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"
35 #include "frontends/Application.h"
36 #include "frontends/alert.h"
38 #include "mathed/InsetMath.h"
39 #include "mathed/InsetMathHull.h"
40 #include "mathed/MathData.h"
41 #include "mathed/MathStream.h"
42 #include "mathed/MathSupport.h"
44 #include "support/debug.h"
45 #include "support/docstream.h"
46 #include "support/FileName.h"
47 #include "support/gettext.h"
48 #include "support/lassert.h"
49 #include "support/lstrings.h"
50 #include "support/textutils.h"
54 #include <QtCore> // sets QT_VERSION
55 #if (QT_VERSION >= 0x050000)
56 #include <QRegularExpression>
60 using namespace lyx::support;
65 // Helper class for deciding what should be ignored
69 IgnoreFormats() = default;
71 bool getFamily() const { return ignoreFamily_; }
73 bool getSeries() const { return ignoreSeries_; }
75 bool getShape() const { return ignoreShape_; }
77 bool getUnderline() const { return ignoreUnderline_; }
79 bool getMarkUp() const { return ignoreMarkUp_; }
81 bool getStrikeOut() const { return ignoreStrikeOut_; }
83 bool getSectioning() const { return ignoreSectioning_; }
85 bool getFrontMatter() const { return ignoreFrontMatter_; }
87 bool getColor() const { return ignoreColor_; }
89 bool getLanguage() const { return ignoreLanguage_; }
91 void setIgnoreFormat(string const & type, bool value);
95 bool ignoreFamily_ = false;
97 bool ignoreSeries_ = false;
99 bool ignoreShape_ = false;
101 bool ignoreUnderline_ = false;
103 bool ignoreMarkUp_ = false;
105 bool ignoreStrikeOut_ = false;
107 bool ignoreSectioning_ = false;
109 bool ignoreFrontMatter_ = false;
111 bool ignoreColor_ = false;
113 bool ignoreLanguage_ = false;
117 void IgnoreFormats::setIgnoreFormat(string const & type, bool value)
119 if (type == "color") {
120 ignoreColor_ = value;
122 else if (type == "language") {
123 ignoreLanguage_ = value;
125 else if (type == "sectioning") {
126 ignoreSectioning_ = value;
127 ignoreFrontMatter_ = value;
129 else if (type == "font") {
130 ignoreSeries_ = value;
131 ignoreShape_ = value;
132 ignoreFamily_ = value;
134 else if (type == "series") {
135 ignoreSeries_ = value;
137 else if (type == "shape") {
138 ignoreShape_ = value;
140 else if (type == "family") {
141 ignoreFamily_ = value;
143 else if (type == "markup") {
144 ignoreMarkUp_ = value;
146 else if (type == "underline") {
147 ignoreUnderline_ = value;
149 else if (type == "strike") {
150 ignoreStrikeOut_ = value;
154 // The global variable that can be changed from outside
155 IgnoreFormats ignoreFormats;
158 void setIgnoreFormat(string const & type, bool value)
160 ignoreFormats.setIgnoreFormat(type, value);
166 bool parse_bool(docstring & howto)
171 howto = split(howto, var, ' ');
179 MatchString(docstring const & s, bool cs, bool mw)
180 : str(s), case_sens(cs), whole_words(mw)
183 // returns true if the specified string is at the specified position
184 // del specifies whether deleted strings in ct mode will be considered
185 int operator()(Paragraph const & par, pos_type pos, bool del = true) const
187 return par.find(str, case_sens, whole_words, pos, del);
195 // match whole words only
200 int findForward(DocIterator & cur, MatchString const & match,
201 bool find_del = true)
203 for (; cur; cur.forwardChar())
204 if (cur.inTexted()) {
205 int len = match(cur.paragraph(), cur.pos(), find_del);
213 int findBackwards(DocIterator & cur, MatchString const & match,
214 bool find_del = true)
218 if (cur.inTexted()) {
219 int len = match(cur.paragraph(), cur.pos(), find_del);
228 bool searchAllowed(docstring const & str)
231 frontend::Alert::error(_("Search error"), _("Search string is empty"));
238 bool findOne(BufferView * bv, docstring const & searchstr,
239 bool case_sens, bool whole, bool forward,
240 bool find_del = true, bool check_wrap = false)
242 if (!searchAllowed(searchstr))
245 DocIterator cur = forward
246 ? bv->cursor().selectionEnd()
247 : bv->cursor().selectionBegin();
249 MatchString const match(searchstr, case_sens, whole);
251 int match_len = forward
252 ? findForward(cur, match, find_del)
253 : findBackwards(cur, match, find_del);
256 bv->putSelectionAt(cur, match_len, !forward);
257 else if (check_wrap) {
258 DocIterator cur_orig(bv->cursor());
261 q = _("End of file reached while searching forward.\n"
262 "Continue searching from the beginning?");
264 q = _("Beginning of file reached while searching backward.\n"
265 "Continue searching from the end?");
266 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
267 q, 0, 1, _("&Yes"), _("&No"));
268 if (wrap_answer == 0) {
270 bv->cursor().clear();
271 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
273 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
274 bv->cursor().backwardPos();
276 bv->clearSelection();
277 if (findOne(bv, searchstr, case_sens, whole, forward, find_del, false))
280 bv->cursor().setCursor(cur_orig);
284 return match_len > 0;
288 int replaceAll(BufferView * bv,
289 docstring const & searchstr, docstring const & replacestr,
290 bool case_sens, bool whole)
292 Buffer & buf = bv->buffer();
294 if (!searchAllowed(searchstr) || buf.isReadonly())
297 DocIterator cur_orig(bv->cursor());
299 MatchString const match(searchstr, case_sens, whole);
302 int const rsize = replacestr.size();
303 int const ssize = searchstr.size();
306 cur.setCursor(doc_iterator_begin(&buf));
307 int match_len = findForward(cur, match, false);
308 while (match_len > 0) {
309 // Backup current cursor position and font.
310 pos_type const pos = cur.pos();
311 Font const font = cur.paragraph().getFontSettings(buf.params(), pos);
313 int striked = ssize -
314 cur.paragraph().eraseChars(pos, pos + match_len,
315 buf.params().track_changes);
316 cur.paragraph().insert(pos, replacestr, font,
317 Change(buf.params().track_changes
319 : Change::UNCHANGED));
320 for (int i = 0; i < rsize + striked; ++i)
323 match_len = findForward(cur, match, false);
326 bv->putSelectionAt(doc_iterator_begin(&buf), 0, false);
328 cur_orig.fixIfBroken();
329 bv->setCursor(cur_orig);
335 // the idea here is that we are going to replace the string that
336 // is selected IF it is the search string.
337 // if there is a selection, but it is not the search string, then
338 // we basically ignore it. (FIXME We ought to replace only within
340 // if there is no selection, then:
341 // (i) if some search string has been provided, then we find it.
342 // (think of how the dialog works when you hit "replace" the
344 // (ii) if no search string has been provided, then we treat the
345 // word the cursor is in as the search string. (why? i have no
346 // idea.) but this only works in text?
348 // returns the number of replacements made (one, if any) and
349 // whether anything at all was done.
350 pair<bool, int> replaceOne(BufferView * bv, docstring searchstr,
351 docstring const & replacestr, bool case_sens,
352 bool whole, bool forward, bool findnext)
354 Cursor & cur = bv->cursor();
355 if (!cur.selection()) {
356 // no selection, non-empty search string: find it
357 if (!searchstr.empty()) {
358 bool const found = findOne(bv, searchstr, case_sens, whole, forward, true, findnext);
359 return make_pair(found, 0);
361 // empty search string
364 return make_pair(false, 0);
365 // select current word and treat it as the search string.
366 // This causes a minor bug as undo will restore this selection,
367 // which the user did not create (#8986).
368 cur.innerText()->selectWord(cur, WHOLE_WORD);
369 searchstr = cur.selectionAsString(false);
372 // if we still don't have a search string, report the error
374 if (!searchAllowed(searchstr))
375 return make_pair(false, 0);
377 bool have_selection = cur.selection();
378 docstring const selected = cur.selectionAsString(false);
381 ? searchstr == selected
382 : compare_no_case(searchstr, selected) == 0;
384 // no selection or current selection is not search word:
385 // just find the search word
386 if (!have_selection || !match) {
387 bool const found = findOne(bv, searchstr, case_sens, whole, forward, true, findnext);
388 return make_pair(found, 0);
391 // we're now actually ready to replace. if the buffer is
392 // read-only, we can't, though.
393 if (bv->buffer().isReadonly())
394 return make_pair(false, 0);
396 cap::replaceSelectionWithString(cur, replacestr);
398 cur.pos() += replacestr.length();
399 LASSERT(cur.pos() <= cur.lastpos(),
400 cur.pos() = cur.lastpos());
403 findOne(bv, searchstr, case_sens, whole, forward, false, findnext);
405 return make_pair(true, 1);
411 docstring const find2string(docstring const & search,
412 bool casesensitive, bool matchword, bool forward)
416 << int(casesensitive) << ' '
417 << int(matchword) << ' '
423 docstring const replace2string(docstring const & replace,
424 docstring const & search,
425 bool casesensitive, bool matchword,
426 bool all, bool forward, bool findnext)
429 ss << replace << '\n'
431 << int(casesensitive) << ' '
432 << int(matchword) << ' '
434 << int(forward) << ' '
440 bool lyxfind(BufferView * bv, FuncRequest const & ev)
442 if (!bv || ev.action() != LFUN_WORD_FIND)
445 //lyxerr << "find called, cmd: " << ev << endl;
447 // data is of the form
449 // <casesensitive> <matchword> <forward>"
451 docstring howto = split(ev.argument(), search, '\n');
453 bool casesensitive = parse_bool(howto);
454 bool matchword = parse_bool(howto);
455 bool forward = parse_bool(howto);
457 return findOne(bv, search, casesensitive, matchword, forward, true, true);
461 bool lyxreplace(BufferView * bv,
462 FuncRequest const & ev, bool has_deleted)
464 if (!bv || ev.action() != LFUN_WORD_REPLACE)
467 // data is of the form
470 // <casesensitive> <matchword> <all> <forward> <findnext>"
473 docstring howto = split(ev.argument(), rplc, '\n');
474 howto = split(howto, search, '\n');
476 bool casesensitive = parse_bool(howto);
477 bool matchword = parse_bool(howto);
478 bool all = parse_bool(howto);
479 bool forward = parse_bool(howto);
480 bool findnext = howto.empty() ? true : parse_bool(howto);
485 int replace_count = 0;
487 replace_count = replaceAll(bv, search, rplc, casesensitive, matchword);
488 update = replace_count > 0;
491 replaceOne(bv, search, rplc, casesensitive, matchword, forward, findnext);
493 replace_count = rv.second;
496 Buffer const & buf = bv->buffer();
498 // emit message signal.
499 buf.message(_("String not found."));
501 if (replace_count == 0) {
502 buf.message(_("String found."));
503 } else if (replace_count == 1) {
504 buf.message(_("String has been replaced."));
506 docstring const str =
507 bformat(_("%1$d strings have been replaced."), replace_count);
511 } else if (findnext) {
512 // if we have deleted characters, we do not replace at all, but
513 // rather search for the next occurence
514 if (findOne(bv, search, casesensitive, matchword, forward, true, findnext))
517 bv->message(_("String not found."));
523 bool findNextChange(BufferView * bv, Cursor & cur, bool const check_wrap)
525 for (; cur; cur.forwardPos())
526 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
530 DocIterator cur_orig(bv->cursor());
531 docstring q = _("End of file reached while searching forward.\n"
532 "Continue searching from the beginning?");
533 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
534 q, 0, 1, _("&Yes"), _("&No"));
535 if (wrap_answer == 0) {
536 bv->cursor().clear();
537 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
538 bv->clearSelection();
539 cur.setCursor(bv->cursor().selectionBegin());
540 if (findNextChange(bv, cur, false))
543 bv->cursor().setCursor(cur_orig);
550 bool findPreviousChange(BufferView * bv, Cursor & cur, bool const check_wrap)
552 for (cur.backwardPos(); cur; cur.backwardPos()) {
553 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
558 DocIterator cur_orig(bv->cursor());
559 docstring q = _("Beginning of file reached while searching backward.\n"
560 "Continue searching from the end?");
561 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
562 q, 0, 1, _("&Yes"), _("&No"));
563 if (wrap_answer == 0) {
564 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
565 bv->cursor().backwardPos();
566 bv->clearSelection();
567 cur.setCursor(bv->cursor().selectionBegin());
568 if (findPreviousChange(bv, cur, false))
571 bv->cursor().setCursor(cur_orig);
578 bool selectChange(Cursor & cur, bool forward)
580 if (!cur.inTexted() || !cur.paragraph().isChanged(cur.pos()))
582 Change ch = cur.paragraph().lookupChange(cur.pos());
584 CursorSlice tip1 = cur.top();
585 for (; tip1.pit() < tip1.lastpit() || tip1.pos() < tip1.lastpos(); tip1.forwardPos()) {
586 Change ch2 = tip1.paragraph().lookupChange(tip1.pos());
587 if (!ch2.isSimilarTo(ch))
590 CursorSlice tip2 = cur.top();
591 for (; tip2.pit() > 0 || tip2.pos() > 0;) {
593 Change ch2 = tip2.paragraph().lookupChange(tip2.pos());
594 if (!ch2.isSimilarTo(ch)) {
595 // take a step forward to correctly set the selection
603 cur.bv().mouseSetCursor(cur, false);
605 cur.bv().mouseSetCursor(cur, true);
613 bool findChange(BufferView * bv, bool forward)
616 cur.setCursor(forward ? bv->cursor().selectionEnd()
617 : bv->cursor().selectionBegin());
618 forward ? findNextChange(bv, cur, true) : findPreviousChange(bv, cur, true);
619 return selectChange(cur, forward);
624 bool findNextChange(BufferView * bv)
626 return findChange(bv, true);
630 bool findPreviousChange(BufferView * bv)
632 return findChange(bv, false);
639 typedef vector<pair<string, string> > Escapes;
641 /// A map of symbols and their escaped equivalent needed within a regex.
