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();
363 docstring const msg = forward
364 ? _("Search reached end of document, continuing from beginning.")
365 : _("Search reached beginning of document, continuing from end.");
368 bv->clearSelection();
369 if (findOne(bv, searchstr, case_sens, whole, forward,
370 find_del, false, false, false, false))
373 bv->cursor().setCursor(cur_orig);
377 return match_len > 0;
383 int replaceAll(BufferView * bv,
384 docstring const & searchstr, docstring const & replacestr,
385 bool case_sens, bool whole, bool onlysel)
387 Buffer & buf = bv->buffer();
389 if (!searchAllowed(searchstr) || buf.isReadonly())
392 DocIterator startcur = bv->cursor().selectionBegin();
393 DocIterator endcur = bv->cursor().selectionEnd();
394 bool const had_selection = bv->cursor().selection();
396 MatchString const match(searchstr, case_sens, whole);
399 int const rsize = replacestr.size();
400 int const ssize = searchstr.size();
403 cur.setCursor(doc_iterator_begin(&buf));
404 int match_len = findForward(cur, endcur, match, false, onlysel);
405 while (match_len > 0) {
406 // Backup current cursor position and font.
407 pos_type const pos = cur.pos();
408 Font const font = cur.paragraph().getFontSettings(buf.params(), pos);
410 int ct_deleted_text = ssize -
411 cur.paragraph().eraseChars(pos, pos + match_len,
412 buf.params().track_changes);
413 cur.paragraph().insert(pos, replacestr, font,
414 Change(buf.params().track_changes
416 : Change::UNCHANGED));
417 for (int i = 0; i < rsize + ct_deleted_text
418 && cur.pos() < cur.lastpos(); ++i)
420 if (onlysel && cur.pit() == endcur.pit() && cur.idx() == endcur.idx()) {
421 // Adjust end of selection for replace-all in selection
423 int const offset = rsize - ssize;
424 for (int i = 0; i < offset + ct_deleted_text
425 && endcur.pos() < endcur.lastpos(); ++i)
428 int const offset = ssize - rsize;
429 for (int i = 0; i < offset && endcur.pos() > 0; ++i)
430 endcur.backwardPos();
431 for (int i = 0; i < ct_deleted_text
432 && endcur.pos() < endcur.lastpos(); ++i)
437 match_len = findForward(cur, endcur, match, false, onlysel);
440 bv->putSelectionAt(doc_iterator_begin(&buf), 0, false);
442 startcur.fixIfBroken();
443 bv->setCursor(startcur);
445 // Reset selection, accounting for changes in selection
447 endcur.fixIfBroken();
448 bv->cursor().resetAnchor();
449 bv->setCursorSelectionTo(endcur);
456 // the idea here is that we are going to replace the string that
457 // is selected IF it is the search string.
458 // if there is a selection, but it is not the search string, then
459 // we basically ignore it. (FIXME We ought to replace only within
461 // if there is no selection, then:
462 // (i) if some search string has been provided, then we find it.
463 // (think of how the dialog works when you hit "replace" the
465 // (ii) if no search string has been provided, then we treat the
466 // word the cursor is in as the search string. (why? i have no
467 // idea.) but this only works in text?
469 // returns the number of replacements made (one, if any) and
470 // whether anything at all was done.
471 pair<bool, int> replaceOne(BufferView * bv, docstring searchstr,
472 docstring const & replacestr, bool case_sens,
473 bool whole, bool forward, bool findnext, bool wrap,
476 Cursor & cur = bv->cursor();
477 if (!cur.selection() || onlysel) {
478 // no selection, non-empty search string: find it
479 if (!searchstr.empty()) {
480 bool const found = findOne(bv, searchstr, case_sens, whole,
481 forward, true, findnext, wrap, false, onlysel);
482 return make_pair(found, 0);
484 // empty search string
487 return make_pair(false, 0);
488 // select current word and treat it as the search string.
489 // This causes a minor bug as undo will restore this selection,
490 // which the user did not create (#8986).
491 cur.innerText()->selectWord(cur, WHOLE_WORD);
492 searchstr = cur.selectionAsString(false, true);
495 // if we still don't have a search string, report the error
497 if (!searchAllowed(searchstr))
498 return make_pair(false, 0);
500 bool have_selection = cur.selection();
501 docstring const selected = cur.selectionAsString(false, true);
504 ? searchstr == selected
505 : compare_no_case(searchstr, selected) == 0;
507 // no selection or current selection is not search word:
508 // just find the search word
509 if (!have_selection || !match) {
510 bool const found = findOne(bv, searchstr, case_sens, whole, forward,
511 true, findnext, wrap, false, onlysel);
512 return make_pair(found, 0);
515 // we're now actually ready to replace. if the buffer is
516 // read-only, we can't, though.
517 if (bv->buffer().isReadonly())
518 return make_pair(false, 0);
520 cap::replaceSelectionWithString(cur, replacestr);
522 cur.pos() += replacestr.length();
523 LASSERT(cur.pos() <= cur.lastpos(),
524 cur.pos() = cur.lastpos());
527 findOne(bv, searchstr, case_sens, whole,
528 forward, false, findnext, wrap, false, onlysel);
530 return make_pair(true, 1);
536 docstring const find2string(docstring const & search,
537 bool casesensitive, bool matchword,
538 bool forward, bool wrap, bool instant,
543 << int(casesensitive) << ' '
544 << int(matchword) << ' '
545 << int(forward) << ' '
547 << int(instant) << ' '
553 docstring const replace2string(docstring const & replace,
554 docstring const & search,
555 bool casesensitive, bool matchword,
556 bool all, bool forward, bool findnext,
557 bool wrap, bool onlysel)
560 ss << replace << '\n'
562 << int(casesensitive) << ' '
563 << int(matchword) << ' '
565 << int(forward) << ' '
566 << int(findnext) << ' '
573 docstring const string2find(docstring const & argument,
581 // data is of the form
583 // <casesensitive> <matchword> <forward> <wrap> <onlysel>"
585 docstring howto = split(argument, search, '\n');
587 casesensitive = parse_bool(howto);
588 matchword = parse_bool(howto);
589 forward = parse_bool(howto, true);
590 wrap = parse_bool(howto);
591 instant = parse_bool(howto);
592 onlysel = parse_bool(howto);
598 bool lyxfind(BufferView * bv, FuncRequest const & ev)
600 if (!bv || ev.action() != LFUN_WORD_FIND)
603 //lyxerr << "find called, cmd: " << ev << endl;
611 docstring search = string2find(ev.argument(), casesensitive,
612 matchword, forward, wrap, instant, onlysel);
614 return findOne(bv, search, casesensitive, matchword, forward,
615 false, true, wrap, instant, onlysel);
619 bool lyxreplace(BufferView * bv, FuncRequest const & ev)
621 if (!bv || ev.action() != LFUN_WORD_REPLACE)
624 // data is of the form
627 // <casesensitive> <matchword> <all> <forward> <findnext> <wrap> <onlysel>"
630 docstring howto = split(ev.argument(), rplc, '\n');
631 howto = split(howto, search, '\n');
633 bool casesensitive = parse_bool(howto);
634 bool matchword = parse_bool(howto);
635 bool all = parse_bool(howto);
636 bool forward = parse_bool(howto, true);
637 bool findnext = parse_bool(howto, true);
638 bool wrap = parse_bool(howto);
639 bool onlysel = parse_bool(howto);
641 if (!bv->cursor().selection())
642 // only selection only makes sense with selection
647 int replace_count = 0;
649 replace_count = replaceAll(bv, search, rplc, casesensitive,
651 update = replace_count > 0;
654 replaceOne(bv, search, rplc, casesensitive, matchword,
655 forward, findnext, wrap, onlysel);
657 replace_count = rv.second;
660 Buffer const & buf = bv->buffer();
662 // emit message signal.
664 buf.message(_("String not found in selection."));
666 buf.message(_("String not found."));
668 if (replace_count == 0) {
669 buf.message(_("String found."));
670 } else if (replace_count == 1) {
671 buf.message(_("String has been replaced."));
673 docstring const str = onlysel
674 ? bformat(_("%1$d strings have been replaced in the selection."), replace_count)
675 : bformat(_("%1$d strings have been replaced."), replace_count);
683 bool findNextChange(BufferView * bv, Cursor & cur, bool const check_wrap)
685 for (; cur; cur.forwardPos())
686 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
690 DocIterator cur_orig(bv->cursor());
691 docstring q = _("End of file reached while searching forward.\n"
692 "Continue searching from the beginning?");
693 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
694 q, 0, 1, _("&Yes"), _("&No"));
695 if (wrap_answer == 0) {
696 bv->cursor().clear();
697 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
698 bv->clearSelection();
699 cur.setCursor(bv->cursor().selectionBegin());
700 if (findNextChange(bv, cur, false))
703 bv->cursor().setCursor(cur_orig);
710 bool findPreviousChange(BufferView * bv, Cursor & cur, bool const check_wrap)
712 for (cur.backwardPos(); cur; cur.backwardPos()) {
713 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
718 DocIterator cur_orig(bv->cursor());
719 docstring q = _("Beginning of file reached while searching backward.\n"
720 "Continue searching from the end?");
721 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
722 q, 0, 1, _("&Yes"), _("&No"));
723 if (wrap_answer == 0) {
724 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
725 bv->cursor().backwardPos();
726 bv->clearSelection();
727 cur.setCursor(bv->cursor().selectionBegin());
728 if (findPreviousChange(bv, cur, false))
731 bv->cursor().setCursor(cur_orig);
738 bool selectChange(Cursor & cur, bool forward)
740 if (!cur.inTexted() || !cur.paragraph().isChanged(cur.pos()))
742 Change ch = cur.paragraph().lookupChange(cur.pos());
744 CursorSlice tip1 = cur.top();
745 for (; tip1.pit() < tip1.lastpit() || tip1.pos() < tip1.lastpos(); tip1.forwardPos()) {
746 Change ch2 = tip1.paragraph().lookupChange(tip1.pos());
747 if (!ch2.isSimilarTo(ch))
750 CursorSlice tip2 = cur.top();
751 for (; tip2.pit() > 0 || tip2.pos() > 0;) {
753 Change ch2 = tip2.paragraph().lookupChange(tip2.pos());
754 if (!ch2.isSimilarTo(ch)) {
755 // take a step forward to correctly set the selection
763 cur.bv().mouseSetCursor(cur, false);
765 cur.bv().mouseSetCursor(cur, true);
773 bool findChange(BufferView * bv, bool forward)
776 cur.setCursor(forward ? bv->cursor().selectionEnd()
777 : bv->cursor().selectionBegin());
778 forward ? findNextChange(bv, cur, true) : findPreviousChange(bv, cur, true);
779 return selectChange(cur, forward);
784 bool findNextChange(BufferView * bv)
786 return findChange(bv, true);
790 bool findPreviousChange(BufferView * bv)
792 return findChange(bv, false);
799 typedef vector<pair<string, string> > Escapes;
801 string string2regex(string in)
803 static std::regex specialChars { R"([-[\]{}()*+?.,\^$|#\s\$\\])" };
804 string temp = std::regex_replace(in, specialChars, R"(\$&)" );
809 while (fl_pos < temp.size()) {
810 fl_pos = temp.find("\\\\foreignlanguage", lastpos + offset);
811 if (fl_pos == string::npos)
814 temp2 += temp.substr(lastpos, fl_pos - lastpos);
820 if (lastpos < temp.size()) {
821 temp2 += temp.substr(lastpos, temp.size() - lastpos);
826 string correctRegex(string t, bool withformat)
828 /* Convert \backslash => \
829 * and \{, \}, \[, \] => {, }, [, ]
832 regex wordre("(\\\\)*(\\\\((backslash|mathcircumflex) ?|[\\[\\]\\{\\}]))");
835 bool backslashed = false;
836 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
839 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
843 if (sub.str(4) == "backslash") {
846 // transforms '\backslash \{' into '\{'
848 string next = t.substr(sub.position(2) + sub.str(2).length(), 2);
849 if ((next == "\\{") || (next == "\\}")) {
855 else if (sub.str(4) == "mathcircumflex")
857 else if (backslashed) {
859 if (withformat && (sub.str(3) == "{"))
860 replace = accents["braceleft"];
861 else if (withformat && (sub.str(3) == "}"))
862 replace = accents["braceright"];
864 // else part should not exist
869 replace = sub.str(3);
871 if (lastpos < (size_t) sub.position(2))
872 s += t.substr(lastpos, sub.position(2) - lastpos);
874 lastpos = sub.position(2) + sub.length(2);
878 else if (lastpos < t.length())
879 s += t.substr(lastpos, t.length() - lastpos);
883 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
884 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
885 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
886 string escape_for_regex(string s, bool withformat)
890 while (lastpos < s.size()) {
891 size_t regex_pos = s.find("\\regexp{", lastpos);
892 if (regex_pos == string::npos) {
893 regex_pos = s.size();
895 if (regex_pos > lastpos) {
896 result += string2regex(s.substr(lastpos, regex_pos-lastpos));
898 if (lastpos == s.size())
901 size_t end_pos = s.find("\\endregexp{}}", regex_pos + 8);
902 result += correctRegex(s.substr(regex_pos + 8, end_pos -(regex_pos + 8)), withformat);
903 lastpos = end_pos + 13;
909 /// Wrapper for lyx::regex_replace with simpler interface
910 bool regex_replace(string const & s, string & t, string const & searchstr,
911 string const & replacestr)
913 regex e(searchstr, regex_constants::ECMAScript);
915 ostream_iterator<char, char> it(oss);
916 regex_replace(it, s.begin(), s.end(), e, replacestr);
917 // tolerate t and s be references to the same variable
918 bool rv = (s != oss.str());
937 vector <string> result = vector <string>();
938 MatchResult(int len = 0): match_len(len),match_prefix(0),match2end(0), pos(0),leadsize(0),pos_len(-1),searched_size(0) {}
941 static MatchResult::range interpretMatch(MatchResult &oldres, MatchResult &newres)
943 if (newres.match2end < oldres.match2end)
944 return MatchResult::newIsTooFar;
945 if (newres.match_len < oldres.match_len)
946 return MatchResult::newIsTooFar;
948 if (newres.match_len == oldres.match_len) {
949 if (newres.match2end == oldres.match2end)
950 return MatchResult::newIsBetter;
952 return MatchResult::newIsInvalid;
955 /** The class performing a match between a position in the document and the FindAdvOptions.
958 class MatchStringAdv {
960 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt);
962 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
963 ** constructor as opt.search, under the opt.* options settings.
966 ** If set, then match is searched only against beginning of text starting at cur.
967 ** If unset, then match is searched anywhere in text starting at cur.
970 ** The length of the matching text, or zero if no match was found.
972 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
981 /// first buffer on which search was started
982 lyx::Buffer * const p_first_buf;
984 FindAndReplaceOptions const & opt;
987 /// Auxiliary find method (does not account for opt.matchword)
988 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
989 void CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string = "");
991 /** Normalize a stringified or latexified LyX paragraph.
995 ** <li>if search is not casesensitive, then lowercase the string;
996 ** <li>remove any newline at begin or end of the string;
997 ** <li>replace any newline in the middle of the string with a simple space;
998 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
1001 ** @todo Normalization should also expand macros, if the corresponding
1002 ** search option was checked.
