterator KeysIterator;
1896 typedef vector< KeyInfo> Entries;
1897 static KeysMap keys = unordered_map<string, KeyInfo>();
1903 Intervall interval_;
1904 void buildKeys(bool);
1905 void buildEntries(bool);
1906 void makeKey(const string &, KeyInfo, bool isPatternString);
1907 void processRegion(int start, int region_end); /* remove {} parts */
1908 void removeHead(KeyInfo const &, int count=0);
1911 LatexInfo(string const & par, bool isPatternString)
1912 : entidx_(-1), interval_(isPatternString, par)
1914 buildKeys(isPatternString);
1915 entries_ = vector<KeyInfo>();
1916 buildEntries(isPatternString);
1920 if (entries_.empty()) {
1923 if (entries_[0].keytype == KeyInfo::isTitle) {
1924 interval_.hasTitle = true;
1925 if (! entries_[0].disabled) {
1926 interval_.titleValue = entries_[0].head;
1929 interval_.titleValue = "";
1931 removeHead(entries_[0]);
1932 if (entries_.size() > 1)
1941 if (int(entries_.size()) > entidx_) {
1948 bool setNextKey(int idx) {
1949 if ((idx == entidx_) && (entidx_ >= 0)) {
1956 int find(int start, KeyInfo::KeyType keytype) const {
1960 while (tmpIdx < int(entries_.size())) {
1961 if (entries_[tmpIdx].keytype == keytype)
1967 int process(ostringstream & os, KeyInfo const & actual);
1968 int dispatch(ostringstream & os, int previousStart, KeyInfo & actual);
1969 // string show(int lastpos) { return interval.show(lastpos);};
1970 int nextNotIgnored(int start) { return interval_.nextNotIgnored(start);};
1971 KeyInfo &getKeyInfo(int keyinfo) {
1972 static KeyInfo invalidInfo = KeyInfo();
1973 if ((keyinfo < 0) || ( keyinfo >= int(entries_.size())))
1976 return entries_[keyinfo];
1978 void setForDefaultLang(KeyInfo const & defLang) {interval_.setForDefaultLang(defLang);};
1979 void addIntervall(int low, int up) { interval_.addIntervall(low, up); };
1983 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
1987 for (int i = start; i < end; i += 1 + skip) {
1991 if (c == '\\') skip = 1;
1995 else if (c == down) {
1998 if ((repeat <= 0) || (par[i+1] != up))
2012 size_t mathpostfixsize;
2014 size_t mathprefixsize;
2018 vector<MathEntry> entries_;
2023 void insert(string const & wait, size_t start, size_t prefixsize, size_t end, size_t postfixsize) {
2024 MathEntry m = MathEntry();
2026 m.mathStart = start;
2027 m.mathprefixsize = prefixsize;
2028 m.mathEnd = end + postfixsize;
2029 m.mathpostfixsize = postfixsize;
2030 m.mathSize = m.mathEnd - m.mathStart;
2031 entries_.push_back(m);
2033 bool empty() const { return entries_.empty(); };
2034 size_t getEndPos() const {
2035 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2038 return entries_[actualIdx_].mathEnd;
2040 size_t getStartPos() const {
2041 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2042 return 100000; /* definitely enough? */
2044 return entries_[actualIdx_].mathStart;
2046 size_t getPrefixSize() const {
2047 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2050 return entries_[actualIdx_].mathprefixsize;
2052 size_t getPostfixSize() const {
2053 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2056 return entries_[actualIdx_].mathpostfixsize;
2058 size_t getFirstPos() {
2060 return getStartPos();
2062 size_t getSize() const {
2063 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2066 return entries_[actualIdx_].mathSize;
2068 void incrEntry() { actualIdx_++; };
2071 void LatexInfo::buildEntries(bool isPatternString)
2073 static regex const rmath("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|x?x?alignat)\\*?\\})(\\{[0-9]+\\})?)");
2074 static regex const rkeys("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?)))");
2075 static bool disableLanguageOverride = false;
2076 smatch sub, submath;
2077 bool evaluatingRegexp = false;
2079 bool evaluatingMath = false;
2080 bool evaluatingCode = false;
2082 bool evaluatingOptional = false;
2083 size_t optionalEnd = 0;
2086 bool math_end_waiting = false;
2087 size_t math_pos = 10000;
2088 size_t math_prefix_size = 1;
2090 static vector<string> usedText = vector<string>();
2091 static bool removeMathHull = false;
2093 interval_.removeAccents();
2095 for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) {
2097 if ((submath.position(2) - submath.position(0)) %2 == 1) {
2098 // prefixed by odd count of '\\'
2101 if (math_end_waiting) {
2102 size_t pos = submath.position(size_t(2));
2103 if ((math_end == "$") &&
2104 (submath.str(2) == "$")) {
2105 mi.insert("$", math_pos, 1, pos, 1);
2106 math_end_waiting = false;
2108 else if ((math_end == "\\]") &&
2109 (submath.str(2) == "\\]")) {
2110 mi.insert("\\]", math_pos, 2, pos, 2);
2111 math_end_waiting = false;
2113 else if ((submath.str(3).compare("end") == 0) &&
2114 (submath.str(5).compare(math_end) == 0)) {
2115 mi.insert(math_end, math_pos, math_prefix_size, pos, submath.str(2).length());
2116 math_end_waiting = false;
2122 if (submath.str(3).compare("begin") == 0) {
2123 math_end_waiting = true;
2124 math_end = submath.str(5);
2125 math_pos = submath.position(size_t(2));
2126 math_prefix_size = submath.str(2).length();
2128 else if (submath.str(2).compare("\\[") == 0) {
2129 math_end_waiting = true;
2131 math_pos = submath.position(size_t(2));
2133 else if (submath.str(2) == "$") {
2134 size_t pos = submath.position(size_t(2));
2135 math_end_waiting = true;
2141 // Ignore language if there is math somewhere in pattern-string
2142 if (isPatternString) {
2143 for (auto s: usedText) {
2144 // Remove entries created in previous search runs
2147 usedText = vector<string>();
2150 keys["foreignlanguage"].disabled = true;
2151 disableLanguageOverride = true;
2152 removeMathHull = false;
2155 removeMathHull = true; // used later if not isPatternString
2156 disableLanguageOverride = false;
2160 if (disableLanguageOverride) {
2161 keys["foreignlanguage"].disabled = true;
2164 math_pos = mi.getFirstPos();
2165 for (sregex_iterator it(interval_.par.begin(), interval_.par.end(), rkeys), end; it != end; ++it) {
2167 if ((sub.position(2) - sub.position(0)) %2 == 1) {
2168 // prefixed by odd count of '\\'
2171 string key = sub.str(5);
2173 if (sub.str(2)[0] == '\\')
2174 key = sub.str(2)[1];
2179 KeysIterator it_key = keys.find(key);
2180 if (it_key != keys.end()) {
2181 if (it_key->second.keytype == KeyInfo::headRemove) {
2182 KeyInfo found1 = it_key->second;
2183 found1.disabled = true;
2184 found1.head = "\\" + key + "{";
2185 found1._tokenstart = sub.position(size_t(2));
2186 found1._tokensize = found1.head.length();
2187 found1._dataStart = found1._tokenstart + found1.head.length();
2188 int endpos = interval_.findclosing(found1._dataStart, interval_.par.length(), '{', '}', 1);
2189 found1._dataEnd = endpos;
2194 if (evaluatingRegexp) {
2195 if (sub.str(3).compare("endregexp") == 0) {
2196 evaluatingRegexp = false;
2197 // found._tokenstart already set
2198 found._dataEnd = sub.position(size_t(2)) + 13;
2199 found._dataStart = found._dataEnd;
2200 found._tokensize = found._dataEnd - found._tokenstart;
2201 found.parenthesiscount = 0;
2202 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2209 if (evaluatingMath) {
2210 if (size_t(sub.position(size_t(2))) < mi.getEndPos())
2212 evaluatingMath = false;
2214 math_pos = mi.getStartPos();
2216 if (it_key == keys.end()) {
2217 found = KeyInfo(KeyInfo::isStandard, 0, true);
2218 LYXERR(Debug::INFO, "Undefined key " << key << " ==> will be used as text");
2219 found = KeyInfo(KeyInfo::isText, 0, false);
2220 if (isPatternString) {
2221 found.keytype = KeyInfo::isChar;
2222 found.disabled = false;
2226 usedText.push_back(key);
2230 if (key.compare("regexp") == 0) {
2231 evaluatingRegexp = true;
2232 found._tokenstart = sub.position(size_t(2));
2233 found._tokensize = 0;
2237 // Handle the other params of key
2238 if (found.keytype == KeyInfo::isIgnored)
2240 else if (found.keytype == KeyInfo::isMath) {
2241 if (size_t(sub.position(size_t(2))) == math_pos) {
2243 found._tokenstart = sub.position(size_t(2));
2244 found._tokensize = mi.getSize();
2245 found._dataEnd = found._tokenstart + found._tokensize;
2246 found._dataStart = found._dataEnd;
2247 found.parenthesiscount = 0;
2248 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2249 if (removeMathHull) {
2250 interval_.addIntervall(found._tokenstart, found._tokenstart + mi.getPrefixSize());
2251 interval_.addIntervall(found._dataEnd - mi.getPostfixSize(), found._dataEnd);
2254 // Treate all math constructs as simple math
2255 interval_.par[found._tokenstart] = '$';
2256 interval_.par[found._dataEnd - mi.getPostfixSize()] = '$';
2257 interval_.addIntervall(found._tokenstart + 1, found._tokenstart + mi.getPrefixSize());
2258 interval_.addIntervall(found._dataEnd - mi.getPostfixSize() + 1, found._dataEnd);
2260 evaluatingMath = true;
2263 // begin|end of unknown env, discard