642 /// @note Beware of order
643 Escapes const & get_regexp_escapes()
645 typedef std::pair<std::string, std::string> P;
647 static Escapes escape_map;
648 if (escape_map.empty()) {
649 escape_map.push_back(P("$", "_x_$"));
650 escape_map.push_back(P("{", "_x_{"));
651 escape_map.push_back(P("}", "_x_}"));
652 escape_map.push_back(P("[", "_x_["));
653 escape_map.push_back(P("]", "_x_]"));
654 escape_map.push_back(P("(", "_x_("));
655 escape_map.push_back(P(")", "_x_)"));
656 escape_map.push_back(P("+", "_x_+"));
657 escape_map.push_back(P("*", "_x_*"));
658 escape_map.push_back(P(".", "_x_."));
659 escape_map.push_back(P("\\", "(?:\\\\|\\\\backslash)"));
660 escape_map.push_back(P("~", "(?:\\\\textasciitilde|\\\\sim)"));
661 escape_map.push_back(P("^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\textasciicircum|\\\\mathcircumflex)"));
662 escape_map.push_back(P("_x_", "\\"));
667 /// A map of lyx escaped strings and their unescaped equivalent.
668 Escapes const & get_lyx_unescapes()
670 typedef std::pair<std::string, std::string> P;
672 static Escapes escape_map;
673 if (escape_map.empty()) {
674 escape_map.push_back(P("\\%", "%"));
675 escape_map.push_back(P("\\{", "{"));
676 escape_map.push_back(P("\\}", "}"));
677 escape_map.push_back(P("\\mathcircumflex ", "^"));
678 escape_map.push_back(P("\\mathcircumflex", "^"));
679 escape_map.push_back(P("\\backslash ", "\\"));
680 escape_map.push_back(P("\\backslash", "\\"));
681 escape_map.push_back(P("\\sim ", "~"));
682 escape_map.push_back(P("\\sim", "~"));
687 /// A map of escapes turning a regexp matching text to one matching latex.
688 Escapes const & get_regexp_latex_escapes()
690 typedef std::pair<std::string, std::string> P;
692 static Escapes escape_map;
693 if (escape_map.empty()) {
694 escape_map.push_back(P("\\\\", "(?:\\\\\\\\|\\\\backslash|\\\\textbackslash\\{\\}|\\\\textbackslash)"));
695 escape_map.push_back(P("(<?!\\\\\\\\textbackslash)\\{", "\\\\\\{"));
696 escape_map.push_back(P("(<?!\\\\\\\\textbackslash\\\\\\{)\\}", "\\\\\\}"));
697 escape_map.push_back(P("\\[", "\\{\\[\\}"));
698 escape_map.push_back(P("\\]", "\\{\\]\\}"));
699 escape_map.push_back(P("\\^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\textasciicircum|\\\\mathcircumflex)"));
700 escape_map.push_back(P("%", "\\\\\\%"));
701 escape_map.push_back(P("#", "\\\\#"));
706 /** @todo Probably the maps need to be migrated to regexps, in order to distinguish if
707 ** the found occurrence were escaped.
709 string apply_escapes(string s, Escapes const & escape_map)
711 LYXERR(Debug::FIND, "Escaping: '" << s << "'");
712 Escapes::const_iterator it;
713 for (it = escape_map.begin(); it != escape_map.end(); ++it) {
714 // LYXERR(Debug::FIND, "Escaping " << it->first << " as " << it->second);
715 unsigned int pos = 0;
716 while (pos < s.length() && (pos = s.find(it->first, pos)) < s.length()) {
717 s.replace(pos, it->first.length(), it->second);
718 LYXERR(Debug::FIND, "After escape: " << s);
719 pos += it->second.length();
720 // LYXERR(Debug::FIND, "pos: " << pos);
723 LYXERR(Debug::FIND, "Escaped : '" << s << "'");
728 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
729 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
730 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
731 string escape_for_regex(string s, bool match_latex)
734 while (pos < s.size()) {
735 size_t new_pos = s.find("\\regexp{", pos);
736 if (new_pos == string::npos)
740 LYXERR(Debug::FIND, "new_pos: " << new_pos);
741 t = apply_escapes(s.substr(pos, new_pos - pos), get_lyx_unescapes());
742 LYXERR(Debug::FIND, "t [lyx]: " << t);
743 t = apply_escapes(t, get_regexp_escapes());
744 LYXERR(Debug::FIND, "t [rxp]: " << t);
745 s.replace(pos, new_pos - pos, t);
746 new_pos = pos + t.size();
747 LYXERR(Debug::FIND, "Regexp after escaping: " << s);
748 LYXERR(Debug::FIND, "new_pos: " << new_pos);
749 if (new_pos == s.size())
752 // Might fail if \\endregexp{} is preceeded by unexpected stuff (weird escapes)
753 size_t end_pos = s.find("\\endregexp{}}", new_pos + 8);
754 LYXERR(Debug::FIND, "end_pos: " << end_pos);
755 t = s.substr(new_pos + 8, end_pos - (new_pos + 8));
756 LYXERR(Debug::FIND, "t in regexp : " << t);
757 t = apply_escapes(t, get_lyx_unescapes());
758 LYXERR(Debug::FIND, "t in regexp after unescapes [lyx]: " << t);
760 t = apply_escapes(t, get_regexp_latex_escapes());
761 LYXERR(Debug::FIND, "t in regexp after latex_escapes [ltx]: " << t);
763 if (end_pos == s.size()) {
764 s.replace(new_pos, end_pos - new_pos, t);
765 LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s);
768 s.replace(new_pos, end_pos + 13 - new_pos, t);
769 LYXERR(Debug::FIND, "Regexp after \\regexp{...\\endregexp{}} removal: " << s);
770 pos = new_pos + t.size();
771 LYXERR(Debug::FIND, "pos: " << pos);
777 /// Wrapper for lyx::regex_replace with simpler interface
778 bool regex_replace(string const & s, string & t, string const & searchstr,
779 string const & replacestr)
781 regex e(searchstr, regex_constants::ECMAScript);
783 ostream_iterator<char, char> it(oss);
784 regex_replace(it, s.begin(), s.end(), e, replacestr);
785 // tolerate t and s be references to the same variable
786 bool rv = (s != oss.str());
792 /** Checks if supplied string segment is well-formed from the standpoint of matching open-closed braces.
794 ** Verify that closed braces exactly match open braces. This avoids that, for example,
795 ** \frac{.*}{x} matches \frac{x+\frac{y}{x}}{z} with .* being 'x+\frac{y'.
798 ** Number of open braces that must remain open at the end for the verification to succeed.
800 #if (QT_VERSION >= 0x050000)
801 bool braces_match(QString const & beg,
804 bool braces_match(string const & beg,
809 #if (QT_VERSION >= 0x050000)
810 LYXERR(Debug::FIND, "Checking " << unmatched << " unmatched braces in '" << beg.toStdString() << "'");
812 LYXERR(Debug::FIND, "Checking " << unmatched << " unmatched braces in '" << beg << "'");
814 int lastidx = beg.size();
815 for (int i=0; i < lastidx; ++i) {
816 // Skip escaped braces in the count
817 #if (QT_VERSION >= 0x050000)
826 } else if (c == '{') {
828 } else if (c == '}') {
829 if (open_pars == 0) {
830 LYXERR(Debug::FIND, "Found unmatched closed brace");
836 if (open_pars != unmatched) {
837 LYXERR(Debug::FIND, "Found " << open_pars
838 << " instead of " << unmatched
839 << " unmatched open braces at the end of count");
842 LYXERR(Debug::FIND, "Braces match as expected");
852 MatchResult(): match_len(0),match2end(0), pos(0) {};
855 /** The class performing a match between a position in the document and the FindAdvOptions.
858 class MatchStringAdv {
860 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt);
862 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
863 ** constructor as opt.search, under the opt.* options settings.
866 ** If set, then match is searched only against beginning of text starting at cur.
867 ** If unset, then match is searched anywhere in text starting at cur.
870 ** The length of the matching text, or zero if no match was found.
872 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
873 #if (QT_VERSION >= 0x050000)
881 /// first buffer on which search was started
882 lyx::Buffer * const p_first_buf;
884 FindAndReplaceOptions const & opt;
887 /// Auxiliary find method (does not account for opt.matchword)
888 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
890 /** Normalize a stringified or latexified LyX paragraph.
894 ** <li>if search is not casesensitive, then lowercase the string;
895 ** <li>remove any newline at begin or end of the string;
896 ** <li>replace any newline in the middle of the string with a simple space;
897 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
900 ** @todo Normalization should also expand macros, if the corresponding
901 ** search option was checked.
903 string normalize(docstring const & s, bool hack_braces) const;
904 // normalized string to search
905 string par_as_string;
906 // regular expression to use for searching
907 // regexp2 is same as regexp, but prefixed with a ".*?"
908 #if (QT_VERSION >= 0x050000)
909 QRegularExpression regexp;
910 QRegularExpression regexp2;
915 // leading format material as string
916 string lead_as_string;
917 // par_as_string after removal of lead_as_string
918 string par_as_string_nolead;
919 // unmatched open braces in the search string/regexp
921 // number of (.*?) subexpressions added at end of search regexp for closing
922 // environments, math mode, styles, etc...
924 // Are we searching with regular expressions ?
929 static docstring buffer_to_latex(Buffer & buffer)
931 //OutputParams runparams(&buffer.params().encoding());
932 OutputParams runparams(encodings.fromLyXName("utf8"));
933 odocstringstream ods;
935 runparams.nice = true;
936 runparams.flavor = Flavor::XeTeX;
937 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
938 // No side effect of file copying and image conversion
939 runparams.dryrun = true;
940 runparams.for_search = true;
941 pit_type const endpit = buffer.paragraphs().size();
942 for (pit_type pit = 0; pit != endpit; ++pit) {
943 TeXOnePar(buffer, buffer.text(), pit, os, runparams);
944 LYXERR(Debug::FIND, "searchString up to here: " << ods.str());
950 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
953 if (!opt.ignoreformat) {
954 str = buffer_to_latex(buffer);
956 // OutputParams runparams(&buffer.params().encoding());
957 OutputParams runparams(encodings.fromLyXName("utf8"));
958 runparams.nice = true;
959 runparams.flavor = Flavor::XeTeX;
960 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
961 runparams.dryrun = true;
962 runparams.for_search = true;
963 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
964 Paragraph const & par = buffer.paragraphs().at(pit);
965 LYXERR(Debug::FIND, "Adding to search string: '"
966 << par.asString(pos_type(0), par.size(),
967 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
970 str += par.asString(pos_type(0), par.size(),
971 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
974 // Even in ignore-format we have to remove "\text{}, \lyxmathsym{}" parts
975 string t = to_utf8(str);
976 while (regex_replace(t, t, "\\\\(text|lyxmathsym)\\{([^\\}]*)\\}", "$2"));
983 /// Return separation pos between the leading material and the rest
984 static size_t identifyLeading(string const & s)
987 // @TODO Support \item[text]
988 // Kornel: Added textsl, textsf, textit, texttt and noun
989 // + allow to search for colored text too
990 while (regex_replace(t, t, "^\\\\(("
991 "(author|title|subtitle|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|"
992 "lyxaddress|lyxrightaddress|"
993 "footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
994 "emph|noun|minisec|text(bf|md|sl|sf|it|tt))|"
995 "((textcolor|foreignlanguage|latexenvironment)\\{[a-z]+\\*?\\})|"
996 "(u|uu)line|(s|x)out|uwave)|((sub)?(((sub)?section)|paragraph)|part|chapter)\\*?)\\{", "")
997 || regex_replace(t, t, "^\\$", "")
998 || regex_replace(t, t, "^\\\\\\[", "")
999 || regex_replace(t, t, "^ ?\\\\item\\{[a-z]+\\}", "")
1000 || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\}", ""))
1002 LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
1007 * Given a latexified string, retrieve some handled features
1008 * The features of the regex will later be compared with the features
1009 * of the searched text. If the regex features are not a
1010 * subset of the analized, then, in not format ignoring search
1011 * we can early stop the search in the relevant inset.