1004 string normalize(docstring const & s) const;
1005 // normalized string to search
1006 string par_as_string;
1007 // regular expression to use for searching
1008 // regexp2 is same as regexp, but prefixed with a ".*?"
1010 QRegularExpression regexp;
1011 QRegularExpression regexp2;
1016 // leading format material as string
1017 string lead_as_string;
1018 // par_as_string after removal of lead_as_string
1019 string par_as_string_nolead;
1020 // unmatched open braces in the search string/regexp
1022 // number of (.*?) subexpressions added at end of search regexp for closing
1023 // environments, math mode, styles, etc...
1024 int close_wildcards;
1026 // Are we searching with regular expressions ?
1028 static int valid_matches;
1029 static vector <string> matches;
1030 void FillResults(MatchResult &found_mr);
1033 int MatchStringAdv::valid_matches = 0;
1034 vector <string> MatchStringAdv::matches = vector <string>(10);
1036 void MatchStringAdv::FillResults(MatchResult &found_mr)
1038 if (found_mr.match_len > 0) {
1039 valid_matches = found_mr.result.size();
1040 for (size_t i = 0; i < found_mr.result.size(); i++)
1041 matches[i] = found_mr.result[i];
1047 static docstring buffer_to_latex(Buffer & buffer)
1049 //OutputParams runparams(&buffer.params().encoding());
1050 OutputParams runparams(encodings.fromLyXName("utf8"));
1051 odocstringstream ods;
1053 runparams.nice = true;
1054 runparams.flavor = Flavor::XeTeX;
1055 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
1056 // No side effect of file copying and image conversion
1057 runparams.dryrun = true;
1058 if (ignoreFormats.getDeleted())
1059 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
1061 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
1062 pit_type const endpit = buffer.paragraphs().size();
1063 for (pit_type pit = 0; pit != endpit; ++pit) {
1064 TeXOnePar(buffer, buffer.text(), pit, os, runparams);
1065 LYXERR(Debug::FIND, "searchString up to here: " << ods.str());
1071 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
1074 if (!opt.ignoreformat) {
1075 str = buffer_to_latex(buffer);
1077 // OutputParams runparams(&buffer.params().encoding());
1078 OutputParams runparams(encodings.fromLyXName("utf8"));
1079 runparams.nice = true;
1080 runparams.flavor = Flavor::XeTeX;
1081 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
1082 runparams.dryrun = true;
1083 int option = AS_STR_INSETS |AS_STR_PLAINTEXT;
1084 if (ignoreFormats.getDeleted()) {
1085 option |= AS_STR_SKIPDELETE;
1086 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
1089 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
1091 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
1092 Paragraph const & par = buffer.paragraphs().at(pit);
1093 LYXERR(Debug::FIND, "Adding to search string: '"
1094 << par.asString(pos_type(0), par.size(),
1098 str += par.asString(pos_type(0), par.size(),
1102 // Even in ignore-format we have to remove "\text{}, \lyxmathsym{}" parts
1103 string t = to_utf8(str);
1104 while (regex_replace(t, t, "\\\\(text|lyxmathsym|ensuremath)\\{([^\\}]*)\\}", "$2"));
1111 /// Return separation pos between the leading material and the rest
1112 static size_t identifyLeading(string const & s)
1115 // @TODO Support \item[text]
1116 // Kornel: Added textsl, textsf, textit, texttt and noun
1117 // + allow to search for colored text too
1118 while (regex_replace(t, t, "^\\\\(("
1119 "(author|title|subtitle|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|"
1120 "lyxaddress|lyxrightaddress|"
1121 "footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1122 "emph|noun|minisec|text(bf|md|sl|sf|it|tt))|"
1123 "((textcolor|foreignlanguage|latexenvironment)\\{[a-z]+\\*?\\})|"
1124 "(u|uu)line|(s|x)out|uwave)|((sub)?(((sub)?section)|paragraph)|part|chapter)\\*?)\\{", "")
1125 || regex_replace(t, t, "^\\$", "")
1126 || regex_replace(t, t, "^\\\\\\[", "")
1127 || regex_replace(t, t, "^ ?\\\\item\\{[a-z]+\\}", "")
1128 || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\}", ""))
1130 LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
1135 * Given a latexified string, retrieve some handled features
1136 * The features of the regex will later be compared with the features
1137 * of the searched text. If the regex features are not a
1138 * subset of the analized, then, in not format ignoring search
1139 * we can early stop the search in the relevant inset.
1141 typedef map<string, bool> Features;
1143 static Features identifyFeatures(string const & s)
1145 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
1146 static regex const valid("^("
1148 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1149 "emph|noun|text(bf|md|sl|sf|it|tt)|"
1150 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
1151 "(u|uu)line|(s|x)out|uwave|"
1152 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
1153 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
1158 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1161 if (sub.str(1).compare("regexp") == 0) {
1165 string token = sub.str(1);
1167 if (regex_match(token, sub2, valid)) {
1175 if (sub.str(1).compare("endregexp") == 0) {
1185 * defines values features of a key "\\[a-z]+{"
1190 /* Char type with content discarded
1191 * like \hspace{1cm} */
1193 /* Char, like \backslash */
1195 /* replace starting backslash with '#' */
1197 /* \part, \section*, ... */
1199 /* title, author etc */
1201 /* \foreignlanguage{ngerman}, ... */
1204 * to discard language in content */
1207 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1209 /* fonts, colors, markups, ... */
1211 /* footnotesize, ... large, ...
1212 * Ignore all of them */
1215 /* inputencoding, ...
1216 * Discard also content, because they do not help in search */
1219 * like remove, but also all arguments */
1221 /* item, listitem */
1223 /* tex, latex, ... like isChar */
1225 /* like \lettrine[lines=5]{}{} */
1227 // like isStandard, but always remove head
1229 /* End of arguments marker for lettrine,
1230 * so that they can be ignored */
1233 KeyInfo() = default;
1234 KeyInfo(KeyType type, int parcount, bool disable)
1236 parenthesiscount(parcount),
1237 disabled(disable) {}
1238 KeyType keytype = invalid;
1240 int _tokensize = -1;
1241 int _tokenstart = -1;
1242 int _dataStart = -1;
1244 int parenthesiscount = 1;
1245 bool disabled = false;
1246 bool used = false; /* by pattern */
1251 Border(int l=0, int u=0) : low(l), upper(u) {}
1256 #define MAXOPENED 30
1258 bool isPatternString_;
1260 explicit Intervall(bool isPattern, string const & p) :
1261 isPatternString_(isPattern), par(p), ignoreidx(-1), actualdeptindex(0),
1262 hasTitle(false), langcount(0)
1270 static vector<Border> borders;
1271 int depts[MAXOPENED];
1272 int closes[MAXOPENED];
1273 int actualdeptindex;
1274 int previousNotIgnored(int) const;
1275 int nextNotIgnored(int) const;
1276 void handleOpenP(int i);
1277 void handleCloseP(int i, bool closingAllowed);
1278 void resetOpenedP(int openPos);
1279 void addIntervall(int upper);
1280 void addIntervall(int low, int upper); /* if explicit */
1281 void removeAccents();
1282 void setForDefaultLang(KeyInfo const & defLang) const;
1283 int findclosing(int start, int end, char up, char down, int repeat);
1284 void handleParentheses(int lastpos, bool closingAllowed);
1286 int langcount; // Number of disabled language specs up to current position in actual interval
1287 int isOpeningPar(int pos) const;
1289 void output(ostringstream &os, int lastpos);
1290 // string show(int lastpos);
1293 vector<Border> Intervall::borders = vector<Border>(30);
1295 int Intervall::isOpeningPar(int pos) const
1297 if ((pos < 0) || (size_t(pos) >= par.size()))
1299 if (par[pos] != '{')
1301 if (size_t(pos) + 2 >= par.size())
1303 if (par[pos+2] != '}')
1305 if (par[pos+1] == '[' || par[pos+1] == ']')
1310 void Intervall::setForDefaultLang(KeyInfo const & defLang) const
1312 // Enable the use of first token again
1313 if (ignoreidx >= 0) {
1314 int value = defLang._tokenstart + defLang._tokensize;
1320 if (borders[borderidx].low < value)
1321 borders[borderidx].low = value;
1322 if (borders[borderidx].upper < value)
1323 borders[borderidx].upper = value;
1328 static void checkDepthIndex(int val)
1330 static int maxdepthidx = MAXOPENED-2;
1331 static int lastmaxdepth = 0;
1332 if (val > lastmaxdepth) {
1333 LYXERR(Debug::INFO, "Depth reached " << val);
1336 if (val > maxdepthidx) {
1338 LYXERR(Debug::INFO, "maxdepthidx now " << val);
1343 // Not needed, because borders are now dynamically expanded
1344 static void checkIgnoreIdx(int val)
1346 static int lastmaxignore = -1;
1347 if ((lastmaxignore < val) && (size_t(val+1) >= borders.size())) {
1348 LYXERR(Debug::INFO, "IgnoreIdx reached " << val);
1349 lastmaxignore = val;
1355 * Expand the region of ignored parts of the input latex string
1356 * The region is only relevant in output()
1358 void Intervall::addIntervall(int low, int upper)
1361 if (low == upper) return;
1362 for (idx = ignoreidx+1; idx > 0; --idx) {
1363 if (low > borders[idx-1].upper) {
1367 Border br(low, upper);
1368 if (idx > ignoreidx) {
1369 if (borders.size() <= size_t(idx)) {
1370 borders.push_back(br);
1376 // checkIgnoreIdx(ignoreidx);
1380 // Expand only if one of the new bound is inside the interwall
1381 // We know here that br.low > borders[idx-1].upper
1382 if (br.upper < borders[idx].low) {
1383 // We have to insert at this pos
1384 if (size_t(ignoreidx+1) >= borders.size()) {
1385 borders.push_back(borders[ignoreidx]);
1388 borders[ignoreidx+1] = borders[ignoreidx];
1390 for (int i = ignoreidx; i > idx; --i) {
1391 borders[i] = borders[i-1];
1395 // checkIgnoreIdx(ignoreidx);
1398 // Here we know, that we are overlapping
1399 if (br.low > borders[idx].low)
1400 br.low = borders[idx].low;
1401 // check what has to be concatenated
1403 for (int i = idx; i <= ignoreidx; i++) {
1404 if (br.upper >= borders[i].low) {
1406 if (br.upper < borders[i].upper)
1407 br.upper = borders[i].upper;
1413 // count should be >= 1 here
1416 for (int i = idx + count; i <= ignoreidx; i++) {
1417 borders[i-count+1] = borders[i];
1419 ignoreidx -= count - 1;
1425 static void buildaccent(string n, string param, string values)
1429 const char delim = '|';
1430 while (getline(s, name, delim)) {
1432 for (char c : param) {
1433 string key = name + "{" + c + "}";
1434 // get the corresponding utf8-value
1435 if ((values[start] & 0xc0) != 0xc0) {
1436 // should not happen, utf8 encoding starts at least with 11xxxxxx
1437 // but value for '\dot{i}' is 'i', which is ascii
1438 if ((values[start] & 0x80) == 0) {
1440 accents[key] = values.substr(start, 1);
1441 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1446 for (int j = 1; ;j++) {
1447 if (start + j >= values.size()) {
1448 accents[key] = values.substr(start, j);
1449 start = values.size() - 1;
1452 else if ((values[start+j] & 0xc0) != 0x80) {
1453 // This is the first byte of following utf8 char
1454 accents[key] = values.substr(start, j);
1456 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1465 static string getutf8(unsigned uchar)
1468 string ret = string();
1470 if (uchar <= 0x7f) {
1471 c[maxc-1] = uchar & 0x7f;
1474 unsigned char rest = 0x40;
1475 unsigned char first = 0x80;
1477 for (int i = start; i >=0; --i) {
1479 c[i] = first + uchar;
1482 c[i] = 0x80 | (uchar & 0x3f);
1489 for (int i = 0; i < maxc; i++) {
1490 if (c[i] == 0) continue;
1496 static void addAccents(string latex_in, string unicode_out)
1498 latex_in = latex_in.substr(1);
1499 AccentsIterator it_ac = accents.find(latex_in);
1500 if (it_ac == accents.end()) {
1501 accents[latex_in] = unicode_out;
1504 LYXERR0("Key " << latex_in << " already set");
1508 void static fillMissingUnicodesymbols()
1510 addAccents("\\textyen", getutf8(0x00a5));
1511 addAccents("\\yen", getutf8(0x00a5));
1512 addAccents("\\textsection", getutf8(0x00a7));
1513 addAccents("\\mathsection", getutf8(0x00a7));
1514 addAccents("\\textlnot", getutf8(0x00ac));
1515 addAccents("\\neg", getutf8(0x00ac));
1516 addAccents("\\textpm", getutf8(0x00b1));
1517 addAccents("\\pm", getutf8(0x00b1));
1518 addAccents("\\textparagraph", getutf8(0x00b6));
1519 addAccents("\\mathparagraph", getutf8(0x00b6));
1520 addAccents("\\textperiodcentered", getutf8(0x00b7));
1521 addAccents("\\texttimes", getutf8(0x00d7));
1522 addAccents("\\times", getutf8(0x00d7));
1523 addAccents("\\dh", getutf8(0x00f0));
1524 addAccents("\\eth", getutf8(0x00f0));
1525 addAccents("\\textdiv", getutf8(0x00f7));
1526 addAccents("\\div", getutf8(0x00f7));
1527 addAccents("\\o", getutf8(0x00f8));
1528 addAccents("\\textcrlambda", getutf8(0x019b));
1529 addAccents("\\j", getutf8(0x0237));
1530 addAccents("\\textalpha", getutf8(0x03b1));
1531 addAccents("\\alpha", getutf8(0x03b1));
1532 addAccents("\\textbeta", getutf8(0x03b2));
1533 addAccents("\\beta", getutf8(0x03b2));
1534 addAccents("\\textgamma", getutf8(0x03b3));
1535 addAccents("\\gamma", getutf8(0x03b3));
1536 addAccents("\\textdelta", getutf8(0x03b4));
1537 addAccents("\\delta", getutf8(0x03b4));
1538 addAccents("\\textepsilon", getutf8(0x03b5));
1539 addAccents("\\varepsilon", getutf8(0x03b5));
1540 addAccents("\\textzeta", getutf8(0x03b6));
1541 addAccents("\\zeta", getutf8(0x03b6));
1542 addAccents("\\texteta", getutf8(0x03b7));
1543 addAccents("\\eta", getutf8(0x03b7));
1544 addAccents("\\texttheta", getutf8(0x03b8));
1545 addAccents("\\theta", getutf8(0x03b8));
1546 addAccents("\\textiota", getutf8(0x03b9));
1547 addAccents("\\iota", getutf8(0x03b9));
1548 addAccents("\\textkappa", getutf8(0x03ba));
1549 addAccents("\\kappa", getutf8(0x03ba));
1550 addAccents("\\textlambda", getutf8(0x03bb));
1551 addAccents("\\lambda", getutf8(0x03bb));
1552 addAccents("\\textmu", getutf8(0x03bc));
1553 addAccents("\\mu", getutf8(0x03bc));
1554 addAccents("\\textnu", getutf8(0x03bd));
1555 addAccents("\\nu", getutf8(0x03bd));
1556 addAccents("\\textxi", getutf8(0x03be));
1557 addAccents("\\xi", getutf8(0x03be));
1558 addAccents("\\textpi", getutf8(0x03c0));
1559 addAccents("\\pi", getutf8(0x03c0));
