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("\\pounds", getutf8(0x00a3));
1511 addAccents("\\textsterling", getutf8(0x00a3));
1512 addAccents("\\textyen", getutf8(0x00a5));
1513 addAccents("\\yen", getutf8(0x00a5));
1514 addAccents("\\textsection", getutf8(0x00a7));
1515 addAccents("\\mathsection", getutf8(0x00a7));
1516 addAccents("\\textcopyright", getutf8(0x00a9));
1517 addAccents("\\copyright", getutf8(0x00a9));
1518 addAccents("\\textlnot", getutf8(0x00ac));
1519 addAccents("\\neg", getutf8(0x00ac));
1520 addAccents("\\textregistered", getutf8(0x00ae));
1521 addAccents("\\circledR", getutf8(0x00ae));
1522 addAccents("\\textpm", getutf8(0x00b1));
1523 addAccents("\\pm", getutf8(0x00b1));
1524 addAccents("\\textparagraph", getutf8(0x00b6));
1525 addAccents("\\mathparagraph", getutf8(0x00b6));
1526 addAccents("\\textperiodcentered", getutf8(0x00b7));
1527 addAccents("\\texttimes", getutf8(0x00d7));
1528 addAccents("\\times", getutf8(0x00d7));
1529 addAccents("\\O", getutf8(0x00d8));
1530 addAccents("\\dh", getutf8(0x00f0));
1531 addAccents("\\eth", getutf8(0x00f0));
1532 addAccents("\\textdiv", getutf8(0x00f7));
1533 addAccents("\\div", getutf8(0x00f7));
1534 addAccents("\\o", getutf8(0x00f8));
1535 addAccents("\\textcrlambda", getutf8(0x019b));
1536 addAccents("\\j", getutf8(0x0237));
1537 addAccents("\\textGamma", getutf8(0x0393));
1538 addAccents("\\Gamma", getutf8(0x0393));
1539 addAccents("\\textDelta", getutf8(0x0394));
1540 addAccents("\\Delta", getutf8(0x0394));
1541 addAccents("\\textTheta", getutf8(0x0398));
1542 addAccents("\\Theta", getutf8(0x0398));
1543 addAccents("\\textLambda", getutf8(0x039b));
1544 addAccents("\\Lambda", getutf8(0x039b));
1545 addAccents("\\textXi", getutf8(0x039e));
1546 addAccents("\\Xi", getutf8(0x039e));
1547 addAccents("\\textPi", getutf8(0x03a0));
1548 addAccents("\\Pi", getutf8(0x03a0));
1549 addAccents("\\textSigma", getutf8(0x03a3));
1550 addAccents("\\Sigma", getutf8(0x03a3));
1551 addAccents("\\textUpsilon", getutf8(0x03a5));
1552 addAccents("\\Upsilon", getutf8(0x03a5));
1553 addAccents("\\textPhi", getutf8(0x03a6));
1554 addAccents("\\Phi", getutf8(0x03a6));
1555 addAccents("\\textPsi", getutf8(0x03a8));
1556 addAccents("\\Psi", getutf8(0x03a8));
1557 addAccents("\\textOmega", getutf8(0x03a9));
1558 addAccents("\\Omega", getutf8(0x03a9));
1559 addAccents("\\textalpha", getutf8(0x03b1));
1560 addAccents("\\alpha", getutf8(0x03b1));
1561 addAccents("\\textbeta", getutf8(0x03b2));
1562 addAccents("\\beta", getutf8(0x03b2));
1563 addAccents("\\textgamma", getutf8(0x03b3));
1564 addAccents("\\gamma", getutf8(0x03b3));
1565 addAccents("\\textdelta", getutf8(0x03b4));
1566 addAccents("\\delta", getutf8(0x03b4));
1567 addAccents("\\textepsilon", getutf8(0x03b5));
1568 addAccents("\\varepsilon", getutf8(0x03b5));
1569 addAccents("\\textzeta", getutf8(0x03b6));
1570 addAccents("\\zeta", getutf8(0x03b6));
1571 addAccents("\\texteta", getutf8(0x03b7));
1572 addAccents("\\eta", getutf8(0x03b7));
1573 addAccents("\\texttheta", getutf8(0x03b8));
1574 addAccents("\\theta", getutf8(0x03b8));
1575 addAccents("\\textiota", getutf8(0x03b9));
1576 addAccents("\\iota", getutf8(0x03b9));
1577 addAccents("\\textkappa", getutf8(0x03ba));
1578 addAccents("\\kappa", getutf8(0x03ba));
1579 addAccents("\\textlambda", getutf8(0x03bb));
1580 addAccents("\\lambda", getutf8(0x03bb));
1581 addAccents("\\textmu", getutf8(0x03bc));
1582 addAccents("\\mu", getutf8(0x03bc));
1583 addAccents("\\textnu", getutf8(0x03bd));
1584 addAccents("\\nu", getutf8(0x03bd));
1585 addAccents("\\textxi", getutf8(0x03be));
1586 addAccents("\\xi", getutf8(0x03be));
1587 addAccents("\\textpi", getutf8(0x03c0));
1588 addAccents("\\pi", getutf8(0x03c0));
1589 addAccents("\\textrho", getutf8(0x03c1));
1590 addAccents("\\rho", getutf8(0x03c1));
1591 addAccents("\\textfinalsigma", getutf8(0x03c2));
1592 addAccents("\\varsigma", getutf8(0x03c2));
1593 addAccents("\\textsigma", getutf8(0x03c3));
1594 addAccents("\\sigma", getutf8(0x03c3));
1595 addAccents("\\texttau", getutf8(0x03c4));
1596 addAccents("\\tau", getutf8(0x03c4));
1597 addAccents("\\textupsilon", getutf8(0x03c5));
1598 addAccents("\\upsilon", getutf8(0x03c5));
1599 addAccents("\\textphi", getutf8(0x03c6));
1600 addAccents("\\varphi", getutf8(0x03c6));
1601 addAccents("\\textchi", getutf8(0x03c7));
1602 addAccents("\\chi", getutf8(0x03c7));
1603 addAccents("\\textpsi", getutf8(0x03c8));
1604 addAccents("\\psi", getutf8(0x03c8));
1605 addAccents("\\textomega", getutf8(0x03c9));
1606 addAccents("\\omega", getutf8(0x03c9));
1607 addAccents("\\textdigamma", getutf8(0x03dd));
1608 addAccents("\\digamma", getutf8(0x03dd));
1609 addAccents("\\hebalef", getutf8(0x05d0));
1610 addAccents("\\aleph", getutf8(0x05d0));
1611 addAccents("\\hebbet", getutf8(0x05d1));
1612 addAccents("\\beth", getutf8(0x05d1));
1613 addAccents("\\hebgimel", getutf8(0x05d2));
1614 addAccents("\\gimel", getutf8(0x05d2));
1615 addAccents("\\hebdalet", getutf8(0x05d3));
1616 addAccents("\\daleth", getutf8(0x05d3));
1617 addAccents("\\hebhe", getutf8(0x05d4));
1618 addAccents("\\hebvav", getutf8(0x05d5));
1619 addAccents("\\hebzayin", getutf8(0x05d6));
1620 addAccents("\\hebhet", getutf8(0x05d7));
1621 addAccents("\\hebtet", getutf8(0x05d8));
1622 addAccents("\\hebyod", getutf8(0x05d9));
1623 addAccents("\\hebfinalkaf", getutf8(0x05da));
1624 addAccents("\\hebkaf", getutf8(0x05db));
1625 addAccents("\\heblamed", getutf8(0x05dc));
1626 addAccents("\\hebfinalmem", getutf8(0x05dd));
1627 addAccents("\\hebmem", getutf8(0x05de));
1628 addAccents("\\hebfinalnun", getutf8(0x05df));
1629 addAccents("\\hebnun", getutf8(0x05e0));
1630 addAccents("\\hebsamekh", getutf8(0x05e1));
1631 addAccents("\\hebayin", getutf8(0x05e2));
1632 addAccents("\\hebfinalpe", getutf8(0x05e3));
1633 addAccents("\\hebpe", getutf8(0x05e4));
1634 addAccents("\\hebfinaltsadi", getutf8(0x05e5));
1635 addAccents("\\hebtsadi", getutf8(0x05e6));
1636 addAccents("\\hebqof", getutf8(0x05e7));
1637 addAccents("\\hebresh", getutf8(0x05e8));
1638 addAccents("\\hebshin", getutf8(0x05e9));
1639 addAccents("\\hebtav", getutf8(0x05ea));
1642 addAccents("\\thaiKoKai", getutf8(0x0e01));
1643 addAccents("\\thaiKhoKhai", getutf8(0x0e02));
1644 addAccents("\\thaiKhoKhuat", getutf8(0x0e03));
1645 addAccents("\\thaiKhoKhwai", getutf8(0x0e04));
1646 addAccents("\\thaiKhoKhon", getutf8(0x0e05));
1647 addAccents("\\thaiKhoRakhang", getutf8(0x0e06));
1648 addAccents("\\thaiNgoNgu", getutf8(0x0e07));
1649 addAccents("\\thaiChoChan", getutf8(0x0e08));
1650 addAccents("\\thaiChoChing", getutf8(0x0e09));
1651 addAccents("\\thaiChoChang", getutf8(0x0e0a));
1652 addAccents("\\thaiSoSo", getutf8(0x0e0b));
1653 addAccents("\\thaiChoChoe", getutf8(0x0e0c));
1654 addAccents("\\thaiYoYing", getutf8(0x0e0d));
1655 addAccents("\\thaiDoChada", getutf8(0x0e0e));
1656 addAccents("\\thaiToPatak", getutf8(0x0e0f));
1657 addAccents("\\thaiThoThan", getutf8(0x0e10));
1658 addAccents("\\thaiThoNangmontho", getutf8(0x0e11));
1659 addAccents("\\thaiThoPhuthao", getutf8(0x0e12));
1660 addAccents("\\thaiNoNen", getutf8(0x0e13));
1661 addAccents("\\thaiDoDek", getutf8(0x0e14));
1662 addAccents("\\thaiToTao", getutf8(0x0e15));
1663 addAccents("\\thaiThoThung", getutf8(0x0e16));
1664 addAccents("\\thaiThoThahan", getutf8(0x0e17));
1665 addAccents("\\thaiThoThong", getutf8(0x0e18));
1666 addAccents("\\thaiNoNu", getutf8(0x0e19));
1667 addAccents("\\thaiBoBaimai", getutf8(0x0e1a));
1668 addAccents("\\thaiPoPla", getutf8(0x0e1b));
1669 addAccents("\\thaiPhoPhung", getutf8(0x0e1c));
1670 addAccents("\\thaiFoFa", getutf8(0x0e1d));
1671 addAccents("\\thaiPhoPhan", getutf8(0x0e1e));
1672 addAccents("\\thaiFoFan", getutf8(0x0e1f));
1673 addAccents("\\thaiPhoSamphao", getutf8(0x0e20));
1674 addAccents("\\thaiMoMa", getutf8(0x0e21));
1675 addAccents("\\thaiYoYak", getutf8(0x0e22));
1676 addAccents("\\thaiRoRua", getutf8(0x0e23));
1677 addAccents("\\thaiRu", getutf8(0x0e24));
1678 addAccents("\\thaiLoLing", getutf8(0x0e25));
1679 addAccents("\\thaiLu", getutf8(0x0e26));
1680 addAccents("\\thaiWoWaen", getutf8(0x0e27));
1681 addAccents("\\thaiSoSala", getutf8(0x0e28));
1682 addAccents("\\thaiSoRusi", getutf8(0x0e29));
1683 addAccents("\\thaiSoSua", getutf8(0x0e2a));
1684 addAccents("\\thaiHoHip", getutf8(0x0e2b));
1685 addAccents("\\thaiLoChula", getutf8(0x0e2c));
1686 addAccents("\\thaiOAng", getutf8(0x0e2d));
1687 addAccents("\\thaiHoNokhuk", getutf8(0x0e2e));
1688 addAccents("\\thaiPaiyannoi", getutf8(0x0e2f));
1689 addAccents("\\thaiSaraA", getutf8(0x0e30));
1690 addAccents("\\thaiMaiHanakat", getutf8(0x0e31));
1691 addAccents("\\thaiSaraAa", getutf8(0x0e32));
1692 addAccents("\\thaiSaraAm", getutf8(0x0e33));
1693 addAccents("\\thaiSaraI", getutf8(0x0e34));
1694 addAccents("\\thaiSaraIi", getutf8(0x0e35));
1695 addAccents("\\thaiSaraUe", getutf8(0x0e36));
1696 addAccents("\\thaiSaraUee", getutf8(0x0e37));
1697 addAccents("\\thaiSaraU", getutf8(0x0e38));
1698 addAccents("\\thaiSaraUu", getutf8(0x0e39));
1699 addAccents("\\thaiPhinthu", getutf8(0x0e3a));
1700 addAccents("\\thaiSaraE", getutf8(0x0e40));
1701 addAccents("\\thaiSaraAe", getutf8(0x0e41));
1702 addAccents("\\thaiSaraO", getutf8(0x0e42));
1703 addAccents("\\thaiSaraAiMaimuan", getutf8(0x0e43));
1704 addAccents("\\thaiSaraAiMaimalai", getutf8(0x0e44));
1705 addAccents("\\thaiLakkhangyao", getutf8(0x0e45));
1706 addAccents("\\thaiMaiyamok", getutf8(0x0e46));
1707 addAccents("\\thaiMaitaikhu", getutf8(0x0e47));
1708 addAccents("\\thaiMaiEk", getutf8(0x0e48));
1709 addAccents("\\thaiMaiTho", getutf8(0x0e49));
1710 addAccents("\\thaiMaiTri", getutf8(0x0e4a));
1711 addAccents("\\thaiMaiChattawa", getutf8(0x0e4b));
1712 addAccents("\\thaiThanthakhat", getutf8(0x0e4c));
1713 addAccents("\\thaiNikhahit", getutf8(0x0e4d));
1714 addAccents("\\thaiYamakkan", getutf8(0x0e4e));
1715 addAccents("\\thaiFongman", getutf8(0x0e4f));
1716 addAccents("\\thaizero", getutf8(0x0e50));
1717 addAccents("\\thaione", getutf8(0x0e51));
1718 addAccents("\\thaitwo", getutf8(0x0e52));
1719 addAccents("\\thaithree", getutf8(0x0e53));
1720 addAccents("\\thaifour", getutf8(0x0e54));
1721 addAccents("\\thaifive", getutf8(0x0e55));
1722 addAccents("\\thaisix", getutf8(0x0e56));
1723 addAccents("\\thaiseven", getutf8(0x0e57));
1724 addAccents("\\thaieight", getutf8(0x0e58));
1725 addAccents("\\thainine", getutf8(0x0e59));
1726 addAccents("\\thaiAngkhankhu", getutf8(0x0e5a));
1727 addAccents("\\thaiKhomut", getutf8(0x0e5b));
1728 addAccents("\\dag", getutf8(0x2020));
1729 addAccents("\\dagger", getutf8(0x2020));
1730 addAccents("\\textdagger", getutf8(0x2020));
1731 addAccents("\\ddag", getutf8(0x2021));
1732 addAccents("\\ddagger", getutf8(0x2021));
1733 addAccents("\\textbullet", getutf8(0x2022));
1734 addAccents("\\bullet", getutf8(0x2022));