1013 typedef map<string, bool> Features;
1015 static Features identifyFeatures(string const & s)
1017 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
1018 static regex const valid("^("
1020 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
1021 "emph|noun|text(bf|md|sl|sf|it|tt)|"
1022 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
1023 "(u|uu)line|(s|x)out|uwave|"
1024 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
1025 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
1030 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1033 if (sub.str(1).compare("regexp") == 0) {
1037 string token = sub.str(1);
1039 if (regex_match(token, sub2, valid)) {
1047 if (sub.str(1).compare("endregexp") == 0) {
1057 * defines values features of a key "\\[a-z]+{"
1062 /* Char type with content discarded
1063 * like \hspace{1cm} */
1065 /* Char, like \backslash */
1067 /* remove starting backslash */
1069 /* \part, \section*, ... */
1071 /* title, author etc */
1073 /* \foreignlanguage{ngerman}, ... */
1076 * to discard language in content */
1079 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1081 /* fonts, colors, markups, ... */
1083 /* footnotesize, ... large, ...
1084 * Ignore all of them */
1087 /* inputencoding, ...
1088 * Discard also content, because they do not help in search */
1091 * like remove, but also all arguments */
1093 /* item, listitem */
1095 /* tex, latex, ... like isChar */
1097 /* like \lettrine[lines=5]{}{} */
1099 // like isStandard, but always remove head
1101 /* End of arguments marker for lettrine,
1102 * so that they can be ignored */
1105 KeyInfo() = default;
1106 KeyInfo(KeyType type, int parcount, bool disable)
1108 parenthesiscount(parcount),
1109 disabled(disable) {}
1110 KeyType keytype = invalid;
1112 int _tokensize = -1;
1113 int _tokenstart = -1;
1114 int _dataStart = -1;
1116 int parenthesiscount = 1;
1117 bool disabled = false;
1118 bool used = false; /* by pattern */
1123 Border(int l=0, int u=0) : low(l), upper(u) {};
1128 #define MAXOPENED 30
1130 bool isPatternString_;
1132 explicit Intervall(bool isPattern, string const & p) :
1133 isPatternString_(isPattern), par(p), ignoreidx(-1), actualdeptindex(0),
1134 hasTitle(false), langcount(0)
1142 static vector<Border> borders;
1143 int depts[MAXOPENED];
1144 int closes[MAXOPENED];
1145 int actualdeptindex;
1146 int previousNotIgnored(int) const;
1147 int nextNotIgnored(int) const;
1148 void handleOpenP(int i);
1149 void handleCloseP(int i, bool closingAllowed);
1150 void resetOpenedP(int openPos);
1151 void addIntervall(int upper);
1152 void addIntervall(int low, int upper); /* if explicit */
1153 void removeAccents();
1154 void setForDefaultLang(KeyInfo const & defLang) const;
1155 int findclosing(int start, int end, char up, char down, int repeat);
1156 void handleParentheses(int lastpos, bool closingAllowed);
1158 int langcount; // Number of disabled language specs up to current position in actual interval
1159 int isOpeningPar(int pos) const;
1161 void output(ostringstream &os, int lastpos);
1162 // string show(int lastpos);
1165 vector<Border> Intervall::borders = vector<Border>(30);
1167 int Intervall::isOpeningPar(int pos) const
1169 if ((pos < 0) || (size_t(pos) >= par.size()))
1171 if (par[pos] != '{')
1173 if (size_t(pos) + 2 >= par.size())
1175 if (par[pos+2] != '}')
1177 if (par[pos+1] == '[' || par[pos+1] == ']')
1182 void Intervall::setForDefaultLang(KeyInfo const & defLang) const
1184 // Enable the use of first token again
1185 if (ignoreidx >= 0) {
1186 int value = defLang._tokenstart + defLang._tokensize;
1192 if (borders[borderidx].low < value)
1193 borders[borderidx].low = value;
1194 if (borders[borderidx].upper < value)
1195 borders[borderidx].upper = value;
1200 static void checkDepthIndex(int val)
1202 static int maxdepthidx = MAXOPENED-2;
1203 static int lastmaxdepth = 0;
1204 if (val > lastmaxdepth) {
1205 LYXERR(Debug::INFO, "Depth reached " << val);
1208 if (val > maxdepthidx) {
1210 LYXERR(Debug::INFO, "maxdepthidx now " << val);
1215 // Not needed, because borders are now dynamically expanded
1216 static void checkIgnoreIdx(int val)
1218 static int lastmaxignore = -1;
1219 if ((lastmaxignore < val) && (size_t(val+1) >= borders.size())) {
1220 LYXERR(Debug::INFO, "IgnoreIdx reached " << val);
1221 lastmaxignore = val;
1227 * Expand the region of ignored parts of the input latex string
1228 * The region is only relevant in output()
1230 void Intervall::addIntervall(int low, int upper)
1233 if (low == upper) return;
1234 for (idx = ignoreidx+1; idx > 0; --idx) {
1235 if (low > borders[idx-1].upper) {
1239 Border br(low, upper);
1240 if (idx > ignoreidx) {
1241 if (borders.size() <= size_t(idx)) {
1242 borders.push_back(br);
1248 // checkIgnoreIdx(ignoreidx);
1252 // Expand only if one of the new bound is inside the interwall
1253 // We know here that br.low > borders[idx-1].upper
1254 if (br.upper < borders[idx].low) {
1255 // We have to insert at this pos
1256 if (size_t(ignoreidx+1) >= borders.size()) {
1257 borders.push_back(borders[ignoreidx]);
1260 borders[ignoreidx+1] = borders[ignoreidx];
1262 for (int i = ignoreidx; i > idx; --i) {
1263 borders[i] = borders[i-1];
1267 // checkIgnoreIdx(ignoreidx);
1270 // Here we know, that we are overlapping
1271 if (br.low > borders[idx].low)
1272 br.low = borders[idx].low;
1273 // check what has to be concatenated
1275 for (int i = idx; i <= ignoreidx; i++) {
1276 if (br.upper >= borders[i].low) {
1278 if (br.upper < borders[i].upper)
1279 br.upper = borders[i].upper;
1285 // count should be >= 1 here
1288 for (int i = idx + count; i <= ignoreidx; i++) {
1289 borders[i-count+1] = borders[i];
1291 ignoreidx -= count - 1;
1297 typedef map<string, string> AccentsMap;
1298 static AccentsMap accents = map<string, string>();
1300 static void buildaccent(string n, string param, string values)
1304 const char delim = '|';
1305 while (getline(s, name, delim)) {
1307 for (char c : param) {
1308 string key = name + "{" + c + "}";
1309 // get the corresponding utf8-value
1310 if ((values[start] & 0xc0) != 0xc0) {
1311 // should not happen, utf8 encoding starts at least with 11xxxxxx
1312 // but value for '\dot{i}' is 'i', which is ascii
1313 if ((values[start] & 0x80) == 0) {
1315 accents[key] = values.substr(start, 1);
1316 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1321 for (int j = 1; ;j++) {
1322 if (start + j >= values.size()) {
1323 accents[key] = values.substr(start, j);
1324 start = values.size() - 1;
1327 else if ((values[start+j] & 0xc0) != 0x80) {
1328 // This is the first byte of following utf8 char
1329 accents[key] = values.substr(start, j);
1331 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1340 static string getutf8(unsigned uchar)
1343 string ret = string();
1345 if (uchar <= 0x7f) {
1346 c[maxc-1] = uchar & 0x7f;
1349 unsigned char rest = 0x40;
1350 unsigned char first = 0x80;
1352 for (int i = start; i >=0; --i) {
1354 c[i] = first + uchar;
1357 c[i] = 0x80 | (uchar & 0x3f);
1364 for (int i = 0; i < maxc; i++) {
1365 if (c[i] == 0) continue;
1371 static void buildAccentsMap()
1373 accents["imath"] = "ı";
1375 accents["jmath"] = "ȷ";
1376 accents["cdot"] = "·";
1377 accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
1378 accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
1379 accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
1380 accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
1381 accents["negmedspace"] = getutf8(0xf0005);
1382 accents["thickspace"] = getutf8(0xf0006);
1383 accents["negthickspace"] = getutf8(0xf0007);
1384 accents["lyx"] = getutf8(0xf0010); // Used logos
1385 accents["LyX"] = getutf8(0xf0010);
1386 accents["tex"] = getutf8(0xf0011);
1387 accents["TeX"] = getutf8(0xf0011);
1388 accents["latex"] = getutf8(0xf0012);
1389 accents["LaTeX"] = getutf8(0xf0012);
1390 accents["latexe"] = getutf8(0xf0013);
1391 accents["LaTeXe"] = getutf8(0xf0013);
1392 accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash
1393 accents["backslash LyX"] = getutf8(0xf0010);
1394 accents["backslash tex"] = getutf8(0xf0011);
1395 accents["backslash TeX"] = getutf8(0xf0011);
1396 accents["backslash latex"] = getutf8(0xf0012);
1397 accents["backslash LaTeX"] = getutf8(0xf0012);
1398 accents["backslash latexe"] = getutf8(0xf0013);
1399 accents["backslash LaTeXe"] = getutf8(0xf0013);
1400 accents["ddot{\\imath}"] = "ï";
1401 buildaccent("ddot", "aAeEhHiIioOtuUwWxXyY",
1402 "äÄëËḧḦïÏïöÖẗüÜẅẄẍẌÿŸ"); // umlaut
1403 buildaccent("dot|.", "aAbBcCdDeEfFGghHIimMnNoOpPrRsStTwWxXyYzZ",
1404 "ȧȦḃḂċĊḋḊėĖḟḞĠġḣḢİİṁṀṅṄȯȮṗṖṙṘṡṠṫṪẇẆẋẊẏẎżŻ"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'İ'
1405 accents["acute{\\imath}"] = "í";
1406 buildaccent("acute", "aAcCeEgGkKlLmMoOnNpPrRsSuUwWyYzZiI",
1407 "áÁćĆéÉǵǴḱḰĺĹḿḾóÓńŃṕṔŕŔśŚúÚẃẂýÝźŹíÍ");
1408 buildaccent("dacute|H|h", "oOuU", "őŐűŰ"); // double acute
1409 buildaccent("mathring|r", "aAuUwy",
1411 accents["check{\\imath}"] = "ǐ";
1412 accents["check{\\jmath}"] = "ǰ";
1413 buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
1414 "čČďĎǎǍěĚǐǏǒǑǔǓǧǦǩǨȟȞľĽňŇřŘšŠŤťžŽ"); // caron
1415 accents["hat{\\imath}"] = "î";
1416 accents["hat{\\jmath}"] = "ĵ";
1417 buildaccent("hat|^", "aAcCeEgGhHiIjJoOsSuUwWyYzZ",
1418 "âÂĉĈêÊĝĜĥĤîÎĵĴôÔŝŜûÛŵŴŷŶẑẐ"); // circ
1419 accents["bar{\\imath}"] = "ī";
1420 buildaccent("bar|=", "aAeEiIoOuUyY",
1421 "āĀēĒīĪōŌūŪȳȲ"); // macron
1422 accents["tilde{\\imath}"] = "ĩ";
1423 buildaccent("tilde", "aAeEiInNoOuUvVyY",
1424 "ãÃẽẼĩĨñÑõÕũŨṽṼỹỸ"); // tilde
1425 accents["breve{\\imath}"] = "ĭ";
1426 buildaccent("breve|u", "aAeEgGiIoOuU",
1427 "ăĂĕĔğĞĭĬŏŎŭŬ"); // breve
1428 accents["grave{\\imath}"] = "ì";
1429 buildaccent("grave|`", "aAeEiIoOuUnNwWyY",
1430 "àÀèÈìÌòÒùÙǹǸẁẀỳỲ"); // grave
1431 buildaccent("subdot|d", "BbDdHhKkLlMmNnRrSsTtVvWwZzAaEeIiOoUuYy",
1432 "ḄḅḌḍḤḥḲḳḶḷṂṃṆṇṚṛṢṣṬṭṾṿẈẉẒẓẠạẸẹỊịỌọỤụỴỵ"); // dot below
1433 buildaccent("ogonek|k", "AaEeIiUuOo",
1434 "ĄąĘęĮįŲųǪǫ"); // ogonek
1435 buildaccent("cedilla|c", "CcGgKkLlNnRrSsTtEeDdHh",
1436 "ÇçĢģĶķĻļŅņŖŗŞşŢţȨȩḐḑḨḩ"); // cedilla
1437 buildaccent("subring|textsubring", "Aa",
1439 buildaccent("subhat|textsubcircum", "DdEeLlNnTtUu",
1440 "ḒḓḘḙḼḽṊṋṰṱṶṷ"); // subcircum
1441 buildaccent("subtilde|textsubtilde", "EeIiUu",
1442 "ḚḛḬḭṴṵ"); // subtilde
1443 accents["dgrave{\\imath}"] = "ȉ";
1444 accents["textdoublegrave{\\i}"] = "ȉ";
1445 buildaccent("dgrave|textdoublegrave", "AaEeIiOoRrUu",
1446 "ȀȁȄȅȈȉȌȍȐȑȔȕ"); // double grave
1447 accents["rcap{\\imath}"] = "ȉ";
1448 accents["textroundcap{\\i}"] = "ȉ";
1449 buildaccent("rcap|textroundcap", "AaEeIiOoRrUu",
1450 "ȂȃȆȇȊȋȎȏȒȓȖȗ"); // inverted breve
1451 buildaccent("slashed", "oO",
1456 * Created accents in math or regexp environment
1457 * are macros, but we need the utf8 equivalent
1459 void Intervall::removeAccents()
1461 if (accents.empty())
1463 static regex const accre("\\\\(([\\S]|grave|breve|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot|ogonek|"
1464 "cedilla|subring|textsubring|subhat|textsubcircum|subtilde|textsubtilde|dgrave|textdoublegrave|rcap|textroundcap|slashed)\\{[^\\{\\}]+\\}"
1465 "|((i|imath|jmath|cdot|[a-z]+space)|((backslash )?([lL]y[xX]|[tT]e[xX]|[lL]a[tT]e[xX]e?)))(?![a-zA-Z]))");
1467 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
1469 string key = sub.str(1);
1470 if (accents.find(key) != accents.end()) {
1471 string val = accents[key];
1472 size_t pos = sub.position(size_t(0));
1473 for (size_t i = 0; i < val.size(); i++) {
1474 par[pos+i] = val[i];
1476 // Remove possibly following space too
1477 if (par[pos+sub.str(0).size()] == ' ')
1478 addIntervall(pos+val.size(), pos + sub.str(0).size()+1);
1480 addIntervall(pos+val.size(), pos + sub.str(0).size());
1481 for (size_t i = pos+val.size(); i < pos + sub.str(0).size(); i++) {
1482 // remove traces of any remaining chars
1487 LYXERR(Debug::INFO, "Not added accent for \"" << key << "\"");
1492 void Intervall::handleOpenP(int i)
1495 depts[actualdeptindex] = i+1;
1496 closes[actualdeptindex] = -1;
1497 checkDepthIndex(actualdeptindex);
1500 void Intervall::handleCloseP(int i, bool closingAllowed)
1502 if (actualdeptindex <= 0) {
1503 if (! closingAllowed)
1504 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1505 // if we are at the very end
1506 addIntervall(i, i+1);
1509 closes[actualdeptindex] = i+1;
1514 void Intervall::resetOpenedP(int openPos)
1516 // Used as initializer for foreignlanguage entry
1517 actualdeptindex = 1;
1518 depts[1] = openPos+1;
1522 int Intervall::previousNotIgnored(int start) const
1524 int idx = 0; /* int intervalls */
1525 for (idx = ignoreidx; idx >= 0; --idx) {
1526 if (start > borders[idx].upper)
1528 if (start >= borders[idx].low)
1529 start = borders[idx].low-1;
1534 int Intervall::nextNotIgnored(int start) const
1536 int idx = 0; /* int intervalls */