2264 // First handle tables
2265 // longtable|tabular
2266 bool discardComment;
2268 found.keytype = KeyInfo::doRemove;
2269 if ((sub.str(7).compare("longtable") == 0) ||
2270 (sub.str(7).compare("tabular") == 0)) {
2271 discardComment = true; /* '%' */
2274 discardComment = false;
2275 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
2277 string token = sub.str(7);
2278 if (regex_match(token, sub2, removeArgs)) {
2279 found.keytype = KeyInfo::removeWithArg;
2282 // discard spaces before pos(2)
2283 int pos = sub.position(size_t(2));
2285 for (count = 0; pos - count > 0; count++) {
2286 char c = interval_.par[pos-count-1];
2287 if (discardComment) {
2288 if ((c != ' ') && (c != '%'))
2294 found._tokenstart = pos - count;
2295 if (sub.str(3).compare(0, 5, "begin") == 0) {
2296 size_t pos1 = pos + sub.str(2).length();
2297 if (sub.str(7).compare("cjk") == 0) {
2298 pos1 = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2299 if ((interval_.par[pos1] == '{') && (interval_.par[pos1+1] == '}'))
2301 found.keytype = KeyInfo::isMain;
2302 found._dataStart = pos1;
2303 found._dataEnd = interval_.par.length();
2304 found.disabled = keys["foreignlanguage"].disabled;
2305 found.used = keys["foreignlanguage"].used;
2306 found._tokensize = pos1 - found._tokenstart;
2307 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2310 // Swallow possible optional params
2311 while (interval_.par[pos1] == '[') {
2312 pos1 = interval_.findclosing(pos1+1, interval_.par.length(), '[', ']')+1;
2314 // Swallow also the eventual parameter
2315 if (interval_.par[pos1] == '{') {
2316 found._dataEnd = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2319 found._dataEnd = pos1;
2321 found._dataStart = found._dataEnd;
2322 found._tokensize = count + found._dataEnd - pos;
2323 found.parenthesiscount = 0;
2324 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2325 found.disabled = true;
2329 // Handle "\end{...}"
2330 found._dataStart = pos + sub.str(2).length();
2331 found._dataEnd = found._dataStart;
2332 found._tokensize = count + found._dataEnd - pos;
2333 found.parenthesiscount = 0;
2334 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2335 found.disabled = true;
2339 else if (found.keytype != KeyInfo::isRegex) {
2340 found._tokenstart = sub.position(size_t(2));
2341 if (found.parenthesiscount == 0) {
2342 // Probably to be discarded
2343 size_t following_pos = sub.position(size_t(2)) + sub.str(5).length() + 1;
2344 char following = interval_.par[following_pos];
2345 if (following == ' ')
2346 found.head = "\\" + sub.str(5) + " ";
2347 else if (following == '=') {
2348 // like \uldepth=1000pt
2349 found.head = sub.str(2);
2352 found.head = "\\" + key;
2353 found._tokensize = found.head.length();
2354 found._dataEnd = found._tokenstart + found._tokensize;
2355 found._dataStart = found._dataEnd;
2358 int params = found._tokenstart + key.length() + 1;
2359 if (evaluatingOptional) {
2360 if (size_t(found._tokenstart) > optionalEnd) {
2361 evaluatingOptional = false;
2364 found.disabled = true;
2367 int optend = params;
2368 while (interval_.par[optend] == '[') {
2369 // discard optional parameters
2370 optend = interval_.findclosing(optend+1, interval_.par.length(), '[', ']') + 1;
2372 if (optend > params) {
2373 key += interval_.par.substr(params, optend-params);
2374 evaluatingOptional = true;
2375 optionalEnd = optend;
2376 if (found.keytype == KeyInfo::isSectioning) {
2377 // Remove optional values (but still keep in header)
2378 interval_.addIntervall(params, optend);
2381 string token = sub.str(7);
2383 if (interval_.par[optend] != '{') {
2385 found.parenthesiscount = 0;
2386 found.head = "\\" + key;
2389 closings = found.parenthesiscount;
2390 if (found.parenthesiscount == 1) {
2391 found.head = "\\" + key + "{";
2393 else if (found.parenthesiscount > 1) {
2395 found.head = sub.str(2) + "{";
2396 closings = found.parenthesiscount - 1;
2399 found.head = "\\" + key + "{";
2402 found._tokensize = found.head.length();
2403 found._dataStart = found._tokenstart + found.head.length();
2404 if (found.keytype == KeyInfo::doRemove) {
2406 size_t endpar = 2 + interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2407 if (endpar >= interval_.par.length())
2408 found._dataStart = interval_.par.length();
2410 found._dataStart = endpar;
2411 found._tokensize = found._dataStart - found._tokenstart;
2414 found._dataStart = found._tokenstart + found._tokensize;
2418 if (interval_.par.substr(found._dataStart, 15).compare("\\endarguments{}") == 0) {
2419 found._dataStart += 15;
2423 endpos = found._dataStart - 1;
2425 endpos = interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2426 if (found.keytype == KeyInfo::isList) {
2427 // Check if it really is list env
2428 static regex const listre("^([a-z]+)$");
2430 if (!regex_match(token, sub2, listre)) {
2431 // Change the key of this entry. It is not in a list/item environment
2432 found.keytype = KeyInfo::endArguments;
2435 if (found.keytype == KeyInfo::noMain) {
2436 evaluatingCode = true;
2438 codeStart = found._dataStart;
2440 else if (evaluatingCode) {
2441 if (size_t(found._dataStart) > codeEnd)
2442 evaluatingCode = false;
2443 else if (found.keytype == KeyInfo::isMain) {
2444 // Disable this key, treate it as standard
2445 found.keytype = KeyInfo::isStandard;
2446 found.disabled = true;
2447 if ((codeEnd +1 >= interval_.par.length()) &&
2448 (found._tokenstart == codeStart)) {
2449 // trickery, because the code inset starts
2450 // with \selectlanguage ...
2452 if (entries_.size() > 1) {
2453 entries_[entries_.size()-1]._dataEnd = codeEnd;
2458 if ((endpos == interval_.par.length()) &&
2459 (found.keytype == KeyInfo::doRemove)) {
2460 // Missing closing => error in latex-input?
2461 // therefore do not delete remaining data
2462 found._dataStart -= 1;
2463 found._dataEnd = found._dataStart;
2466 found._dataEnd = endpos;
2468 if (isPatternString) {
2469 keys[key].used = true;
2472 entries_.push_back(found);
2476 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
2478 stringstream s(keysstring);
2480 const char delim = '|';
2481 while (getline(s, key, delim)) {
2482 KeyInfo keyII(keyI);
2483 if (isPatternString) {
2486 else if ( !keys[key].used)
2487 keyII.disabled = true;
2492 void LatexInfo::buildKeys(bool isPatternString)
2495 static bool keysBuilt = false;
2496 if (keysBuilt && !isPatternString) return;
2498 // Keys to ignore in any case
2499 makeKey("text|textcyrillic|lyxmathsym|ensuremath", KeyInfo(KeyInfo::headRemove, 1, true), true);
2500 // Known standard keys with 1 parameter.
2501 // Split is done, if not at start of region
2502 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
2503 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
2504 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
2505 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
2506 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
2507 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
2509 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
2510 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2511 makeKey("section*|subsection*|subsubsection*|paragraph*",
2512 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2513 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2514 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
2516 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
2518 // Split is done, if not at start of region
2519 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
2520 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
2522 // Split is done always.
2523 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
2527 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2528 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2529 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2530 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2532 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2533 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2534 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2535 makeKey("thickspace|medspace|thinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2537 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2538 // Custom space/skip, remove the content (== length value)
2539 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
2540 // Found in fr/UserGuide.lyx
2541 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2543 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2544 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2545 // Known macros to remove (including their parameter)
2547 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
2548 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
2549 // handle like standard keys with 1 parameter.