1735 addAccents("\\dots", getutf8(0x2026));
1736 addAccents("\\ldots", getutf8(0x2026));
1737 addAccents("\\textellipsis", getutf8(0x2026));
1738 addAccents("\\textasciiacute", getutf8(0x2032));
1739 addAccents("\\prime", getutf8(0x2032));
1740 addAccents("\\textacutedbl", getutf8(0x2033));
1741 addAccents("\\dprime", getutf8(0x2033));
1742 addAccents("\\textasciigrave", getutf8(0x2035));
1743 addAccents("\\backprime", getutf8(0x2035));
1744 addAccents("\\textsubcircum{ }", getutf8(0x2038));
1745 addAccents("\\caretinsert", getutf8(0x2038));
1746 addAccents("\\textasteriskcentered", getutf8(0x204e));
1747 addAccents("\\ast", getutf8(0x204e));
1748 addAccents("\\textmho", getutf8(0x2127));
1749 addAccents("\\mho", getutf8(0x2127));
1750 addAccents("\\textleftarrow", getutf8(0x2190));
1751 addAccents("\\leftarrow", getutf8(0x2190));
1752 addAccents("\\textuparrow", getutf8(0x2191));
1753 addAccents("\\uparrow", getutf8(0x2191));
1754 addAccents("\\textrightarrow", getutf8(0x2192));
1755 addAccents("\\rightarrow", getutf8(0x2192));
1756 addAccents("\\textdownarrow", getutf8(0x2193));
1757 addAccents("\\downarrow", getutf8(0x2193));
1758 addAccents("\\textglobrise", getutf8(0x2197));
1759 addAccents("\\nearrow", getutf8(0x2197));
1760 addAccents("\\textglobfall", getutf8(0x2198));
1761 addAccents("\\searrow", getutf8(0x2198));
1762 addAccents("\\textsurd", getutf8(0x221a));
1763 addAccents("\\surd", getutf8(0x221a));
1764 addAccents("\\textbigcircle", getutf8(0x25ef));
1765 addAccents("\\bigcirc", getutf8(0x25ef));
1766 addAccents("\\FiveStar", getutf8(0x2605));
1767 addAccents("\\bigstar", getutf8(0x2605));
1768 addAccents("\\FiveStarOpen", getutf8(0x2606));
1769 addAccents("\\bigwhitestar", getutf8(0x2606));
1770 addAccents("\\Checkmark", getutf8(0x2713));
1771 addAccents("\\checkmark", getutf8(0x2713));
1772 addAccents("\\CrossMaltese", getutf8(0x2720));
1773 addAccents("\\maltese", getutf8(0x2720));
1774 addAccents("\\textlangle", getutf8(0x27e8));
1775 addAccents("\\langle", getutf8(0x27e8));
1776 addAccents("\\textrangle", getutf8(0x27e9));
1777 addAccents("\\rangle", getutf8(0x27e9));
1780 static void buildAccentsMap()
1782 accents["imath"] = "ı";
1784 accents["jmath"] = "ȷ";
1785 accents["cdot"] = "·";
1786 accents["textasciicircum"] = "^";
1787 accents["mathcircumflex"] = "^";
1788 accents["sim"] = "~";
1789 accents["guillemotright"] = "»";
1790 accents["guillemotleft"] = "«";
1791 accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
1792 accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
1793 accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
1794 accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
1795 accents["negmedspace"] = getutf8(0xf0005);
1796 accents["thickspace"] = getutf8(0xf0006);
1797 accents["negthickspace"] = getutf8(0xf0007);
1798 accents["lyx"] = getutf8(0xf0010); // Used logos
1799 accents["LyX"] = getutf8(0xf0010);
1800 accents["tex"] = getutf8(0xf0011);
1801 accents["TeX"] = getutf8(0xf0011);
1802 accents["latex"] = getutf8(0xf0012);
1803 accents["LaTeX"] = getutf8(0xf0012);
1804 accents["latexe"] = getutf8(0xf0013);
1805 accents["LaTeXe"] = getutf8(0xf0013);
1806 accents["lyxarrow"] = getutf8(0xf0020);
1807 accents["braceleft"] = getutf8(0xf0030);
1808 accents["braceright"] = getutf8(0xf0031);
1809 accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash
1810 accents["backslash LyX"] = getutf8(0xf0010);
1811 accents["backslash tex"] = getutf8(0xf0011);
1812 accents["backslash TeX"] = getutf8(0xf0011);
1813 accents["backslash latex"] = getutf8(0xf0012);
1814 accents["backslash LaTeX"] = getutf8(0xf0012);
1815 accents["backslash latexe"] = getutf8(0xf0013);
1816 accents["backslash LaTeXe"] = getutf8(0xf0013);
1817 accents["backslash lyxarrow"] = getutf8(0xf0020);
1818 accents["ddot{\\imath}"] = "ï";
1819 buildaccent("ddot", "aAeEhHiIoOtuUwWxXyY",
1820 "äÄëËḧḦïÏöÖẗüÜẅẄẍẌÿŸ"); // umlaut
1821 buildaccent("dot|.", "aAbBcCdDeEfFGghHIimMnNoOpPrRsStTwWxXyYzZ",
1822 "ȧȦḃḂċĊḋḊėĖḟḞĠġḣḢİİṁṀṅṄȯȮṗṖṙṘṡṠṫṪẇẆẋẊẏẎżŻ"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'İ'
1823 accents["acute{\\imath}"] = "í";
1824 buildaccent("acute", "aAcCeEgGkKlLmMoOnNpPrRsSuUwWyYzZiI",
1825 "áÁćĆéÉǵǴḱḰĺĹḿḾóÓńŃṕṔŕŔśŚúÚẃẂýÝźŹíÍ");
1826 buildaccent("dacute|H|h", "oOuU", "őŐűŰ"); // double acute
1827 buildaccent("mathring|r", "aAuUwy",
1829 accents["check{\\imath}"] = "ǐ";
1830 accents["check{\\jmath}"] = "ǰ";
1831 buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
1832 "čČďĎǎǍěĚǐǏǒǑǔǓǧǦǩǨȟȞľĽňŇřŘšŠŤťžŽ"); // caron
1833 accents["hat{\\imath}"] = "î";
1834 accents["hat{\\jmath}"] = "ĵ";
1835 buildaccent("hat|^", "aAcCeEgGhHiIjJoOsSuUwWyYzZ",
1836 "âÂĉĈêÊĝĜĥĤîÎĵĴôÔŝŜûÛŵŴŷŶẑẐ"); // circ
1837 accents["bar{\\imath}"] = "ī";
1838 buildaccent("bar|=", "aAeEiIoOuUyY",
1839 "āĀēĒīĪōŌūŪȳȲ"); // macron
1840 accents["tilde{\\imath}"] = "ĩ";
1841 buildaccent("tilde", "aAeEiInNoOuUvVyY",
1842 "ãÃẽẼĩĨñÑõÕũŨṽṼỹỸ"); // tilde
1843 accents["breve{\\imath}"] = "ĭ";
1844 buildaccent("breve|u", "aAeEgGiIoOuU",
1845 "ăĂĕĔğĞĭĬŏŎŭŬ"); // breve
1846 accents["grave{\\imath}"] = "ì";
1847 buildaccent("grave|`", "aAeEiIoOuUnNwWyY",
1848 "àÀèÈìÌòÒùÙǹǸẁẀỳỲ"); // grave
1849 buildaccent("subdot|d", "BbDdHhKkLlMmNnRrSsTtVvWwZzAaEeIiOoUuYy",
1850 "ḄḅḌḍḤḥḲḳḶḷṂṃṆṇṚṛṢṣṬṭṾṿẈẉẒẓẠạẸẹỊịỌọỤụỴỵ"); // dot below
1851 buildaccent("ogonek|k", "AaEeIiUuOo",
1852 "ĄąĘęĮįŲųǪǫ"); // ogonek
1853 buildaccent("cedilla|c", "CcGgKkLlNnRrSsTtEeDdHh",
1854 "ÇçĢģĶķĻļŅņŖŗŞşŢţȨȩḐḑḨḩ"); // cedilla
1855 buildaccent("subring|textsubring", "Aa",
1857 buildaccent("subhat|textsubcircum", "DdEeLlNnTtUu",
1858 "ḒḓḘḙḼḽṊṋṰṱṶṷ"); // subcircum
1859 buildaccent("subtilde|textsubtilde", "EeIiUu",
1860 "ḚḛḬḭṴṵ"); // subtilde
1861 accents["dgrave{\\imath}"] = "ȉ";
1862 accents["textdoublegrave{\\i}"] = "ȉ";
1863 buildaccent("dgrave|textdoublegrave", "AaEeIiOoRrUu",
1864 "ȀȁȄȅȈȉȌȍȐȑȔȕ"); // double grave
1865 accents["rcap{\\imath}"] = "ȋ";
1866 accents["textroundcap{\\i}"] = "ȋ";
1867 buildaccent("rcap|textroundcap", "AaEeIiOoRrUu",
1868 "ȂȃȆȇȊȋȎȏȒȓȖȗ"); // inverted breve
1869 buildaccent("slashed", "oO",
1871 fillMissingUnicodesymbols(); // Add some still not handled entries contained in 'unicodesynbols'
1872 // LYXERR0("Number of accents " << accents.size());
1876 * Created accents in math or regexp environment
1877 * are macros, but we need the utf8 equivalent
1879 void Intervall::removeAccents()
1881 if (accents.empty())
1883 static regex const accre("\\\\(([\\S]|grave|breve|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot|ogonek|"
1884 "cedilla|subring|textsubring|subhat|textsubcircum|subtilde|textsubtilde|dgrave|textdoublegrave|rcap|textroundcap|slashed)\\{[^\\{\\}]+\\}"
1885 "|((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]))");
1887 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
1889 string key = sub.str(1);
1890 AccentsIterator it_ac = accents.find(key);
1891 if (it_ac != accents.end()) {
1892 string val = it_ac->second;
1893 size_t pos = sub.position(size_t(0));
1894 for (size_t i = 0; i < val.size(); i++) {
1895 par[pos+i] = val[i];
1897 // Remove possibly following space too
1898 if (par[pos+sub.str(0).size()] == ' ')
1899 addIntervall(pos+val.size(), pos + sub.str(0).size()+1);
1901 addIntervall(pos+val.size(), pos + sub.str(0).size());
1902 for (size_t i = pos+val.size(); i < pos + sub.str(0).size(); i++) {
1903 // remove traces of any remaining chars
1908 LYXERR(Debug::INFO, "Not added accent for \"" << key << "\"");
1913 void Intervall::handleOpenP(int i)
1916 depts[actualdeptindex] = i+1;
1917 closes[actualdeptindex] = -1;
1918 checkDepthIndex(actualdeptindex);
1921 void Intervall::handleCloseP(int i, bool closingAllowed)
1923 if (actualdeptindex <= 0) {
1924 if (! closingAllowed)
1925 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1926 // if we are at the very end
1927 addIntervall(i, i+1);
1930 closes[actualdeptindex] = i+1;
1935 void Intervall::resetOpenedP(int openPos)
1937 // Used as initializer for foreignlanguage entry
1938 actualdeptindex = 1;
1939 depts[1] = openPos+1;
1943 int Intervall::previousNotIgnored(int start) const
1945 int idx = 0; /* int intervalls */
1946 for (idx = ignoreidx; idx >= 0; --idx) {
1947 if (start > borders[idx].upper)
1949 if (start >= borders[idx].low)
1950 start = borders[idx].low-1;
1955 int Intervall::nextNotIgnored(int start) const
1957 int idx = 0; /* int intervalls */
1958 for (idx = 0; idx <= ignoreidx; idx++) {
1959 if (start < borders[idx].low)
1961 if (start < borders[idx].upper)
1962 start = borders[idx].upper;
1967 typedef unordered_map<string, KeyInfo> KeysMap;
1968 typedef unordered_map<string, KeyInfo>::const_iterator KeysIterator;
1969 typedef vector< KeyInfo> Entries;
1970 static KeysMap keys = unordered_map<string, KeyInfo>();
1976 Intervall interval_;
1977 void buildKeys(bool);
1978 void buildEntries(bool);
1979 void makeKey(const string &, KeyInfo, bool isPatternString);
1980 void processRegion(int start, int region_end); /* remove {} parts */
1981 void removeHead(KeyInfo const &, int count=0);
1984 LatexInfo(string const & par, bool isPatternString)
1985 : entidx_(-1), interval_(isPatternString, par)
1987 buildKeys(isPatternString);
1988 entries_ = vector<KeyInfo>();
1989 buildEntries(isPatternString);
1993 if (entries_.empty()) {
1996 if (entries_[0].keytype == KeyInfo::isTitle) {
1997 interval_.hasTitle = true;
1998 if (! entries_[0].disabled) {
1999 interval_.titleValue = entries_[0].head;
2002 interval_.titleValue = "";
2004 removeHead(entries_[0]);
2005 if (entries_.size() > 1)
2014 if (int(entries_.size()) > entidx_) {
2021 bool setNextKey(int idx) {
2022 if ((idx == entidx_) && (entidx_ >= 0)) {
2029 int find(int start, KeyInfo::KeyType keytype) const {
2033 while (tmpIdx < int(entries_.size())) {
2034 if (entries_[tmpIdx].keytype == keytype)
2040 int process(ostringstream & os, KeyInfo const & actual);
2041 int dispatch(ostringstream & os, int previousStart, KeyInfo & actual);
2042 // string show(int lastpos) { return interval.show(lastpos);}
2043 int nextNotIgnored(int start) { return interval_.nextNotIgnored(start);}
2044 KeyInfo &getKeyInfo(int keyinfo) {
2045 static KeyInfo invalidInfo = KeyInfo();
2046 if ((keyinfo < 0) || ( keyinfo >= int(entries_.size())))
2049 return entries_[keyinfo];
2051 void setForDefaultLang(KeyInfo const & defLang) {interval_.setForDefaultLang(defLang);}
2052 void addIntervall(int low, int up) { interval_.addIntervall(low, up); }
2056 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
2060 for (int i = start; i < end; i += 1 + skip) {
2064 if (c == '\\') skip = 1;
2068 else if (c == down) {
2071 if ((repeat <= 0) || (par[i+1] != up))
2085 size_t mathpostfixsize;
2087 size_t mathprefixsize;
2091 vector<MathEntry> entries_;
2096 void insert(string const & wait, size_t start, size_t prefixsize, size_t end, size_t postfixsize) {