1537 for (idx = 0; idx <= ignoreidx; idx++) {
1538 if (start < borders[idx].low)
1540 if (start < borders[idx].upper)
1541 start = borders[idx].upper;
1546 typedef map<string, KeyInfo> KeysMap;
1547 typedef vector< KeyInfo> Entries;
1548 static KeysMap keys = map<string, KeyInfo>();
1554 Intervall interval_;
1555 void buildKeys(bool);
1556 void buildEntries(bool);
1557 void makeKey(const string &, KeyInfo, bool isPatternString);
1558 void processRegion(int start, int region_end); /* remove {} parts */
1559 void removeHead(KeyInfo const &, int count=0);
1562 LatexInfo(string const & par, bool isPatternString)
1563 : entidx_(-1), interval_(isPatternString, par)
1565 buildKeys(isPatternString);
1566 entries_ = vector<KeyInfo>();
1567 buildEntries(isPatternString);
1571 if (entries_.empty()) {
1574 if (entries_[0].keytype == KeyInfo::isTitle) {
1575 interval_.hasTitle = true;
1576 if (! entries_[0].disabled) {
1577 interval_.titleValue = entries_[0].head;
1580 interval_.titleValue = "";
1582 removeHead(entries_[0]);
1583 if (entries_.size() > 1)
1592 if (int(entries_.size()) > entidx_) {
1599 bool setNextKey(int idx) {
1600 if ((idx == entidx_) && (entidx_ >= 0)) {
1607 int find(int start, KeyInfo::KeyType keytype) const {
1611 while (tmpIdx < int(entries_.size())) {
1612 if (entries_[tmpIdx].keytype == keytype)
1618 int process(ostringstream & os, KeyInfo const & actual);
1619 int dispatch(ostringstream & os, int previousStart, KeyInfo & actual);
1620 // string show(int lastpos) { return interval.show(lastpos);};
1621 int nextNotIgnored(int start) { return interval_.nextNotIgnored(start);};
1622 KeyInfo &getKeyInfo(int keyinfo) {
1623 static KeyInfo invalidInfo = KeyInfo();
1624 if ((keyinfo < 0) || ( keyinfo >= int(entries_.size())))
1627 return entries_[keyinfo];
1629 void setForDefaultLang(KeyInfo const & defLang) {interval_.setForDefaultLang(defLang);};
1630 void addIntervall(int low, int up) { interval_.addIntervall(low, up); };
1634 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
1638 for (int i = start; i < end; i += 1 + skip) {
1642 if (c == '\\') skip = 1;
1646 else if (c == down) {
1649 if ((repeat <= 0) || (par[i+1] != up))
1667 vector<MathEntry> entries_;
1672 void insert(string const & wait, size_t start, size_t end) {
1673 MathEntry m = MathEntry();
1675 m.mathStart = start;
1677 m.mathSize = end - start;
1678 entries_.push_back(m);
1680 bool empty() const { return entries_.empty(); };
1681 size_t getEndPos() const {
1682 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1685 return entries_[actualIdx_].mathEnd;
1687 size_t getStartPos() const {
1688 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1689 return 100000; /* definitely enough? */
1691 return entries_[actualIdx_].mathStart;
1693 size_t getFirstPos() {
1695 return getStartPos();
1697 size_t getSize() const {
1698 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1701 return entries_[actualIdx_].mathSize;
1703 void incrEntry() { actualIdx_++; };
1706 void LatexInfo::buildEntries(bool isPatternString)
1708 static regex const rmath("\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|alignat)\\*?)\\}");
1709 static regex const rkeys("\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?))");
1710 static bool disableLanguageOverride = false;
1711 smatch sub, submath;
1712 bool evaluatingRegexp = false;
1714 bool evaluatingMath = false;
1715 bool evaluatingCode = false;
1717 bool evaluatingOptional = false;
1718 size_t optionalEnd = 0;
1721 bool math_end_waiting = false;
1722 size_t math_pos = 10000;
1724 static vector<string> usedText = vector<string>();
1726 interval_.removeAccents();
1728 for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) {
1730 if (math_end_waiting) {
1731 size_t pos = submath.position(size_t(0));
1732 if ((math_end == "$") &&
1733 (submath.str(0) == "$") &&
1734 (interval_.par[pos-1] != '\\')) {
1735 mi.insert("$", math_pos, pos + 1);
1736 math_end_waiting = false;
1738 else if ((math_end == "\\]") &&
1739 (submath.str(0) == "\\]")) {
1740 mi.insert("\\]", math_pos, pos + 2);
1741 math_end_waiting = false;
1743 else if ((submath.str(1).compare("end") == 0) &&
1744 (submath.str(2).compare(math_end) == 0)) {
1745 mi.insert(math_end, math_pos, pos + submath.str(0).length());
1746 math_end_waiting = false;
1752 if (submath.str(1).compare("begin") == 0) {
1753 math_end_waiting = true;
1754 math_end = submath.str(2);
1755 math_pos = submath.position(size_t(0));
1757 else if (submath.str(0).compare("\\[") == 0) {
1758 math_end_waiting = true;
1760 math_pos = submath.position(size_t(0));
1762 else if (submath.str(0) == "$") {
1763 size_t pos = submath.position(size_t(0));
1764 if ((pos == 0) || (interval_.par[pos-1] != '\\')) {
1765 math_end_waiting = true;
1772 // Ignore language if there is math somewhere in pattern-string
1773 if (isPatternString) {
1774 for (auto s: usedText) {
1775 // Remove entries created in previous search runs
1778 usedText = vector<string>();
1781 keys["foreignlanguage"].disabled = true;
1782 disableLanguageOverride = true;
1785 disableLanguageOverride = false;
1788 if (disableLanguageOverride) {
1789 keys["foreignlanguage"].disabled = true;
1792 math_pos = mi.getFirstPos();
1793 for (sregex_iterator it(interval_.par.begin(), interval_.par.end(), rkeys), end; it != end; ++it) {
1795 string key = sub.str(3);
1797 if (sub.str(0)[0] == '\\')
1798 key = sub.str(0)[1];
1802 size_t k_pos = sub.position(size_t(0));
1803 if ((k_pos > 0) && (interval_.par[k_pos - 1] == '\\')) {
1804 // Escaped '$', ignoring
1810 if (keys.find(key) != keys.end()) {
1811 if (keys[key].keytype == KeyInfo::headRemove) {
1812 KeyInfo found1 = keys[key];
1813 found1.disabled = true;
1814 found1.head = "\\" + key + "{";
1815 found1._tokenstart = sub.position(size_t(0));
1816 found1._tokensize = found1.head.length();
1817 found1._dataStart = found1._tokenstart + found1.head.length();
1818 int endpos = interval_.findclosing(found1._dataStart, interval_.par.length(), '{', '}', 1);
1819 found1._dataEnd = endpos;
1824 if (evaluatingRegexp) {
1825 if (sub.str(1).compare("endregexp") == 0) {
1826 evaluatingRegexp = false;
1827 // found._tokenstart already set
1828 found._dataEnd = sub.position(size_t(0)) + 13;
1829 found._dataStart = found._dataEnd;
1830 found._tokensize = found._dataEnd - found._tokenstart;
1831 found.parenthesiscount = 0;
1832 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1839 if (evaluatingMath) {
1840 if (size_t(sub.position(size_t(0))) < mi.getEndPos())
1842 evaluatingMath = false;
1844 math_pos = mi.getStartPos();
1846 if (keys.find(key) == keys.end()) {
1847 found = KeyInfo(KeyInfo::isStandard, 0, true);
1848 LYXERR(Debug::INFO, "Undefined key " << key << " ==> will be used as text");
1849 found = KeyInfo(KeyInfo::isText, 0, false);
1850 if (isPatternString) {
1851 found.keytype = KeyInfo::isChar;
1852 found.disabled = false;
1856 usedText.push_back(key);
1860 if (key.compare("regexp") == 0) {
1861 evaluatingRegexp = true;
1862 found._tokenstart = sub.position(size_t(0));
1863 found._tokensize = 0;
1867 // Handle the other params of key
1868 if (found.keytype == KeyInfo::isIgnored)
1870 else if (found.keytype == KeyInfo::isMath) {
1871 if (size_t(sub.position(size_t(0))) == math_pos) {
1873 found._tokenstart = sub.position(size_t(0));
1874 found._tokensize = mi.getSize();
1875 found._dataEnd = found._tokenstart + found._tokensize;
1876 found._dataStart = found._dataEnd;
1877 found.parenthesiscount = 0;
1878 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1879 evaluatingMath = true;
1882 // begin|end of unknown env, discard
1883 // First handle tables
1884 // longtable|tabular
1885 bool discardComment;
1887 found.keytype = KeyInfo::doRemove;
1888 if ((sub.str(5).compare("longtable") == 0) ||
1889 (sub.str(5).compare("tabular") == 0)) {
1890 discardComment = true; /* '%' */
1893 discardComment = false;
1894 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
1896 string token = sub.str(5);
1897 if (regex_match(token, sub2, removeArgs)) {
1898 found.keytype = KeyInfo::removeWithArg;
1901 // discard spaces before pos(0)
1902 int pos = sub.position(size_t(0));
1904 for (count = 0; pos - count > 0; count++) {
1905 char c = interval_.par[pos-count-1];
1906 if (discardComment) {
1907 if ((c != ' ') && (c != '%'))
1913 found._tokenstart = pos - count;
1914 if (sub.str(1).compare(0, 5, "begin") == 0) {
1915 size_t pos1 = pos + sub.str(0).length();
1916 if (sub.str(5).compare("cjk") == 0) {
1917 pos1 = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
1918 if ((interval_.par[pos1] == '{') && (interval_.par[pos1+1] == '}'))
1920 found.keytype = KeyInfo::isMain;
1921 found._dataStart = pos1;
1922 found._dataEnd = interval_.par.length();
1923 found.disabled = keys["foreignlanguage"].disabled;
1924 found.used = keys["foreignlanguage"].used;
1925 found._tokensize = pos1 - found._tokenstart;
1926 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1929 // Swallow possible optional params
1930 while (interval_.par[pos1] == '[') {
1931 pos1 = interval_.findclosing(pos1+1, interval_.par.length(), '[', ']')+1;
1933 // Swallow also the eventual parameter
1934 if (interval_.par[pos1] == '{') {
1935 found._dataEnd = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
1938 found._dataEnd = pos1;
1940 found._dataStart = found._dataEnd;
1941 found._tokensize = count + found._dataEnd - pos;
1942 found.parenthesiscount = 0;
1943 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1944 found.disabled = true;
1948 // Handle "\end{...}"
1949 found._dataStart = pos + sub.str(0).length();
1950 found._dataEnd = found._dataStart;
1951 found._tokensize = count + found._dataEnd - pos;
1952 found.parenthesiscount = 0;
1953 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1954 found.disabled = true;
1958 else if (found.keytype != KeyInfo::isRegex) {
1959 found._tokenstart = sub.position(size_t(0));
1960 if (found.parenthesiscount == 0) {
1961 // Probably to be discarded
1962 size_t following_pos = sub.position(size_t(0)) + sub.str(3).length() + 1;
1963 char following = interval_.par[following_pos];
1964 if (following == ' ')
1965 found.head = "\\" + sub.str(3) + " ";
1966 else if (following == '=') {
1967 // like \uldepth=1000pt
1968 found.head = sub.str(0);
1971 found.head = "\\" + key;
1972 found._tokensize = found.head.length();
1973 found._dataEnd = found._tokenstart + found._tokensize;
1974 found._dataStart = found._dataEnd;
1977 int params = found._tokenstart + key.length() + 1;
1978 if (evaluatingOptional) {
1979 if (size_t(found._tokenstart) > optionalEnd) {
1980 evaluatingOptional = false;
1983 found.disabled = true;
1986 int optend = params;
1987 while (interval_.par[optend] == '[') {
1988 // discard optional parameters
1989 optend = interval_.findclosing(optend+1, interval_.par.length(), '[', ']') + 1;
1991 if (optend > params) {
1992 key += interval_.par.substr(params, optend-params);
1993 evaluatingOptional = true;
1994 optionalEnd = optend;
1996 string token = sub.str(5);
1997 int closings = found.parenthesiscount;
1998 if (found.parenthesiscount == 1) {
1999 found.head = "\\" + key + "{";
2001 else if (found.parenthesiscount > 1) {
2003 found.head = sub.str(0) + "{";
2004 closings = found.parenthesiscount - 1;
2007 found.head = "\\" + key + "{";
2010 found._tokensize = found.head.length();
2011 found._dataStart = found._tokenstart + found.head.length();
2012 if (found.keytype == KeyInfo::doRemove) {
2013 int endpar = 2 + interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2014 found._dataStart = endpar;
2015 found._tokensize = found._dataStart - found._tokenstart;
2018 if (interval_.par.substr(found._dataStart-1, 15).compare("\\endarguments{}") == 0) {
2019 found._dataStart += 15;
2023 endpos = found._dataStart - 1;
2025 endpos = interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
2026 if (found.keytype == KeyInfo::isList) {
2027 // Check if it really is list env
2028 static regex const listre("^([a-z]+)$");
2030 if (!regex_match(token, sub2, listre)) {
2031 // Change the key of this entry. It is not in a list/item environment
2032 found.keytype = KeyInfo::endArguments;
2035 if (found.keytype == KeyInfo::noMain) {
2036 evaluatingCode = true;
2038 codeStart = found._dataStart;
2040 else if (evaluatingCode) {
2041 if (size_t(found._dataStart) > codeEnd)
2042 evaluatingCode = false;
2043 else if (found.keytype == KeyInfo::isMain) {
2044 // Disable this key, treate it as standard
2045 found.keytype = KeyInfo::isStandard;
2046 found.disabled = true;
2047 if ((codeEnd +1 >= interval_.par.length()) &&
2048 (found._tokenstart == codeStart)) {
2049 // trickery, because the code inset starts
2050 // with \selectlanguage ...
2052 if (entries_.size() > 1) {