2550 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
2552 // Ignore deleted text
2553 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString);
2554 // but preserve added text
2555 makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2557 // Macros to remove, but let the parameter survive
2559 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2561 // Remove language spec from content of these insets
2562 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
2564 // Same effect as previous, parameter will survive (because there is no one anyway)
2566 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2567 // Remove table decorations
2568 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
2569 // Discard shape-header.
2570 // For footnote or shortcut too, because of lang settings
2571 // and wrong handling if used 'KeyInfo::noMain'
2572 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2573 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2574 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2575 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2576 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2577 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
2578 // like ('tiny{}' or '\tiny ' ... )
2579 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
2581 // Survives, like known character
2582 // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2583 makeKey("tableofcontents", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2584 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
2586 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2587 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2588 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2590 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2591 // Remove RTL/LTR marker
2592 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2593 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
2594 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
2595 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
2596 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
2597 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
2598 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2599 if (isPatternString) {
2600 // Allow the first searched string to rebuild the keys too
2604 // no need to rebuild again
2610 * Keep the list of actual opened parentheses actual
2611 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2613 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2616 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2620 if (c == '\\') skip = 1;
2621 else if (c == '{') {
2624 else if (c == '}') {
2625 handleCloseP(i, closingAllowed);
2631 string Intervall::show(int lastpos)
2633 int idx = 0; /* int intervalls */
2636 for (idx = 0; idx <= ignoreidx; idx++) {
2637 while (i < lastpos) {
2639 if (i <= borders[idx].low) {
2640 if (borders[idx].low > lastpos)
2641 printsize = lastpos - i;
2643 printsize = borders[idx].low - i;
2644 s += par.substr(i, printsize);
2646 if (i >= borders[idx].low)
2647 i = borders[idx].upper;
2650 i = borders[idx].upper;
2656 s += par.substr(i, lastpos-i);
2662 void Intervall::output(ostringstream &os, int lastpos)
2664 // get number of chars to output
2665 int idx = 0; /* int intervalls */
2668 string startTitle = titleValue;
2669 for (idx = 0; idx <= ignoreidx; idx++) {
2671 if (i <= borders[idx].low) {
2673 if (borders[idx].low > lastpos)
2674 printsize = lastpos - i;
2676 printsize = borders[idx].low - i;
2677 if (printsize > 0) {
2678 os << startTitle << par.substr(i, printsize);
2680 printed += printsize;
2683 handleParentheses(i, false);
2684 if (i >= borders[idx].low)
2685 i = borders[idx].upper;
2688 i = borders[idx].upper;
2695 os << startTitle << par.substr(i, lastpos-i);
2696 printed += lastpos-i;
2698 handleParentheses(lastpos, false);
2700 if (keys["foreignlanguage"].disabled)
2701 startindex = actualdeptindex-langcount;
2703 startindex = actualdeptindex;
2704 for (int i = startindex; i > 0; --i) {
2707 if (hasTitle && (printed > 0))
2709 if (! isPatternString_)
2711 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2714 void LatexInfo::processRegion(int start, int region_end)
2716 while (start < region_end) { /* Let {[} and {]} survive */
2717 int cnt = interval_.isOpeningPar(start);
2719 // Closing is allowed past the region
2720 int closing = interval_.findclosing(start+1, interval_.par.length());
2721 interval_.addIntervall(start, start+1);
2722 interval_.addIntervall(closing, closing+1);
2726 start = interval_.nextNotIgnored(start+1);
2730 void LatexInfo::removeHead(KeyInfo const & actual, int count)
2732 if (actual.parenthesiscount == 0) {
2733 // "{\tiny{} ...}" ==> "{{} ...}"
2734 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2737 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2738 interval_.addIntervall(actual._tokenstart - count, actual._dataStart);
2739 interval_.addIntervall(actual._dataEnd, actual._dataEnd+1);
2743 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2746 switch (actual.keytype)
2748 case KeyInfo::isTitle: {
2750 nextKeyIdx = getNextKey();
2753 case KeyInfo::cleanToStart: {
2754 actual._dataEnd = actual._dataStart;
2755 nextKeyIdx = getNextKey();
2756 // Search for end of arguments
2757 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2759 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2760 entries_[i].disabled = true;
2762 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2764 while (interval_.par[actual._dataEnd] == ' ')
2766 interval_.addIntervall(0, actual._dataEnd+1);
2767 interval_.actualdeptindex = 0;
2768 interval_.depts[0] = actual._dataEnd+1;
2769 interval_.closes[0] = -1;
2772 case KeyInfo::isText:
2773 interval_.par[actual._tokenstart] = '#';
2774 //interval_.addIntervall(actual._tokenstart, actual._tokenstart+1);
2775 nextKeyIdx = getNextKey();
2777 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2778 if (actual.disabled)
2779 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2781 interval_.addIntervall(actual._dataStart, actual._dataEnd);
2784 case KeyInfo::isChar: {
2785 nextKeyIdx = getNextKey();
2788 case KeyInfo::isSize: {
2789 if (actual.disabled || (interval_.par[actual._dataStart] != '{') || (interval_.par[actual._dataStart-1] == ' ')) {
2790 if (actual.parenthesiscount == 0)
2791 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2793 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2795 nextKeyIdx = getNextKey();
2797 // Here _dataStart points to '{', so correct it
2798 actual._dataStart += 1;
2799 actual._tokensize += 1;
2800 actual.parenthesiscount = 1;
2801 if (interval_.par[actual._dataStart] == '}') {
2802 // Determine the end if used like '{\tiny{}...}'
2803 actual._dataEnd = interval_.findclosing(actual._dataStart+1, interval_.par.length()) + 1;
2804 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2807 // Determine the end if used like '\tiny{...}'
2808 actual._dataEnd = interval_.findclosing(actual._dataStart, interval_.par.length()) + 1;
2810 // Split on this key if not at start
2811 int start = interval_.nextNotIgnored(previousStart);
2812 if (start < actual._tokenstart) {
2813 interval_.output(os, actual._tokenstart);
2814 interval_.addIntervall(start, actual._tokenstart);
2816 // discard entry if at end of actual
2817 nextKeyIdx = process(os, actual);
2821 case KeyInfo::endArguments: {
2822 // Remove trailing '{}' too
2823 actual._dataStart += 1;
2824 actual._dataEnd += 1;
2825 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2826 nextKeyIdx = getNextKey();
2829 case KeyInfo::noMain:
2831 case KeyInfo::isStandard: {
2832 if (actual.disabled) {
2834 processRegion(actual._dataStart, actual._dataStart+1);
2835 nextKeyIdx = getNextKey();
2837 // Split on this key if not at datastart of calling entry
2838 int start = interval_.nextNotIgnored(previousStart);
2839 if (start < actual._tokenstart) {
2840 interval_.output(os, actual._tokenstart);
2841 interval_.addIntervall(start, actual._tokenstart);
2843 // discard entry if at end of actual
2844 nextKeyIdx = process(os, actual);
2848 case KeyInfo::removeWithArg: {
2849 nextKeyIdx = getNextKey();
2850 // Search for end of arguments
2851 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2853 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2854 entries_[i].disabled = true;
2856 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2858 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2861 case KeyInfo::doRemove: {
2862 // Remove the key with all parameters and following spaces
2865 if (interval_.par[actual._dataEnd-1] == ' ')
2866 start = actual._dataEnd;
2868 start = actual._dataEnd+1;
2869 for (pos = start; pos < interval_.par.length(); pos++) {
2870 if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
2873 // Remove also enclosing parentheses [] and {}
2876 while (actual._tokenstart > numpars) {
2877 if (pos+numpars >= interval_.par.size())
2879 else if (interval_.par[pos+numpars] == ']' && interval_.par[actual._tokenstart-numpars-1] == '[')
2881 else if (interval_.par[pos+numpars] == '}' && interval_.par[actual._tokenstart-numpars-1] == '{')
2887 if (interval_.par[pos+numpars] == ' ')
2891 interval_.addIntervall(actual._tokenstart-numpars, pos+numpars+spaces);
2892 nextKeyIdx = getNextKey();
2895 case KeyInfo::isList: {
2896 // Discard space before _tokenstart
2898 for (count = 0; count < actual._tokenstart; count++) {
2899 if (interval_.par[actual._tokenstart-count-1] != ' ')
2902 nextKeyIdx = getNextKey();
2903 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2905 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2907 // How else can we catch this one?