2097 MathEntry m = MathEntry();
2099 m.mathStart = start;
2100 m.mathprefixsize = prefixsize;
2101 m.mathEnd = end + postfixsize;
2102 m.mathpostfixsize = postfixsize;
2103 m.mathSize = m.mathEnd - m.mathStart;
2104 entries_.push_back(m);
2106 bool empty() const { return entries_.empty(); }
2107 size_t getEndPos() const {
2108 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2111 return entries_[actualIdx_].mathEnd;
2113 size_t getStartPos() const {
2114 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2115 return 100000; /* definitely enough? */
2117 return entries_[actualIdx_].mathStart;
2119 size_t getPrefixSize() const {
2120 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2123 return entries_[actualIdx_].mathprefixsize;
2125 size_t getPostfixSize() const {
2126 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2129 return entries_[actualIdx_].mathpostfixsize;
2131 size_t getFirstPos() {
2133 return getStartPos();
2135 size_t getSize() const {
2136 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
2139 return entries_[actualIdx_].mathSize;
2141 void incrEntry() { actualIdx_++; }
2144 void LatexInfo::buildEntries(bool isPatternString)
2146 static regex const rmath("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|x?x?alignat)\\*?\\})(\\{[0-9]+\\})?)");
2147 static regex const rkeys("(\\\\)*(\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?)))");
2148 static bool disableLanguageOverride = false;
2149 smatch sub, submath;
2150 bool evaluatingRegexp = false;
2152 bool evaluatingMath = false;
2153 bool evaluatingCode = false;
2155 bool evaluatingOptional = false;
2156 size_t optionalEnd = 0;
2159 bool math_end_waiting = false;
2160 size_t math_pos = 10000;
2161 size_t math_prefix_size = 1;
2163 static vector<string> usedText = vector<string>();
2164 static bool removeMathHull = false;
2166 interval_.removeAccents();
2168 for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) {
2170 if ((submath.position(2) - submath.position(0)) %2 == 1) {
2171 // prefixed by odd count of '\\'
2174 if (math_end_waiting) {
2175 size_t pos = submath.position(size_t(2));
2176 if ((math_end == "$") &&
2177 (submath.str(2) == "$")) {
2178 mi.insert("$", math_pos, 1, pos, 1);
2179 math_end_waiting = false;
2181 else if ((math_end == "\\]") &&
2182 (submath.str(2) == "\\]")) {
2183 mi.insert("\\]", math_pos, 2, pos, 2);
2184 math_end_waiting = false;
2186 else if ((submath.str(3).compare("end") == 0) &&
2187 (submath.str(5).compare(math_end) == 0)) {
2188 mi.insert(math_end, math_pos, math_prefix_size, pos, submath.str(2).length());
2189 math_end_waiting = false;
2195 if (submath.str(3).compare("begin") == 0) {
2196 math_end_waiting = true;
2197 math_end = submath.str(5);
2198 math_pos = submath.position(size_t(2));
2199 math_prefix_size = submath.str(2).length();
2201 else if (submath.str(2).compare("\\[") == 0) {
2202 math_end_waiting = true;
2204 math_pos = submath.position(size_t(2));
2206 else if (submath.str(2) == "$") {
2207 size_t pos = submath.position(size_t(2));
2208 math_end_waiting = true;
2214 // Ignore language if there is math somewhere in pattern-string
2215 if (isPatternString) {
2216 for (auto s: usedText) {
2217 // Remove entries created in previous search runs
2220 usedText = vector<string>();
2223 keys["foreignlanguage"].disabled = true;
2224 disableLanguageOverride = true;
2225 removeMathHull = false;
2228 removeMathHull = true; // used later if not isPatternString
2229 disableLanguageOverride = false;
2233 if (disableLanguageOverride) {
2234 keys["foreignlanguage"].disabled = true;
2237 math_pos = mi.getFirstPos();
2238 for (sregex_iterator it(interval_.par.begin(), interval_.par.end(), rkeys), end; it != end; ++it) {
2240 if ((sub.position(2) - sub.position(0)) %2 == 1) {
2241 // prefixed by odd count of '\\'
2244 string key = sub.str(5);
2246 if (sub.str(2)[0] == '\\')
2247 key = sub.str(2)[1];
2252 KeysIterator it_key = keys.find(key);
2253 if (it_key != keys.end()) {
2254 if (it_key->second.keytype == KeyInfo::headRemove) {
2255 KeyInfo found1 = it_key->second;
2256 found1.disabled = true;
2257 found1.head = "\\" + key + "{";
2258 found1._tokenstart = sub.position(size_t(2));
2259 found1._tokensize = found1.head.length();
2260 found1._dataStart = found1._tokenstart + found1.head.length();
2261 int endpos = interval_.findclosing(found1._dataStart, interval_.par.length(), '{', '}', 1);
2262 found1._dataEnd = endpos;
2267 if (evaluatingRegexp) {
2268 if (sub.str(3).compare("endregexp") == 0) {
2269 evaluatingRegexp = false;
2270 // found._tokenstart already set
2271 found._dataEnd = sub.position(size_t(2)) + 13;
2272 found._dataStart = found._dataEnd;
2273 found._tokensize = found._dataEnd - found._tokenstart;
2274 found.parenthesiscount = 0;
2275 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2282 if (evaluatingMath) {
2283 if (size_t(sub.position(size_t(2))) < mi.getEndPos())
2285 evaluatingMath = false;
2287 math_pos = mi.getStartPos();
2289 if (it_key == keys.end()) {
2290 found = KeyInfo(KeyInfo::isStandard, 0, true);
2291 LYXERR(Debug::INFO, "Undefined key " << key << " ==> will be used as text");
2292 found = KeyInfo(KeyInfo::isText, 0, false);
2293 if (isPatternString) {
2294 found.keytype = KeyInfo::isChar;
2295 found.disabled = false;
2299 usedText.push_back(key);
2303 if (key.compare("regexp") == 0) {
2304 evaluatingRegexp = true;
2305 found._tokenstart = sub.position(size_t(2));
2306 found._tokensize = 0;
2310 // Handle the other params of key
2311 if (found.keytype == KeyInfo::isIgnored)
2313 else if (found.keytype == KeyInfo::isMath) {
2314 if (size_t(sub.position(size_t(2))) == math_pos) {
2316 found._tokenstart = sub.position(size_t(2));
2317 found._tokensize = mi.getSize();
2318 found._dataEnd = found._tokenstart + found._tokensize;
2319 found._dataStart = found._dataEnd;
2320 found.parenthesiscount = 0;
2321 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2322 if (removeMathHull) {
2323 interval_.addIntervall(found._tokenstart, found._tokenstart + mi.getPrefixSize());
2324 interval_.addIntervall(found._dataEnd - mi.getPostfixSize(), found._dataEnd);
2327 // Treate all math constructs as simple math
2328 interval_.par[found._tokenstart] = '$';
2329 interval_.par[found._dataEnd - mi.getPostfixSize()] = '$';
2330 interval_.addIntervall(found._tokenstart + 1, found._tokenstart + mi.getPrefixSize());
2331 interval_.addIntervall(found._dataEnd - mi.getPostfixSize() + 1, found._dataEnd);
2333 evaluatingMath = true;
2336 // begin|end of unknown env, discard
2337 // First handle tables
2338 // longtable|tabular
2339 bool discardComment;
2341 found.keytype = KeyInfo::doRemove;
2342 if ((sub.str(7).compare("longtable") == 0) ||
2343 (sub.str(7).compare("tabular") == 0)) {
2344 discardComment = true; /* '%' */
2347 discardComment = false;
2348 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
2350 string token = sub.str(7);
2351 if (regex_match(token, sub2, removeArgs)) {
2352 found.keytype = KeyInfo::removeWithArg;
2355 // discard spaces before pos(2)
2356 int pos = sub.position(size_t(2));
2358 for (count = 0; pos - count > 0; count++) {
2359 char c = interval_.par[pos-count-1];
2360 if (discardComment) {
2361 if ((c != ' ') && (c != '%'))
2367 found._tokenstart = pos - count;
2368 if (sub.str(3).compare(0, 5, "begin") == 0) {
2369 size_t pos1 = pos + sub.str(2).length();
2370 if (sub.str(7).compare("cjk") == 0) {
2371 pos1 = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2372 if ((interval_.par[pos1] == '{') && (interval_.par[pos1+1] == '}'))
2374 found.keytype = KeyInfo::isMain;
2375 found._dataStart = pos1;
2376 found._dataEnd = interval_.par.length();
2377 found.disabled = keys["foreignlanguage"].disabled;
2378 found.used = keys["foreignlanguage"].used;
2379 found._tokensize = pos1 - found._tokenstart;
2380 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2383 // Swallow possible optional params
2384 while (interval_.par[pos1] == '[') {
2385 pos1 = interval_.findclosing(pos1+1, interval_.par.length(), '[', ']')+1;
2387 // Swallow also the eventual parameter
2388 if (interval_.par[pos1] == '{') {
2389 found._dataEnd = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
2392 found._dataEnd = pos1;
2394 found._dataStart = found._dataEnd;
2395 found._tokensize = count + found._dataEnd - pos;
2396 found.parenthesiscount = 0;
2397 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2398 found.disabled = true;
2402 // Handle "\end{...}"
2403 found._dataStart = pos + sub.str(2).length();
2404 found._dataEnd = found._dataStart;
2405 found._tokensize = count + found._dataEnd - pos;
2406 found.parenthesiscount = 0;
2407 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
2408 found.disabled = true;
2412 else if (found.keytype != KeyInfo::isRegex) {
2413 found._tokenstart = sub.position(size_t(2));
2414 if (found.parenthesiscount == 0) {
2415 // Probably to be discarded
2416 size_t following_pos = sub.position(size_t(2)) + sub.str(5).length() + 1;
2417 char following = interval_.par[following_pos];
2418 if (following == ' ')
2419 found.head = "\\" + sub.str(5) + " ";
2420 else if (following == '=') {
2421 // like \uldepth=1000pt
2422 found.head = sub.str(2);
2425 found.head = "\\" + key;
2426 found._tokensize = found.head.length();
2427 found._dataEnd = found._tokenstart + found._tokensize;
2428 found._dataStart = found._dataEnd;
2431 int params = found._tokenstart + key.length() + 1;
2432 if (evaluatingOptional) {
2433 if (size_t(found._tokenstart) > optionalEnd) {
2434 evaluatingOptional = false;
2437 found.disabled = true;
2440 int optend = params;
2441 while (interval_.par[optend] == '[') {
2442 // discard optional parameters
2443 optend = interval_.findclosing(optend+1, interval_.par.length(), '[', ']') + 1;
2445 if (optend > params) {
2446 key += interval_.par.substr(params, optend-params);
2447 evaluatingOptional = true;
2448 optionalEnd = optend;
2449 if (found.keytype == KeyInfo::isSectioning) {
2450 // Remove optional values (but still keep in header)
2451 interval_.addIntervall(params, optend);
2454 string token = sub.str(7);
2456 if (interval_.par[optend] != '{') {
2458 found.parenthesiscount = 0;
2459 found.head = "\\" + key;
2462 closings = found.parenthesiscount;
2463 if (found.parenthesiscount == 1) {
2464 found.head = "\\" + key + "{";
2466 else if (found.parenthesiscount > 1) {
2468 found.head = sub.str(2) + "{";
2469 closings = found.parenthesiscount - 1;
2472 found.head = "\\" + key + "{";
2475 found._tokensize = found.head.length();
2476 found._dataStart = found._tokenstart + found.head.length();
2477 if (found.keytype == KeyInfo::doRemove) {
2479 size_t endpar = 2 + interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2480 if (endpar >= interval_.par.length())
2481 found._dataStart = interval_.par.length();