2053 entries_[entries_.size()-1]._dataEnd = codeEnd;
2058 if ((endpos == interval_.par.length()) &&
2059 (found.keytype == KeyInfo::doRemove)) {
2060 // Missing closing => error in latex-input?
2061 // therefore do not delete remaining data
2062 found._dataStart -= 1;
2063 found._dataEnd = found._dataStart;
2066 found._dataEnd = endpos;
2068 if (isPatternString) {
2069 keys[key].used = true;
2072 entries_.push_back(found);
2076 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
2078 stringstream s(keysstring);
2080 const char delim = '|';
2081 while (getline(s, key, delim)) {
2082 KeyInfo keyII(keyI);
2083 if (isPatternString) {
2086 else if ( !keys[key].used)
2087 keyII.disabled = true;
2092 void LatexInfo::buildKeys(bool isPatternString)
2095 static bool keysBuilt = false;
2096 if (keysBuilt && !isPatternString) return;
2098 // Keys to ignore in any case
2099 makeKey("text|textcyrillic|lyxmathsym", KeyInfo(KeyInfo::headRemove, 1, true), true);
2100 // Known standard keys with 1 parameter.
2101 // Split is done, if not at start of region
2102 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
2103 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
2104 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
2105 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
2106 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
2107 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
2109 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
2110 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2111 makeKey("section*|subsection*|subsubsection*|paragraph*",
2112 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2113 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2114 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
2116 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
2118 // Split is done, if not at start of region
2119 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
2120 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
2122 // Split is done always.
2123 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
2127 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2128 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2129 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2130 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2131 makeKey("guillemotright|guillemotleft", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2133 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2134 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2135 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2136 makeKey("thickspace|medspace|thinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2138 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2139 // Custom space/skip, remove the content (== length value)
2140 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
2141 // Found in fr/UserGuide.lyx
2142 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2144 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2145 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2146 // Known macros to remove (including their parameter)
2148 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
2149 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
2150 // handle like standard keys with 1 parameter.
2151 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
2153 // Ignore deleted text
2154 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString);
2155 // but preserve added text
2156 makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2158 // Macros to remove, but let the parameter survive
2160 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2162 // Remove language spec from content of these insets
2163 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
2165 // Same effect as previous, parameter will survive (because there is no one anyway)
2167 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2168 // Remove table decorations
2169 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
2170 // Discard shape-header.
2171 // For footnote or shortcut too, because of lang settings
2172 // and wrong handling if used 'KeyInfo::noMain'
2173 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2174 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2175 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2176 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2177 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2178 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
2179 // like ('tiny{}' or '\tiny ' ... )
2180 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
2182 // Survives, like known character
2183 // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2184 makeKey("tableofcontents", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2185 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
2187 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2188 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2189 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2191 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2192 // Remove RTL/LTR marker
2193 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2194 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
2195 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
2196 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
2197 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
2198 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
2199 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2200 if (isPatternString) {
2201 // Allow the first searched string to rebuild the keys too
2205 // no need to rebuild again
2211 * Keep the list of actual opened parentheses actual
2212 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2214 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2217 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2221 if (c == '\\') skip = 1;
2222 else if (c == '{') {
2225 else if (c == '}') {
2226 handleCloseP(i, closingAllowed);
2232 string Intervall::show(int lastpos)
2234 int idx = 0; /* int intervalls */
2237 for (idx = 0; idx <= ignoreidx; idx++) {
2238 while (i < lastpos) {
2240 if (i <= borders[idx].low) {
2241 if (borders[idx].low > lastpos)
2242 printsize = lastpos - i;
2244 printsize = borders[idx].low - i;
2245 s += par.substr(i, printsize);
2247 if (i >= borders[idx].low)
2248 i = borders[idx].upper;
2251 i = borders[idx].upper;
2257 s += par.substr(i, lastpos-i);
2263 void Intervall::output(ostringstream &os, int lastpos)
2265 // get number of chars to output
2266 int idx = 0; /* int intervalls */
2269 string startTitle = titleValue;
2270 for (idx = 0; idx <= ignoreidx; idx++) {
2272 if (i <= borders[idx].low) {
2274 if (borders[idx].low > lastpos)
2275 printsize = lastpos - i;
2277 printsize = borders[idx].low - i;
2278 if (printsize > 0) {
2279 os << startTitle << par.substr(i, printsize);
2281 printed += printsize;
2284 handleParentheses(i, false);
2285 if (i >= borders[idx].low)
2286 i = borders[idx].upper;
2289 i = borders[idx].upper;
2296 os << startTitle << par.substr(i, lastpos-i);
2297 printed += lastpos-i;
2299 handleParentheses(lastpos, false);
2301 if (keys["foreignlanguage"].disabled)
2302 startindex = actualdeptindex-langcount;
2304 startindex = actualdeptindex;
2305 for (int i = startindex; i > 0; --i) {
2308 if (hasTitle && (printed > 0))
2310 if (! isPatternString_)
2312 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2315 void LatexInfo::processRegion(int start, int region_end)
2317 while (start < region_end) { /* Let {[} and {]} survive */
2318 int cnt = interval_.isOpeningPar(start);
2320 // Closing is allowed past the region
2321 int closing = interval_.findclosing(start+1, interval_.par.length());
2322 interval_.addIntervall(start, start+1);
2323 interval_.addIntervall(closing, closing+1);
2327 start = interval_.nextNotIgnored(start+1);
2331 void LatexInfo::removeHead(KeyInfo const & actual, int count)
2333 if (actual.parenthesiscount == 0) {
2334 // "{\tiny{} ...}" ==> "{{} ...}"
2335 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2338 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2339 interval_.addIntervall(actual._tokenstart - count, actual._dataStart);
2340 interval_.addIntervall(actual._dataEnd, actual._dataEnd+1);
2344 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2347 switch (actual.keytype)
2349 case KeyInfo::isTitle: {
2351 nextKeyIdx = getNextKey();
2354 case KeyInfo::cleanToStart: {
2355 actual._dataEnd = actual._dataStart;
2356 nextKeyIdx = getNextKey();
2357 // Search for end of arguments
2358 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2360 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2361 entries_[i].disabled = true;
2363 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2365 while (interval_.par[actual._dataEnd] == ' ')
2367 interval_.addIntervall(0, actual._dataEnd+1);
2368 interval_.actualdeptindex = 0;
2369 interval_.depts[0] = actual._dataEnd+1;
2370 interval_.closes[0] = -1;
2373 case KeyInfo::isText:
2374 interval_.addIntervall(actual._tokenstart, actual._tokenstart+1);
2376 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2377 if (actual.disabled)
2378 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2380 interval_.addIntervall(actual._dataStart, actual._dataEnd);
2383 case KeyInfo::isChar: {
2384 nextKeyIdx = getNextKey();
2387 case KeyInfo::isSize: {
2388 if (actual.disabled || (interval_.par[actual._dataStart] != '{') || (interval_.par[actual._dataStart-1] == ' ')) {
2389 processRegion(actual._dataEnd, actual._dataEnd+1); /* remove possibly following {} */
2390 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2391 nextKeyIdx = getNextKey();
2393 // Here _dataStart points to '{', so correct it
2394 actual._dataStart += 1;
2395 actual._tokensize += 1;
2396 actual.parenthesiscount = 1;
2397 if (interval_.par[actual._dataStart] == '}') {
2398 // Determine the end if used like '{\tiny{}...}'
2399 actual._dataEnd = interval_.findclosing(actual._dataStart+1, interval_.par.length()) + 1;
2400 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2403 // Determine the end if used like '\tiny{...}'
2404 actual._dataEnd = interval_.findclosing(actual._dataStart, interval_.par.length()) + 1;
2406 // Split on this key if not at start
2407 int start = interval_.nextNotIgnored(previousStart);
2408 if (start < actual._tokenstart) {
2409 interval_.output(os, actual._tokenstart);
2410 interval_.addIntervall(start, actual._tokenstart);
2412 // discard entry if at end of actual
2413 nextKeyIdx = process(os, actual);
2417 case KeyInfo::endArguments: {
2418 // Remove trailing '{}' too
2419 actual._dataStart += 1;
2420 actual._dataEnd += 1;
2421 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2422 nextKeyIdx = getNextKey();
2425 case KeyInfo::noMain:
2427 case KeyInfo::isStandard: {
2428 if (actual.disabled) {
2430 processRegion(actual._dataStart, actual._dataStart+1);
2431 nextKeyIdx = getNextKey();
2433 // Split on this key if not at datastart of calling entry
2434 int start = interval_.nextNotIgnored(previousStart);
2435 if (start < actual._tokenstart) {
2436 interval_.output(os, actual._tokenstart);
2437 interval_.addIntervall(start, actual._tokenstart);
2439 // discard entry if at end of actual
2440 nextKeyIdx = process(os, actual);
2444 case KeyInfo::removeWithArg: {
2445 nextKeyIdx = getNextKey();
2446 // Search for end of arguments
2447 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2449 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2450 entries_[i].disabled = true;
2452 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2454 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2457 case KeyInfo::doRemove: {
2458 // Remove the key with all parameters and following spaces
2460 for (pos = actual._dataEnd+1; pos < interval_.par.length(); pos++) {
2461 if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
2464 // Remove also enclosing parentheses [] and {}
2467 while (actual._tokenstart > numpars) {
2468 if (pos+numpars >= interval_.par.size())
2470 else if (interval_.par[pos+numpars] == ']' && interval_.par[actual._tokenstart-numpars-1] == '[')
2472 else if (interval_.par[pos+numpars] == '}' && interval_.par[actual._tokenstart-numpars-1] == '{')
2478 if (interval_.par[pos+numpars] == ' ')
2482 interval_.addIntervall(actual._tokenstart-numpars, pos+numpars+spaces);
2483 nextKeyIdx = getNextKey();
2486 case KeyInfo::isList: {
2487 // Discard space before _tokenstart
2489 for (count = 0; count < actual._tokenstart; count++) {
2490 if (interval_.par[actual._tokenstart-count-1] != ' ')
2493 nextKeyIdx = getNextKey();
2494 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2496 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2498 // How else can we catch this one?