1560 addAccents("\\textrho", getutf8(0x03c1));
1561 addAccents("\\rho", getutf8(0x03c1));
1562 addAccents("\\textfinalsigma", getutf8(0x03c2));
1563 addAccents("\\varsigma", getutf8(0x03c2));
1564 addAccents("\\textsigma", getutf8(0x03c3));
1565 addAccents("\\sigma", getutf8(0x03c3));
1566 addAccents("\\texttau", getutf8(0x03c4));
1567 addAccents("\\tau", getutf8(0x03c4));
1568 addAccents("\\textupsilon", getutf8(0x03c5));
1569 addAccents("\\upsilon", getutf8(0x03c5));
1570 addAccents("\\textphi", getutf8(0x03c6));
1571 addAccents("\\varphi", getutf8(0x03c6));
1572 addAccents("\\textchi", getutf8(0x03c7));
1573 addAccents("\\chi", getutf8(0x03c7));
1574 addAccents("\\textpsi", getutf8(0x03c8));
1575 addAccents("\\psi", getutf8(0x03c8));
1576 addAccents("\\textomega", getutf8(0x03c9));
1577 addAccents("\\omega", getutf8(0x03c9));
1578 addAccents("\\textdigamma", getutf8(0x03dd));
1579 addAccents("\\digamma", getutf8(0x03dd));
1580 addAccents("\\hebalef", getutf8(0x05d0));
1581 addAccents("\\aleph", getutf8(0x05d0));
1582 addAccents("\\hebbet", getutf8(0x05d1));
1583 addAccents("\\beth", getutf8(0x05d1));
1584 addAccents("\\hebgimel", getutf8(0x05d2));
1585 addAccents("\\gimel", getutf8(0x05d2));
1586 addAccents("\\hebdalet", getutf8(0x05d3));
1587 addAccents("\\daleth", getutf8(0x05d3));
1588 addAccents("\\hebhe", getutf8(0x05d4));
1589 addAccents("\\hebvav", getutf8(0x05d5));
1590 addAccents("\\hebzayin", getutf8(0x05d6));
1591 addAccents("\\hebhet", getutf8(0x05d7));
1592 addAccents("\\hebtet", getutf8(0x05d8));
1593 addAccents("\\hebyod", getutf8(0x05d9));
1594 addAccents("\\hebfinalkaf", getutf8(0x05da));
1595 addAccents("\\hebkaf", getutf8(0x05db));
1596 addAccents("\\heblamed", getutf8(0x05dc));
1597 addAccents("\\hebfinalmem", getutf8(0x05dd));
1598 addAccents("\\hebmem", getutf8(0x05de));
1599 addAccents("\\hebfinalnun", getutf8(0x05df));
1600 addAccents("\\hebnun", getutf8(0x05e0));
1601 addAccents("\\hebsamekh", getutf8(0x05e1));
1602 addAccents("\\hebayin", getutf8(0x05e2));
1603 addAccents("\\hebfinalpe", getutf8(0x05e3));
1604 addAccents("\\hebpe", getutf8(0x05e4));
1605 addAccents("\\hebfinaltsadi", getutf8(0x05e5));
1606 addAccents("\\hebtsadi", getutf8(0x05e6));
1607 addAccents("\\hebqof", getutf8(0x05e7));
1608 addAccents("\\hebresh", getutf8(0x05e8));
1609 addAccents("\\hebshin", getutf8(0x05e9));
1610 addAccents("\\hebtav", getutf8(0x05ea));
1613 addAccents("\\thaiKoKai", getutf8(0x0e01));
1614 addAccents("\\thaiKhoKhai", getutf8(0x0e02));
1615 addAccents("\\thaiKhoKhuat", getutf8(0x0e03));
1616 addAccents("\\thaiKhoKhwai", getutf8(0x0e04));
1617 addAccents("\\thaiKhoKhon", getutf8(0x0e05));
1618 addAccents("\\thaiKhoRakhang", getutf8(0x0e06));
1619 addAccents("\\thaiNgoNgu", getutf8(0x0e07));
1620 addAccents("\\thaiChoChan", getutf8(0x0e08));
1621 addAccents("\\thaiChoChing", getutf8(0x0e09));
1622 addAccents("\\thaiChoChang", getutf8(0x0e0a));
1623 addAccents("\\thaiSoSo", getutf8(0x0e0b));
1624 addAccents("\\thaiChoChoe", getutf8(0x0e0c));
1625 addAccents("\\thaiYoYing", getutf8(0x0e0d));
1626 addAccents("\\thaiDoChada", getutf8(0x0e0e));
1627 addAccents("\\thaiToPatak", getutf8(0x0e0f));
1628 addAccents("\\thaiThoThan", getutf8(0x0e10));
1629 addAccents("\\thaiThoNangmontho", getutf8(0x0e11));
1630 addAccents("\\thaiThoPhuthao", getutf8(0x0e12));
1631 addAccents("\\thaiNoNen", getutf8(0x0e13));
1632 addAccents("\\thaiDoDek", getutf8(0x0e14));
1633 addAccents("\\thaiToTao", getutf8(0x0e15));
1634 addAccents("\\thaiThoThung", getutf8(0x0e16));
1635 addAccents("\\thaiThoThahan", getutf8(0x0e17));
1636 addAccents("\\thaiThoThong", getutf8(0x0e18));
1637 addAccents("\\thaiNoNu", getutf8(0x0e19));
1638 addAccents("\\thaiBoBaimai", getutf8(0x0e1a));
1639 addAccents("\\thaiPoPla", getutf8(0x0e1b));
1640 addAccents("\\thaiPhoPhung", getutf8(0x0e1c));
1641 addAccents("\\thaiFoFa", getutf8(0x0e1d));
1642 addAccents("\\thaiPhoPhan", getutf8(0x0e1e));
1643 addAccents("\\thaiFoFan", getutf8(0x0e1f));
1644 addAccents("\\thaiPhoSamphao", getutf8(0x0e20));
1645 addAccents("\\thaiMoMa", getutf8(0x0e21));
1646 addAccents("\\thaiYoYak", getutf8(0x0e22));
1647 addAccents("\\thaiRoRua", getutf8(0x0e23));
1648 addAccents("\\thaiRu", getutf8(0x0e24));
1649 addAccents("\\thaiLoLing", getutf8(0x0e25));
1650 addAccents("\\thaiLu", getutf8(0x0e26));
1651 addAccents("\\thaiWoWaen", getutf8(0x0e27));
1652 addAccents("\\thaiSoSala", getutf8(0x0e28));
1653 addAccents("\\thaiSoRusi", getutf8(0x0e29));
1654 addAccents("\\thaiSoSua", getutf8(0x0e2a));
1655 addAccents("\\thaiHoHip", getutf8(0x0e2b));
1656 addAccents("\\thaiLoChula", getutf8(0x0e2c));
1657 addAccents("\\thaiOAng", getutf8(0x0e2d));
1658 addAccents("\\thaiHoNokhuk", getutf8(0x0e2e));
1659 addAccents("\\thaiPaiyannoi", getutf8(0x0e2f));
1660 addAccents("\\thaiSaraA", getutf8(0x0e30));
1661 addAccents("\\thaiMaiHanakat", getutf8(0x0e31));
1662 addAccents("\\thaiSaraAa", getutf8(0x0e32));
1663 addAccents("\\thaiSaraAm", getutf8(0x0e33));
1664 addAccents("\\thaiSaraI", getutf8(0x0e34));
1665 addAccents("\\thaiSaraIi", getutf8(0x0e35));
1666 addAccents("\\thaiSaraUe", getutf8(0x0e36));
1667 addAccents("\\thaiSaraUee", getutf8(0x0e37));
1668 addAccents("\\thaiSaraU", getutf8(0x0e38));
1669 addAccents("\\thaiSaraUu", getutf8(0x0e39));
1670 addAccents("\\thaiPhinthu", getutf8(0x0e3a));
1671 addAccents("\\thaiSaraE", getutf8(0x0e40));
1672 addAccents("\\thaiSaraAe", getutf8(0x0e41));
1673 addAccents("\\thaiSaraO", getutf8(0x0e42));
1674 addAccents("\\thaiSaraAiMaimuan", getutf8(0x0e43));
1675 addAccents("\\thaiSaraAiMaimalai", getutf8(0x0e44));
1676 addAccents("\\thaiLakkhangyao", getutf8(0x0e45));
1677 addAccents("\\thaiMaiyamok", getutf8(0x0e46));
1678 addAccents("\\thaiMaitaikhu", getutf8(0x0e47));
1679 addAccents("\\thaiMaiEk", getutf8(0x0e48));
1680 addAccents("\\thaiMaiTho", getutf8(0x0e49));
1681 addAccents("\\thaiMaiTri", getutf8(0x0e4a));
1682 addAccents("\\thaiMaiChattawa", getutf8(0x0e4b));
1683 addAccents("\\thaiThanthakhat", getutf8(0x0e4c));
1684 addAccents("\\thaiNikhahit", getutf8(0x0e4d));
1685 addAccents("\\thaiYamakkan", getutf8(0x0e4e));
1686 addAccents("\\thaiFongman", getutf8(0x0e4f));
1687 addAccents("\\thaizero", getutf8(0x0e50));
1688 addAccents("\\thaione", getutf8(0x0e51));
1689 addAccents("\\thaitwo", getutf8(0x0e52));
1690 addAccents("\\thaithree", getutf8(0x0e53));
1691 addAccents("\\thaifour", getutf8(0x0e54));
1692 addAccents("\\thaifive", getutf8(0x0e55));
1693 addAccents("\\thaisix", getutf8(0x0e56));
1694 addAccents("\\thaiseven", getutf8(0x0e57));
1695 addAccents("\\thaieight", getutf8(0x0e58));
1696 addAccents("\\thainine", getutf8(0x0e59));
1697 addAccents("\\thaiAngkhankhu", getutf8(0x0e5a));
1698 addAccents("\\thaiKhomut", getutf8(0x0e5b));
1700 addAccents("\\dag", getutf8(0x2020));
1701 addAccents("\\dagger", getutf8(0x2020));
1702 addAccents("\\ddag", getutf8(0x2021));
1703 addAccents("\\ddagger", getutf8(0x2021));
1704 addAccents("\\textbullet", getutf8(0x2022));
1705 addAccents("\\bullet", getutf8(0x2022));
1706 addAccents("\\dots", getutf8(0x2026));
1707 addAccents("\\ldots", getutf8(0x2026));
1708 addAccents("\\textasciiacute", getutf8(0x2032));
1709 addAccents("\\prime", getutf8(0x2032));
1710 addAccents("\\textasciigrave", getutf8(0x2035));
1711 addAccents("\\backprime", getutf8(0x2035));
1712 addAccents("\\textasteriskcentered", getutf8(0x204e));
1713 addAccents("\\ast", getutf8(0x204e));
1714 addAccents("\\textmho", getutf8(0x2127));
1715 addAccents("\\mho", getutf8(0x2127));
1716 addAccents("\\textleftarrow", getutf8(0x2190));
1717 addAccents("\\leftarrow", getutf8(0x2190));
1718 addAccents("\\textuparrow", getutf8(0x2191));
1719 addAccents("\\uparrow", getutf8(0x2191));
1720 addAccents("\\textrightarrow", getutf8(0x2192));
1721 addAccents("\\rightarrow", getutf8(0x2192));
1722 addAccents("\\textdownarrow", getutf8(0x2193));
1723 addAccents("\\downarrow", getutf8(0x2193));
1724 addAccents("\\textglobrise", getutf8(0x2197));
1725 addAccents("\\nearrow", getutf8(0x2197));
1726 addAccents("\\textglobfall", getutf8(0x2198));
1727 addAccents("\\searrow", getutf8(0x2198));
1728 addAccents("\\textsurd", getutf8(0x221a));
1729 addAccents("\\surd", getutf8(0x221a));
1730 addAccents("\\textbigcircle", getutf8(0x25ef));
1731 addAccents("\\bigcirc", getutf8(0x25ef));
1732 addAccents("\\textlangle", getutf8(0x27e8));
1733 addAccents("\\langle", getutf8(0x27e8));
1734 addAccents("\\textrangle", getutf8(0x27e9));
1735 addAccents("\\rangle", getutf8(0x27e9));
1738 static void buildAccentsMap()
1740 accents["imath"] = "ı";
1742 accents["jmath"] = "ȷ";
1743 accents["cdot"] = "·";
1744 accents["textasciicircum"] = "^";
1745 accents["mathcircumflex"] = "^";
1746 accents["sim"] = "~";
1747 accents["guillemotright"] = "»";
1748 accents["guillemotleft"] = "«";
1749 accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
1750 accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
1751 accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
1752 accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
1753 accents["negmedspace"] = getutf8(0xf0005);
1754 accents["thickspace"] = getutf8(0xf0006);
1755 accents["negthickspace"] = getutf8(0xf0007);
1756 accents["lyx"] = getutf8(0xf0010); // Used logos
1757 accents["LyX"] = getutf8(0xf0010);
1758 accents["tex"] = getutf8(0xf0011);
1759 accents["TeX"] = getutf8(0xf0011);
1760 accents["latex"] = getutf8(0xf0012);
1761 accents["LaTeX"] = getutf8(0xf0012);
1762 accents["latexe"] = getutf8(0xf0013);
1763 accents["LaTeXe"] = getutf8(0xf0013);
1764 accents["lyxarrow"] = getutf8(0xf0020);
1765 accents["braceleft"] = getutf8(0xf0030);
1766 accents["braceright"] = getutf8(0xf0031);
1767 accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash
1768 accents["backslash LyX"] = getutf8(0xf0010);
1769 accents["backslash tex"] = getutf8(0xf0011);
1770 accents["backslash TeX"] = getutf8(0xf0011);
1771 accents["backslash latex"] = getutf8(0xf0012);
1772 accents["backslash LaTeX"] = getutf8(0xf0012);
1773 accents["backslash latexe"] = getutf8(0xf0013);
1774 accents["backslash LaTeXe"] = getutf8(0xf0013);
1775 accents["backslash lyxarrow"] = getutf8(0xf0020);
1776 accents["ddot{\\imath}"] = "ï";
1777 buildaccent("ddot", "aAeEhHiIioOtuUwWxXyY",
1778 "äÄëËḧḦïÏïöÖẗüÜẅẄẍẌÿŸ"); // umlaut
1779 buildaccent("dot|.", "aAbBcCdDeEfFGghHIimMnNoOpPrRsStTwWxXyYzZ",
1780 "ȧȦḃḂċĊḋḊėĖḟḞĠġḣḢİİṁṀṅṄȯȮṗṖṙṘṡṠṫṪẇẆẋẊẏẎżŻ"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'İ'
1781 accents["acute{\\imath}"] = "í";
1782 buildaccent("acute", "aAcCeEgGkKlLmMoOnNpPrRsSuUwWyYzZiI",
1783 "áÁćĆéÉǵǴḱḰĺĹḿḾóÓńŃṕṔŕŔśŚúÚẃẂýÝźŹíÍ");
1784 buildaccent("dacute|H|h", "oOuU", "őŐűŰ"); // double acute
1785 buildaccent("mathring|r", "aAuUwy",
1787 accents["check{\\imath}"] = "ǐ";
1788 accents["check{\\jmath}"] = "ǰ";
1789 buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
1790 "čČďĎǎǍěĚǐǏǒǑǔǓǧǦǩǨȟȞľĽňŇřŘšŠŤťžŽ"); // caron
1791 accents["hat{\\imath}"] = "î";
1792 accents["hat{\\jmath}"] = "ĵ";
1793 buildaccent("hat|^", "aAcCeEgGhHiIjJoOsSuUwWyYzZ",
1794 "âÂĉĈêÊĝĜĥĤîÎĵĴôÔŝŜûÛŵŴŷŶẑẐ"); // circ
1795 accents["bar{\\imath}"] = "ī";
1796 buildaccent("bar|=", "aAeEiIoOuUyY",
1797 "āĀēĒīĪōŌūŪȳȲ"); // macron
1798 accents["tilde{\\imath}"] = "ĩ";
1799 buildaccent("tilde", "aAeEiInNoOuUvVyY",
1800 "ãÃẽẼĩĨñÑõÕũŨṽṼỹỸ"); // tilde
1801 accents["breve{\\imath}"] = "ĭ";
1802 buildaccent("breve|u", "aAeEgGiIoOuU",
1803 "ăĂĕĔğĞĭĬŏŎŭŬ"); // breve
1804 accents["grave{\\imath}"] = "ì";
1805 buildaccent("grave|`", "aAeEiIoOuUnNwWyY",
1806 "àÀèÈìÌòÒùÙǹǸẁẀỳỲ"); // grave
1807 buildaccent("subdot|d", "BbDdHhKkLlMmNnRrSsTtVvWwZzAaEeIiOoUuYy",
1808 "ḄḅḌḍḤḥḲḳḶḷṂṃṆṇṚṛṢṣṬṭṾṿẈẉẒẓẠạẸẹỊịỌọỤụỴỵ"); // dot below
1809 buildaccent("ogonek|k", "AaEeIiUuOo",
1810 "ĄąĘęĮįŲųǪǫ"); // ogonek
1811 buildaccent("cedilla|c", "CcGgKkLlNnRrSsTtEeDdHh",
1812 "ÇçĢģĶķĻļŅņŖŗŞşŢţȨȩḐḑḨḩ"); // cedilla
1813 buildaccent("subring|textsubring", "Aa",
1815 buildaccent("subhat|textsubcircum", "DdEeLlNnTtUu",
1816 "ḒḓḘḙḼḽṊṋṰṱṶṷ"); // subcircum
1817 buildaccent("subtilde|textsubtilde", "EeIiUu",
1818 "ḚḛḬḭṴṵ"); // subtilde
1819 accents["dgrave{\\imath}"] = "ȉ";
1820 accents["textdoublegrave{\\i}"] = "ȉ";
1821 buildaccent("dgrave|textdoublegrave", "AaEeIiOoRrUu",
1822 "ȀȁȄȅȈȉȌȍȐȑȔȕ"); // double grave
1823 accents["rcap{\\imath}"] = "ȉ";
1824 accents["textroundcap{\\i}"] = "ȉ";
1825 buildaccent("rcap|textroundcap", "AaEeIiOoRrUu",
1826 "ȂȃȆȇȊȋȎȏȒȓȖȗ"); // inverted breve
1827 buildaccent("slashed", "oO",
1829 fillMissingUnicodesymbols(); // Add some still not handled entries contained in 'unicodesynbols'
1830 // LYXERR0("Number of accents " << accents.size());
1834 * Created accents in math or regexp environment
1835 * are macros, but we need the utf8 equivalent
1837 void Intervall::removeAccents()
1839 if (accents.empty())
1841 static regex const accre("\\\\(([\\S]|grave|breve|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot|ogonek|"
1842 "cedilla|subring|textsubring|subhat|textsubcircum|subtilde|textsubtilde|dgrave|textdoublegrave|rcap|textroundcap|slashed)\\{[^\\{\\}]+\\}"
1843 "|((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]))");
1845 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
1847 string key = sub.str(1);
1848 AccentsIterator it_ac = accents.find(key);
1849 if (it_ac != accents.end()) {
1850 string val = it_ac->second;
1851 size_t pos = sub.position(size_t(0));
1852 for (size_t i = 0; i < val.size(); i++) {
1853 par[pos+i] = val[i];
1855 // Remove possibly following space too
1856 if (par[pos+sub.str(0).size()] == ' ')
1857 addIntervall(pos+val.size(), pos + sub.str(0).size()+1);
1859 addIntervall(pos+val.size(), pos + sub.str(0).size());
1860 for (size_t i = pos+val.size(); i < pos + sub.str(0).size(); i++) {
1861 // remove traces of any remaining chars
1866 LYXERR(Debug::INFO, "Not added accent for \"" << key << "\"");
1871 void Intervall::handleOpenP(int i)
1874 depts[actualdeptindex] = i+1;
1875 closes[actualdeptindex] = -1;
1876 checkDepthIndex(actualdeptindex);
1879 void Intervall::handleCloseP(int i, bool closingAllowed)
1881 if (actualdeptindex <= 0) {