2483 found._dataStart = endpar;
2484 found._tokensize = found._dataStart - found._tokenstart;
2487 found._dataStart = found._tokenstart + found._tokensize;
2491 if (interval_.par.substr(found._dataStart, 15).compare("\\endarguments{}") == 0) {
2492 found._dataStart += 15;
2496 endpos = found._dataStart - 1;
2498 endpos = interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2499 if (found.keytype == KeyInfo::isList) {
2500 // Check if it really is list env
2501 static regex const listre("^([a-z]+)$");
2503 if (!regex_match(token, sub2, listre)) {
2504 // Change the key of this entry. It is not in a list/item environment
2505 found.keytype = KeyInfo::endArguments;
2508 if (found.keytype == KeyInfo::noMain) {
2509 evaluatingCode = true;
2511 codeStart = found._dataStart;
2513 else if (evaluatingCode) {
2514 if (size_t(found._dataStart) > codeEnd)
2515 evaluatingCode = false;
2516 else if (found.keytype == KeyInfo::isMain) {
2517 // Disable this key, treate it as standard
2518 found.keytype = KeyInfo::isStandard;
2519 found.disabled = true;
2520 if ((codeEnd +1 >= interval_.par.length()) &&
2521 (found._tokenstart == codeStart)) {
2522 // trickery, because the code inset starts
2523 // with \selectlanguage ...
2525 if (entries_.size() > 1) {
2526 entries_[entries_.size()-1]._dataEnd = codeEnd;
2531 if ((endpos == interval_.par.length()) &&
2532 (found.keytype == KeyInfo::doRemove)) {
2533 // Missing closing => error in latex-input?
2534 // therefore do not delete remaining data
2535 found._dataStart -= 1;
2536 found._dataEnd = found._dataStart;
2539 found._dataEnd = endpos;
2541 if (isPatternString) {
2542 keys[key].used = true;
2545 entries_.push_back(found);
2549 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
2551 stringstream s(keysstring);
2553 const char delim = '|';
2554 while (getline(s, key, delim)) {
2555 KeyInfo keyII(keyI);
2556 if (isPatternString) {
2559 else if ( !keys[key].used)
2560 keyII.disabled = true;
2565 void LatexInfo::buildKeys(bool isPatternString)
2568 static bool keysBuilt = false;
2569 if (keysBuilt && !isPatternString) return;
2571 // Keys to ignore in any case
2572 makeKey("text|textcyrillic|lyxmathsym|ensuremath", KeyInfo(KeyInfo::headRemove, 1, true), true);
2573 // Known standard keys with 1 parameter.
2574 // Split is done, if not at start of region
2575 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
2576 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
2577 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
2578 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
2579 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
2580 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
2582 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
2583 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2584 makeKey("section*|subsection*|subsubsection*|paragraph*",
2585 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2586 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2587 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
2589 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
2591 // Split is done, if not at start of region
2592 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
2593 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
2595 // Split is done always.
2596 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
2600 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2601 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2602 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2603 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2605 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2606 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2607 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2608 makeKey("thickspace|medspace|thinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2610 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2611 // Custom space/skip, remove the content (== length value)
2612 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
2613 // Found in fr/UserGuide.lyx
2614 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2616 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2617 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2618 // Known macros to remove (including their parameter)
2620 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
2621 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
2622 // handle like standard keys with 1 parameter.
2623 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
2625 // Ignore deleted text
2626 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString);
2627 // but preserve added text
2628 makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2630 // Macros to remove, but let the parameter survive
2632 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2634 // Remove language spec from content of these insets
2635 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
2637 // Same effect as previous, parameter will survive (because there is no one anyway)
2639 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2640 // Remove table decorations
2641 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
2642 // Discard shape-header.
2643 // For footnote or shortcut too, because of lang settings
2644 // and wrong handling if used 'KeyInfo::noMain'
2645 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2646 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2647 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2648 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2649 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2650 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
2651 // like ('tiny{}' or '\tiny ' ... )
2652 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
2654 // Survives, like known character
2655 // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2656 makeKey("tableofcontents", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2657 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
2659 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2660 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2661 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2663 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2664 // Remove RTL/LTR marker
2665 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2666 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
2667 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
2668 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
2669 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
2670 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
2671 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2672 if (isPatternString) {
2673 // Allow the first searched string to rebuild the keys too
2677 // no need to rebuild again
2683 * Keep the list of actual opened parentheses actual
2684 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2686 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2689 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2693 if (c == '\\') skip = 1;
2694 else if (c == '{') {
2697 else if (c == '}') {
2698 handleCloseP(i, closingAllowed);
2704 string Intervall::show(int lastpos)
2706 int idx = 0; /* int intervalls */
2709 for (idx = 0; idx <= ignoreidx; idx++) {
2710 while (i < lastpos) {
2712 if (i <= borders[idx].low) {
2713 if (borders[idx].low > lastpos)
2714 printsize = lastpos - i;
2716 printsize = borders[idx].low - i;
2717 s += par.substr(i, printsize);
2719 if (i >= borders[idx].low)
2720 i = borders[idx].upper;
2723 i = borders[idx].upper;
2729 s += par.substr(i, lastpos-i);
2735 void Intervall::output(ostringstream &os, int lastpos)
2737 // get number of chars to output
2738 int idx = 0; /* int intervalls */
2741 string startTitle = titleValue;
2742 for (idx = 0; idx <= ignoreidx; idx++) {
2744 if (i <= borders[idx].low) {
2746 if (borders[idx].low > lastpos)
2747 printsize = lastpos - i;
2749 printsize = borders[idx].low - i;
2750 if (printsize > 0) {
2751 os << startTitle << par.substr(i, printsize);
2753 printed += printsize;
2756 handleParentheses(i, false);
2757 if (i >= borders[idx].low)
2758 i = borders[idx].upper;
2761 i = borders[idx].upper;
2768 os << startTitle << par.substr(i, lastpos-i);
2769 printed += lastpos-i;
2771 handleParentheses(lastpos, false);
2773 if (keys["foreignlanguage"].disabled)
2774 startindex = actualdeptindex-langcount;
2776 startindex = actualdeptindex;
2777 for (int i = startindex; i > 0; --i) {
2780 if (hasTitle && (printed > 0))
2782 if (! isPatternString_)
2784 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2787 void LatexInfo::processRegion(int start, int region_end)
2789 while (start < region_end) { /* Let {[} and {]} survive */
2790 int cnt = interval_.isOpeningPar(start);
2792 // Closing is allowed past the region
2793 int closing = interval_.findclosing(start+1, interval_.par.length());
2794 interval_.addIntervall(start, start+1);
2795 interval_.addIntervall(closing, closing+1);
2799 start = interval_.nextNotIgnored(start+1);
2803 void LatexInfo::removeHead(KeyInfo const & actual, int count)
2805 if (actual.parenthesiscount == 0) {
2806 // "{\tiny{} ...}" ==> "{{} ...}"
2807 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2810 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2811 interval_.addIntervall(actual._tokenstart - count, actual._dataStart);
2812 interval_.addIntervall(actual._dataEnd, actual._dataEnd+1);
2816 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2819 switch (actual.keytype)
2821 case KeyInfo::isTitle: {
2823 nextKeyIdx = getNextKey();
2826 case KeyInfo::cleanToStart: {
2827 actual._dataEnd = actual._dataStart;
2828 nextKeyIdx = getNextKey();
2829 // Search for end of arguments
2830 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2832 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2833 entries_[i].disabled = true;
2835 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2837 while (interval_.par[actual._dataEnd] == ' ')
2839 interval_.addIntervall(0, actual._dataEnd+1);
2840 interval_.actualdeptindex = 0;
2841 interval_.depts[0] = actual._dataEnd+1;
2842 interval_.closes[0] = -1;
2845 case KeyInfo::isText:
2846 interval_.par[actual._tokenstart] = '#';
2847 //interval_.addIntervall(actual._tokenstart, actual._tokenstart+1);
2848 nextKeyIdx = getNextKey();
2850 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2851 if (actual.disabled)
2852 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2854 interval_.addIntervall(actual._dataStart, actual._dataEnd);
2857 case KeyInfo::isChar: {
2858 nextKeyIdx = getNextKey();
2861 case KeyInfo::isSize: {
2862 if (actual.disabled || (interval_.par[actual._dataStart] != '{') || (interval_.par[actual._dataStart-1] == ' ')) {
2863 if (actual.parenthesiscount == 0)