2499 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2500 entries_[i].disabled = true;
2502 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2504 else if (nextKeyIdx > 0) {
2505 // Ignore any lang entries inside data region
2506 for (int i = nextKeyIdx; i < int(entries_.size()) && entries_[i]._tokenstart < actual._dataEnd; i++) {
2507 if (entries_[i].keytype == KeyInfo::isMain)
2508 entries_[i].disabled = true;
2511 if (actual.disabled) {
2512 interval_.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2515 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2517 if (interval_.par[actual._dataEnd+1] == '[') {
2518 int posdown = interval_.findclosing(actual._dataEnd+2, interval_.par.length(), '[', ']');
2519 if ((interval_.par[actual._dataEnd+2] == '{') &&
2520 (interval_.par[posdown-1] == '}')) {
2521 interval_.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
2522 interval_.addIntervall(posdown-1, posdown+1);
2525 interval_.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
2526 interval_.addIntervall(posdown, posdown+1);
2528 int blk = interval_.nextNotIgnored(actual._dataEnd+1);
2529 if (blk > posdown) {
2530 // Discard at most 1 space after empty item
2532 for (count = 0; count < 1; count++) {
2533 if (interval_.par[blk+count] != ' ')
2537 interval_.addIntervall(blk, blk+count);
2542 case KeyInfo::isSectioning: {
2543 // Discard spaces before _tokenstart
2545 int val = actual._tokenstart;
2546 for (count = 0; count < actual._tokenstart;) {
2547 val = interval_.previousNotIgnored(val-1);
2548 if (val < 0 || interval_.par[val] != ' ')
2551 count = actual._tokenstart - val;
2554 if (actual.disabled) {
2555 removeHead(actual, count);
2556 nextKeyIdx = getNextKey();
2558 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2559 nextKeyIdx = process(os, actual);
2563 case KeyInfo::isMath: {
2564 // Same as regex, use the content unchanged
2565 nextKeyIdx = getNextKey();
2568 case KeyInfo::isRegex: {
2569 // DO NOT SPLIT ON REGEX
2571 nextKeyIdx = getNextKey();
2574 case KeyInfo::isIgnored: {
2575 // Treat like a character for now
2576 nextKeyIdx = getNextKey();
2579 case KeyInfo::isMain: {
2580 if (interval_.par.substr(actual._dataStart, 2) == "% ")
2581 interval_.addIntervall(actual._dataStart, actual._dataStart+2);
2582 if (actual._tokenstart > 0) {
2583 int prev = interval_.previousNotIgnored(actual._tokenstart - 1);
2584 if ((prev >= 0) && interval_.par[prev] == '%')
2585 interval_.addIntervall(prev, prev+1);
2587 if (actual.disabled) {
2589 interval_.langcount++;
2590 if ((interval_.par.substr(actual._dataStart, 3) == " \\[") ||
2591 (interval_.par.substr(actual._dataStart, 8) == " \\begin{")) {
2592 // Discard also the space before math-equation
2593 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2595 nextKeyIdx = getNextKey();
2596 // interval.resetOpenedP(actual._dataStart-1);
2599 if (actual._tokenstart < 26) {
2600 // for the first (and maybe dummy) language
2601 interval_.setForDefaultLang(actual);
2603 interval_.resetOpenedP(actual._dataStart-1);
2607 case KeyInfo::invalid:
2608 case KeyInfo::headRemove:
2609 // These two cases cannot happen, already handled
2612 // LYXERR(Debug::INFO, "Unhandled keytype");
2613 nextKeyIdx = getNextKey();
2620 int LatexInfo::process(ostringstream & os, KeyInfo const & actual )
2622 int end = interval_.nextNotIgnored(actual._dataEnd);
2623 int oldStart = actual._dataStart;
2624 int nextKeyIdx = getNextKey();
2626 if ((nextKeyIdx < 0) ||
2627 (entries_[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
2628 (entries_[nextKeyIdx].keytype == KeyInfo::invalid)) {
2629 if (oldStart <= end) {
2630 processRegion(oldStart, end);
2635 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
2637 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
2638 (void) dispatch(os, actual._dataStart, nextKey);
2639 end = nextKey._tokenstart;
2642 processRegion(oldStart, nextKey._tokenstart);
2643 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
2645 oldStart = nextKey._dataEnd+1;
2647 // now nextKey is either invalid or is outside of actual._dataEnd
2648 // output the remaining and discard myself
2649 if (oldStart <= end) {
2650 processRegion(oldStart, end);
2652 if (interval_.par.size() > (size_t) end && interval_.par[end] == '}') {
2654 // This is the normal case.
2655 // But if using the firstlanguage, the closing may be missing
2657 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
2659 if (actual._dataEnd < end)
2660 output_end = interval_.nextNotIgnored(actual._dataEnd);
2661 else if (interval_.par.size() > (size_t) end)
2662 output_end = interval_.nextNotIgnored(end);
2664 output_end = interval_.par.size();
2665 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
2666 interval_.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
2668 // Remove possible empty data
2669 int dstart = interval_.nextNotIgnored(actual._dataStart);
2670 while (interval_.isOpeningPar(dstart) == 1) {
2671 interval_.addIntervall(dstart, dstart+1);
2672 int dend = interval_.findclosing(dstart+1, output_end);
2673 interval_.addIntervall(dend, dend+1);
2674 dstart = interval_.nextNotIgnored(dstart+1);
2676 if (dstart < output_end)
2677 interval_.output(os, output_end);
2678 interval_.addIntervall(actual._tokenstart, end);
2682 string splitOnKnownMacros(string par, bool isPatternString)
2685 LatexInfo li(par, isPatternString);
2686 // LYXERR(Debug::INFO, "Berfore split: " << par);
2687 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
2689 DummyKey._tokensize = 0;
2690 DummyKey._dataStart = 0;
2691 DummyKey._dataEnd = par.length();
2692 DummyKey.disabled = true;
2693 int firstkeyIdx = li.getFirstKey();
2695 if (firstkeyIdx >= 0) {
2696 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
2697 DummyKey._tokenstart = firstKey._tokenstart;
2699 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
2700 // Use dummy firstKey
2701 firstKey = DummyKey;
2702 (void) li.setNextKey(firstkeyIdx);
2705 if (par.substr(firstKey._dataStart, 2) == "% ")
2706 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
2708 nextkeyIdx = li.process(os, firstKey);
2709 while (nextkeyIdx >= 0) {
2710 // Check for a possible gap between the last
2711 // entry and this one
2712 int datastart = li.nextNotIgnored(firstKey._dataStart);
2713 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
2714 if ((nextKey._tokenstart > datastart)) {
2716 firstKey._dataStart = datastart;
2717 firstKey._dataEnd = par.length();
2718 (void) li.setNextKey(nextkeyIdx);
2719 // Fake the last opened parenthesis
2720 li.setForDefaultLang(firstKey);
2721 nextkeyIdx = li.process(os, firstKey);
2724 if (nextKey.keytype != KeyInfo::isMain) {
2725 firstKey._dataStart = datastart;
2726 firstKey._dataEnd = nextKey._dataEnd+1;
2727 (void) li.setNextKey(nextkeyIdx);
2728 li.setForDefaultLang(firstKey);
2729 nextkeyIdx = li.process(os, firstKey);
2732 nextkeyIdx = li.process(os, nextKey);
2736 // Handle the remaining
2737 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
2738 firstKey._dataEnd = par.length();
2740 if ((firstKey._dataStart < firstKey._dataEnd) &&
2741 (par[firstKey._dataStart] != '}')) {
2742 li.setForDefaultLang(firstKey);
2743 (void) li.process(os, firstKey);
2747 // return string definitelly impossible to match
2748 s = "\\foreignlanguage{ignore}{ }";
2752 s = par; /* no known macros found */
2753 // LYXERR(Debug::INFO, "After split: " << s);
2758 * Try to unify the language specs in the latexified text.
2759 * Resulting modified string is set to "", if
2760 * the searched tex does not contain all the features in the search pattern
2762 static string correctlanguagesetting(string par, bool isPatternString, bool withformat)
2764 static Features regex_f;
2765 static int missed = 0;
2766 static bool regex_with_format = false;
2768 int parlen = par.length();
2770 while ((parlen > 0) && (par[parlen-1] == '\n')) {
2773 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
2774 // Happens to be there in case of description or labeling environment
2779 // Split the latex input into pieces which
2780 // can be digested by our search engine
2781 LYXERR(Debug::FIND, "input: \"" << par << "\"");
2782 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
2783 LYXERR(Debug::FIND, "After split: \"" << result << "\"");
2786 result = par.substr(0, parlen);
2787 if (isPatternString) {
2790 regex_f = identifyFeatures(result);
2791 string features = "";
2792 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2793 string a = it->first;
2794 regex_with_format = true;
2795 features += " " + a;
2796 // LYXERR(Debug::INFO, "Identified regex format:" << a);
2798 LYXERR(Debug::FIND, "Identified Features" << features);
2801 } else if (regex_with_format) {
2802 Features info = identifyFeatures(result);
2803 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2804 string a = it->first;
2805 bool b = it->second;
2806 if (b && ! info[a]) {
2808 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
2814 // LYXERR(Debug::INFO, "No regex formats");
2820 // Remove trailing closure of math, macros and environments, so to catch parts of them.