1882 if (! closingAllowed)
1883 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1884 // if we are at the very end
1885 addIntervall(i, i+1);
1888 closes[actualdeptindex] = i+1;
1893 void Intervall::resetOpenedP(int openPos)
1895 // Used as initializer for foreignlanguage entry
1896 actualdeptindex = 1;
1897 depts[1] = openPos+1;
1901 int Intervall::previousNotIgnored(int start) const
1903 int idx = 0; /* int intervalls */
1904 for (idx = ignoreidx; idx >= 0; --idx) {
1905 if (start > borders[idx].upper)
1907 if (start >= borders[idx].low)
1908 start = borders[idx].low-1;
1913 int Intervall::nextNotIgnored(int start) const
1915 int idx = 0; /* int intervalls */
1916 for (idx = 0; idx <= ignoreidx; idx++) {
1917 if (start < borders[idx].low)
1919 if (start < borders[idx].upper)
1920 start = borders[idx].upper;
1925 typedef unordered_map<string, KeyInfo> KeysMap;
1926 typedef unordered_map<string, KeyInfo>::const_iterator KeysIterator;
1927 typedef vector< KeyInfo> Entries;
1928 static KeysMap keys = unordered_map<string, KeyInfo>();
1934 Intervall interval_;
1935 void buildKeys(bool);
1936 void buildEntries(bool);
1937 void makeKey(const string &, KeyInfo, bool isPatternString);
1938 void processRegion(int start, int region_end); /* remove {} parts */
1939 void removeHead(KeyInfo const &, int count=0);
1942 LatexInfo(string const & par, bool isPatternString)
1943 : entidx_(-1), interval_(isPatternString, par)
1945 buildKeys(isPatternString);
1946 entries_ = vector<KeyInfo>();
1947 buildEntries(isPatternString);
1951 if (entries_.empty()) {
1954 if (entries_[0].keytype == KeyInfo::isTitle) {
1955 interval_.hasTitle = true;
1956 if (! entries_[0].disabled) {
1957 interval_.titleValue = entries_[0].head;
1960 interval_.titleValue = "";
1962 removeHead(entries_[0]);
1963 if (entries_.size() > 1)
1972 if (int(entries_.size()) > entidx_) {
1979 bool setNextKey(int idx) {
1980 if ((idx == entidx_) && (entidx_ >= 0)) {
1987 int find(int start, KeyInfo::KeyType keytype) const {
1991 while (tmpIdx < int(entries_.size())) {
1992 if (entries_[tmpIdx].keytype == keytype)
1998 int process(ostringstream & os, KeyInfo const & actual);
1999 int dispatch(ostringstream & os, int previousStart, KeyInfo & actual);
2000 // string show(int lastpos) { return interval.show(lastpos);}
2001 int nextNotIgnored(int start) { return interval_.nextNotIgnored(start);}
2002 KeyInfo &getKeyInfo(int keyinfo) {
2003 static KeyInfo invalidInfo = KeyInfo();
2004 if ((keyinfo < 0) || ( keyinfo >= int(entries_.size())))
2007 return entries_[keyinfo];
2009 void setForDefaultLang(KeyInfo const & defLang) {interval_.setForDefaultLang(defLang);}
2010 void addIntervall(int low, int up) { interval_.addIntervall(low, up); }
2014 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
2018 for (int i = start; i < end; i += 1 + skip) {
2022 if (c == '\\') skip = 1;
2026 else if (c == down) {
2029 if ((repeat <= 0) || (par[i+1] != up))
2043 size_t mathpostfixsize;
2045 size_t mathprefixsize;
2049 vector<MathEntry> entries_;
2054 void insert(string const & wait, size_t start, size_t prefixsize, size_t end, size_t postfixsize) {
2055 MathEntry m = MathEntry();
2057 m.mathStart = start;
2058 m.mathprefixsize = prefixsize;
2059 m.mathEnd = end + postfixsize;
2060 m.mathpostfixsize = postfixsize;
2061 m.mathSize = m.mathEnd - m.mathStart;
2062 entries_.push_back(m);
2064 bool empty() const { return entries_.empty(); }
2065 size_t getEndPos() const {
2066 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2069 return entries_[actualIdx_].mathEnd;
2071 size_t getStartPos() const {
2072 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2073 return 100000; /* definitely enough? */
2075 return entries_[actualIdx_].mathStart;
2077 size_t getPrefixSize() const {
2078 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2081 return entries_[actualIdx_].mathprefixsize;
2083 size_t getPostfixSize() const {
2084 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2087 return entries_[actualIdx_].mathpostfixsize;
2089 size_t getFirstPos() {
2091 return getStartPos();
2093 size_t getSize() const {
2094 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2097 return entries_[actualIdx_].mathSize;
2099 void incrEntry() { actualIdx_++; }
2102 void LatexInfo::buildEntries(bool isPatternString)
2104 static regex const rmath("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|x?x?alignat)\\*?\\})(\\{[0-9]+\\})?)");
2105 static regex const rkeys("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?)))");
2106 static bool disableLanguageOverride = false;
2107 smatch sub, submath;
2108 bool evaluatingRegexp = false;
2110 bool evaluatingMath = false;
2111 bool evaluatingCode = false;
2113 bool evaluatingOptional = false;
2114 size_t optionalEnd = 0;
2117 bool math_end_waiting = false;
2118 size_t math_pos = 10000;
2119 size_t math_prefix_size = 1;
2121 static vector<string> usedText = vector<string>();
2122 static bool removeMathHull = false;
2124 interval_.removeAccents();
2126 for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) {
2128 if ((submath.position(2) - submath.position(0)) %2 == 1) {
2129 // prefixed by odd count of '\\'
2132 if (math_end_waiting) {
2133 size_t pos = submath.position(size_t(2));
2134 if ((math_end == "$") &&
2135 (submath.str(2) == "$")) {
2136 mi.insert("$", math_pos, 1, pos, 1);
2137 math_end_waiting = false;
2139 else if ((math_end == "\\]") &&
2140 (submath.str(2) == "\\]")) {
2141 mi.insert("\\]", math_pos, 2, pos, 2);
2142 math_end_waiting = false;
2144 else if ((submath.str(3).compare("end") == 0) &&
2145 (submath.str(5).compare(math_end) == 0)) {
2146 mi.insert(math_end, math_pos, math_prefix_size, pos, submath.str(2).length());
2147 math_end_waiting = false;
2153 if (submath.str(3).compare("begin") == 0) {
2154 math_end_waiting = true;
2155 math_end = submath.str(5);
2156 math_pos = submath.position(size_t(2));
2157 math_prefix_size = submath.str(2).length();
2159 else if (submath.str(2).compare("\\[") == 0) {
2160 math_end_waiting = true;
2162 math_pos = submath.position(size_t(2));
2164 else if (submath.str(2) == "$") {
2165 size_t pos = submath.position(size_t(2));
2166 math_end_waiting = true;
2172 // Ignore language if there is math somewhere in pattern-string
2173 if (isPatternString) {
2174 for (auto s: usedText) {
2175 // Remove entries created in previous search runs
2178 usedText = vector<string>();
2181 keys["foreignlanguage"].disabled = true;
2182 disableLanguageOverride = true;
2183 removeMathHull = false;
2186 removeMathHull = true; // used later if not isPatternString
2187 disableLanguageOverride = false;
2191 if (disableLanguageOverride) {
2192 keys["foreignlanguage"].disabled = true;
2195 math_pos = mi.getFirstPos();
2196 for (sregex_iterator it(interval_.par.begin(), interval_.par.end(), rkeys), end; it != end; ++it) {
2198 if ((sub.position(2) - sub.position(0)) %2 == 1) {
2199 // prefixed by odd count of '\\'
2202 string key = sub.str(5);
2204 if (sub.str(2)[0] == '\\')
2205 key = sub.str(2)[1];
2210 KeysIterator it_key = keys.find(key);
2211 if (it_key != keys.end()) {
2212 if (it_key->second.keytype == KeyInfo::headRemove) {
2213 KeyInfo found1 = it_key->second;
2214 found1.disabled = true;
2215 found1.head = "\\" + key + "{";
2216 found1._tokenstart = sub.position(size_t(2));
2217 found1._tokensize = found1.head.length();
2218 found1._dataStart = found1._tokenstart + found1.head.length();
2219 int endpos = interval_.findclosing(found1._dataStart, interval_.par.length(), '{', '}', 1);
2220 found1._dataEnd = endpos;
2225 if (evaluatingRegexp) {
2226 if (sub.str(3).compare("endregexp") == 0) {
2227 evaluatingRegexp = false;
2228 // found._tokenstart already set
2229 found._dataEnd = sub.position(size_t(2)) + 13;
2230 found._dataStart = found._dataEnd;
2231 found._tokensize = found._dataEnd - found._tokenstart;
2232 found.parenthesiscount = 0;
2233 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2240 if (evaluatingMath) {
2241 if (size_t(sub.position(size_t(2))) < mi.getEndPos())
2243 evaluatingMath = false;
2245 math_pos = mi.getStartPos();
2247 if (it_key == keys.end()) {
2248 found = KeyInfo(KeyInfo::isStandard, 0, true);
2249 LYXERR(Debug::INFO, "Undefined key " << key << " ==> will be used as text");
2250 found = KeyInfo(KeyInfo::isText, 0, false);
2251 if (isPatternString) {
2252 found.keytype = KeyInfo::isChar;
2253 found.disabled = false;
2257 usedText.push_back(key);
2261 if (key.compare("regexp") == 0) {
2262 evaluatingRegexp = true;
2263 found._tokenstart = sub.position(size_t(2));
2264 found._tokensize = 0;
2268 // Handle the other params of key
2269 if (found.keytype == KeyInfo::isIgnored)
2271 else if (found.keytype == KeyInfo::isMath) {
2272 if (size_t(sub.position(size_t(2))) == math_pos) {
2274 found._tokenstart = sub.position(size_t(2));
2275 found._tokensize = mi.getSize();
2276 found._dataEnd = found._tokenstart + found._tokensize;
2277 found._dataStart = found._dataEnd;
2278 found.parenthesiscount = 0;
2279 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2280 if (removeMathHull) {
2281 interval_.addIntervall(found._tokenstart, found._tokenstart + mi.getPrefixSize());
2282 interval_.addIntervall(found._dataEnd - mi.getPostfixSize(), found._dataEnd);
2285 // Treate all math constructs as simple math
2286 interval_.par[found._tokenstart] = '$';
2287 interval_.par[found._dataEnd - mi.getPostfixSize()] = '$';
2288 interval_.addIntervall(found._tokenstart + 1, found._tokenstart + mi.getPrefixSize());
2289 interval_.addIntervall(found._dataEnd - mi.getPostfixSize() + 1, found._dataEnd);
2291 evaluatingMath = true;
2294 // begin|end of unknown env, discard
2295 // First handle tables
2296 // longtable|tabular
2297 bool discardComment;
2299 found.keytype = KeyInfo::doRemove;
2300 if ((sub.str(7).compare("longtable") == 0) ||
2301 (sub.str(7).compare("tabular") == 0)) {
2302 discardComment = true; /* '%' */
2305 discardComment = false;
2306 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
2308 string token = sub.str(7);
2309 if (regex_match(token, sub2, removeArgs)) {
2310 found.keytype = KeyInfo::removeWithArg;
2313 // discard spaces before pos(2)
2314 int pos = sub.position(size_t(2));
2316 for (count = 0; pos - count > 0; count++) {
2317 char c = interval_.par[pos-count-1];
2318 if (discardComment) {
2319 if ((c != ' ') && (c != '%'))
2325 found._tokenstart = pos - count;
2326 if (sub.str(3).compare(0, 5, "begin") == 0) {
2327 size_t pos1 = pos + sub.str(2).length();
2328 if (sub.str(7).compare("cjk") == 0) {
2329 pos1 = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2330 if ((interval_.par[pos1] == '{') && (interval_.par[pos1+1] == '}'))
2332 found.keytype = KeyInfo::isMain;
2333 found._dataStart = pos1;
2334 found._dataEnd = interval_.par.length();
2335 found.disabled = keys["foreignlanguage"].disabled;
2336 found.used = keys["foreignlanguage"].used;
2337 found._tokensize = pos1 - found._tokenstart;
2338 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2341 // Swallow possible optional params
2342 while (interval_.par[pos1] == '[') {
2343 pos1 = interval_.findclosing(pos1+1, interval_.par.length(), '[', ']')+1;
2345 // Swallow also the eventual parameter
2346 if (interval_.par[pos1] == '{') {
2347 found._dataEnd = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2350 found._dataEnd = pos1;
2352 found._dataStart = found._dataEnd;
2353 found._tokensize = count + found._dataEnd - pos;
2354 found.parenthesiscount = 0;
2355 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2356 found.disabled = true;
2360 // Handle "\end{...}"
2361 found._dataStart = pos + sub.str(2).length();
2362 found._dataEnd = found._dataStart;
2363 found._tokensize = count + found._dataEnd - pos;
2364 found.parenthesiscount = 0;
2365 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2366 found.disabled = true;
2370 else if (found.keytype != KeyInfo::isRegex) {
2371 found._tokenstart = sub.position(size_t(2));
2372 if (found.parenthesiscount == 0) {
2373 // Probably to be discarded
2374 size_t following_pos = sub.position(size_t(2)) + sub.str(5).length() + 1;
2375 char following = interval_.par[following_pos];
2376 if (following == ' ')
2377 found.head = "\\" + sub.str(5) + " ";
2378 else if (following == '=') {
2379 // like \uldepth=1000pt
2380 found.head = sub.str(2);
2383 found.head = "\\" + key;
2384 found._tokensize = found.head.length();
2385 found._dataEnd = found._tokenstart + found._tokensize;
2386 found._dataStart = found._dataEnd;
2389 int params = found._tokenstart + key.length() + 1;
2390 if (evaluatingOptional) {
2391 if (size_t(found._tokenstart) > optionalEnd) {
2392 evaluatingOptional = false;
2395 found.disabled = true;
2398 int optend = params;
2399 while (interval_.par[optend] == '[') {
2400 // discard optional parameters
2401 optend = interval_.findclosing(optend+1, interval_.par.length(), '[', ']') + 1;
2403 if (optend > params) {
2404 key += interval_.par.substr(params, optend-params);
2405 evaluatingOptional = true;
2406 optionalEnd = optend;
2407 if (found.keytype == KeyInfo::isSectioning) {
2408 // Remove optional values (but still keep in header)
2409 interval_.addIntervall(params, optend);
2412 string token = sub.str(7);
2414 if (interval_.par[optend] != '{') {
2416 found.parenthesiscount = 0;
2417 found.head = "\\" + key;
2420 closings = found.parenthesiscount;
2421 if (found.parenthesiscount == 1) {
2422 found.head = "\\" + key + "{";
2424 else if (found.parenthesiscount > 1) {
2426 found.head = sub.str(2) + "{";
2427 closings = found.parenthesiscount - 1;
2430 found.head = "\\" + key + "{";
2433 found._tokensize = found.head.length();
2434 found._dataStart = found._tokenstart + found.head.length();
2435 if (found.keytype == KeyInfo::doRemove) {
2437 size_t endpar = 2 + interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2438 if (endpar >= interval_.par.length())
2439 found._dataStart = interval_.par.length();
2441 found._dataStart = endpar;
2442 found._tokensize = found._dataStart - found._tokenstart;