2864 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2866 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2868 nextKeyIdx = getNextKey();
2870 // Here _dataStart points to '{', so correct it
2871 actual._dataStart += 1;
2872 actual._tokensize += 1;
2873 actual.parenthesiscount = 1;
2874 if (interval_.par[actual._dataStart] == '}') {
2875 // Determine the end if used like '{\tiny{}...}'
2876 actual._dataEnd = interval_.findclosing(actual._dataStart+1, interval_.par.length()) + 1;
2877 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2880 // Determine the end if used like '\tiny{...}'
2881 actual._dataEnd = interval_.findclosing(actual._dataStart, interval_.par.length()) + 1;
2883 // Split on this key if not at start
2884 int start = interval_.nextNotIgnored(previousStart);
2885 if (start < actual._tokenstart) {
2886 interval_.output(os, actual._tokenstart);
2887 interval_.addIntervall(start, actual._tokenstart);
2889 // discard entry if at end of actual
2890 nextKeyIdx = process(os, actual);
2894 case KeyInfo::endArguments: {
2895 // Remove trailing '{}' too
2896 actual._dataStart += 1;
2897 actual._dataEnd += 1;
2898 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2899 nextKeyIdx = getNextKey();
2902 case KeyInfo::noMain:
2904 case KeyInfo::isStandard: {
2905 if (actual.disabled) {
2907 processRegion(actual._dataStart, actual._dataStart+1);
2908 nextKeyIdx = getNextKey();
2910 // Split on this key if not at datastart of calling entry
2911 int start = interval_.nextNotIgnored(previousStart);
2912 if (start < actual._tokenstart) {
2913 interval_.output(os, actual._tokenstart);
2914 interval_.addIntervall(start, actual._tokenstart);
2916 // discard entry if at end of actual
2917 nextKeyIdx = process(os, actual);
2921 case KeyInfo::removeWithArg: {
2922 nextKeyIdx = getNextKey();
2923 // Search for end of arguments
2924 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2926 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2927 entries_[i].disabled = true;
2929 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2931 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2934 case KeyInfo::doRemove: {
2935 // Remove the key with all parameters and following spaces
2938 if (interval_.par[actual._dataEnd-1] == ' ')
2939 start = actual._dataEnd;
2941 start = actual._dataEnd+1;
2942 for (pos = start; pos < interval_.par.length(); pos++) {
2943 if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
2946 // Remove also enclosing parentheses [] and {}
2949 while (actual._tokenstart > numpars) {
2950 if (pos+numpars >= interval_.par.size())
2952 else if (interval_.par[pos+numpars] == ']' && interval_.par[actual._tokenstart-numpars-1] == '[')
2954 else if (interval_.par[pos+numpars] == '}' && interval_.par[actual._tokenstart-numpars-1] == '{')
2960 if (interval_.par[pos+numpars] == ' ')
2964 interval_.addIntervall(actual._tokenstart-numpars, pos+numpars+spaces);
2965 nextKeyIdx = getNextKey();
2968 case KeyInfo::isList: {
2969 // Discard space before _tokenstart
2971 for (count = 0; count < actual._tokenstart; count++) {
2972 if (interval_.par[actual._tokenstart-count-1] != ' ')
2975 nextKeyIdx = getNextKey();
2976 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2978 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2980 // How else can we catch this one?
2981 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2982 entries_[i].disabled = true;
2984 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2986 else if (nextKeyIdx > 0) {
2987 // Ignore any lang entries inside data region
2988 for (int i = nextKeyIdx; i < int(entries_.size()) && entries_[i]._tokenstart < actual._dataEnd; i++) {
2989 if (entries_[i].keytype == KeyInfo::isMain)
2990 entries_[i].disabled = true;
2993 if (actual.disabled) {
2994 interval_.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2997 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2999 if (interval_.par[actual._dataEnd+1] == '[') {
3000 int posdown = interval_.findclosing(actual._dataEnd+2, interval_.par.length(), '[', ']');
3001 if ((interval_.par[actual._dataEnd+2] == '{') &&
3002 (interval_.par[posdown-1] == '}')) {
3003 interval_.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
3004 interval_.addIntervall(posdown-1, posdown+1);
3007 interval_.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
3008 interval_.addIntervall(posdown, posdown+1);
3010 int blk = interval_.nextNotIgnored(actual._dataEnd+1);
3011 if (blk > posdown) {
3012 // Discard at most 1 space after empty item
3014 for (count = 0; count < 1; count++) {
3015 if (interval_.par[blk+count] != ' ')
3019 interval_.addIntervall(blk, blk+count);
3024 case KeyInfo::isSectioning: {
3025 // Discard spaces before _tokenstart
3027 int val = actual._tokenstart;
3028 for (count = 0; count < actual._tokenstart;) {
3029 val = interval_.previousNotIgnored(val-1);
3030 if (val < 0 || interval_.par[val] != ' ')
3033 count = actual._tokenstart - val;
3036 if (actual.disabled) {
3037 removeHead(actual, count);
3038 nextKeyIdx = getNextKey();
3040 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
3041 nextKeyIdx = process(os, actual);
3045 case KeyInfo::isMath: {
3046 // Same as regex, use the content unchanged
3047 nextKeyIdx = getNextKey();
3050 case KeyInfo::isRegex: {
3051 // DO NOT SPLIT ON REGEX
3053 nextKeyIdx = getNextKey();
3056 case KeyInfo::isIgnored: {
3057 // Treat like a character for now
3058 nextKeyIdx = getNextKey();
3061 case KeyInfo::isMain: {
3062 if (interval_.par.substr(actual._dataStart, 2) == "% ")
3063 interval_.addIntervall(actual._dataStart, actual._dataStart+2);
3064 if (actual._tokenstart > 0) {
3065 int prev = interval_.previousNotIgnored(actual._tokenstart - 1);
3066 if ((prev >= 0) && interval_.par[prev] == '%')
3067 interval_.addIntervall(prev, prev+1);
3069 if (actual.disabled) {
3071 interval_.langcount++;
3072 if ((interval_.par.substr(actual._dataStart, 3) == " \\[") ||
3073 (interval_.par.substr(actual._dataStart, 8) == " \\begin{")) {
3074 // Discard also the space before math-equation
3075 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
3077 nextKeyIdx = getNextKey();
3078 // interval.resetOpenedP(actual._dataStart-1);
3081 if (actual._tokenstart < 26) {
3082 // for the first (and maybe dummy) language
3083 interval_.setForDefaultLang(actual);
3085 interval_.resetOpenedP(actual._dataStart-1);
3089 case KeyInfo::invalid:
3090 case KeyInfo::headRemove:
3091 // These two cases cannot happen, already handled
3094 // LYXERR(Debug::INFO, "Unhandled keytype");
3095 nextKeyIdx = getNextKey();
3102 int LatexInfo::process(ostringstream & os, KeyInfo const & actual )
3104 int end = interval_.nextNotIgnored(actual._dataEnd);
3105 int oldStart = actual._dataStart;
3106 int nextKeyIdx = getNextKey();
3108 if ((nextKeyIdx < 0) ||
3109 (entries_[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
3110 (entries_[nextKeyIdx].keytype == KeyInfo::invalid)) {
3111 if (oldStart <= end) {
3112 processRegion(oldStart, end);
3117 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
3119 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
3120 (void) dispatch(os, actual._dataStart, nextKey);
3121 end = nextKey._tokenstart;
3124 processRegion(oldStart, nextKey._tokenstart);
3125 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
3127 oldStart = nextKey._dataEnd+1;
3129 // now nextKey is either invalid or is outside of actual._dataEnd
3130 // output the remaining and discard myself
3131 if (oldStart <= end) {
3132 processRegion(oldStart, end);
3134 if (interval_.par.size() > (size_t) end && interval_.par[end] == '}') {
3136 // This is the normal case.
3137 // But if using the firstlanguage, the closing may be missing
3139 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
3141 if (actual._dataEnd < end)
3142 output_end = interval_.nextNotIgnored(actual._dataEnd);
3143 else if (interval_.par.size() > (size_t) end)
3144 output_end = interval_.nextNotIgnored(end);
3146 output_end = interval_.par.size();
3147 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
3148 interval_.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
3150 // Remove possible empty data
3151 int dstart = interval_.nextNotIgnored(actual._dataStart);
3152 while (interval_.isOpeningPar(dstart) == 1) {
3153 interval_.addIntervall(dstart, dstart+1);
3154 int dend = interval_.findclosing(dstart+1, output_end);
3155 interval_.addIntervall(dend, dend+1);
3156 dstart = interval_.nextNotIgnored(dstart+1);
3158 if (dstart < output_end)
3159 interval_.output(os, output_end);
3161 interval_.addIntervall(0, end);
3163 interval_.addIntervall(actual._tokenstart, end);
3167 string splitOnKnownMacros(string par, bool isPatternString)
3170 LatexInfo li(par, isPatternString);
3171 // LYXERR(Debug::INFO, "Berfore split: " << par);
3172 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
3174 DummyKey._tokensize = 0;
3175 DummyKey._dataStart = 0;
3176 DummyKey._dataEnd = par.length();
3177 DummyKey.disabled = true;
3178 int firstkeyIdx = li.getFirstKey();
3180 if (firstkeyIdx >= 0) {
3181 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
3182 DummyKey._tokenstart = firstKey._tokenstart;
3184 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
3185 // Use dummy firstKey
3186 firstKey = DummyKey;
3187 (void) li.setNextKey(firstkeyIdx);
3190 if (par.substr(firstKey._dataStart, 2) == "% ")
3191 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
3193 nextkeyIdx = li.process(os, firstKey);
3194 while (nextkeyIdx >= 0) {
3195 // Check for a possible gap between the last
3196 // entry and this one
3197 int datastart = li.nextNotIgnored(firstKey._dataStart);
3198 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
3199 if ((nextKey._tokenstart > datastart)) {
3201 firstKey._dataStart = datastart;
3202 firstKey._dataEnd = par.length();
3203 (void) li.setNextKey(nextkeyIdx);
3204 // Fake the last opened parenthesis
3205 li.setForDefaultLang(firstKey);
3206 nextkeyIdx = li.process(os, firstKey);
3209 if (nextKey.keytype != KeyInfo::isMain) {
3210 firstKey._dataStart = datastart;
3211 firstKey._dataEnd = nextKey._dataEnd+1;
3212 (void) li.setNextKey(nextkeyIdx);
3213 li.setForDefaultLang(firstKey);
3214 nextkeyIdx = li.process(os, firstKey);
3217 nextkeyIdx = li.process(os, nextKey);
3221 // Handle the remaining
3222 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
3223 firstKey._dataEnd = par.length();
3225 if ((firstKey._dataStart < firstKey._dataEnd) &&
3226 (par[firstKey._dataStart] != '}')) {
3227 li.setForDefaultLang(firstKey);
3228 (void) li.process(os, firstKey);
3232 // return string definitelly impossible to match
3233 s = "\\foreignlanguage{ignore}{ }";
3237 s = par; /* no known macros found */
3238 // LYXERR(Debug::INFO, "After split: " << s);
3243 * Try to unify the language specs in the latexified text.