2821 static int identifyClosing(string & t)
2823 int open_braces = 0;
2825 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
2826 if (regex_replace(t, t, "(.*[^\\\\])\\$$", "$1"))
2828 if (regex_replace(t, t, "(.*[^\\\\])\\\\\\]$", "$1"))
2830 if (regex_replace(t, t, "(.*[^\\\\])\\\\end\\{[a-zA-Z_]*\\*?\\}$", "$1"))
2832 if (regex_replace(t, t, "(.*[^\\\\])\\}$", "$1")) {
2841 static int num_replaced = 0;
2842 static bool previous_single_replace = true;
2844 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt)
2845 : p_buf(&buf), p_first_buf(&buf), opt(opt)
2847 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
2848 docstring const & ds = stringifySearchBuffer(find_buf, opt);
2849 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
2850 if (opt.replace_all && previous_single_replace) {
2851 previous_single_replace = false;
2854 else if (!opt.replace_all) {
2855 num_replaced = 0; // count number of replaced strings
2856 previous_single_replace = true;
2858 // When using regexp, braces are hacked already by escape_for_regex()
2859 par_as_string = normalize(ds, !use_regexp);
2861 close_wildcards = 0;
2863 size_t lead_size = 0;
2864 // correct the language settings
2865 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat);
2866 if (opt.ignoreformat) {
2868 // if par_as_string_nolead were emty,
2869 // the following call to findAux will always *find* the string
2870 // in the checked data, and thus always using the slow
2871 // examining of the current text part.
2872 par_as_string_nolead = par_as_string;
2875 lead_size = identifyLeading(par_as_string);
2876 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
2877 lead_as_string = par_as_string.substr(0, lead_size);
2878 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
2882 open_braces = identifyClosing(par_as_string);
2883 identifyClosing(par_as_string_nolead);
2884 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2885 LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'");
2887 string lead_as_regexp;
2888 if (lead_size > 0) {
2889 // @todo No need to search for \regexp{} insets in leading material
2890 lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat);
2891 par_as_string = par_as_string_nolead;
2892 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
2893 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2895 LYXERR(Debug::FIND, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
2896 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
2897 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
2898 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2900 // Insert .* before trailing '\$' ('$' has been escaped by escape_for_regex)
2901 regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\$)\\'", "$1(.*?)$2")
2902 // Insert .* before trailing '\\\]' ('\]' has been escaped by escape_for_regex)
2903 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])( \\\\\\\\\\\\\\])\\'", "$1(.*?)$2")
2904 // Insert .* before trailing '\\end\{...}' ('\end{...}' has been escaped by escape_for_regex)
2905 || regex_replace(par_as_string, par_as_string,
2906 "(.*[^\\\\])( \\\\\\\\end\\\\\\{[a-zA-Z_]*)(\\\\\\*)?(\\\\\\})\\'", "$1(.*?)$2$3$4")
2907 // Insert .* before trailing '\}' ('}' has been escaped by escape_for_regex)
2908 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\})\\'", "$1(.*?)$2")
2912 if (!opt.ignoreformat) {
2913 // Remove extra '\}' at end if not part of \{\.\}
2914 size_t lng = par_as_string.size();
2916 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
2918 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
2927 if (lng < par_as_string.size())
2928 par_as_string = par_as_string.substr(0,lng);
2931 regex_replace(par_as_string, par_as_string, "\\\\\\.", "_xxbdotxx_");
2932 // handle '.' -> '[^]', replace later as '[^\}\{\\]'
2933 regex_replace(par_as_string, par_as_string, "\\.", "[^]");
2934 // replace '[^...]' with '[^...\}\{\\]'
2935 regex_replace(par_as_string, par_as_string, "\\[\\^([^\\\\\\]]*)\\]", "_xxbrlxx_$1\\}\\{\\\\_xxbrrxx_");
2936 regex_replace(par_as_string, par_as_string, "_xxbrlxx_", "[^");
2937 regex_replace(par_as_string, par_as_string, "_xxbrrxx_", "]");
2939 regex_replace(par_as_string, par_as_string, "_xxbdotxx_", "\\.");
2942 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2943 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2944 LYXERR(Debug::FIND, "Close .*? : " << close_wildcards);
2945 LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
2947 // If entered regexp must match at begin of searched string buffer
2948 // Kornel: Added parentheses to use $1 for size of the leading string
2952 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
2953 // Unfortunately is '\1', '\2', etc not working for strings with extra format
2954 // so the convert has no effect in that case
2955 for (int i = 8; i > 0; --i) {
2956 string orig = "\\\\" + std::to_string(i);
2957 string dest = "\\" + std::to_string(i+1);
2958 while (regex_replace(par_as_string, par_as_string, orig, dest));
2960 regexp_str = "(" + lead_as_regexp + ")" + par_as_string;
2961 regexp2_str = "(" + lead_as_regexp + ").*?" + par_as_string;
2963 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
2964 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
2965 #if (QT_VERSION >= 0x050000)
2966 // Handle \w properly
2967 QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
2968 if (! opt.casesensitive) {
2969 popts |= QRegularExpression::CaseInsensitiveOption;
2971 regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
2972 regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
2974 if (regexp.isValid() && regexp2.isValid()) {
2975 regexIsValid = true;
2976 // Check '{', '}' pairs inside the regex
2979 for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
2980 char c = par_as_string[i];
2987 else if (c == '}') {
2994 if (balanced != 0) {
2995 regexIsValid = false;
2996 regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
3000 regexIsValid = false;
3001 if (!regexp.isValid())
3002 regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
3003 if (!regexp2.isValid())
3004 regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
3007 if (opt.casesensitive) {
3008 regexp = regex(regexp_str);
3009 regexp2 = regex(regexp2_str);
3012 regexp = regex(regexp_str, std::regex_constants::icase);
3013 regexp2 = regex(regexp2_str, std::regex_constants::icase);
3020 // Count number of characters in string
3024 // \\[a-zA-Z]+ ==> 1
3025 #if (QT_VERSION >= 0x050000)
3026 static int computeSize(QStringRef s, int len)
3027 #define isLyxAlpha(arg) arg.isLetter()
3029 static int computeSize(string s, int len)
3030 #define isLyxAlpha(arg) isalpha(arg)
3037 for (int i = 0; i < len; i += skip, count++) {
3038 if (s.at(i) == '\\') {
3040 if (i + 1 < len && isLyxAlpha(s.at(i+1))) {
3041 for (int j = 2; i+j < len; j++) {
3042 if (! isLyxAlpha(s.at(i+j))) {
3043 if (s.at(i+j) == ' ')
3045 else if (s.at(i+j) == '{') {
3046 if (i+j+1 < len && s.at(i+j+1) == '}')
3048 else if (i + j + 1 >= len)
3057 else if (s.at(i) == '{') {
3058 if (i + 1 < len && s.at(i+1) == '}')
3063 else if (s.at(i) == '-') {
3064 if (i+1 < len && s.at(i+1) == '-') {
3065 if (i + 2 < len && s.at(i+2) == '-')
3080 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
3085 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
3088 docstring docstr = stringifyFromForSearch(opt, cur, len);
3090 if (use_regexp || opt.casesensitive)
3091 str = normalize(docstr, true);
3093 str = normalize(lowercase(docstr), true);
3094 if (!opt.ignoreformat) {
3095 str = correctlanguagesetting(str, false, !opt.ignoreformat);
3098 mres.match_len = -1;
3101 LYXERR(Debug::FIND, "Matching against '" << lyx::to_utf8(docstr) << "'");
3102 LYXERR(Debug::FIND, "After normalization: '" << str << "'");
3105 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
3106 #if (QT_VERSION >= 0x050000)
3107 QString qstr = QString::fromStdString(str);
3108 QRegularExpression const *p_regexp;
3109 QRegularExpression::MatchType flags = QRegularExpression::NormalMatch;
3113 p_regexp = ®exp2;
3115 QRegularExpressionMatch match = p_regexp->match(qstr, 0, flags);
3116 if (!match.hasMatch())
3118 // Check braces on segments that matched all (.*?) subexpressions,
3119 // except the last "padding" one inserted by lyx.
3120 for (int i = 1; i < match.lastCapturedIndex(); ++i)
3121 if (!braces_match(match.captured(i), open_braces))
3124 regex const *p_regexp;
3125 regex_constants::match_flag_type flags;
3127 flags = regex_constants::match_continuous;
3130 flags = regex_constants::match_default;
3131 p_regexp = ®exp2;
3133 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
3134 if (re_it == sregex_iterator())
3136 match_results<string::const_iterator> const & m = *re_it;
3137 // Check braces on segments that matched all (.*?) subexpressions,
3138 // except the last "padding" one inserted by lyx.
3139 for (size_t i = 1; i < m.size() - 1; ++i)
3140 if (!braces_match(m[i], open_braces))
3143 // Exclude from the returned match length any length
3144 // due to close wildcards added at end of regexp
3145 // and also the length of the leading (e.g. '\emph{}')
3147 // Whole found string, including the leading: m[0].second - m[0].first
3148 // Size of the leading string: m[1].second - m[1].first
3149 int leadingsize = 0;
3152 #if (QT_VERSION >= 0x050000)
3153 if (match.lastCapturedIndex() > 0)
3154 leadingsize = match.capturedEnd(1) - match.capturedStart(1);
3156 int lastidx = match.lastCapturedIndex();
3157 for (int i = 0; i <= lastidx; i++) {
3158 LYXERR(Debug::FIND, "Match " << i << " is " << match.capturedEnd(i) - match.capturedStart(i) << " long");
3160 if (close_wildcards == 0)
3161 result = match.capturedEnd(0) - match.capturedStart(0);
3163 result = match.capturedStart(lastidx + 1 - close_wildcards) - match.capturedStart(0);
3165 pos = match.capturedStart(0);
3166 // Ignore last closing characters
3167 while (result > 0) {
3168 if (qstr.at(pos+result-1) == '}')
3175 leadingsize = m[1].second - m[1].first;
3176 for (size_t i = 0; i < m.size(); i++) {
3177 LYXERR(Debug::FIND, "Match " << i << " is " << m[i].second - m[i].first << " long");
3179 if (close_wildcards == 0)
3180 result = m[0].second - m[0].first;
3182 result = m[m.size() - close_wildcards].first - m[0].first;
3183 pos = m.position(size_t(0));
3184 // Ignore last closing characters
3185 while (result > 0) {
3186 if (str[pos+result-1] == '}')
3192 if (result > leadingsize)
3193 result -= leadingsize;
3196 #if (QT_VERSION >= 0x050000)
3197 mres.match_len = computeSize(QStringRef(&qstr, pos+leadingsize,result), result);
3198 mres.match2end = qstr.size() - pos - leadingsize;
3200 mres.match_len = computeSize(str.substr(pos+leadingsize,result), result);
3201 mres.match2end = str.size() - pos - leadingsize;
3203 mres.pos = pos+leadingsize;
3207 // else !use_regexp: but all code paths above return
3208 LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='"
3209 << par_as_string << "', str='" << str << "'");
3210 LYXERR(Debug::FIND, "Searching in normal mode: lead_as_string='"
3211 << lead_as_string << "', par_as_string_nolead='"
3212 << par_as_string_nolead << "'");
3215 LYXERR(Debug::FIND, "size=" << par_as_string.size()
3216 << ", substr='" << str.substr(0, par_as_string.size()) << "'");
3217 if (str.substr(0, par_as_string.size()) == par_as_string) {
3218 mres.match_len = par_as_string.size();
3219 mres.match2end = str.size();
3224 // Start the search _after_ the leading part
3225 size_t pos = str.find(par_as_string_nolead, lead_as_string.size());
3226 if (pos != string::npos) {
3227 mres.match_len = par_as_string.size();
3228 mres.match2end = str.size() - pos;
3237 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
3239 MatchResult mres = findAux(cur, len, at_begin);
3240 int res = mres.match_len;
3242 "res=" << res << ", at_begin=" << at_begin
3243 << ", matchword=" << opt.matchword
3244 << ", inTexted=" << cur.inTexted());
3245 if (res == 0 || !at_begin || !opt.matchword || !cur.inTexted())
3247 if ((len > 0) && (res < len)) {
3251 Paragraph const & par = cur.paragraph();
3252 bool ws_left = (cur.pos() > 0)
3253 ? par.isWordSeparator(cur.pos() - 1)
3259 ws_right = (cur.pos() + len < par.size())
3260 ? par.isWordSeparator(cur.pos() + len)
3264 "cur.pos()=" << cur.pos() << ", res=" << res
3265 << ", separ: " << ws_left << ", " << ws_right
3268 if (ws_left && ws_right) {
3269 // Check for word separators inside the found 'word'
3270 for (int i = 0; i < len; i++) {
3271 if (par.isWordSeparator(cur.pos() + i)) {
3283 string MatchStringAdv::normalize(docstring const & s, bool hack_braces) const
3286 t = lyx::to_utf8(s);
3287 // Remove \n at begin
3288 while (!t.empty() && t[0] == '\n')
3291 while (!t.empty() && t[t.size() - 1] == '\n')
3292 t = t.substr(0, t.size() - 1);
3294 // Handle all other '\n'
3295 while ((pos = t.find("\n")) != string::npos) {
3296 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
3298 if (isAlnumASCII(t[pos+1])) {
3299 t.replace(pos-2, 3, " ");
3302 t.replace(pos-2, 3, "");
3305 else if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
3306 // '\n' adjacent to non-alpha-numerics, discard
3307 t.replace(pos, 1, "");
3310 // Replace all other \n with spaces
3311 t.replace(pos, 1, " ");
3314 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
3315 // Kornel: Added textsl, textsf, textit, texttt and noun
3316 // + allow to seach for colored text too
3317 LYXERR(Debug::FIND, "Removing stale empty \\emph{}, \\textbf{}, \\*section{} macros from: " << t);
3318 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
3319 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3320 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
3321 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3323 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
3324 // FIXME - check what preceeds the brace
3326 if (opt.ignoreformat)
3327 while (regex_replace(t, t, "\\{", "_x_<")
3328 || regex_replace(t, t, "\\}", "_x_>"))
3329 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
3331 while (regex_replace(t, t, "\\\\\\{", "_x_<")
3332 || regex_replace(t, t, "\\\\\\}", "_x_>"))
3333 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
3340 docstring stringifyFromCursor(DocIterator const & cur, int len)
3342 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
3343 if (cur.inTexted()) {
3344 Paragraph const & par = cur.paragraph();
3345 // TODO what about searching beyond/across paragraph breaks ?