2908 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2909 entries_[i].disabled = true;
2911 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2913 else if (nextKeyIdx > 0) {
2914 // Ignore any lang entries inside data region
2915 for (int i = nextKeyIdx; i < int(entries_.size()) && entries_[i]._tokenstart < actual._dataEnd; i++) {
2916 if (entries_[i].keytype == KeyInfo::isMain)
2917 entries_[i].disabled = true;
2920 if (actual.disabled) {
2921 interval_.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2924 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2926 if (interval_.par[actual._dataEnd+1] == '[') {
2927 int posdown = interval_.findclosing(actual._dataEnd+2, interval_.par.length(), '[', ']');
2928 if ((interval_.par[actual._dataEnd+2] == '{') &&
2929 (interval_.par[posdown-1] == '}')) {
2930 interval_.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
2931 interval_.addIntervall(posdown-1, posdown+1);
2934 interval_.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
2935 interval_.addIntervall(posdown, posdown+1);
2937 int blk = interval_.nextNotIgnored(actual._dataEnd+1);
2938 if (blk > posdown) {
2939 // Discard at most 1 space after empty item
2941 for (count = 0; count < 1; count++) {
2942 if (interval_.par[blk+count] != ' ')
2946 interval_.addIntervall(blk, blk+count);
2951 case KeyInfo::isSectioning: {
2952 // Discard spaces before _tokenstart
2954 int val = actual._tokenstart;
2955 for (count = 0; count < actual._tokenstart;) {
2956 val = interval_.previousNotIgnored(val-1);
2957 if (val < 0 || interval_.par[val] != ' ')
2960 count = actual._tokenstart - val;
2963 if (actual.disabled) {
2964 removeHead(actual, count);
2965 nextKeyIdx = getNextKey();
2967 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2968 nextKeyIdx = process(os, actual);
2972 case KeyInfo::isMath: {
2973 // Same as regex, use the content unchanged
2974 nextKeyIdx = getNextKey();
2977 case KeyInfo::isRegex: {
2978 // DO NOT SPLIT ON REGEX
2980 nextKeyIdx = getNextKey();
2983 case KeyInfo::isIgnored: {
2984 // Treat like a character for now
2985 nextKeyIdx = getNextKey();
2988 case KeyInfo::isMain: {
2989 if (interval_.par.substr(actual._dataStart, 2) == "% ")
2990 interval_.addIntervall(actual._dataStart, actual._dataStart+2);
2991 if (actual._tokenstart > 0) {
2992 int prev = interval_.previousNotIgnored(actual._tokenstart - 1);
2993 if ((prev >= 0) && interval_.par[prev] == '%')
2994 interval_.addIntervall(prev, prev+1);
2996 if (actual.disabled) {
2998 interval_.langcount++;
2999 if ((interval_.par.substr(actual._dataStart, 3) == " \\[") ||
3000 (interval_.par.substr(actual._dataStart, 8) == " \\begin{")) {
3001 // Discard also the space before math-equation
3002 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
3004 nextKeyIdx = getNextKey();
3005 // interval.resetOpenedP(actual._dataStart-1);
3008 if (actual._tokenstart < 26) {
3009 // for the first (and maybe dummy) language
3010 interval_.setForDefaultLang(actual);
3012 interval_.resetOpenedP(actual._dataStart-1);
3016 case KeyInfo::invalid:
3017 case KeyInfo::headRemove:
3018 // These two cases cannot happen, already handled
3021 // LYXERR(Debug::INFO, "Unhandled keytype");
3022 nextKeyIdx = getNextKey();
3029 int LatexInfo::process(ostringstream & os, KeyInfo const & actual )
3031 int end = interval_.nextNotIgnored(actual._dataEnd);
3032 int oldStart = actual._dataStart;
3033 int nextKeyIdx = getNextKey();
3035 if ((nextKeyIdx < 0) ||
3036 (entries_[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
3037 (entries_[nextKeyIdx].keytype == KeyInfo::invalid)) {
3038 if (oldStart <= end) {
3039 processRegion(oldStart, end);
3044 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
3046 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
3047 (void) dispatch(os, actual._dataStart, nextKey);
3048 end = nextKey._tokenstart;
3051 processRegion(oldStart, nextKey._tokenstart);
3052 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
3054 oldStart = nextKey._dataEnd+1;
3056 // now nextKey is either invalid or is outside of actual._dataEnd
3057 // output the remaining and discard myself
3058 if (oldStart <= end) {
3059 processRegion(oldStart, end);
3061 if (interval_.par.size() > (size_t) end && interval_.par[end] == '}') {
3063 // This is the normal case.
3064 // But if using the firstlanguage, the closing may be missing
3066 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
3068 if (actual._dataEnd < end)
3069 output_end = interval_.nextNotIgnored(actual._dataEnd);
3070 else if (interval_.par.size() > (size_t) end)
3071 output_end = interval_.nextNotIgnored(end);
3073 output_end = interval_.par.size();
3074 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
3075 interval_.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
3077 // Remove possible empty data
3078 int dstart = interval_.nextNotIgnored(actual._dataStart);
3079 while (interval_.isOpeningPar(dstart) == 1) {
3080 interval_.addIntervall(dstart, dstart+1);
3081 int dend = interval_.findclosing(dstart+1, output_end);
3082 interval_.addIntervall(dend, dend+1);
3083 dstart = interval_.nextNotIgnored(dstart+1);
3085 if (dstart < output_end)
3086 interval_.output(os, output_end);
3088 interval_.addIntervall(0, end);
3090 interval_.addIntervall(actual._tokenstart, end);
3094 string splitOnKnownMacros(string par, bool isPatternString)
3097 LatexInfo li(par, isPatternString);
3098 // LYXERR(Debug::INFO, "Berfore split: " << par);
3099 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
3101 DummyKey._tokensize = 0;
3102 DummyKey._dataStart = 0;
3103 DummyKey._dataEnd = par.length();
3104 DummyKey.disabled = true;
3105 int firstkeyIdx = li.getFirstKey();
3107 if (firstkeyIdx >= 0) {
3108 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
3109 DummyKey._tokenstart = firstKey._tokenstart;
3111 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
3112 // Use dummy firstKey
3113 firstKey = DummyKey;
3114 (void) li.setNextKey(firstkeyIdx);
3117 if (par.substr(firstKey._dataStart, 2) == "% ")
3118 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
3120 nextkeyIdx = li.process(os, firstKey);
3121 while (nextkeyIdx >= 0) {
3122 // Check for a possible gap between the last
3123 // entry and this one
3124 int datastart = li.nextNotIgnored(firstKey._dataStart);
3125 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
3126 if ((nextKey._tokenstart > datastart)) {
3128 firstKey._dataStart = datastart;
3129 firstKey._dataEnd = par.length();
3130 (void) li.setNextKey(nextkeyIdx);
3131 // Fake the last opened parenthesis
3132 li.setForDefaultLang(firstKey);
3133 nextkeyIdx = li.process(os, firstKey);
3136 if (nextKey.keytype != KeyInfo::isMain) {
3137 firstKey._dataStart = datastart;
3138 firstKey._dataEnd = nextKey._dataEnd+1;
3139 (void) li.setNextKey(nextkeyIdx);
3140 li.setForDefaultLang(firstKey);
3141 nextkeyIdx = li.process(os, firstKey);
3144 nextkeyIdx = li.process(os, nextKey);
3148 // Handle the remaining
3149 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
3150 firstKey._dataEnd = par.length();
3152 if ((firstKey._dataStart < firstKey._dataEnd) &&
3153 (par[firstKey._dataStart] != '}')) {
3154 li.setForDefaultLang(firstKey);
3155 (void) li.process(os, firstKey);
3159 // return string definitelly impossible to match
3160 s = "\\foreignlanguage{ignore}{ }";
3164 s = par; /* no known macros found */
3165 // LYXERR(Debug::INFO, "After split: " << s);
3170 * Try to unify the language specs in the latexified text.
3171 * Resulting modified string is set to "", if
3172 * the searched tex does not contain all the features in the search pattern
3174 static string correctlanguagesetting(string par, bool isPatternString, bool withformat, lyx::Buffer *pbuf = nullptr)
3176 static Features regex_f;
3177 static int missed = 0;
3178 static bool regex_with_format = false;
3180 int parlen = par.length();
3182 while ((parlen > 0) && (par[parlen-1] == '\n')) {
3185 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
3186 // Happens to be there in case of description or labeling environment
3191 // Split the latex input into pieces which
3192 // can be digested by our search engine
3193 LYXERR(Debug::FIND, "input: \"" << par << "\"");
3194 if (isPatternString && (pbuf != nullptr)) { // Check if we should disable/enable test for language
3195 // We check for polyglossia, because in runparams.flavor we use Flavor::XeTeX
3196 string doclang = pbuf->params().language->polyglossia();
3197 static regex langre("\\\\(foreignlanguage)\\{([^\\}]+)\\}");
3199 bool toIgnoreLang = true;
3200 for (sregex_iterator it(par.begin(), par.end(), langre), end; it != end; ++it) {
3202 if (sub.str(2) != doclang) {
3203 toIgnoreLang = false;
3207 setIgnoreFormat("language", toIgnoreLang, false);
3210 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
3211 LYXERR(Debug::FIND, "After splitOnKnownMacros:\n\"" << result << "\"");
3214 result = par.substr(0, parlen);
3215 if (isPatternString) {
3218 regex_f = identifyFeatures(result);
3219 string features = "";
3220 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3221 string a = it->first;
3222 regex_with_format = true;
3223 features += " " + a;
3224 // LYXERR(Debug::INFO, "Identified regex format:" << a);
3226 LYXERR(Debug::FIND, "Identified Features" << features);
3229 } else if (regex_with_format) {
3230 Features info = identifyFeatures(result);
3231 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3232 string a = it->first;
3233 bool b = it->second;
3234 if (b && ! info[a]) {
3236 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
3243 // LYXERR(Debug::INFO, "No regex formats");
3249 // Remove trailing closure of math, macros and environments, so to catch parts of them.