2445 found._dataStart = found._tokenstart + found._tokensize;
2449 if (interval_.par.substr(found._dataStart, 15).compare("\\endarguments{}") == 0) {
2450 found._dataStart += 15;
2454 endpos = found._dataStart - 1;
2456 endpos = interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2457 if (found.keytype == KeyInfo::isList) {
2458 // Check if it really is list env
2459 static regex const listre("^([a-z]+)$");
2461 if (!regex_match(token, sub2, listre)) {
2462 // Change the key of this entry. It is not in a list/item environment
2463 found.keytype = KeyInfo::endArguments;
2466 if (found.keytype == KeyInfo::noMain) {
2467 evaluatingCode = true;
2469 codeStart = found._dataStart;
2471 else if (evaluatingCode) {
2472 if (size_t(found._dataStart) > codeEnd)
2473 evaluatingCode = false;
2474 else if (found.keytype == KeyInfo::isMain) {
2475 // Disable this key, treate it as standard
2476 found.keytype = KeyInfo::isStandard;
2477 found.disabled = true;
2478 if ((codeEnd +1 >= interval_.par.length()) &&
2479 (found._tokenstart == codeStart)) {
2480 // trickery, because the code inset starts
2481 // with \selectlanguage ...
2483 if (entries_.size() > 1) {
2484 entries_[entries_.size()-1]._dataEnd = codeEnd;
2489 if ((endpos == interval_.par.length()) &&
2490 (found.keytype == KeyInfo::doRemove)) {
2491 // Missing closing => error in latex-input?
2492 // therefore do not delete remaining data
2493 found._dataStart -= 1;
2494 found._dataEnd = found._dataStart;
2497 found._dataEnd = endpos;
2499 if (isPatternString) {
2500 keys[key].used = true;
2503 entries_.push_back(found);
2507 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
2509 stringstream s(keysstring);
2511 const char delim = '|';
2512 while (getline(s, key, delim)) {
2513 KeyInfo keyII(keyI);
2514 if (isPatternString) {
2517 else if ( !keys[key].used)
2518 keyII.disabled = true;
2523 void LatexInfo::buildKeys(bool isPatternString)
2526 static bool keysBuilt = false;
2527 if (keysBuilt && !isPatternString) return;
2529 // Keys to ignore in any case
2530 makeKey("text|textcyrillic|lyxmathsym|ensuremath", KeyInfo(KeyInfo::headRemove, 1, true), true);
2531 // Known standard keys with 1 parameter.
2532 // Split is done, if not at start of region
2533 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
2534 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
2535 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
2536 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
2537 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
2538 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
2540 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
2541 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2542 makeKey("section*|subsection*|subsubsection*|paragraph*",
2543 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2544 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2545 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
2547 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
2549 // Split is done, if not at start of region
2550 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
2551 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
2553 // Split is done always.
2554 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
2558 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2559 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2560 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2561 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2563 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2564 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2565 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2566 makeKey("thickspace|medspace|thinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2568 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2569 // Custom space/skip, remove the content (== length value)
2570 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
2571 // Found in fr/UserGuide.lyx
2572 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2574 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2575 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2576 // Known macros to remove (including their parameter)
2578 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
2579 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
2580 // handle like standard keys with 1 parameter.
2581 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
2583 // Ignore deleted text
2584 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString);
2585 // but preserve added text
2586 makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2588 // Macros to remove, but let the parameter survive
2590 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2592 // Remove language spec from content of these insets
2593 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
2595 // Same effect as previous, parameter will survive (because there is no one anyway)
2597 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2598 // Remove table decorations
2599 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
2600 // Discard shape-header.
2601 // For footnote or shortcut too, because of lang settings
2602 // and wrong handling if used 'KeyInfo::noMain'
2603 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2604 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2605 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2606 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2607 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2608 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
2609 // like ('tiny{}' or '\tiny ' ... )
2610 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
2612 // Survives, like known character
2613 // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2614 makeKey("tableofcontents", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2615 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
2617 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2618 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2619 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2621 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2622 // Remove RTL/LTR marker
2623 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2624 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
2625 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
2626 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
2627 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
2628 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
2629 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2630 if (isPatternString) {
2631 // Allow the first searched string to rebuild the keys too
2635 // no need to rebuild again
2641 * Keep the list of actual opened parentheses actual
2642 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2644 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2647 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2651 if (c == '\\') skip = 1;
2652 else if (c == '{') {
2655 else if (c == '}') {
2656 handleCloseP(i, closingAllowed);
2662 string Intervall::show(int lastpos)
2664 int idx = 0; /* int intervalls */
2667 for (idx = 0; idx <= ignoreidx; idx++) {
2668 while (i < lastpos) {
2670 if (i <= borders[idx].low) {
2671 if (borders[idx].low > lastpos)
2672 printsize = lastpos - i;
2674 printsize = borders[idx].low - i;
2675 s += par.substr(i, printsize);
2677 if (i >= borders[idx].low)
2678 i = borders[idx].upper;
2681 i = borders[idx].upper;
2687 s += par.substr(i, lastpos-i);
2693 void Intervall::output(ostringstream &os, int lastpos)
2695 // get number of chars to output
2696 int idx = 0; /* int intervalls */
2699 string startTitle = titleValue;
2700 for (idx = 0; idx <= ignoreidx; idx++) {
2702 if (i <= borders[idx].low) {
2704 if (borders[idx].low > lastpos)
2705 printsize = lastpos - i;
2707 printsize = borders[idx].low - i;
2708 if (printsize > 0) {
2709 os << startTitle << par.substr(i, printsize);
2711 printed += printsize;
2714 handleParentheses(i, false);
2715 if (i >= borders[idx].low)
2716 i = borders[idx].upper;
2719 i = borders[idx].upper;
2726 os << startTitle << par.substr(i, lastpos-i);
2727 printed += lastpos-i;
2729 handleParentheses(lastpos, false);
2731 if (keys["foreignlanguage"].disabled)
2732 startindex = actualdeptindex-langcount;
2734 startindex = actualdeptindex;
2735 for (int i = startindex; i > 0; --i) {
2738 if (hasTitle && (printed > 0))
2740 if (! isPatternString_)
2742 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2745 void LatexInfo::processRegion(int start, int region_end)
2747 while (start < region_end) { /* Let {[} and {]} survive */
2748 int cnt = interval_.isOpeningPar(start);
2750 // Closing is allowed past the region
2751 int closing = interval_.findclosing(start+1, interval_.par.length());
2752 interval_.addIntervall(start, start+1);
2753 interval_.addIntervall(closing, closing+1);
2757 start = interval_.nextNotIgnored(start+1);
2761 void LatexInfo::removeHead(KeyInfo const & actual, int count)
2763 if (actual.parenthesiscount == 0) {
2764 // "{\tiny{} ...}" ==> "{{} ...}"
2765 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2768 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2769 interval_.addIntervall(actual._tokenstart - count, actual._dataStart);
2770 interval_.addIntervall(actual._dataEnd, actual._dataEnd+1);
2774 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2777 switch (actual.keytype)
2779 case KeyInfo::isTitle: {
2781 nextKeyIdx = getNextKey();
2784 case KeyInfo::cleanToStart: {
2785 actual._dataEnd = actual._dataStart;
2786 nextKeyIdx = getNextKey();
2787 // Search for end of arguments
2788 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2790 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2791 entries_[i].disabled = true;
2793 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2795 while (interval_.par[actual._dataEnd] == ' ')
2797 interval_.addIntervall(0, actual._dataEnd+1);
2798 interval_.actualdeptindex = 0;
2799 interval_.depts[0] = actual._dataEnd+1;
2800 interval_.closes[0] = -1;
2803 case KeyInfo::isText:
2804 interval_.par[actual._tokenstart] = '#';
2805 //interval_.addIntervall(actual._tokenstart, actual._tokenstart+1);
2806 nextKeyIdx = getNextKey();
2808 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2809 if (actual.disabled)
2810 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2812 interval_.addIntervall(actual._dataStart, actual._dataEnd);
2815 case KeyInfo::isChar: {
2816 nextKeyIdx = getNextKey();
2819 case KeyInfo::isSize: {
2820 if (actual.disabled || (interval_.par[actual._dataStart] != '{') || (interval_.par[actual._dataStart-1] == ' ')) {
2821 if (actual.parenthesiscount == 0)
2822 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2824 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2826 nextKeyIdx = getNextKey();
2828 // Here _dataStart points to '{', so correct it
2829 actual._dataStart += 1;
2830 actual._tokensize += 1;
2831 actual.parenthesiscount = 1;
2832 if (interval_.par[actual._dataStart] == '}') {
2833 // Determine the end if used like '{\tiny{}...}'
2834 actual._dataEnd = interval_.findclosing(actual._dataStart+1, interval_.par.length()) + 1;
2835 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2838 // Determine the end if used like '\tiny{...}'
2839 actual._dataEnd = interval_.findclosing(actual._dataStart, interval_.par.length()) + 1;
2841 // Split on this key if not at start
2842 int start = interval_.nextNotIgnored(previousStart);
2843 if (start < actual._tokenstart) {
2844 interval_.output(os, actual._tokenstart);
2845 interval_.addIntervall(start, actual._tokenstart);
2847 // discard entry if at end of actual
2848 nextKeyIdx = process(os, actual);
2852 case KeyInfo::endArguments: {
2853 // Remove trailing '{}' too
2854 actual._dataStart += 1;
2855 actual._dataEnd += 1;
2856 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2857 nextKeyIdx = getNextKey();
2860 case KeyInfo::noMain:
2862 case KeyInfo::isStandard: {
2863 if (actual.disabled) {
2865 processRegion(actual._dataStart, actual._dataStart+1);
2866 nextKeyIdx = getNextKey();
2868 // Split on this key if not at datastart of calling entry
2869 int start = interval_.nextNotIgnored(previousStart);
2870 if (start < actual._tokenstart) {
2871 interval_.output(os, actual._tokenstart);
2872 interval_.addIntervall(start, actual._tokenstart);
2874 // discard entry if at end of actual
2875 nextKeyIdx = process(os, actual);
2879 case KeyInfo::removeWithArg: {
2880 nextKeyIdx = getNextKey();
2881 // Search for end of arguments
2882 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2884 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2885 entries_[i].disabled = true;
2887 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2889 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2892 case KeyInfo::doRemove: {
2893 // Remove the key with all parameters and following spaces
2896 if (interval_.par[actual._dataEnd-1] == ' ')
2897 start = actual._dataEnd;
2899 start = actual._dataEnd+1;
2900 for (pos = start; pos < interval_.par.length(); pos++) {
2901 if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
2904 // Remove also enclosing parentheses [] and {}
2907 while (actual._tokenstart > numpars) {
2908 if (pos+numpars >= interval_.par.size())
2910 else if (interval_.par[pos+numpars] == ']' && interval_.par[actual._tokenstart-numpars-1] == '[')
2912 else if (interval_.par[pos+numpars] == '}' && interval_.par[actual._tokenstart-numpars-1] == '{')
2918 if (interval_.par[pos+numpars] == ' ')
2922 interval_.addIntervall(actual._tokenstart-numpars, pos+numpars+spaces);
2923 nextKeyIdx = getNextKey();
2926 case KeyInfo::isList: {
2927 // Discard space before _tokenstart
2929 for (count = 0; count < actual._tokenstart; count++) {
2930 if (interval_.par[actual._tokenstart-count-1] != ' ')
2933 nextKeyIdx = getNextKey();
2934 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2936 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2938 // How else can we catch this one?