3244 * Resulting modified string is set to "", if
3245 * the searched tex does not contain all the features in the search pattern
3247 static string correctlanguagesetting(string par, bool isPatternString, bool withformat, lyx::Buffer *pbuf = nullptr)
3249 static Features regex_f;
3250 static int missed = 0;
3251 static bool regex_with_format = false;
3253 int parlen = par.length();
3255 while ((parlen > 0) && (par[parlen-1] == '\n')) {
3258 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
3259 // Happens to be there in case of description or labeling environment
3264 // Split the latex input into pieces which
3265 // can be digested by our search engine
3266 LYXERR(Debug::FIND, "input: \"" << par << "\"");
3267 if (isPatternString && (pbuf != nullptr)) { // Check if we should disable/enable test for language
3268 // We check for polyglossia, because in runparams.flavor we use Flavor::XeTeX
3269 string doclang = pbuf->params().language->polyglossia();
3270 static regex langre("\\\\(foreignlanguage)\\{([^\\}]+)\\}");
3272 bool toIgnoreLang = true;
3273 for (sregex_iterator it(par.begin(), par.end(), langre), end; it != end; ++it) {
3275 if (sub.str(2) != doclang) {
3276 toIgnoreLang = false;
3280 setIgnoreFormat("language", toIgnoreLang, false);
3283 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
3284 LYXERR(Debug::FIND, "After splitOnKnownMacros:\n\"" << result << "\"");
3287 result = par.substr(0, parlen);
3288 if (isPatternString) {
3291 regex_f = identifyFeatures(result);
3292 string features = "";
3293 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3294 string a = it->first;
3295 regex_with_format = true;
3296 features += " " + a;
3297 // LYXERR(Debug::INFO, "Identified regex format:" << a);
3299 LYXERR(Debug::FIND, "Identified Features" << features);
3302 } else if (regex_with_format) {
3303 Features info = identifyFeatures(result);
3304 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
3305 string a = it->first;
3306 bool b = it->second;
3307 if (b && ! info[a]) {
3309 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
3316 // LYXERR(Debug::INFO, "No regex formats");
3322 // Remove trailing closure of math, macros and environments, so to catch parts of them.
3323 static int identifyClosing(string & t)
3325 int open_braces = 0;
3327 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
3328 if (regex_replace(t, t, "(.*[^\\\\])\\$$", "$1"))
3330 if (regex_replace(t, t, "(.*[^\\\\])\\\\\\]$", "$1"))
3332 if (regex_replace(t, t, "(.*[^\\\\])\\\\end\\{[a-zA-Z_]*\\*?\\}$", "$1"))
3334 if (regex_replace(t, t, "(.*[^\\\\])\\}$", "$1")) {
3343 static int num_replaced = 0;
3344 static bool previous_single_replace = true;
3346 void MatchStringAdv::CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string)
3349 // Handle \w properly
3350 QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
3351 if (! opt.casesensitive) {
3352 popts |= QRegularExpression::CaseInsensitiveOption;
3354 regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
3355 regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
3357 if (regexp.isValid() && regexp2.isValid()) {
3358 regexIsValid = true;
3359 // Check '{', '}' pairs inside the regex
3362 for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
3363 char c = par_as_string[i];
3370 else if (c == '}') {
3377 if (balanced != 0) {
3378 regexIsValid = false;
3379 regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
3383 regexIsValid = false;
3384 if (!regexp.isValid())
3385 regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
3387 regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
3390 (void)par_as_string;
3391 if (opt.casesensitive) {
3392 regexp = regex(regexp_str);
3393 regexp2 = regex(regexp2_str);
3396 regexp = regex(regexp_str, std::regex_constants::icase);
3397 regexp2 = regex(regexp2_str, std::regex_constants::icase);
3402 static void modifyRegexForMatchWord(string &t)
3405 regex wordre("(\\\\)*((\\.|\\\\b))");
3408 for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
3410 if ((sub.position(2) - sub.position(0)) % 2 == 1) {
3413 else if (sub.str(2) == "\\\\b")
3415 if (lastpos < (size_t) sub.position(2))
3416 s += t.substr(lastpos, sub.position(2) - lastpos);
3418 lastpos = sub.position(2) + sub.length(2);
3421 s = "\\b" + t + "\\b";
3425 else if (lastpos < t.length())
3426 s += t.substr(lastpos, t.length() - lastpos);
3427 t = "\\b" + s + "\\b";
3430 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt)
3431 : p_buf(&buf), p_first_buf(&buf), opt(opt)
3433 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
3434 docstring const & ds = stringifySearchBuffer(find_buf, opt);
3435 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
3436 if (opt.replace_all && previous_single_replace) {
3437 previous_single_replace = false;
3440 else if (!opt.replace_all) {
3441 num_replaced = 0; // count number of replaced strings
3442 previous_single_replace = true;
3444 // When using regexp, braces are hacked already by escape_for_regex()
3445 par_as_string = normalize(ds);
3447 close_wildcards = 0;
3449 size_t lead_size = 0;
3450 // correct the language settings
3451 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat, &buf);
3452 opt.matchAtStart = false;
3454 identifyClosing(par_as_string); // Removes math closings ($, ], ...) at end of string
3455 if (opt.ignoreformat) {
3459 lead_size = identifyLeading(par_as_string);
3461 lead_as_string = par_as_string.substr(0, lead_size);
3462 string lead_as_regex_string = string2regex(lead_as_string);
3463 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3464 string par_as_regex_string_nolead = string2regex(par_as_string_nolead);
3465 /* Handle whole words too in this case
3467 if (opt.matchword) {
3468 par_as_regex_string_nolead = "\\b" + par_as_regex_string_nolead + "\\b";
3469 opt.matchword = false;
3471 string regexp_str = "(" + lead_as_regex_string + ")()" + par_as_regex_string_nolead;
3472 string regexp2_str = "(" + lead_as_regex_string + ")(.*?)" + par_as_regex_string_nolead;
3473 CreateRegexp(opt, regexp_str, regexp2_str);
3475 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3476 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3480 if (!opt.ignoreformat) {
3481 lead_size = identifyLeading(par_as_string);
3482 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
3483 lead_as_string = par_as_string.substr(0, lead_size);
3484 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
3487 // Here we are using regexp
3488 LASSERT(use_regexp, /**/);
3490 string lead_as_regexp;
3491 if (lead_size > 0) {
3492 lead_as_regexp = string2regex(par_as_string.substr(0, lead_size));
3493 (void)regex_replace(par_as_string_nolead, par_as_string_nolead, "}$", "");
3494 par_as_string = par_as_string_nolead;
3495 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
3496 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3498 // LYXERR(Debug::FIND, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
3499 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
3500 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
3501 // LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3503 size_t lng = par_as_string.size();
3504 if (!opt.ignoreformat) {
3505 // Remove extra '\}' at end if not part of \{\.\}
3507 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
3509 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
3518 if (lng < par_as_string.size())
3519 par_as_string = par_as_string.substr(0,lng);
3521 LYXERR(Debug::FIND, "par_as_string after correctRegex is '" << par_as_string << "'");
3522 if ((lng > 0) && (par_as_string[0] == '^')) {
3523 par_as_string = par_as_string.substr(1);
3525 opt.matchAtStart = true;
3527 // LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
3528 // LYXERR(Debug::FIND, "Open braces: " << open_braces);
3529 // LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
3531 // If entered regexp must match at begin of searched string buffer
3532 // Kornel: Added parentheses to use $1 for size of the leading string
3536 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
3537 // Unfortunately is '\1', '\2', etc not working for strings with extra format
3538 // so the convert has no effect in that case
3539 for (int i = 7; i > 0; --i) {
3540 string orig = "\\\\" + std::to_string(i);
3541 string dest = "\\" + std::to_string(i+2);
3542 while (regex_replace(par_as_string, par_as_string, orig, dest));
3544 if (opt.matchword) {
3545 modifyRegexForMatchWord(par_as_string);
3546 opt.matchword = false;
3548 regexp_str = "(" + lead_as_regexp + ")()" + par_as_string;
3549 regexp2_str = "(" + lead_as_regexp + ")(.*?)" + par_as_string;
3551 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
3552 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
3553 CreateRegexp(opt, regexp_str, regexp2_str, par_as_string);
3557 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
3561 mres.searched_size = len;
3563 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
3566 docstring docstr = stringifyFromForSearch(opt, cur, len);
3568 str = normalize(docstr);
3569 if (!opt.ignoreformat) {
3570 str = correctlanguagesetting(str, false, !opt.ignoreformat);
3571 // remove closing '}' and '\n' to allow for use of '$' in regex
3572 size_t lng = str.size();
3573 while ((lng > 1) && ((str[lng -1] == '}') || (str[lng -1] == '\n')))
3575 if (lng != str.size()) {
3576 str = str.substr(0, lng);
3580 mres.match_len = -1;
3583 LYXERR(Debug::FIND, "After normalization: Matching against:\n'" << str << "'");
3585 LASSERT(use_regexp, /**/);
3587 // use_regexp always true
3588 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
3590 QString qstr = QString::fromStdString(str);
3591 QRegularExpression const *p_regexp;
3592 QRegularExpression::MatchType flags = QRegularExpression::NormalMatch;
3596 p_regexp = ®exp2;
3598 QRegularExpressionMatch match = p_regexp->match(qstr, 0, flags);
3599 if (!match.hasMatch())
3602 regex const *p_regexp;
3603 regex_constants::match_flag_type flags;
3605 flags = regex_constants::match_continuous;
3608 flags = regex_constants::match_default;
3609 p_regexp = ®exp2;
3611 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
3612 if (re_it == sregex_iterator())
3614 match_results<string::const_iterator> const & m = *re_it;
3616 // Whole found string, including the leading
3617 // std: m[0].second - m[0].first
3618 // Qt: match.capturedEnd(0) - match.capturedStart(0)
3620 // Size of the leading string
3621 // std: m[1].second - m[1].first
3622 // Qt: match.capturedEnd(1) - match.capturedStart(1)
3623 int leadingsize = 0;
3625 if (match.lastCapturedIndex() > 0) {
3626 leadingsize = match.capturedEnd(1) - match.capturedStart(1);
3631 leadingsize = m[1].second - m[1].first;
3635 mres.match_prefix = match.capturedEnd(2) - match.capturedStart(2);
3636 mres.match_len = match.capturedEnd(0) - match.capturedEnd(2);
3637 // because of different number of closing at end of string
3638 // we have to 'unify' the length of the post-match.