3250 static int identifyClosing(string & t)
3252 int open_braces = 0;
3254 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
3255 if (regex_replace(t, t, "(.*[^\\\\])\\$$", "$1"))
3257 if (regex_replace(t, t, "(.*[^\\\\])\\\\\\]$", "$1"))
3259 if (regex_replace(t, t, "(.*[^\\\\])\\\\end\\{[a-zA-Z_]*\\*?\\}$", "$1"))
3261 if (regex_replace(t, t, "(.*[^\\\\])\\}$", "$1")) {
3270 static int num_replaced = 0;
3271 static bool previous_single_replace = true;
3273 void MatchStringAdv::CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string)
3276 // Handle \w properly
3277 QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
3278 if (! opt.casesensitive) {
3279 popts |= QRegularExpression::CaseInsensitiveOption;
3281 regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
3282 regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
3284 if (regexp.isValid() && regexp2.isValid()) {
3285 regexIsValid = true;
3286 // Check '{', '}' pairs inside the regex
3289 for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
3290 char c = par_as_string[i];
3297 else if (c == '}') {
3304 if (balanced != 0) {
3305 regexIsValid = false;
3306 regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
3310 regexIsValid = false;
3311 if (!regexp.isValid())
3312 regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
3314 regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
3317 (void)par_as_string;
3318 if (opt.casesensitive) {
3319 regexp = regex(regexp_str);
3320 regexp2 = regex(regexp2_str);
3323 regexp = regex(regexp_str, std::regex_constants::icase);
3324 regexp2 = regex(regexp2_str, std::regex_constants::icase);
3329 static void modifyRegexForMatchWord(string &t)
3332 regex wordre("(\\\\)*((\\.|\\\\b))");
3335 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
3337 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
3340 else if (sub.str(2) == "\\\\b")
3342 if (lastpos < (size_t) sub.position(2))
3343 s += t.substr(lastpos, sub.position(2) - lastpos);
3345 lastpos = sub.position(2) + sub.length(2);
3348 s = "\\b" + t + "\\b";
3352 else if (lastpos < t.length())
3353 s += t.substr(lastpos, t.length() - lastpos);
3354 t = "\\b" + s + "\\b";
3357 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt)
3358 : p_buf(&buf), p_first_buf(&buf), opt(opt)
3360 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
3361 docstring const & ds = stringifySearchBuffer(find_buf, opt);
3362 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
3363 if (opt.replace_all && previous_single_replace) {
3364 previous_single_replace = false;
3367 else if (!opt.replace_all) {
3368 num_replaced = 0; // count number of replaced strings
3369 previous_single_replace = true;
3371 // When using regexp, braces are hacked already by escape_for_regex()
3372 par_as_string = normalize(ds);
3374 close_wildcards = 0;
3376 size_t lead_size = 0;
3377 // correct the language settings
3378 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat, &buf);
3379 opt.matchAtStart = false;
3381 identifyClosing(par_as_string); // Removes math closings ($, ], ...) at end of string
3382 if (opt.ignoreformat) {
3386 lead_size = identifyLeading(par_as_string);
3388 lead_as_string = par_as_string.substr(0, lead_size);
3389 string lead_as_regex_string = string2regex(lead_as_string);
3390 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3391 string par_as_regex_string_nolead = string2regex(par_as_string_nolead);
3392 /* Handle whole words too in this case
3394 if (opt.matchword) {
3395 par_as_regex_string_nolead = "\\b" + par_as_regex_string_nolead + "\\b";
3396 opt.matchword = false;
3398 string regexp_str = "(" + lead_as_regex_string + ")()" + par_as_regex_string_nolead;
3399 string regexp2_str = "(" + lead_as_regex_string + ")(.*?)" + par_as_regex_string_nolead;
3400 CreateRegexp(opt, regexp_str, regexp2_str);
3402 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3403 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3407 if (!opt.ignoreformat) {
3408 lead_size = identifyLeading(par_as_string);
3409 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
3410 lead_as_string = par_as_string.substr(0, lead_size);
3411 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3414 // Here we are using regexp
3415 LASSERT(use_regexp, /**/);
3417 string lead_as_regexp;
3418 if (lead_size > 0) {
3419 lead_as_regexp = string2regex(par_as_string.substr(0, lead_size));
3420 (void)regex_replace(par_as_string_nolead, par_as_string_nolead, "}$", "");
3421 par_as_string = par_as_string_nolead;
3422 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
3423 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3425 // LYXERR(Debug::FIND, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
3426 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
3427 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
3428 // LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3430 size_t lng = par_as_string.size();
3431 if (!opt.ignoreformat) {
3432 // Remove extra '\}' at end if not part of \{\.\}
3434 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
3436 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
3445 if (lng < par_as_string.size())
3446 par_as_string = par_as_string.substr(0,lng);
3448 LYXERR(Debug::FIND, "par_as_string after correctRegex is '" << par_as_string << "'");
3449 if ((lng > 0) && (par_as_string[0] == '^')) {
3450 par_as_string = par_as_string.substr(1);
3452 opt.matchAtStart = true;
3454 // LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3455 // LYXERR(Debug::FIND, "Open braces: " << open_braces);
3456 // LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
3458 // If entered regexp must match at begin of searched string buffer
3459 // Kornel: Added parentheses to use $1 for size of the leading string
3463 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
3464 // Unfortunately is '\1', '\2', etc not working for strings with extra format
3465 // so the convert has no effect in that case
3466 for (int i = 7; i > 0; --i) {
3467 string orig = "\\\\" + std::to_string(i);
3468 string dest = "\\" + std::to_string(i+2);
3469 while (regex_replace(par_as_string, par_as_string, orig, dest));
3471 if (opt.matchword) {
3472 modifyRegexForMatchWord(par_as_string);
3473 opt.matchword = false;
3475 regexp_str = "(" + lead_as_regexp + ")()" + par_as_string;
3476 regexp2_str = "(" + lead_as_regexp + ")(.*?)" + par_as_string;
3478 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3479 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3480 CreateRegexp(opt, regexp_str, regexp2_str, par_as_string);
3484 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
3488 mres.searched_size = len;
3490 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
3493 docstring docstr = stringifyFromForSearch(opt, cur, len);
3495 str = normalize(docstr);
3496 if (!opt.ignoreformat) {
3497 str = correctlanguagesetting(str, false, !opt.ignoreformat);
3498 // remove closing '}' and '\n' to allow for use of '$' in regex
3499 size_t lng = str.size();
3500 while ((lng > 1) && ((str[lng -1] == '}') || (str[lng -1] == '\n')))
3502 if (lng != str.size()) {
3503 str = str.substr(0, lng);
3507 mres.match_len = -1;
3510 LYXERR(Debug::FIND, "After normalization: Matching against:\n'" << str << "'");
3512 LASSERT(use_regexp, /**/);
3514 // use_regexp always true
3515 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
3517 QString qstr = QString::fromStdString(str);
3518 QRegularExpression const *p_regexp;
3519 QRegularExpression::MatchType flags = QRegularExpression::NormalMatch;
3523 p_regexp = ®exp2;
3525 QRegularExpressionMatch match = p_regexp->match(qstr, 0, flags);
3526 if (!match.hasMatch())
3529 regex const *p_regexp;
3530 regex_constants::match_flag_type flags;
3532 flags = regex_constants::match_continuous;
3535 flags = regex_constants::match_default;
3536 p_regexp = ®exp2;
3538 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
3539 if (re_it == sregex_iterator())
3541 match_results<string::const_iterator> const & m = *re_it;
3543 // Whole found string, including the leading
3544 // std: m[0].second - m[0].first
3545 // Qt: match.capturedEnd(0) - match.capturedStart(0)
3547 // Size of the leading string
3548 // std: m[1].second - m[1].first
3549 // Qt: match.capturedEnd(1) - match.capturedStart(1)
3550 int leadingsize = 0;
3552 if (match.lastCapturedIndex() > 0) {
3553 leadingsize = match.capturedEnd(1) - match.capturedStart(1);
3558 leadingsize = m[1].second - m[1].first;
3562 mres.match_prefix = match.capturedEnd(2) - match.capturedStart(2);
3563 mres.match_len = match.capturedEnd(0) - match.capturedEnd(2);
3564 // because of different number of closing at end of string
3565 // we have to 'unify' the length of the post-match.