2939 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2940 entries_[i].disabled = true;
2942 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2944 else if (nextKeyIdx > 0) {
2945 // Ignore any lang entries inside data region
2946 for (int i = nextKeyIdx; i < int(entries_.size()) && entries_[i]._tokenstart < actual._dataEnd; i++) {
2947 if (entries_[i].keytype == KeyInfo::isMain)
2948 entries_[i].disabled = true;
2951 if (actual.disabled) {
2952 interval_.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2955 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2957 if (interval_.par[actual._dataEnd+1] == '[') {
2958 int posdown = interval_.findclosing(actual._dataEnd+2, interval_.par.length(), '[', ']');
2959 if ((interval_.par[actual._dataEnd+2] == '{') &&
2960 (interval_.par[posdown-1] == '}')) {
2961 interval_.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
2962 interval_.addIntervall(posdown-1, posdown+1);
2965 interval_.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
2966 interval_.addIntervall(posdown, posdown+1);
2968 int blk = interval_.nextNotIgnored(actual._dataEnd+1);
2969 if (blk > posdown) {
2970 // Discard at most 1 space after empty item
2972 for (count = 0; count < 1; count++) {
2973 if (interval_.par[blk+count] != ' ')
2977 interval_.addIntervall(blk, blk+count);
2982 case KeyInfo::isSectioning: {
2983 // Discard spaces before _tokenstart
2985 int val = actual._tokenstart;
2986 for (count = 0; count < actual._tokenstart;) {
2987 val = interval_.previousNotIgnored(val-1);
2988 if (val < 0 || interval_.par[val] != ' ')
2991 count = actual._tokenstart - val;
2994 if (actual.disabled) {
2995 removeHead(actual, count);
2996 nextKeyIdx = getNextKey();
2998 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2999 nextKeyIdx = process(os, actual);
3003 case KeyInfo::isMath: {
3004 // Same as regex, use the content unchanged
3005 nextKeyIdx = getNextKey();
3008 case KeyInfo::isRegex: {
3009 // DO NOT SPLIT ON REGEX
3011 nextKeyIdx = getNextKey();
3014 case KeyInfo::isIgnored: {
3015 // Treat like a character for now
3016 nextKeyIdx = getNextKey();
3019 case KeyInfo::isMain: {
3020 if (interval_.par.substr(actual._dataStart, 2) == "% ")
3021 interval_.addIntervall(actual._dataStart, actual._dataStart+2);
3022 if (actual._tokenstart > 0) {
3023 int prev = interval_.previousNotIgnored(actual._tokenstart - 1);
3024 if ((prev >= 0) && interval_.par[prev] == '%')
3025 interval_.addIntervall(prev, prev+1);
3027 if (actual.disabled) {
3029 interval_.langcount++;
3030 if ((interval_.par.substr(actual._dataStart, 3) == " \\[") ||
3031 (interval_.par.substr(actual._dataStart, 8) == " \\begin{")) {
3032 // Discard also the space before math-equation
3033 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
3035 nextKeyIdx = getNextKey();
3036 // interval.resetOpenedP(actual._dataStart-1);
3039 if (actual._tokenstart < 26) {
3040 // for the first (and maybe dummy) language
3041 interval_.setForDefaultLang(actual);
3043 interval_.resetOpenedP(actual._dataStart-1);
3047 case KeyInfo::invalid:
3048 case KeyInfo::headRemove:
3049 // These two cases cannot happen, already handled
3052 // LYXERR(Debug::INFO, "Unhandled keytype");
3053 nextKeyIdx = getNextKey();
3060 int LatexInfo::process(ostringstream & os, KeyInfo const & actual )
3062 int end = interval_.nextNotIgnored(actual._dataEnd);
3063 int oldStart = actual._dataStart;
3064 int nextKeyIdx = getNextKey();
3066 if ((nextKeyIdx < 0) ||
3067 (entries_[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
3068 (entries_[nextKeyIdx].keytype == KeyInfo::invalid)) {
3069 if (oldStart <= end) {
3070 processRegion(oldStart, end);
3075 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
3077 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
3078 (void) dispatch(os, actual._dataStart, nextKey);
3079 end = nextKey._tokenstart;
3082 processRegion(oldStart, nextKey._tokenstart);
3083 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
3085 oldStart = nextKey._dataEnd+1;
3087 // now nextKey is either invalid or is outside of actual._dataEnd
3088 // output the remaining and discard myself
3089 if (oldStart <= end) {
3090 processRegion(oldStart, end);
3092 if (interval_.par.size() > (size_t) end && interval_.par[end] == '}') {
3094 // This is the normal case.
3095 // But if using the firstlanguage, the closing may be missing
3097 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
3099 if (actual._dataEnd < end)
3100 output_end = interval_.nextNotIgnored(actual._dataEnd);
3101 else if (interval_.par.size() > (size_t) end)
3102 output_end = interval_.nextNotIgnored(end);
3104 output_end = interval_.par.size();
3105 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
3106 interval_.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
3108 // Remove possible empty data
3109 int dstart = interval_.nextNotIgnored(actual._dataStart);
3110 while (interval_.isOpeningPar(dstart) == 1) {
3111 interval_.addIntervall(dstart, dstart+1);
3112 int dend = interval_.findclosing(dstart+1, output_end);
3113 interval_.addIntervall(dend, dend+1);
3114 dstart = interval_.nextNotIgnored(dstart+1);
3116 if (dstart < output_end)
3117 interval_.output(os, output_end);
3119 interval_.addIntervall(0, end);
3121 interval_.addIntervall(actual._tokenstart, end);
3125 string splitOnKnownMacros(string par, bool isPatternString)
3128 LatexInfo li(par, isPatternString);
3129 // LYXERR(Debug::INFO, "Berfore split: " << par);
3130 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
3132 DummyKey._tokensize = 0;
3133 DummyKey._dataStart = 0;
3134 DummyKey._dataEnd = par.length();
3135 DummyKey.disabled = true;
3136 int firstkeyIdx = li.getFirstKey();
3138 if (firstkeyIdx >= 0) {
3139 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
3140 DummyKey._tokenstart = firstKey._tokenstart;
3142 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
3143 // Use dummy firstKey
3144 firstKey = DummyKey;
3145 (void) li.setNextKey(firstkeyIdx);
3148 if (par.substr(firstKey._dataStart, 2) == "% ")
3149 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
3151 nextkeyIdx = li.process(os, firstKey);
3152 while (nextkeyIdx >= 0) {
3153 // Check for a possible gap between the last
3154 // entry and this one
3155 int datastart = li.nextNotIgnored(firstKey._dataStart);
3156 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
3157 if ((nextKey._tokenstart > datastart)) {
3159 firstKey._dataStart = datastart;
3160 firstKey._dataEnd = par.length();
3161 (void) li.setNextKey(nextkeyIdx);
3162 // Fake the last opened parenthesis
3163 li.setForDefaultLang(firstKey);
3164 nextkeyIdx = li.process(os, firstKey);
3167 if (nextKey.keytype != KeyInfo::isMain) {
3168 firstKey._dataStart = datastart;
3169 firstKey._dataEnd = nextKey._dataEnd+1;
3170 (void) li.setNextKey(nextkeyIdx);
3171 li.setForDefaultLang(firstKey);
3172 nextkeyIdx = li.process(os, firstKey);
3175 nextkeyIdx = li.process(os, nextKey);
3179 // Handle the remaining
3180 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
3181 firstKey._dataEnd = par.length();
3183 if ((firstKey._dataStart < firstKey._dataEnd) &&
3184 (par[firstKey._dataStart] != '}')) {
3185 li.setForDefaultLang(firstKey);
3186 (void) li.process(os, firstKey);
3190 // return string definitelly impossible to match
3191 s = "\\foreignlanguage{ignore}{ }";
3195 s = par; /* no known macros found */
3196 // LYXERR(Debug::INFO, "After split: " << s);
3201 * Try to unify the language specs in the latexified text.
3202 * Resulting modified string is set to "", if
3203 * the searched tex does not contain all the features in the search pattern
3205 static string correctlanguagesetting(string par, bool isPatternString, bool withformat, lyx::Buffer *pbuf = nullptr)
3207 static Features regex_f;
3208 static int missed = 0;
3209 static bool regex_with_format = false;
3211 int parlen = par.length();
3213 while ((parlen > 0) && (par[parlen-1] == '\n')) {
3216 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
3217 // Happens to be there in case of description or labeling environment
3222 // Split the latex input into pieces which
3223 // can be digested by our search engine
3224 LYXERR(Debug::FIND, "input: \"" << par << "\"");
3225 if (isPatternString && (pbuf != nullptr)) { // Check if we should disable/enable test for language
3226 // We check for polyglossia, because in runparams.flavor we use Flavor::XeTeX
3227 string doclang = pbuf->params().language->polyglossia();
3228 static regex langre("\\\\(foreignlanguage)\\{([^\\}]+)\\}");
3230 bool toIgnoreLang = true;
3231 for (sregex_iterator it(par.begin(), par.end(), langre), end; it != end; ++it) {
3233 if (sub.str(2) != doclang) {
3234 toIgnoreLang = false;
3238 setIgnoreFormat("language", toIgnoreLang, false);
3241 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
3242 LYXERR(Debug::FIND, "After splitOnKnownMacros:\n\"" << result << "\"");
3245 result = par.substr(0, parlen);
3246 if (isPatternString) {
3249 regex_f = identifyFeatures(result);
3250 string features = "";
3251 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3252 string a = it->first;
3253 regex_with_format = true;
3254 features += " " + a;
3255 // LYXERR(Debug::INFO, "Identified regex format:" << a);
3257 LYXERR(Debug::FIND, "Identified Features" << features);
3260 } else if (regex_with_format) {
3261 Features info = identifyFeatures(result);
3262 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3263 string a = it->first;
3264 bool b = it->second;
3265 if (b && ! info[a]) {
3267 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
3274 // LYXERR(Debug::INFO, "No regex formats");
3280 // Remove trailing closure of math, macros and environments, so to catch parts of them.