3639 // Done by ignoring closing parenthesis and linefeeds at string end
3640 int matchend = match.capturedEnd(0);
3641 size_t strsize = qstr.size();
3642 if (!opt.ignoreformat) {
3643 while (mres.match_len > 0) {
3644 QChar c = qstr.at(matchend - 1);
3645 if ((c == '\n') || (c == '}') || (c == '{')) {
3652 while (strsize > (size_t) match.capturedEnd(0)) {
3653 QChar c = qstr.at(strsize-1);
3654 if ((c == '\n') || (c == '}')) {
3661 // LYXERR0(qstr.toStdString());
3662 mres.match2end = strsize - matchend;
3663 mres.pos = match.capturedStart(2);
3665 mres.match_prefix = m[2].second - m[2].first;
3666 mres.match_len = m[0].second - m[2].second;
3667 // ignore closing parenthesis and linefeeds at string end
3668 size_t strend = m[0].second - m[0].first;
3669 int matchend = strend;
3670 size_t strsize = str.size();
3671 if (!opt.ignoreformat) {
3672 while (mres.match_len > 0) {
3673 char c = str.at(matchend - 1);
3674 if ((c == '\n') || (c == '}') || (c == '{')) {
3681 while (strsize > strend) {
3682 if ((str.at(strsize-1) == '}') || (str.at(strsize-1) == '\n')) {
3690 mres.match2end = strsize - matchend;
3691 mres.pos = m[2].first - m[0].first;;
3693 if (mres.match2end < 0)
3695 mres.leadsize = leadingsize;
3697 if (mres.match_len > 0) {
3698 string a0 = match.captured(0).mid(mres.pos + mres.match_prefix, mres.match_len).toStdString();
3699 mres.result.push_back(a0);
3700 for (int i = 3; i <= match.lastCapturedIndex(); i++) {
3701 mres.result.push_back(match.captured(i).toStdString());
3705 if (mres.match_len > 0) {
3706 string a0 = m[0].str().substr(mres.pos + mres.match_prefix, mres.match_len);
3707 mres.result.push_back(a0);
3708 for (size_t i = 3; i < m.size(); i++) {
3709 mres.result.push_back(m[i]);
3718 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
3720 MatchResult mres = findAux(cur, len, at_begin);
3721 int res = mres.match_len;
3723 "res=" << res << ", at_begin=" << at_begin
3724 << ", matchAtStart=" << opt.matchAtStart
3725 << ", inTexted=" << cur.inTexted());
3726 if (opt.matchAtStart) {
3729 else if (mres.match_prefix > 0)
3738 static bool simple_replace(string &t, string from, string to)
3740 regex repl("(\\\\)*(" + from + ")");
3744 for (sregex_iterator it(t.begin(), t.end(), repl), end; it != end; ++it) {
3746 if ((sub.position(2) - sub.position(0)) % 2 == 1)
3748 if (lastpos < (size_t) sub.position(2))
3749 s += t.substr(lastpos, sub.position(2) - lastpos);
3751 lastpos = sub.position(2) + sub.length(2);
3755 else if (lastpos < t.length())
3756 s += t.substr(lastpos, t.length() - lastpos);
3762 string MatchStringAdv::normalize(docstring const & s) const
3765 t = lyx::to_utf8(s);
3766 // Remove \n at begin
3767 while (!t.empty() && t[0] == '\n')
3770 while (!t.empty() && t[t.size() - 1] == '\n')
3771 t = t.substr(0, t.size() - 1);
3773 // Handle all other '\n'
3774 while ((pos = t.find("\n")) != string::npos) {
3775 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
3777 if (isAlnumASCII(t[pos+1])) {
3778 t.replace(pos-2, 3, " ");
3781 t.replace(pos-2, 3, "");
3784 else if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
3785 // '\n' adjacent to non-alpha-numerics, discard
3786 t.replace(pos, 1, "");
3789 // Replace all other \n with spaces
3790 t.replace(pos, 1, " ");
3793 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
3794 // Kornel: Added textsl, textsf, textit, texttt and noun
3795 // + allow to seach for colored text too
3796 LYXERR(Debug::FIND, "Removing stale empty macros from: " << t);
3797 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
3798 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3799 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
3800 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3801 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
3807 docstring stringifyFromCursor(DocIterator const & cur, int len)
3809 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
3810 if (cur.inTexted()) {
3811 Paragraph const & par = cur.paragraph();
3812 // TODO what about searching beyond/across paragraph breaks ?
3813 // TODO Try adding a AS_STR_INSERTS as last arg
3814 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
3815 int(par.size()) : cur.pos() + len;
3816 // OutputParams runparams(&cur.buffer()->params().encoding());
3817 OutputParams runparams(encodings.fromLyXName("utf8"));
3818 runparams.nice = true;
3819 runparams.flavor = Flavor::XeTeX;
3820 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
3821 // No side effect of file copying and image conversion
3822 runparams.dryrun = true;
3823 int option = AS_STR_INSETS | AS_STR_PLAINTEXT;
3824 if (ignoreFormats.getDeleted()) {
3825 option |= AS_STR_SKIPDELETE;
3826 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
3829 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
3831 LYXERR(Debug::FIND, "Stringifying with cur: "
3832 << cur << ", from pos: " << cur.pos() << ", end: " << end);
3833 return par.asString(cur.pos(), end,
3836 } else if (cur.inMathed()) {
3837 CursorSlice cs = cur.top();
3838 MathData md = cs.cell();
3839 MathData::const_iterator it_end =
3840 (( len == -1 || cs.pos() + len > int(md.size()))
3842 : md.begin() + cs.pos() + len );
3844 for (MathData::const_iterator it = md.begin() + cs.pos();
3847 docstring s = asString(md2);
3848 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
3851 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3856 /** Computes the LaTeX export of buf starting from cur and ending len positions
3857 * after cur, if len is positive, or at the paragraph or innermost inset end
3860 docstring latexifyFromCursor(DocIterator const & cur, int len)
3863 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
3864 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
3865 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
3867 Buffer const & buf = *cur.buffer();
3869 odocstringstream ods;
3871 //OutputParams runparams(&buf.params().encoding());
3872 OutputParams runparams(encodings.fromLyXName("utf8"));
3873 runparams.nice = false;
3874 runparams.flavor = Flavor::XeTeX;
3875 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3876 // No side effect of file copying and image conversion
3877 runparams.dryrun = true;
3878 if (ignoreFormats.getDeleted()) {
3879 runparams.for_searchAdv = OutputParams::SearchWithoutDeleted;
3882 runparams.for_searchAdv = OutputParams::SearchWithDeleted;
3885 if (cur.inTexted()) {
3886 // @TODO what about searching beyond/across paragraph breaks ?
3887 pos_type endpos = cur.paragraph().size();
3888 if (len != -1 && endpos > cur.pos() + len)
3889 endpos = cur.pos() + len;
3890 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
3891 string(), cur.pos(), endpos);
3892 string s = lyx::to_utf8(ods.str());
3893 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
3894 return(lyx::from_utf8(s));
3895 } else if (cur.inMathed()) {
3896 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
3897 for (int s = cur.depth() - 1; s >= 0; --s) {
3898 CursorSlice const & cs = cur[s];
3899 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
3900 TeXMathStream ws(os);
3901 cs.asInsetMath()->asHullInset()->header_write(ws);
3906 CursorSlice const & cs = cur.top();
3907 MathData md = cs.cell();
3908 MathData::const_iterator it_end =
3909 ((len == -1 || cs.pos() + len > int(md.size()))
3911 : md.begin() + cs.pos() + len);
3913 for (MathData::const_iterator it = md.begin() + cs.pos();
3917 ods << asString(md2);
3918 // Retrieve the math environment type, and add '$' or '$]'
3919 // or others (\end{equation}) accordingly
3920 for (int s = cur.depth() - 1; s >= 0; --s) {
3921 CursorSlice const & cs2 = cur[s];
3922 InsetMath * inset = cs2.asInsetMath();
3923 if (inset && inset->asHullInset()) {
3924 TeXMathStream ws(os);
3925 inset->asHullInset()->footer_write(ws);
3929 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3931 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3936 #if defined(ResultsDebug)
3938 static void displayMResult(MatchResult &mres, string from, DocIterator & cur)
3940 LYXERR0( "from:\t\t\t" << from);
3942 if (mres.pos_len > 0) {
3944 status = "FINALSEARCH";
3947 if (mres.match_len > 0) {
3948 if ((mres.match_prefix == 0) && (mres.pos == mres.leadsize))
3949 status = "Good Match";
3951 status = "Matched in";
3954 status = "MissedSearch";
3957 LYXERR0( status << "(" << cur.pos() << " ... " << mres.searched_size + cur.pos() << ") cur.lastpos(" << cur.lastpos() << ")");
3958 if ((mres.leadsize > 0) || (mres.match_len > 0) || (mres.match2end > 0))
3959 LYXERR0( "leadsize(" << mres.leadsize << ") match_len(" << mres.match_len << ") match2end(" << mres.match2end << ")");
3960 if ((mres.pos > 0) || (mres.match_prefix > 0))
3961 LYXERR0( "pos(" << mres.pos << ") match_prefix(" << mres.match_prefix << ")");
3962 for (size_t i = 0; i < mres.result.size(); i++)
3963 LYXERR0( "Match " << i << " = \"" << mres.result[i] << "\"");
3965 #define displayMres(s, txt, cur) displayMResult(s, txt, cur);
3967 #define displayMres(s, txt, cur)
3970 /** Finalize an advanced find operation, advancing the cursor to the innermost
3971 ** position that matches, plus computing the length of the matching text to
3973 ** Return the cur.pos() difference between start and end of found match
3975 MatchResult findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, MatchResult const & expected = MatchResult(-1))
3977 // Search the foremost position that matches (avoids find of entire math
3978 // inset when match at start of it)
3979 DocIterator old_cur(cur.buffer());
3981 static MatchResult fail = MatchResult();
3982 MatchResult max_match;
3983 // If (prefix_len > 0) means that forwarding 1 position will remove the complete entry
3984 // Happens with e.g. hyperlinks
3985 // either one sees "http://www.bla.bla" or nothing
3986 // so the search for "www" gives prefix_len = 7 (== sizeof("http://")
3987 // and although we search for only 3 chars, we find the whole hyperlink inset
3988 bool at_begin = (expected.match_prefix == 0);
3989 if (!match.opt.forward && match.opt.ignoreformat) {
3990 if (expected.pos > 0)
3993 LASSERT(at_begin, /**/);
3994 if (expected.match_len > 0 && at_begin) {
3995 // Search for deepest match
3997 max_match = expected;
3999 size_t d = cur.depth();
4003 if (cur.depth() < d)
4005 if (cur.depth() == d)
4008 while (cur && cur.depth() > lastd) {
4009 lastd = cur.depth();
4010 mres = match(cur, -1, at_begin);
4011 displayMres(mres, "Checking innermost", cur);
4012 if (mres.match_len > 0)
4017 if (mres.match_len < expected.match_len)
4025 // (expected.match_len <= 0)
4026 mres = match(cur); /* match valid only if not searching whole words */
4027 displayMres(mres, "Start with negative match", cur);
4030 if (max_match.match_len <= 0) return fail;
4031 LYXERR(Debug::FIND, "Ok");
4033 // Compute the match length
4035 if (cur.pos() + len > cur.lastpos())
4038 LASSERT(match.use_regexp, /**/);
4041 int maxl = cur.lastpos() - cur.pos();
4042 // Greedy behaviour while matching regexps
4043 while (maxl > minl) {
4045 mres2 = match(cur, len, at_begin);
4046 displayMres(mres2, "Finalize loop", cur);
4047 int actual_match_len = mres2.match_len;
4048 if (actual_match_len >= max_match.match_len) {
4049 // actual_match_len > max_match _can_ happen,
4050 // if the search area splits
4051 // some following word so that the regex
4052 // (e.g. 'r.*r\b' matches 'r' from the middle of the
4054 // This means, the len value is too big
4055 actual_match_len = max_match.match_len;
4057 max_match.match_len = actual_match_len;
4059 if (maxl - minl < 4)
4060 len = (maxl + minl)/2;
4062 len = minl + (maxl - minl + 3)/4;
4065 // (actual_match_len < max_match.match_len)
4067 len = (maxl + minl)/2;
4072 // Search for real start of matched characters
4074 MatchResult actual_match;
4077 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
4078 if (cur.depth() < old_cur.depth()) {
4080 LYXERR(Debug::INFO, "cur.depth() < old_cur.depth(), this should never happen");
4083 if (cur.pos() != old_cur.pos()) {
4084 // OK, forwarded 1 pos in actual inset
4085 actual_match = match(cur, len-1, at_begin);
4086 if (actual_match.match_len == max_match.match_len) {
4087 // Ha, got it! The shorter selection has the same match length
4090 max_match = actual_match;
4093 // OK, the shorter selection matches less chars, revert to previous value
4099 LYXERR(Debug::INFO, "cur.pos() == old_cur.pos(), this should never happen");
4100 actual_match = match(cur, len, at_begin);
4101 if (actual_match.match_len == max_match.match_len) {
4103 max_match = actual_match;
4110 max_match.pos_len = len;
4111 displayMres(max_match, "SEARCH RESULT", cur)
4118 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
4122 bool repeat = false;
4123 DocIterator orig_cur; // to be used if repeat not successful
4124 MatchResult orig_mres;
4125 while (!theApp()->longOperationCancelled() && cur) {
4126 //(void) findAdvForwardInnermost(cur);
4127 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
4128 MatchResult mres = match(cur, -1, false);
4129 string msg = "Starting";
4132 displayMres(mres, msg + " findForwardAdv", cur)
4133 int match_len = mres.match_len;
4134 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
4135 LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
4138 if (match_len <= 0) {
4139 // This should exit nested insets, if any, or otherwise undefine the currsor.