3566 // Done by ignoring closing parenthesis and linefeeds at string end
3567 int matchend = match.capturedEnd(0);
3568 size_t strsize = qstr.size();
3569 if (!opt.ignoreformat) {
3570 while (mres.match_len > 0) {
3571 QChar c = qstr.at(matchend - 1);
3572 if ((c == '\n') || (c == '}') || (c == '{')) {
3579 while (strsize > (size_t) match.capturedEnd(0)) {
3580 QChar c = qstr.at(strsize-1);
3581 if ((c == '\n') || (c == '}')) {
3588 // LYXERR0(qstr.toStdString());
3589 mres.match2end = strsize - matchend;
3590 mres.pos = match.capturedStart(2);
3592 mres.match_prefix = m[2].second - m[2].first;
3593 mres.match_len = m[0].second - m[2].second;
3594 // ignore closing parenthesis and linefeeds at string end
3595 size_t strend = m[0].second - m[0].first;
3596 int matchend = strend;
3597 size_t strsize = str.size();
3598 if (!opt.ignoreformat) {
3599 while (mres.match_len > 0) {
3600 char c = str.at(matchend - 1);
3601 if ((c == '\n') || (c == '}') || (c == '{')) {
3608 while (strsize > strend) {
3609 if ((str.at(strsize-1) == '}') || (str.at(strsize-1) == '\n')) {
3617 mres.match2end = strsize - matchend;
3618 mres.pos = m[2].first - m[0].first;;
3620 if (mres.match2end < 0)
3622 mres.leadsize = leadingsize;
3624 if (mres.match_len > 0) {
3625 string a0 = match.captured(0).mid(mres.pos + mres.match_prefix, mres.match_len).toStdString();
3626 mres.result.push_back(a0);
3627 for (int i = 3; i <= match.lastCapturedIndex(); i++) {
3628 mres.result.push_back(match.captured(i).toStdString());
3632 if (mres.match_len > 0) {
3633 string a0 = m[0].str().substr(mres.pos + mres.match_prefix, mres.match_len);
3634 mres.result.push_back(a0);
3635 for (size_t i = 3; i < m.size(); i++) {
3636 mres.result.push_back(m[i]);
3645 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
3647 MatchResult mres = findAux(cur, len, at_begin);
3648 int res = mres.match_len;
3650 "res=" << res << ", at_begin=" << at_begin
3651 << ", matchAtStart=" << opt.matchAtStart
3652 << ", inTexted=" << cur.inTexted());
3653 if (opt.matchAtStart) {
3656 else if (mres.match_prefix > 0)
3665 static bool simple_replace(string &t, string from, string to)
3667 regex repl("(\\\\)*(" + from + ")");
3671 for (sregex_iterator it(t.begin(), t.end(), repl), end; it != end; ++it) {
3673 if ((sub.position(2) - sub.position(0)) % 2 == 1)
3675 if (lastpos < (size_t) sub.position(2))
3676 s += t.substr(lastpos, sub.position(2) - lastpos);
3678 lastpos = sub.position(2) + sub.length(2);
3682 else if (lastpos < t.length())
3683 s += t.substr(lastpos, t.length() - lastpos);
3689 string MatchStringAdv::normalize(docstring const & s) const
3692 t = lyx::to_utf8(s);
3693 // Remove \n at begin
3694 while (!t.empty() && t[0] == '\n')
3697 while (!t.empty() && t[t.size() - 1] == '\n')
3698 t = t.substr(0, t.size() - 1);
3700 // Handle all other '\n'
3701 while ((pos = t.find("\n")) != string::npos) {
3702 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
3704 if (isAlnumASCII(t[pos+1])) {
3705 t.replace(pos-2, 3, " ");
3708 t.replace(pos-2, 3, "");
3711 else if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
3712 // '\n' adjacent to non-alpha-numerics, discard
3713 t.replace(pos, 1, "");
3716 // Replace all other \n with spaces
3717 t.replace(pos, 1, " ");
3720 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
3721 // Kornel: Added textsl, textsf, textit, texttt and noun
3722 // + allow to seach for colored text too
3723 LYXERR(Debug::FIND, "Removing stale empty macros from: " << t);
3724 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
3725 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3726 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
3727 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3728 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
3734 docstring stringifyFromCursor(DocIterator const & cur, int len)
3736 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
3737 if (cur.inTexted()) {
3738 Paragraph const & par = cur.paragraph();
3739 // TODO what about searching beyond/across paragraph breaks ?
3740 // TODO Try adding a AS_STR_INSERTS as last arg
3741 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
3742 int(par.size()) : cur.pos() + len;
3743 // OutputParams runparams(&cur.buffer()->params().encoding());
3744 OutputParams runparams(encodings.fromLyXName("utf8"));
3745 runparams.nice = true;
3746 runparams.flavor = Flavor::XeTeX;
3747 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
3748 // No side effect of file copying and image conversion
3749 runparams.dryrun = true;
3750 int option = AS_STR_INSETS | AS_STR_PLAINTEXT;
3751 if (ignoreFormats.getDeleted()) {
3752 option |= AS_STR_SKIPDELETE;
3753 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
3756 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
3758 LYXERR(Debug::FIND, "Stringifying with cur: "
3759 << cur << ", from pos: " << cur.pos() << ", end: " << end);
3760 return par.asString(cur.pos(), end,
3763 } else if (cur.inMathed()) {
3764 CursorSlice cs = cur.top();
3765 MathData md = cs.cell();
3766 MathData::const_iterator it_end =
3767 (( len == -1 || cs.pos() + len > int(md.size()))
3769 : md.begin() + cs.pos() + len );
3771 for (MathData::const_iterator it = md.begin() + cs.pos();
3774 docstring s = asString(md2);
3775 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
3778 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3783 /** Computes the LaTeX export of buf starting from cur and ending len positions
3784 * after cur, if len is positive, or at the paragraph or innermost inset end
3787 docstring latexifyFromCursor(DocIterator const & cur, int len)
3790 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
3791 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
3792 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
3794 Buffer const & buf = *cur.buffer();
3796 odocstringstream ods;
3798 //OutputParams runparams(&buf.params().encoding());
3799 OutputParams runparams(encodings.fromLyXName("utf8"));
3800 runparams.nice = false;
3801 runparams.flavor = Flavor::XeTeX;
3802 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3803 // No side effect of file copying and image conversion
3804 runparams.dryrun = true;
3805 if (ignoreFormats.getDeleted()) {
3806 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
3809 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
3812 if (cur.inTexted()) {
3813 // @TODO what about searching beyond/across paragraph breaks ?