3281 static int identifyClosing(string & t)
3283 int open_braces = 0;
3285 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
3286 if (regex_replace(t, t, "(.*[^\\\\])\\$$", "$1"))
3288 if (regex_replace(t, t, "(.*[^\\\\])\\\\\\]$", "$1"))
3290 if (regex_replace(t, t, "(.*[^\\\\])\\\\end\\{[a-zA-Z_]*\\*?\\}$", "$1"))
3292 if (regex_replace(t, t, "(.*[^\\\\])\\}$", "$1")) {
3301 static int num_replaced = 0;
3302 static bool previous_single_replace = true;
3304 void MatchStringAdv::CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string)
3307 // Handle \w properly
3308 QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
3309 if (! opt.casesensitive) {
3310 popts |= QRegularExpression::CaseInsensitiveOption;
3312 regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
3313 regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
3315 if (regexp.isValid() && regexp2.isValid()) {
3316 regexIsValid = true;
3317 // Check '{', '}' pairs inside the regex
3320 for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
3321 char c = par_as_string[i];
3328 else if (c == '}') {
3335 if (balanced != 0) {
3336 regexIsValid = false;
3337 regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
3341 regexIsValid = false;
3342 if (!regexp.isValid())
3343 regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
3345 regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
3348 (void)par_as_string;
3349 if (opt.casesensitive) {
3350 regexp = regex(regexp_str);
3351 regexp2 = regex(regexp2_str);
3354 regexp = regex(regexp_str, std::regex_constants::icase);
3355 regexp2 = regex(regexp2_str, std::regex_constants::icase);
3360 static void modifyRegexForMatchWord(string &t)
3363 regex wordre("(\\\\)*((\\.|\\\\b))");
3366 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
3368 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
3371 else if (sub.str(2) == "\\\\b")
3373 if (lastpos < (size_t) sub.position(2))
3374 s += t.substr(lastpos, sub.position(2) - lastpos);
3376 lastpos = sub.position(2) + sub.length(2);
3379 s = "\\b" + t + "\\b";
3383 else if (lastpos < t.length())
3384 s += t.substr(lastpos, t.length() - lastpos);
3385 t = "\\b" + s + "\\b";
3388 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt)
3389 : p_buf(&buf), p_first_buf(&buf), opt(opt)
3391 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
3392 docstring const & ds = stringifySearchBuffer(find_buf, opt);
3393 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
3394 if (opt.replace_all && previous_single_replace) {
3395 previous_single_replace = false;
3398 else if (!opt.replace_all) {
3399 num_replaced = 0; // count number of replaced strings
3400 previous_single_replace = true;
3402 // When using regexp, braces are hacked already by escape_for_regex()
3403 par_as_string = normalize(ds);
3405 close_wildcards = 0;
3407 size_t lead_size = 0;
3408 // correct the language settings
3409 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat, &buf);
3410 opt.matchAtStart = false;
3412 identifyClosing(par_as_string); // Removes math closings ($, ], ...) at end of string
3413 if (opt.ignoreformat) {
3417 lead_size = identifyLeading(par_as_string);
3419 lead_as_string = par_as_string.substr(0, lead_size);
3420 string lead_as_regex_string = string2regex(lead_as_string);
3421 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3422 string par_as_regex_string_nolead = string2regex(par_as_string_nolead);
3423 /* Handle whole words too in this case
3425 if (opt.matchword) {
3426 par_as_regex_string_nolead = "\\b" + par_as_regex_string_nolead + "\\b";
3427 opt.matchword = false;
3429 string regexp_str = "(" + lead_as_regex_string + ")()" + par_as_regex_string_nolead;
3430 string regexp2_str = "(" + lead_as_regex_string + ")(.*?)" + par_as_regex_string_nolead;
3431 CreateRegexp(opt, regexp_str, regexp2_str);
3433 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3434 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3438 if (!opt.ignoreformat) {
3439 lead_size = identifyLeading(par_as_string);
3440 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
3441 lead_as_string = par_as_string.substr(0, lead_size);
3442 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3445 // Here we are using regexp
3446 LASSERT(use_regexp, /**/);
3448 string lead_as_regexp;
3449 if (lead_size > 0) {
3450 lead_as_regexp = string2regex(par_as_string.substr(0, lead_size));
3451 (void)regex_replace(par_as_string_nolead, par_as_string_nolead, "}$", "");
3452 par_as_string = par_as_string_nolead;
3453 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
3454 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3456 // LYXERR(Debug::FIND, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
3457 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
3458 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
3459 // LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3461 size_t lng = par_as_string.size();
3462 if (!opt.ignoreformat) {
3463 // Remove extra '\}' at end if not part of \{\.\}
3465 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
3467 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
3476 if (lng < par_as_string.size())
3477 par_as_string = par_as_string.substr(0,lng);
3479 LYXERR(Debug::FIND, "par_as_string after correctRegex is '" << par_as_string << "'");
3480 if ((lng > 0) && (par_as_string[0] == '^')) {
3481 par_as_string = par_as_string.substr(1);
3483 opt.matchAtStart = true;
3485 // LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3486 // LYXERR(Debug::FIND, "Open braces: " << open_braces);
3487 // LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
3489 // If entered regexp must match at begin of searched string buffer
3490 // Kornel: Added parentheses to use $1 for size of the leading string
3494 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
3495 // Unfortunately is '\1', '\2', etc not working for strings with extra format
3496 // so the convert has no effect in that case
3497 for (int i = 7; i > 0; --i) {
3498 string orig = "\\\\" + std::to_string(i);
3499 string dest = "\\" + std::to_string(i+2);
3500 while (regex_replace(par_as_string, par_as_string, orig, dest));
3502 if (opt.matchword) {
3503 modifyRegexForMatchWord(par_as_string);
3504 opt.matchword = false;
3506 regexp_str = "(" + lead_as_regexp + ")()" + par_as_string;
3507 regexp2_str = "(" + lead_as_regexp + ")(.*?)" + par_as_string;
3509 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3510 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3511 CreateRegexp(opt, regexp_str, regexp2_str, par_as_string);
3515 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
3519 mres.searched_size = len;
3521 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
3524 docstring docstr = stringifyFromForSearch(opt, cur, len);
3526 str = normalize(docstr);
3527 if (!opt.ignoreformat) {
3528 str = correctlanguagesetting(str, false, !opt.ignoreformat);
3529 // remove closing '}' and '\n' to allow for use of '$' in regex
3530 size_t lng = str.size();
3531 while ((lng > 1) && ((str[lng -1] == '}') || (str[lng -1] == '\n')))
3533 if (lng != str.size()) {
3534 str = str.substr(0, lng);
3538 mres.match_len = -1;
3541 LYXERR(Debug::FIND, "After normalization: Matching against:\n'" << str << "'");
3543 LASSERT(use_regexp, /**/);
3545 // use_regexp always true
3546 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
3548 QString qstr = QString::fromStdString(str);
3549 QRegularExpression const *p_regexp;
3550 QRegularExpression::MatchType flags = QRegularExpression::NormalMatch;
3554 p_regexp = ®exp2;
3556 QRegularExpressionMatch match = p_regexp->match(qstr, 0, flags);
3557 if (!match.hasMatch())
3560 regex const *p_regexp;
3561 regex_constants::match_flag_type flags;
3563 flags = regex_constants::match_continuous;
3566 flags = regex_constants::match_default;
3567 p_regexp = ®exp2;
3569 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
3570 if (re_it == sregex_iterator())
3572 match_results<string::const_iterator> const & m = *re_it;
3574 // Whole found string, including the leading
3575 // std: m[0].second - m[0].first
3576 // Qt: match.capturedEnd(0) - match.capturedStart(0)
3578 // Size of the leading string
3579 // std: m[1].second - m[1].first
3580 // Qt: match.capturedEnd(1) - match.capturedStart(1)
3581 int leadingsize = 0;
3583 if (match.lastCapturedIndex() > 0) {
3584 leadingsize = match.capturedEnd(1) - match.capturedStart(1);
3589 leadingsize = m[1].second - m[1].first;
3593 mres.match_prefix = match.capturedEnd(2) - match.capturedStart(2);
3594 mres.match_len = match.capturedEnd(0) - match.capturedEnd(2);
3595 // because of different number of closing at end of string
3596 // we have to 'unify' the length of the post-match.
3597 // Done by ignoring closing parenthesis and linefeeds at string end
3598 int matchend = match.capturedEnd(0);
3599 size_t strsize = qstr.size();
3600 if (!opt.ignoreformat) {
3601 while (mres.match_len > 0) {
3602 QChar c = qstr.at(matchend - 1);
3603 if ((c == '\n') || (c == '}') || (c == '{')) {
3610 while (strsize > (size_t) match.capturedEnd(0)) {
3611 QChar c = qstr.at(strsize-1);
3612 if ((c == '\n') || (c == '}')) {
3619 // LYXERR0(qstr.toStdString());
3620 mres.match2end = strsize - matchend;
3621 mres.pos = match.capturedStart(2);
3623 mres.match_prefix = m[2].second - m[2].first;
3624 mres.match_len = m[0].second - m[2].second;
3625 // ignore closing parenthesis and linefeeds at string end
3626 size_t strend = m[0].second - m[0].first;
3627 int matchend = strend;
3628 size_t strsize = str.size();
3629 if (!opt.ignoreformat) {
3630 while (mres.match_len > 0) {
3631 char c = str.at(matchend - 1);
3632 if ((c == '\n') || (c == '}') || (c == '{')) {
3639 while (strsize > strend) {
3640 if ((str.at(strsize-1) == '}') || (str.at(strsize-1) == '\n')) {
3648 mres.match2end = strsize - matchend;
3649 mres.pos = m[2].first - m[0].first;;
3651 if (mres.match2end < 0)
3653 mres.leadsize = leadingsize;
3655 if (mres.match_len > 0) {
3656 string a0 = match.captured(0).mid(mres.pos + mres.match_prefix, mres.match_len).toStdString();
3657 mres.result.push_back(a0);
3658 for (int i = 3; i <= match.lastCapturedIndex(); i++) {
3659 mres.result.push_back(match.captured(i).toStdString());
3663 if (mres.match_len > 0) {
3664 string a0 = m[0].str().substr(mres.pos + mres.match_prefix, mres.match_len);
3665 mres.result.push_back(a0);
3666 for (size_t i = 3; i < m.size(); i++) {
3667 mres.result.push_back(m[i]);
3676 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
3678 MatchResult mres = findAux(cur, len, at_begin);
3679 int res = mres.match_len;
3681 "res=" << res << ", at_begin=" << at_begin
3682 << ", matchAtStart=" << opt.matchAtStart
3683 << ", inTexted=" << cur.inTexted());
3684 if (opt.matchAtStart) {
3687 else if (mres.match_prefix > 0)
3696 static bool simple_replace(string &t, string from, string to)
3698 regex repl("(\\\\)*(" + from + ")");
3702 for (sregex_iterator it(t.begin(), t.end(), repl), end; it != end; ++it) {
3704 if ((sub.position(2) - sub.position(0)) % 2 == 1)
3706 if (lastpos < (size_t) sub.position(2))
3707 s += t.substr(lastpos, sub.position(2) - lastpos);
3709 lastpos = sub.position(2) + sub.length(2);
3713 else if (lastpos < t.length())
3714 s += t.substr(lastpos, t.length() - lastpos);
3720 string MatchStringAdv::normalize(docstring const & s) const
3723 t = lyx::to_utf8(s);
3724 // Remove \n at begin
3725 while (!t.empty() && t[0] == '\n')
3728 while (!t.empty() && t[t.size() - 1] == '\n')
3729 t = t.substr(0, t.size() - 1);
3731 // Handle all other '\n'
3732 while ((pos = t.find("\n")) != string::npos) {
3733 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
3735 if (isAlnumASCII(t[pos+1])) {
3736 t.replace(pos-2, 3, " ");
3739 t.replace(pos-2, 3, "");
3742 else if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
3743 // '\n' adjacent to non-alpha-numerics, discard
3744 t.replace(pos, 1, "");
3747 // Replace all other \n with spaces
3748 t.replace(pos, 1, " ");
3751 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
3752 // Kornel: Added textsl, textsf, textit, texttt and noun
3753 // + allow to seach for colored text too
3754 LYXERR(Debug::FIND, "Removing stale empty macros from: " << t);
3755 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
3756 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3757 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
3758 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3759 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
3765 docstring stringifyFromCursor(DocIterator const & cur, int len)
3767 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
3768 if (cur.inTexted()) {
3769 Paragraph const & par = cur.paragraph();
3770 // TODO what about searching beyond/across paragraph breaks ?
3771 // TODO Try adding a AS_STR_INSERTS as last arg
3772 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
3773 int(par.size()) : cur.pos() + len;
3774 // OutputParams runparams(&cur.buffer()->params().encoding());
3775 OutputParams runparams(encodings.fromLyXName("utf8"));
3776 runparams.nice = true;
3777 runparams.flavor = Flavor::XeTeX;
3778 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
3779 // No side effect of file copying and image conversion
3780 runparams.dryrun = true;
3781 int option = AS_STR_INSETS | AS_STR_PLAINTEXT;
3782 if (ignoreFormats.getDeleted()) {
3783 option |= AS_STR_SKIPDELETE;
3784 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
3787 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
3789 LYXERR(Debug::FIND, "Stringifying with cur: "
3790 << cur << ", from pos: " << cur.pos() << ", end: " << end);
3791 return par.asString(cur.pos(), end,
3794 } else if (cur.inMathed()) {
3795 CursorSlice cs = cur.top();
3796 MathData md = cs.cell();
3797 MathData::const_iterator it_end =
3798 (( len == -1 || cs.pos() + len > int(md.size()))
3800 : md.begin() + cs.pos() + len );
3802 for (MathData::const_iterator it = md.begin() + cs.pos();
3805 docstring s = asString(md2);
3806 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
3809 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3814 /** Computes the LaTeX export of buf starting from cur and ending len positions
3815 * after cur, if len is positive, or at the paragraph or innermost inset end
3818 docstring latexifyFromCursor(DocIterator const & cur, int len)
3821 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
3822 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
3823 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
3825 Buffer const & buf = *cur.buffer();
3827 odocstringstream ods;
3829 //OutputParams runparams(&buf.params().encoding());
3830 OutputParams runparams(encodings.fromLyXName("utf8"));
3831 runparams.nice = false;
3832 runparams.flavor = Flavor::XeTeX;
3833 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3834 // No side effect of file copying and image conversion
3835 runparams.dryrun = true;
3836 if (ignoreFormats.getDeleted()) {
3837 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
3840 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
3843 if (cur.inTexted()) {
3844 // @TODO what about searching beyond/across paragraph breaks ?
3845 pos_type endpos = cur.paragraph().size();
3846 if (len != -1 && endpos > cur.pos() + len)
3847 endpos = cur.pos() + len;
3848 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
3849 string(), cur.pos(), endpos);
3850 string s = lyx::to_utf8(ods.str());
3851 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
3852 return(lyx::from_utf8(s));
3853 } else if (cur.inMathed()) {
3854 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
3855 for (int s = cur.depth() - 1; s >= 0; --s) {
3856 CursorSlice const & cs = cur[s];
3857 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
3858 TeXMathStream ws(os);
3859 cs.asInsetMath()->asHullInset()->header_write(ws);
3864 CursorSlice const & cs = cur.top();
3865 MathData md = cs.cell();
3866 MathData::const_iterator it_end =
3867 ((len == -1 || cs.pos() + len > int(md.size()))
3869 : md.begin() + cs.pos() + len);
3871 for (MathData::const_iterator it = md.begin() + cs.pos();
3875 ods << asString(md2);
3876 // Retrieve the math environment type, and add '$' or '$]'
3877 // or others (\end{equation}) accordingly
3878 for (int s = cur.depth() - 1; s >= 0; --s) {
3879 CursorSlice const & cs2 = cur[s];
3880 InsetMath * inset = cs2.asInsetMath();
3881 if (inset && inset->asHullInset()) {
3882 TeXMathStream ws(os);
3883 inset->asHullInset()->footer_write(ws);
3887 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3889 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3894 #if defined(ResultsDebug)
3896 static void displayMResult(MatchResult &mres, string from, DocIterator & cur)
3898 LYXERR0( "from:\t\t\t" << from);
3900 if (mres.pos_len > 0) {
3902 status = "FINALSEARCH";
3905 if (mres.match_len > 0) {
3906 if ((mres.match_prefix == 0) && (mres.pos == mres.leadsize))
3907 status = "Good Match";
3909 status = "Matched in";
3912 status = "MissedSearch";
3915 LYXERR0( status << "(" << cur.pos() << " ... " << mres.searched_size + cur.pos() << ") cur.lastpos(" << cur.lastpos() << ")");
3916 if ((mres.leadsize > 0) || (mres.match_len > 0) || (mres.match2end > 0))
3917 LYXERR0( "leadsize(" << mres.leadsize << ") match_len(" << mres.match_len << ") match2end(" << mres.match2end << ")");
3918 if ((mres.pos > 0) || (mres.match_prefix > 0))
3919 LYXERR0( "pos(" << mres.pos << ") match_prefix(" << mres.match_prefix << ")");
3920 for (size_t i = 0; i < mres.result.size(); i++)
3921 LYXERR0( "Match " << i << " = \"" << mres.result[i] << "\"");
3923 #define displayMres(s, txt, cur) displayMResult(s, txt, cur);
3925 #define displayMres(s, txt, cur)
3928 /** Finalize an advanced find operation, advancing the cursor to the innermost
3929 ** position that matches, plus computing the length of the matching text to
3931 ** Return the cur.pos() difference between start and end of found match
3933 MatchResult findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, MatchResult const & expected = MatchResult(-1))
3935 // Search the foremost position that matches (avoids find of entire math
3936 // inset when match at start of it)
3937 DocIterator old_cur(cur.buffer());
3939 static MatchResult fail = MatchResult();
3940 MatchResult max_match;
3941 // If (prefix_len > 0) means that forwarding 1 position will remove the complete entry
3942 // Happens with e.g. hyperlinks
3943 // either one sees "http://www.bla.bla" or nothing
3944 // so the search for "www" gives prefix_len = 7 (== sizeof("http://")
3945 // and although we search for only 3 chars, we find the whole hyperlink inset
3946 bool at_begin = (expected.match_prefix == 0);
3947 if (!match.opt.forward && match.opt.ignoreformat) {
3948 if (expected.pos > 0)
3951 LASSERT(at_begin, /**/);
3952 if (expected.match_len > 0 && at_begin) {
3953 // Search for deepest match
3955 max_match = expected;
3957 size_t d = cur.depth();
3961 if (cur.depth() < d)
3963 if (cur.depth() == d)
3966 while (cur && cur.depth() > lastd) {
3967 lastd = cur.depth();
3968 mres = match(cur, -1, at_begin);
3969 displayMres(mres, "Checking innermost", cur);
3970 if (mres.match_len > 0)
3975 if (mres.match_len < expected.match_len)
3983 // (expected.match_len <= 0)
3984 mres = match(cur); /* match valid only if not searching whole words */
3985 displayMres(mres, "Start with negative match", cur);
3988 if (max_match.match_len <= 0) return fail;
3989 LYXERR(Debug::FIND, "Ok");
3991 // Compute the match length
3993 if (cur.pos() + len > cur.lastpos())
3996 LASSERT(match.use_regexp, /**/);
3999 int maxl = cur.lastpos() - cur.pos();
4000 // Greedy behaviour while matching regexps
4001 while (maxl > minl) {
4003 mres2 = match(cur, len, at_begin);
4004 displayMres(mres2, "Finalize loop", cur);
4005 int actual_match_len = mres2.match_len;
4006 if (actual_match_len >= max_match.match_len) {
4007 // actual_match_len > max_match _can_ happen,
4008 // if the search area splits
4009 // some following word so that the regex
4010 // (e.g. 'r.*r\b' matches 'r' from the middle of the
4012 // This means, the len value is too big
4013 actual_match_len = max_match.match_len;
4015 max_match.match_len = actual_match_len;
4017 if (maxl - minl < 4)
4018 len = (maxl + minl)/2;
4020 len = minl + (maxl - minl + 3)/4;
4023 // (actual_match_len < max_match.match_len)
4025 len = (maxl + minl)/2;
4030 // Search for real start of matched characters
4032 MatchResult actual_match;
4035 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
4036 if (cur.depth() < old_cur.depth()) {
4038 LYXERR(Debug::INFO, "cur.depth() < old_cur.depth(), this should never happen");
4041 if (cur.pos() != old_cur.pos()) {
4042 // OK, forwarded 1 pos in actual inset
4043 actual_match = match(cur, len-1, at_begin);
4044 if (actual_match.match_len == max_match.match_len) {
4045 // Ha, got it! The shorter selection has the same match length
4048 max_match = actual_match;
4051 // OK, the shorter selection matches less chars, revert to previous value
4057 LYXERR(Debug::INFO, "cur.pos() == old_cur.pos(), this should never happen");
4058 actual_match = match(cur, len, at_begin);
4059 if (actual_match.match_len == max_match.match_len) {
4061 max_match = actual_match;
4068 max_match.pos_len = len;
4069 displayMres(max_match, "SEARCH RESULT", cur)
4076 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
4080 bool repeat = false;
4081 DocIterator orig_cur; // to be used if repeat not successful
4082 MatchResult orig_mres;
4083 while (!theApp()->longOperationCancelled() && cur) {
4084 //(void) findAdvForwardInnermost(cur);
4085 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
4086 MatchResult mres = match(cur, -1, false);
4087 string msg = "Starting";
4090 displayMres(mres, msg + " findForwardAdv", cur)
4091 int match_len = mres.match_len;
4092 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
4093 LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
4096 if (match_len <= 0) {
4097 // This should exit nested insets, if any, or otherwise undefine the currsor.