4140 cur.pos() = cur.lastpos();
4141 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
4144 else { // match_len > 0
4145 // Try to find the begin of searched string
4147 int firstInvalid = cur.lastpos() - cur.pos();
4149 int incrmatch = (mres.match_prefix + mres.pos - mres.leadsize + 1)*3/4;
4150 int incrcur = (firstInvalid + 1 )*3/4;
4151 if (incrcur < incrmatch)
4152 increment = incrcur;
4154 increment = incrmatch;
4158 LYXERR(Debug::FIND, "Set increment to " << increment);
4159 while (increment > 0) {
4160 DocIterator old_cur = cur;
4161 if (cur.pos() + increment >= cur.lastpos()) {
4165 cur.pos() = cur.pos() + increment;
4166 MatchResult mres2 = match(cur, -1, false);
4167 displayMres(mres2, "findForwardAdv loop", cur)
4168 switch (interpretMatch(mres, mres2)) {
4169 case MatchResult::newIsTooFar:
4170 // behind the expected match
4171 firstInvalid = increment;
4175 case MatchResult::newIsBetter:
4176 // not reached yet, but cur.pos()+increment is bettert
4178 firstInvalid -= increment;
4179 if (increment > firstInvalid*3/4)
4180 increment = firstInvalid*3/4;
4181 if ((mres2.pos == mres2.leadsize) && (increment >= mres2.match_prefix)) {
4182 if (increment >= mres2.match_prefix)
4183 increment = (mres2.match_prefix+1)*3/4;
4188 // Handle not like MatchResult::newIsTooFar
4189 LYXERR0( "Probably too far: Increment = " << increment << " match_prefix = " << mres.match_prefix);
4191 increment = increment*3/4;
4196 if (mres.match_len > 0) {
4197 if (mres.match_prefix + mres.pos - mres.leadsize > 0) {
4198 // The match seems to indicate some deeper level
4207 // should never be reached.
4211 // LYXERR0("Leaving first loop");
4212 LYXERR(Debug::FIND, "Finalizing 1");
4213 MatchResult found_match = findAdvFinalize(cur, match, mres);
4214 if (found_match.match_len > 0) {
4215 LASSERT(found_match.pos_len > 0, /**/);
4216 match.FillResults(found_match);
4217 return found_match.pos_len;
4220 // try next possible match
4231 /// Find the most backward consecutive match within same paragraph while searching backwards.
4232 MatchResult findMostBackwards(DocIterator & cur, MatchStringAdv const & match, MatchResult &expected)
4234 DocIterator cur_begin = cur;
4235 cur_begin.pos() = 0;
4236 DocIterator tmp_cur = cur;
4237 MatchResult mr = findAdvFinalize(tmp_cur, match, expected);
4238 Inset & inset = cur.inset();
4239 for (; cur != cur_begin; cur.backwardPos()) {
4240 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
4241 DocIterator new_cur = cur;
4242 new_cur.backwardPos();
4243 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
4245 MatchResult new_mr = findAdvFinalize(new_cur, match, expected);
4246 if (new_mr.match_len == mr.match_len)
4250 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
4256 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
4260 // Backup of original position
4261 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
4262 if (cur == cur_begin)
4265 DocIterator cur_orig(cur);
4266 bool pit_changed = false;
4269 MatchResult found_match = match(cur, -1, false);
4271 if (found_match.match_len > 0) {
4273 cur.pos() = cur.lastpos();
4275 cur.pos() = cur_orig.pos();
4276 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
4277 DocIterator cur_prev_iter;
4279 found_match = match(cur);
4280 LYXERR(Debug::FIND, "findBackAdv3: found_match="
4281 << (found_match.match_len > 0) << ", cur: " << cur);
4282 if (found_match.match_len > 0) {
4283 MatchResult found_mr = findMostBackwards(cur, match, found_match);
4284 if (found_mr.pos_len > 0) {
4285 match.FillResults(found_mr);
4286 return found_mr.pos_len;
4290 // Stop if begin of document reached
4291 if (cur == cur_begin)
4293 cur_prev_iter = cur;
4297 if (cur == cur_begin)
4304 } while (!theApp()->longOperationCancelled());
4312 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
4313 DocIterator const & cur, int len)
4315 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
4317 if (!opt.ignoreformat)
4318 return latexifyFromCursor(cur, len);
4320 return stringifyFromCursor(cur, len);
4324 FindAndReplaceOptions::FindAndReplaceOptions(
4325 docstring const & _find_buf_name, bool _casesensitive,
4326 bool _matchword, bool _forward, bool _expandmacros, bool _ignoreformat,
4327 docstring const & _repl_buf_name, bool _keep_case,
4328 SearchScope _scope, SearchRestriction _restr, bool _replace_all)
4329 : find_buf_name(_find_buf_name), casesensitive(_casesensitive), matchword(_matchword),
4330 forward(_forward), expandmacros(_expandmacros), ignoreformat(_ignoreformat),
4331 repl_buf_name(_repl_buf_name), keep_case(_keep_case), scope(_scope), restr(_restr), replace_all(_replace_all)
4339 /** Check if 'len' letters following cursor are all non-lowercase */
4340 static bool allNonLowercase(Cursor const & cur, int len)
4342 pos_type beg_pos = cur.selectionBegin().pos();
4343 pos_type end_pos = cur.selectionBegin().pos() + len;
4344 if (len > cur.lastpos() + 1 - beg_pos) {
4345 LYXERR(Debug::FIND, "This should not happen, more debug needed");
4346 len = cur.lastpos() + 1 - beg_pos;
4347 end_pos = beg_pos + len;
4349 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
4350 if (isLowerCase(cur.paragraph().getChar(pos)))
4356 /** Check if first letter is upper case and second one is lower case */
4357 static bool firstUppercase(Cursor const & cur)
4360 pos_type pos = cur.selectionBegin().pos();
4361 if (pos >= cur.lastpos() - 1) {
4362 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
4365 ch1 = cur.paragraph().getChar(pos);
4366 ch2 = cur.paragraph().getChar(pos + 1);
4367 bool result = isUpperCase(ch1) && isLowerCase(ch2);
4368 LYXERR(Debug::FIND, "firstUppercase(): "
4369 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
4370 << ch2 << "(" << char(ch2) << ")"
4371 << ", result=" << result << ", cur=" << cur);
4376 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
4378 ** \fixme What to do with possible further paragraphs in replace buffer ?
4380 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
4382 ParagraphList::iterator pit = buffer.paragraphs().begin();
4383 LASSERT(!pit->empty(), /**/);
4384 pos_type right = pos_type(1);
4385 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
4386 right = pit->size();
4387 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
4391 static bool replaceMatches(string &t, int maxmatchnum, vector <string> const & replacements)
4393 // Should replace the string "$" + std::to_string(matchnum) with replacement
4394 // if the char '$' is not prefixed with odd number of char '\\'
4395 static regex const rematch("(\\\\)*(\\$\\$([0-9]))");
4399 for (sregex_iterator it(t.begin(), t.end(), rematch), end; it != end; ++it) {
4401 if ((sub.position(2) - sub.position(0)) % 2 == 1)
4403 int num = stoi(sub.str(3), nullptr, 10);
4404 if (num >= maxmatchnum)
4406 if (lastpos < (size_t) sub.position(2))
4407 s += t.substr(lastpos, sub.position(2) - lastpos);
4408 s += replacements[num];
4409 lastpos = sub.position(2) + sub.length(2);
4413 else if (lastpos < t.length())
4414 s += t.substr(lastpos, t.length() - lastpos);
4420 static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
4422 Cursor & cur = bv->cursor();
4423 if (opt.repl_buf_name.empty()
4424 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
4425 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4428 DocIterator sel_beg = cur.selectionBegin();
4429 DocIterator sel_end = cur.selectionEnd();
4430 if (&sel_beg.inset() != &sel_end.inset()
4431 || sel_beg.pit() != sel_end.pit()
4432 || sel_beg.idx() != sel_end.idx())
4434 int sel_len = sel_end.pos() - sel_beg.pos();
4435 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
4436 << ", sel_len: " << sel_len << endl);
4439 LASSERT(sel_len > 0, return 0);
4441 if (!matchAdv(sel_beg, sel_len).match_len)
4444 // Build a copy of the replace buffer, adapted to the KeepCase option
4445 Buffer const & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
4447 repl_buffer_orig.write(oss);
4448 string lyx = oss.str();
4449 if (matchAdv.valid_matches > 0)
4450 replaceMatches(lyx, matchAdv.valid_matches, matchAdv.matches);
4451 Buffer repl_buffer(string(), false);
4452 repl_buffer.setInternal(true);
4453 repl_buffer.setUnnamed(true);
4454 LASSERT(repl_buffer.readString(lyx), return 0);
4455 if (opt.keep_case && sel_len >= 2) {
4456 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
4457 if (cur.inTexted()) {
4458 if (firstUppercase(cur))
4459 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
4460 else if (allNonLowercase(cur, sel_len))
4461 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
4464 cap::cutSelection(cur, false);
4465 if (cur.inTexted()) {
4466 repl_buffer.changeLanguage(
4467 repl_buffer.language(),
4468 cur.getFont().language());
4469 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
4470 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
4471 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
4472 repl_buffer.params().documentClassPtr(),
4473 repl_buffer.params().authors(),
4474 bv->buffer().errorList("Paste"));
4475 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
4476 sel_len = repl_buffer.paragraphs().begin()->size();
4477 } else if (cur.inMathed()) {
4478 odocstringstream ods;
4480 // OutputParams runparams(&repl_buffer.params().encoding());
4481 OutputParams runparams(encodings.fromLyXName("utf8"));
4482 runparams.nice = false;
4483 runparams.flavor = Flavor::XeTeX;
4484 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
4485 runparams.dryrun = true;
4486 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
4487 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
4488 docstring repl_latex = ods.str();
4489 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
4491 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
4492 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
4493 repl_latex = from_utf8(s);
4494 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
4495 MathData ar(cur.buffer());
4496 asArray(repl_latex, ar, Parse::NORMAL);
4498 sel_len = ar.size();
4499 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4501 if (cur.pos() >= sel_len)
4502 cur.pos() -= sel_len;
4505 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
4506 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
4507 bv->processUpdateFlags(Update::Force);
4512 /// Perform a FindAdv operation.
4513 bool findAdv(BufferView * bv, FindAndReplaceOptions & opt)
4518 // e.g., when invoking word-findadv from mini-buffer wither with
4519 // wrong options syntax or before ever opening advanced F&R pane
4520 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
4524 MatchStringAdv matchAdv(bv->buffer(), opt);
4526 if (!matchAdv.regexIsValid) {
4527 bv->message(lyx::from_utf8(matchAdv.regexError));
4531 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
4533 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
4534 num_replaced += findAdvReplace(bv, opt, matchAdv);
4537 pos_len = findForwardAdv(cur, matchAdv);
4539 pos_len = findBackwardsAdv(cur, matchAdv);
4540 } catch (exception & ex) {
4541 bv->message(from_utf8(ex.what()));
4546 if (num_replaced > 0) {
4547 switch (num_replaced)
4550 bv->message(_("One match has been replaced."));
4553 bv->message(_("Two matches have been replaced."));
4556 bv->message(bformat(_("%1$d matches have been replaced."), num_replaced));
4562 bv->message(_("Match not found."));
4567 if (num_replaced > 0)
4568 bv->message(_("Match has been replaced."));
4570 bv->message(_("Match found."));
4572 if (cur.pos() + pos_len > cur.lastpos()) {
4573 // Prevent crash in bv->putSelectionAt()
4574 // Should never happen, maybe LASSERT() here?
4575 pos_len = cur.lastpos() - cur.pos();
4577 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << pos_len);
4578 bv->putSelectionAt(cur, pos_len, !opt.forward);
4584 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
4586 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
4587 << opt.casesensitive << ' '
4588 << opt.matchword << ' '
4589 << opt.forward << ' '
4590 << opt.expandmacros << ' '
4591 << opt.ignoreformat << ' '
4592 << opt.replace_all << ' '
4593 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
4594 << opt.keep_case << ' '
4595 << int(opt.scope) << ' '
4598 LYXERR(Debug::FIND, "built: " << os.str());
4604 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
4606 // LYXERR(Debug::FIND, "parsing");
4610 while (line != "EOSS") {
4614 if (is.eof()) // Tolerate malformed request
4618 // LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
4619 opt.find_buf_name = from_utf8(s);
4620 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.replace_all;
4621 is.get(); // Waste space before replace string
4624 while (line != "EOSS") {
4628 if (is.eof()) // Tolerate malformed request
4632 // LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
4633 opt.repl_buf_name = from_utf8(s);
4634 is >> opt.keep_case;
4637 opt.scope = FindAndReplaceOptions::SearchScope(i);
4639 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
4642 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
4643 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
4644 << opt.scope << ' ' << opt.restr);