3814 pos_type endpos = cur.paragraph().size();
3815 if (len != -1 && endpos > cur.pos() + len)
3816 endpos = cur.pos() + len;
3817 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
3818 string(), cur.pos(), endpos);
3819 string s = lyx::to_utf8(ods.str());
3820 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
3821 return(lyx::from_utf8(s));
3822 } else if (cur.inMathed()) {
3823 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
3824 for (int s = cur.depth() - 1; s >= 0; --s) {
3825 CursorSlice const & cs = cur[s];
3826 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
3827 TeXMathStream ws(os);
3828 cs.asInsetMath()->asHullInset()->header_write(ws);
3833 CursorSlice const & cs = cur.top();
3834 MathData md = cs.cell();
3835 MathData::const_iterator it_end =
3836 ((len == -1 || cs.pos() + len > int(md.size()))
3838 : md.begin() + cs.pos() + len);
3840 for (MathData::const_iterator it = md.begin() + cs.pos();
3844 ods << asString(md2);
3845 // Retrieve the math environment type, and add '$' or '$]'
3846 // or others (\end{equation}) accordingly
3847 for (int s = cur.depth() - 1; s >= 0; --s) {
3848 CursorSlice const & cs2 = cur[s];
3849 InsetMath * inset = cs2.asInsetMath();
3850 if (inset && inset->asHullInset()) {
3851 TeXMathStream ws(os);
3852 inset->asHullInset()->footer_write(ws);
3856 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3858 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3863 #if defined(ResultsDebug)
3865 static void displayMResult(MatchResult &mres, string from, DocIterator & cur)
3867 LYXERR0( "from:\t\t\t" << from);
3869 if (mres.pos_len > 0) {
3871 status = "FINALSEARCH";
3874 if (mres.match_len > 0) {
3875 if ((mres.match_prefix == 0) && (mres.pos == mres.leadsize))
3876 status = "Good Match";
3878 status = "Matched in";
3881 status = "MissedSearch";
3884 LYXERR0( status << "(" << cur.pos() << " ... " << mres.searched_size + cur.pos() << ") cur.lastpos(" << cur.lastpos() << ")");
3885 if ((mres.leadsize > 0) || (mres.match_len > 0) || (mres.match2end > 0))
3886 LYXERR0( "leadsize(" << mres.leadsize << ") match_len(" << mres.match_len << ") match2end(" << mres.match2end << ")");
3887 if ((mres.pos > 0) || (mres.match_prefix > 0))
3888 LYXERR0( "pos(" << mres.pos << ") match_prefix(" << mres.match_prefix << ")");
3889 for (size_t i = 0; i < mres.result.size(); i++)
3890 LYXERR0( "Match " << i << " = \"" << mres.result[i] << "\"");
3892 #define displayMres(s, txt, cur) displayMResult(s, txt, cur);
3894 #define displayMres(s, txt, cur)
3897 /** Finalize an advanced find operation, advancing the cursor to the innermost
3898 ** position that matches, plus computing the length of the matching text to
3900 ** Return the cur.pos() difference between start and end of found match
3902 MatchResult findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, MatchResult const & expected = MatchResult(-1))
3904 // Search the foremost position that matches (avoids find of entire math
3905 // inset when match at start of it)
3906 DocIterator old_cur(cur.buffer());
3908 static MatchResult fail = MatchResult();
3909 MatchResult max_match;
3910 // If (prefix_len > 0) means that forwarding 1 position will remove the complete entry
3911 // Happens with e.g. hyperlinks
3912 // either one sees "http://www.bla.bla" or nothing
3913 // so the search for "www" gives prefix_len = 7 (== sizeof("http://")
3914 // and although we search for only 3 chars, we find the whole hyperlink inset
3915 bool at_begin = (expected.match_prefix == 0);
3916 if (!match.opt.forward && match.opt.ignoreformat) {
3917 if (expected.pos > 0)
3920 LASSERT(at_begin, /**/);
3921 if (expected.match_len > 0 && at_begin) {
3922 // Search for deepest match
3924 max_match = expected;
3926 size_t d = cur.depth();
3930 if (cur.depth() < d)
3932 if (cur.depth() == d)
3935 while (cur && cur.depth() > lastd) {
3936 lastd = cur.depth();
3937 mres = match(cur, -1, at_begin);
3938 displayMres(mres, "Checking innermost", cur);
3939 if (mres.match_len > 0)
3944 if (mres.match_len < expected.match_len)
3952 // (expected.match_len <= 0)
3953 mres = match(cur); /* match valid only if not searching whole words */
3954 displayMres(mres, "Start with negative match", cur);
3957 if (max_match.match_len <= 0) return fail;
3958 LYXERR(Debug::FIND, "Ok");
3960 // Compute the match length
3962 if (cur.pos() + len > cur.lastpos())
3965 LASSERT(match.use_regexp, /**/);
3968 int maxl = cur.lastpos() - cur.pos();
3969 // Greedy behaviour while matching regexps
3970 while (maxl > minl) {
3972 mres2 = match(cur, len, at_begin);
3973 displayMres(mres2, "Finalize loop", cur);
3974 int actual_match_len = mres2.match_len;
3975 if (actual_match_len >= max_match.match_len) {
3976 // actual_match_len > max_match _can_ happen,
3977 // if the search area splits
3978 // some following word so that the regex
3979 // (e.g. 'r.*r\b' matches 'r' from the middle of the
3981 // This means, the len value is too big
3982 actual_match_len = max_match.match_len;
3984 max_match.match_len = actual_match_len;
3986 if (maxl - minl < 4)
3987 len = (int)((maxl + minl)/2);
3989 len = (int)(minl + (maxl - minl + 3)/4);
3992 // (actual_match_len < max_match.match_len)
3994 len = (int)((maxl + minl)/2);
3999 // Search for real start of matched characters
4001 MatchResult actual_match;
4004 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
4005 if (cur.depth() < old_cur.depth()) {
4007 LYXERR(Debug::INFO, "cur.depth() < old_cur.depth(), this should never happen");
4010 if (cur.pos() != old_cur.pos()) {
4011 // OK, forwarded 1 pos in actual inset
4012 actual_match = match(cur, len-1, at_begin);
4013 if (actual_match.match_len == max_match.match_len) {
4014 // Ha, got it! The shorter selection has the same match length
4017 max_match = actual_match;
4020 // OK, the shorter selection matches less chars, revert to previous value
4026 LYXERR(Debug::INFO, "cur.pos() == old_cur.pos(), this should never happen");
4027 actual_match = match(cur, len, at_begin);
4028 if (actual_match.match_len == max_match.match_len) {
4030 max_match = actual_match;
4037 max_match.pos_len = len;
4038 displayMres(max_match, "SEARCH RESULT", cur)
4045 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
4049 bool repeat = false;
4050 DocIterator orig_cur; // to be used if repeat not successful
4051 MatchResult orig_mres;
4052 while (!theApp()->longOperationCancelled() && cur) {
4053 //(void) findAdvForwardInnermost(cur);
4054 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
4055 MatchResult mres = match(cur, -1, false);
4056 string msg = "Starting";
4059 displayMres(mres, msg + " findForwardAdv", cur)
4060 int match_len = mres.match_len;
4061 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
4062 LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
4065 if (match_len <= 0) {
4066 // This should exit nested insets, if any, or otherwise undefine the currsor.
4067 cur.pos() = cur.lastpos();
4068 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
4071 else { // match_len > 0
4072 // Try to find the begin of searched string
4074 int firstInvalid = cur.lastpos() - cur.pos();
4076 int incrmatch = (mres.match_prefix + mres.pos - mres.leadsize + 1)*3/4;
4077 int incrcur = (firstInvalid + 1 )*3/4;
4078 if (incrcur < incrmatch)
4079 increment = incrcur;
4081 increment = incrmatch;
4085 LYXERR(Debug::FIND, "Set increment to " << increment);
4086 while (increment > 0) {
4087 DocIterator old_cur = cur;
4088 if (cur.pos() + increment >= cur.lastpos()) {
4092 cur.pos() = cur.pos() + increment;
4093 MatchResult mres2 = match(cur, -1, false);
4094 displayMres(mres2, "findForwardAdv loop", cur)
4095 switch (interpretMatch(mres, mres2)) {
4096 case MatchResult::newIsTooFar:
4097 // behind the expected match
4098 firstInvalid = increment;
4102 case MatchResult::newIsBetter:
4103 // not reached yet, but cur.pos()+increment is bettert
4105 firstInvalid -= increment;
4106 if (increment > firstInvalid*3/4)
4107 increment = firstInvalid*3/4;
4108 if ((mres2.pos == mres2.leadsize) && (increment >= mres2.match_prefix)) {
4109 if (increment >= mres2.match_prefix)
4110 increment = (mres2.match_prefix+1)*3/4;
4115 // Handle not like MatchResult::newIsTooFar
4116 LYXERR0( "Probably too far: Increment = " << increment << " match_prefix = " << mres.match_prefix);
4118 increment = increment*3/4;
4123 if (mres.match_len > 0) {
4124 if (mres.match_prefix + mres.pos - mres.leadsize > 0) {
4125 // The match seems to indicate some deeper level
4134 // should never be reached.
4138 // LYXERR0("Leaving first loop");
4139 LYXERR(Debug::FIND, "Finalizing 1");
4140 MatchResult found_match = findAdvFinalize(cur, match, mres);
4141 if (found_match.match_len > 0) {
4142 LASSERT(found_match.pos_len > 0, /**/);
4143 match.FillResults(found_match);
4144 return found_match.pos_len;
4147 // try next possible match
4158 /// Find the most backward consecutive match within same paragraph while searching backwards.