4098 cur.pos() = cur.lastpos();
4099 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
4102 else { // match_len > 0
4103 // Try to find the begin of searched string
4105 int firstInvalid = cur.lastpos() - cur.pos();
4107 int incrmatch = (mres.match_prefix + mres.pos - mres.leadsize + 1)*3/4;
4108 int incrcur = (firstInvalid + 1 )*3/4;
4109 if (incrcur < incrmatch)
4110 increment = incrcur;
4112 increment = incrmatch;
4116 LYXERR(Debug::FIND, "Set increment to " << increment);
4117 while (increment > 0) {
4118 DocIterator old_cur = cur;
4119 if (cur.pos() + increment >= cur.lastpos()) {
4123 cur.pos() = cur.pos() + increment;
4124 MatchResult mres2 = match(cur, -1, false);
4125 displayMres(mres2, "findForwardAdv loop", cur)
4126 switch (interpretMatch(mres, mres2)) {
4127 case MatchResult::newIsTooFar:
4128 // behind the expected match
4129 firstInvalid = increment;
4133 case MatchResult::newIsBetter:
4134 // not reached yet, but cur.pos()+increment is bettert
4136 firstInvalid -= increment;
4137 if (increment > firstInvalid*3/4)
4138 increment = firstInvalid*3/4;
4139 if ((mres2.pos == mres2.leadsize) && (increment >= mres2.match_prefix)) {
4140 if (increment >= mres2.match_prefix)
4141 increment = (mres2.match_prefix+1)*3/4;
4146 // Handle not like MatchResult::newIsTooFar
4147 LYXERR0( "Probably too far: Increment = " << increment << " match_prefix = " << mres.match_prefix);
4149 increment = increment*3/4;
4154 if (mres.match_len > 0) {
4155 if (mres.match_prefix + mres.pos - mres.leadsize > 0) {
4156 // The match seems to indicate some deeper level
4165 // should never be reached.
4169 // LYXERR0("Leaving first loop");
4170 LYXERR(Debug::FIND, "Finalizing 1");
4171 MatchResult found_match = findAdvFinalize(cur, match, mres);
4172 if (found_match.match_len > 0) {
4173 LASSERT(found_match.pos_len > 0, /**/);
4174 match.FillResults(found_match);
4175 return found_match.pos_len;
4178 // try next possible match
4189 /// Find the most backward consecutive match within same paragraph while searching backwards.
4190 MatchResult findMostBackwards(DocIterator & cur, MatchStringAdv const & match, MatchResult &expected)
4192 DocIterator cur_begin = cur;
4193 cur_begin.pos() = 0;
4194 DocIterator tmp_cur = cur;
4195 MatchResult mr = findAdvFinalize(tmp_cur, match, expected);
4196 Inset & inset = cur.inset();
4197 for (; cur != cur_begin; cur.backwardPos()) {
4198 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
4199 DocIterator new_cur = cur;
4200 new_cur.backwardPos();
4201 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
4203 MatchResult new_mr = findAdvFinalize(new_cur, match, expected);
4204 if (new_mr.match_len == mr.match_len)
4208 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
4214 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
4218 // Backup of original position
4219 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
4220 if (cur == cur_begin)
4223 DocIterator cur_orig(cur);
4224 bool pit_changed = false;
4227 MatchResult found_match = match(cur, -1, false);
4229 if (found_match.match_len > 0) {
4231 cur.pos() = cur.lastpos();
4233 cur.pos() = cur_orig.pos();
4234 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
4235 DocIterator cur_prev_iter;
4237 found_match = match(cur);
4238 LYXERR(Debug::FIND, "findBackAdv3: found_match="
4239 << (found_match.match_len > 0) << ", cur: " << cur);
4240 if (found_match.match_len > 0) {
4241 MatchResult found_mr = findMostBackwards(cur, match, found_match);
4242 if (found_mr.pos_len > 0) {
4243 match.FillResults(found_mr);
4244 return found_mr.pos_len;
4248 // Stop if begin of document reached
4249 if (cur == cur_begin)
4251 cur_prev_iter = cur;
4255 if (cur == cur_begin)
4262 } while (!theApp()->longOperationCancelled());
4270 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
4271 DocIterator const & cur, int len)
4273 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
4275 if (!opt.ignoreformat)
4276 return latexifyFromCursor(cur, len);
4278 return stringifyFromCursor(cur, len);
4282 FindAndReplaceOptions::FindAndReplaceOptions(
4283 docstring const & _find_buf_name, bool _casesensitive,
4284 bool _matchword, bool _forward, bool _expandmacros, bool _ignoreformat,
4285 docstring const & _repl_buf_name, bool _keep_case,
4286 SearchScope _scope, SearchRestriction _restr, bool _replace_all)
4287 : find_buf_name(_find_buf_name), casesensitive(_casesensitive), matchword(_matchword),
4288 forward(_forward), expandmacros(_expandmacros), ignoreformat(_ignoreformat),
4289 repl_buf_name(_repl_buf_name), keep_case(_keep_case), scope(_scope), restr(_restr), replace_all(_replace_all)
4297 /** Check if 'len' letters following cursor are all non-lowercase */
4298 static bool allNonLowercase(Cursor const & cur, int len)
4300 pos_type beg_pos = cur.selectionBegin().pos();
4301 pos_type end_pos = cur.selectionBegin().pos() + len;
4302 if (len > cur.lastpos() + 1 - beg_pos) {
4303 LYXERR(Debug::FIND, "This should not happen, more debug needed");
4304 len = cur.lastpos() + 1 - beg_pos;
4305 end_pos = beg_pos + len;
4307 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
4308 if (isLowerCase(cur.paragraph().getChar(pos)))
4314 /** Check if first letter is upper case and second one is lower case */
4315 static bool firstUppercase(Cursor const & cur)
4318 pos_type pos = cur.selectionBegin().pos();
4319 if (pos >= cur.lastpos() - 1) {
4320 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
4323 ch1 = cur.paragraph().getChar(pos);
4324 ch2 = cur.paragraph().getChar(pos + 1);
4325 bool result = isUpperCase(ch1) && isLowerCase(ch2);
4326 LYXERR(Debug::FIND, "firstUppercase(): "
4327 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
4328 << ch2 << "(" << char(ch2) << ")"
4329 << ", result=" << result << ", cur=" << cur);
4334 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
4336 ** \fixme What to do with possible further paragraphs in replace buffer ?
4338 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
4340 ParagraphList::iterator pit = buffer.paragraphs().begin();
4341 LASSERT(!pit->empty(), /**/);
4342 pos_type right = pos_type(1);
4343 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
4344 right = pit->size();
4345 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
4349 static bool replaceMatches(string &t, int maxmatchnum, vector <string> const & replacements)
4351 // Should replace the string "$" + std::to_string(matchnum) with replacement
4352 // if the char '$' is not prefixed with odd number of char '\\'
4353 static regex const rematch("(\\\\)*(\\$\\$([0-9]))");
4357 for (sregex_iterator it(t.begin(), t.end(), rematch), end; it != end; ++it) {
4359 if ((sub.position(2) - sub.position(0)) % 2 == 1)
4361 int num = stoi(sub.str(3), nullptr, 10);
4362 if (num >= maxmatchnum)
4364 if (lastpos < (size_t) sub.position(2))
4365 s += t.substr(lastpos, sub.position(2) - lastpos);
4366 s += replacements[num];
4367 lastpos = sub.position(2) + sub.length(2);
4371 else if (lastpos < t.length())
4372 s += t.substr(lastpos, t.length() - lastpos);
4378 static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
4380 Cursor & cur = bv->cursor();
4381 if (opt.repl_buf_name.empty()
4382 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
4383 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4386 DocIterator sel_beg = cur.selectionBegin();
4387 DocIterator sel_end = cur.selectionEnd();
4388 if (&sel_beg.inset() != &sel_end.inset()
4389 || sel_beg.pit() != sel_end.pit()
4390 || sel_beg.idx() != sel_end.idx())
4392 int sel_len = sel_end.pos() - sel_beg.pos();
4393 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
4394 << ", sel_len: " << sel_len << endl);
4397 LASSERT(sel_len > 0, return 0);
4399 if (!matchAdv(sel_beg, sel_len).match_len)
4402 // Build a copy of the replace buffer, adapted to the KeepCase option
4403 Buffer const & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
4405 repl_buffer_orig.write(oss);
4406 string lyx = oss.str();
4407 if (matchAdv.valid_matches > 0)
4408 replaceMatches(lyx, matchAdv.valid_matches, matchAdv.matches);
4409 Buffer repl_buffer(string(), false);
4410 repl_buffer.setInternal(true);
4411 repl_buffer.setUnnamed(true);
4412 LASSERT(repl_buffer.readString(lyx), return 0);
4413 if (opt.keep_case && sel_len >= 2) {
4414 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
4415 if (cur.inTexted()) {
4416 if (firstUppercase(cur))
4417 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
4418 else if (allNonLowercase(cur, sel_len))
4419 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
4422 cap::cutSelection(cur, false);
4423 if (cur.inTexted()) {
4424 repl_buffer.changeLanguage(
4425 repl_buffer.language(),
4426 cur.getFont().language());
4427 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
4428 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
4429 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
4430 repl_buffer.params().documentClassPtr(),
4431 repl_buffer.params().authors(),
4432 bv->buffer().errorList("Paste"));
4433 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
4434 sel_len = repl_buffer.paragraphs().begin()->size();
4435 } else if (cur.inMathed()) {
4436 odocstringstream ods;
4438 // OutputParams runparams(&repl_buffer.params().encoding());
4439 OutputParams runparams(encodings.fromLyXName("utf8"));
4440 runparams.nice = false;
4441 runparams.flavor = Flavor::XeTeX;
4442 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
4443 runparams.dryrun = true;
4444 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
4445 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
4446 docstring repl_latex = ods.str();
4447 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
4449 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
4450 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
4451 repl_latex = from_utf8(s);
4452 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
4453 MathData ar(cur.buffer());
4454 asArray(repl_latex, ar, Parse::NORMAL);
4456 sel_len = ar.size();
4457 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4459 if (cur.pos() >= sel_len)
4460 cur.pos() -= sel_len;
4463 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4464 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
4465 bv->processUpdateFlags(Update::Force);
4470 /// Perform a FindAdv operation.
4471 bool findAdv(BufferView * bv, FindAndReplaceOptions & opt)
4476 // e.g., when invoking word-findadv from mini-buffer wither with
4477 // wrong options syntax or before ever opening advanced F&R pane
4478 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4482 MatchStringAdv matchAdv(bv->buffer(), opt);
4484 if (!matchAdv.regexIsValid) {
4485 bv->message(lyx::from_utf8(matchAdv.regexError));
4489 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
4491 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
4492 num_replaced += findAdvReplace(bv, opt, matchAdv);
4495 pos_len = findForwardAdv(cur, matchAdv);
4497 pos_len = findBackwardsAdv(cur, matchAdv);
4498 } catch (exception & ex) {
4499 bv->message(from_utf8(ex.what()));
4504 if (num_replaced > 0) {
4505 switch (num_replaced)
4508 bv->message(_("One match has been replaced."));
4511 bv->message(_("Two matches have been replaced."));
4514 bv->message(bformat(_("%1$d matches have been replaced."), num_replaced));
4520 bv->message(_("Match not found."));
4525 if (num_replaced > 0)
4526 bv->message(_("Match has been replaced."));
4528 bv->message(_("Match found."));
4530 if (cur.pos() + pos_len > cur.lastpos()) {
4531 // Prevent crash in bv->putSelectionAt()
4532 // Should never happen, maybe LASSERT() here?
4533 pos_len = cur.lastpos() - cur.pos();
4535 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << pos_len);
4536 bv->putSelectionAt(cur, pos_len, !opt.forward);
4542 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
4544 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
4545 << opt.casesensitive << ' '
4546 << opt.matchword << ' '
4547 << opt.forward << ' '
4548 << opt.expandmacros << ' '
4549 << opt.ignoreformat << ' '
4550 << opt.replace_all << ' '
4551 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
4552 << opt.keep_case << ' '
4553 << int(opt.scope) << ' '
4556 LYXERR(Debug::FIND, "built: " << os.str());
4562 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
4564 // LYXERR(Debug::FIND, "parsing");
4568 while (line != "EOSS") {
4572 if (is.eof()) // Tolerate malformed request
4576 // LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
4577 opt.find_buf_name = from_utf8(s);
4578 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.replace_all;
4579 is.get(); // Waste space before replace string
4582 while (line != "EOSS") {
4586 if (is.eof()) // Tolerate malformed request
4590 // LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
4591 opt.repl_buf_name = from_utf8(s);
4592 is >> opt.keep_case;
4595 opt.scope = FindAndReplaceOptions::SearchScope(i);
4597 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
4600 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
4601 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
4602 << opt.scope << ' ' << opt.restr);