4159 MatchResult findMostBackwards(DocIterator & cur, MatchStringAdv const & match, MatchResult &expected)
4161 DocIterator cur_begin = cur;
4162 cur_begin.pos() = 0;
4163 DocIterator tmp_cur = cur;
4164 MatchResult mr = findAdvFinalize(tmp_cur, match, expected);
4165 Inset & inset = cur.inset();
4166 for (; cur != cur_begin; cur.backwardPos()) {
4167 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
4168 DocIterator new_cur = cur;
4169 new_cur.backwardPos();
4170 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
4172 MatchResult new_mr = findAdvFinalize(new_cur, match, expected);
4173 if (new_mr.match_len == mr.match_len)
4177 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
4183 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
4187 // Backup of original position
4188 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
4189 if (cur == cur_begin)
4192 DocIterator cur_orig(cur);
4193 bool pit_changed = false;
4196 MatchResult found_match = match(cur, -1, false);
4198 if (found_match.match_len > 0) {
4200 cur.pos() = cur.lastpos();
4202 cur.pos() = cur_orig.pos();
4203 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
4204 DocIterator cur_prev_iter;
4206 found_match = match(cur);
4207 LYXERR(Debug::FIND, "findBackAdv3: found_match="
4208 << (found_match.match_len > 0) << ", cur: " << cur);
4209 if (found_match.match_len > 0) {
4210 MatchResult found_mr = findMostBackwards(cur, match, found_match);
4211 if (found_mr.pos_len > 0) {
4212 match.FillResults(found_mr);
4213 return found_mr.pos_len;
4217 // Stop if begin of document reached
4218 if (cur == cur_begin)
4220 cur_prev_iter = cur;
4224 if (cur == cur_begin)
4231 } while (!theApp()->longOperationCancelled());
4239 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
4240 DocIterator const & cur, int len)
4242 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
4244 if (!opt.ignoreformat)
4245 return latexifyFromCursor(cur, len);
4247 return stringifyFromCursor(cur, len);
4251 FindAndReplaceOptions::FindAndReplaceOptions(
4252 docstring const & _find_buf_name, bool _casesensitive,
4253 bool _matchword, bool _forward, bool _expandmacros, bool _ignoreformat,
4254 docstring const & _repl_buf_name, bool _keep_case,
4255 SearchScope _scope, SearchRestriction _restr, bool _replace_all)
4256 : find_buf_name(_find_buf_name), casesensitive(_casesensitive), matchword(_matchword),
4257 forward(_forward), expandmacros(_expandmacros), ignoreformat(_ignoreformat),
4258 repl_buf_name(_repl_buf_name), keep_case(_keep_case), scope(_scope), restr(_restr), replace_all(_replace_all)
4266 /** Check if 'len' letters following cursor are all non-lowercase */
4267 static bool allNonLowercase(Cursor const & cur, int len)
4269 pos_type beg_pos = cur.selectionBegin().pos();
4270 pos_type end_pos = cur.selectionBegin().pos() + len;
4271 if (len > cur.lastpos() + 1 - beg_pos) {
4272 LYXERR(Debug::FIND, "This should not happen, more debug needed");
4273 len = cur.lastpos() + 1 - beg_pos;
4274 end_pos = beg_pos + len;
4276 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
4277 if (isLowerCase(cur.paragraph().getChar(pos)))
4283 /** Check if first letter is upper case and second one is lower case */
4284 static bool firstUppercase(Cursor const & cur)
4287 pos_type pos = cur.selectionBegin().pos();
4288 if (pos >= cur.lastpos() - 1) {
4289 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
4292 ch1 = cur.paragraph().getChar(pos);
4293 ch2 = cur.paragraph().getChar(pos + 1);
4294 bool result = isUpperCase(ch1) && isLowerCase(ch2);
4295 LYXERR(Debug::FIND, "firstUppercase(): "
4296 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
4297 << ch2 << "(" << char(ch2) << ")"
4298 << ", result=" << result << ", cur=" << cur);
4303 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
4305 ** \fixme What to do with possible further paragraphs in replace buffer ?
4307 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
4309 ParagraphList::iterator pit = buffer.paragraphs().begin();
4310 LASSERT(!pit->empty(), /**/);
4311 pos_type right = pos_type(1);
4312 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
4313 right = pit->size();
4314 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
4318 static bool replaceMatches(string &t, int maxmatchnum, vector <string> const & replacements)
4320 // Should replace the string "$" + std::to_string(matchnum) with replacement
4321 // if the char '$' is not prefixed with odd number of char '\\'
4322 static regex const rematch("(\\\\)*(\\$\\$([0-9]))");
4326 for (sregex_iterator it(t.begin(), t.end(), rematch), end; it != end; ++it) {
4328 if ((sub.position(2) - sub.position(0)) % 2 == 1)
4330 int num = stoi(sub.str(3), nullptr, 10);
4331 if (num >= maxmatchnum)
4333 if (lastpos < (size_t) sub.position(2))
4334 s += t.substr(lastpos, sub.position(2) - lastpos);
4335 s += replacements[num];
4336 lastpos = sub.position(2) + sub.length(2);
4340 else if (lastpos < t.length())
4341 s += t.substr(lastpos, t.length() - lastpos);
4347 static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
4349 Cursor & cur = bv->cursor();
4350 if (opt.repl_buf_name.empty()
4351 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
4352 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4355 DocIterator sel_beg = cur.selectionBegin();
4356 DocIterator sel_end = cur.selectionEnd();
4357 if (&sel_beg.inset() != &sel_end.inset()
4358 || sel_beg.pit() != sel_end.pit()
4359 || sel_beg.idx() != sel_end.idx())
4361 int sel_len = sel_end.pos() - sel_beg.pos();
4362 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
4363 << ", sel_len: " << sel_len << endl);
4366 LASSERT(sel_len > 0, return 0);
4368 if (!matchAdv(sel_beg, sel_len).match_len)
4371 // Build a copy of the replace buffer, adapted to the KeepCase option
4372 Buffer const & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
4374 repl_buffer_orig.write(oss);
4375 string lyx = oss.str();
4376 if (matchAdv.valid_matches > 0)
4377 replaceMatches(lyx, matchAdv.valid_matches, matchAdv.matches);
4378 Buffer repl_buffer(string(), false);
4379 repl_buffer.setInternal(true);
4380 repl_buffer.setUnnamed(true);
4381 LASSERT(repl_buffer.readString(lyx), return 0);
4382 if (opt.keep_case && sel_len >= 2) {
4383 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
4384 if (cur.inTexted()) {
4385 if (firstUppercase(cur))
4386 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
4387 else if (allNonLowercase(cur, sel_len))
4388 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
4391 cap::cutSelection(cur, false);
4392 if (cur.inTexted()) {
4393 repl_buffer.changeLanguage(
4394 repl_buffer.language(),
4395 cur.getFont().language());
4396 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
4397 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
4398 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
4399 repl_buffer.params().documentClassPtr(),
4400 repl_buffer.params().authors(),
4401 bv->buffer().errorList("Paste"));
4402 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
4403 sel_len = repl_buffer.paragraphs().begin()->size();
4404 } else if (cur.inMathed()) {
4405 odocstringstream ods;
4407 // OutputParams runparams(&repl_buffer.params().encoding());
4408 OutputParams runparams(encodings.fromLyXName("utf8"));
4409 runparams.nice = false;
4410 runparams.flavor = Flavor::XeTeX;
4411 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
4412 runparams.dryrun = true;
4413 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
4414 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
4415 docstring repl_latex = ods.str();
4416 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
4418 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
4419 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
4420 repl_latex = from_utf8(s);
4421 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
4422 MathData ar(cur.buffer());
4423 asArray(repl_latex, ar, Parse::NORMAL);
4425 sel_len = ar.size();
4426 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4428 if (cur.pos() >= sel_len)
4429 cur.pos() -= sel_len;
4432 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4433 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
4434 bv->processUpdateFlags(Update::Force);
4439 /// Perform a FindAdv operation.
4440 bool findAdv(BufferView * bv, FindAndReplaceOptions & opt)
4445 // e.g., when invoking word-findadv from mini-buffer wither with
4446 // wrong options syntax or before ever opening advanced F&R pane
4447 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4451 MatchStringAdv matchAdv(bv->buffer(), opt);
4453 if (!matchAdv.regexIsValid) {
4454 bv->message(lyx::from_utf8(matchAdv.regexError));
4458 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
4460 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
4461 num_replaced += findAdvReplace(bv, opt, matchAdv);
4464 pos_len = findForwardAdv(cur, matchAdv);
4466 pos_len = findBackwardsAdv(cur, matchAdv);
4467 } catch (exception & ex) {
4468 bv->message(from_utf8(ex.what()));
4473 if (num_replaced > 0) {
4474 switch (num_replaced)
4477 bv->message(_("One match has been replaced."));
4480 bv->message(_("Two matches have been replaced."));
4483 bv->message(bformat(_("%1$d matches have been replaced."), num_replaced));
4489 bv->message(_("Match not found."));
4494 if (num_replaced > 0)
4495 bv->message(_("Match has been replaced."));
4497 bv->message(_("Match found."));
4499 if (cur.pos() + pos_len > cur.lastpos()) {
4500 // Prevent crash in bv->putSelectionAt()
4501 // Should never happen, maybe LASSERT() here?
4502 pos_len = cur.lastpos() - cur.pos();
4504 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << pos_len);
4505 bv->putSelectionAt(cur, pos_len, !opt.forward);
4511 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
4513 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
4514 << opt.casesensitive << ' '
4515 << opt.matchword << ' '
4516 << opt.forward << ' '
4517 << opt.expandmacros << ' '
4518 << opt.ignoreformat << ' '
4519 << opt.replace_all << ' '
4520 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
4521 << opt.keep_case << ' '
4522 << int(opt.scope) << ' '
4525 LYXERR(Debug::FIND, "built: " << os.str());
4531 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
4533 // LYXERR(Debug::FIND, "parsing");
4537 while (line != "EOSS") {
4541 if (is.eof()) // Tolerate malformed request
4545 // LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
4546 opt.find_buf_name = from_utf8(s);
4547 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.replace_all;
4548 is.get(); // Waste space before replace string
4551 while (line != "EOSS") {
4555 if (is.eof()) // Tolerate malformed request
4559 // LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
4560 opt.repl_buf_name = from_utf8(s);
4561 is >> opt.keep_case;
4564 opt.scope = FindAndReplaceOptions::SearchScope(i);
4566 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
4569 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
4570 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
4571 << opt.scope << ' ' << opt.restr);