3346 // TODO Try adding a AS_STR_INSERTS as last arg
3347 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
3348 int(par.size()) : cur.pos() + len;
3349 // OutputParams runparams(&cur.buffer()->params().encoding());
3350 OutputParams runparams(encodings.fromLyXName("utf8"));
3351 runparams.nice = true;
3352 runparams.flavor = Flavor::XeTeX;
3353 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
3354 // No side effect of file copying and image conversion
3355 runparams.dryrun = true;
3356 runparams.for_search = true;
3357 LYXERR(Debug::FIND, "Stringifying with cur: "
3358 << cur << ", from pos: " << cur.pos() << ", end: " << end);
3359 return par.asString(cur.pos(), end,
3360 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
3362 } else if (cur.inMathed()) {
3363 CursorSlice cs = cur.top();
3364 MathData md = cs.cell();
3365 MathData::const_iterator it_end =
3366 (( len == -1 || cs.pos() + len > int(md.size()))
3368 : md.begin() + cs.pos() + len );
3370 for (MathData::const_iterator it = md.begin() + cs.pos();
3373 docstring s = asString(md2);
3374 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
3377 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3382 /** Computes the LaTeX export of buf starting from cur and ending len positions
3383 * after cur, if len is positive, or at the paragraph or innermost inset end
3386 docstring latexifyFromCursor(DocIterator const & cur, int len)
3388 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
3389 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
3390 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
3391 Buffer const & buf = *cur.buffer();
3393 odocstringstream ods;
3395 //OutputParams runparams(&buf.params().encoding());
3396 OutputParams runparams(encodings.fromLyXName("utf8"));
3397 runparams.nice = false;
3398 runparams.flavor = Flavor::XeTeX;
3399 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3400 // No side effect of file copying and image conversion
3401 runparams.dryrun = true;
3402 runparams.for_search = true;
3404 if (cur.inTexted()) {
3405 // @TODO what about searching beyond/across paragraph breaks ?
3406 pos_type endpos = cur.paragraph().size();
3407 if (len != -1 && endpos > cur.pos() + len)
3408 endpos = cur.pos() + len;
3409 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
3410 string(), cur.pos(), endpos);
3411 string s = lyx::to_utf8(ods.str());
3412 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
3413 return(lyx::from_utf8(s));
3414 } else if (cur.inMathed()) {
3415 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
3416 for (int s = cur.depth() - 1; s >= 0; --s) {
3417 CursorSlice const & cs = cur[s];
3418 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
3420 cs.asInsetMath()->asHullInset()->header_write(ws);
3425 CursorSlice const & cs = cur.top();
3426 MathData md = cs.cell();
3427 MathData::const_iterator it_end =
3428 ((len == -1 || cs.pos() + len > int(md.size()))
3430 : md.begin() + cs.pos() + len);
3432 for (MathData::const_iterator it = md.begin() + cs.pos();
3436 ods << asString(md2);
3437 // Retrieve the math environment type, and add '$' or '$]'
3438 // or others (\end{equation}) accordingly
3439 for (int s = cur.depth() - 1; s >= 0; --s) {
3440 CursorSlice const & cs2 = cur[s];
3441 InsetMath * inset = cs2.asInsetMath();
3442 if (inset && inset->asHullInset()) {
3444 inset->asHullInset()->footer_write(ws);
3448 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3450 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3456 /** Finalize an advanced find operation, advancing the cursor to the innermost
3457 ** position that matches, plus computing the length of the matching text to
3460 int findAdvFinalize(DocIterator & cur, MatchStringAdv const & match)
3462 // Search the foremost position that matches (avoids find of entire math
3463 // inset when match at start of it)
3465 DocIterator old_cur(cur.buffer());
3467 LYXERR(Debug::FIND, "Forwarding one step (searching for innermost match)");
3471 } while (cur && cur.depth() > d && match(cur).match_len > 0);
3473 int max_match = match(cur).match_len; /* match valid only if not searching whole words */
3474 if (max_match <= 0) return 0;
3475 LYXERR(Debug::FIND, "Ok");
3477 // Compute the match length
3479 if (cur.pos() + len > cur.lastpos())
3481 if (match.opt.matchword) {
3482 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3483 while (cur.pos() + len <= cur.lastpos() && match(cur, len).match_len <= 0) {
3485 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3487 // Length of matched text (different from len param)
3488 int old_match = match(cur, len).match_len;
3492 // Greedy behaviour while matching regexps
3493 while ((new_match = match(cur, len + 1).match_len) > old_match) {
3495 old_match = new_match;
3496 LYXERR(Debug::FIND, "verifying match with len = " << len);
3503 int maxl = cur.lastpos() - cur.pos();
3504 // Greedy behaviour while matching regexps
3505 while (maxl > minl) {
3506 int actual_match = match(cur, len).match_len;
3507 if (actual_match >= max_match) {
3508 // actual_match > max_match _can_ happen,
3509 // if the search area splits
3510 // some following word so that the regex
3511 // (e.g. 'r.*r\b' matches 'r' from the middle of the
3513 // This means, the len value is too big
3515 len = (int)((maxl + minl)/2);
3518 // (actual_match < max_match)
3520 len = (int)((maxl + minl)/2);
3524 // Search for real start of matched characters
3529 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
3530 if (cur.depth() < old_cur.depth()) {
3532 LYXERR(Debug::INFO, "cur.depth() < old_cur.depth(), this should never happen");
3535 if (cur.pos() != old_cur.pos()) {
3536 // OK, forwarded 1 pos in actual inset
3537 actual_match = match(cur, len-1).match_len;
3538 if (actual_match == max_match) {
3539 // Ha, got it! The shorter selection has the same match length
3544 // OK, the shorter selection matches less chars, revert to previous value
3550 LYXERR(Debug::INFO, "cur.pos() == old_cur.pos(), this should never happen");
3551 actual_match = match(cur, len).match_len;
3552 if (actual_match == max_match)
3562 int findForwardAdv(DocIterator & cur, MatchStringAdv const & match)
3566 while (!theApp()->longOperationCancelled() && cur) {
3567 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
3568 MatchResult mres = match(cur, -1, false);
3569 int match_len = mres.match_len;
3570 LYXERR(Debug::FIND, "match_len: " << match_len);
3571 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
3572 LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
3575 if (match_len > 0) {
3576 // Try to find the begin of searched string
3577 int increment = mres.pos/2;
3578 while (mres.pos > 5 && (increment > 5)) {
3579 DocIterator old_cur = cur;
3580 for (int i = 0; i < increment && cur; cur.forwardPos(), i++) {
3582 if (! cur || (cur.pit() > old_cur.pit())) {
3583 // Are we outside of the paragraph?
3584 // This can happen if moving past some UTF8-encoded chars
3589 MatchResult mres2 = match(cur, -1, false);
3590 if ((mres2.match2end < mres.match2end) ||
3591 (mres2.match_len < mres.match_len)) {
3598 if (increment > mres.pos/2)
3599 increment = mres.pos/2;
3603 int match_len_zero_count = 0;
3604 for (int i = 0; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) {
3606 int remaining_len = match(cur, -1, false).match_len;
3607 if (remaining_len <= 0) {
3608 // Apparently the searched string is not in the remaining part
3615 LYXERR(Debug::FIND, "Advancing cur: " << cur);
3616 int match_len3 = match(cur, 1).match_len;
3619 int match_len2 = match(cur).match_len;
3620 LYXERR(Debug::FIND, "match_len2: " << match_len2);
3621 if (match_len2 > 0) {
3622 // Sometimes in finalize we understand it wasn't a match
3623 // and we need to continue the outest loop
3624 int len = findAdvFinalize(cur, match);
3629 if (match_len2 >= 0) {
3630 if (match_len2 == 0)
3631 match_len_zero_count++;
3633 match_len_zero_count = 0;
3636 if (++match_len_zero_count > 3) {
3637 LYXERR(Debug::FIND, "match_len2_zero_count: " << match_len_zero_count << ", match_len was " << match_len);
3645 if (match_len >= 0 && cur.pit() < cur.lastpit()) {
3646 LYXERR(Debug::FIND, "Advancing par: cur=" << cur);
3649 // This should exit nested insets, if any, or otherwise undefine the currsor.
3650 cur.pos() = cur.lastpos();
3651 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
3659 /// Find the most backward consecutive match within same paragraph while searching backwards.
3660 int findMostBackwards(DocIterator & cur, MatchStringAdv const & match)
3662 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3663 DocIterator tmp_cur = cur;
3664 int len = findAdvFinalize(tmp_cur, match);
3665 Inset & inset = cur.inset();
3666 for (; cur != cur_begin; cur.backwardPos()) {
3667 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
3668 DocIterator new_cur = cur;
3669 new_cur.backwardPos();
3670 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
3672 int new_len = findAdvFinalize(new_cur, match);
3677 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
3683 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
3687 // Backup of original position
3688 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3689 if (cur == cur_begin)
3692 DocIterator cur_orig(cur);
3693 bool pit_changed = false;
3696 bool found_match = (match(cur, -1, false).match_len > 0);
3700 cur.pos() = cur.lastpos();
3702 cur.pos() = cur_orig.pos();
3703 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
3704 DocIterator cur_prev_iter;
3706 found_match = (match(cur).match_len > 0);
3707 LYXERR(Debug::FIND, "findBackAdv3: found_match="
3708 << found_match << ", cur: " << cur);
3710 return findMostBackwards(cur, match);
3712 // Stop if begin of document reached
3713 if (cur == cur_begin)
3715 cur_prev_iter = cur;
3719 if (cur == cur_begin)
3726 } while (!theApp()->longOperationCancelled());
3734 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
3735 DocIterator const & cur, int len)
3737 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
3739 if (!opt.ignoreformat)
3740 return latexifyFromCursor(cur, len);
3742 return stringifyFromCursor(cur, len);
3746 FindAndReplaceOptions::FindAndReplaceOptions(
3747 docstring const & _find_buf_name, bool _casesensitive,
3748 bool _matchword, bool _forward, bool _expandmacros, bool _ignoreformat,
3749 docstring const & _repl_buf_name, bool _keep_case,
3750 SearchScope _scope, SearchRestriction _restr, bool _replace_all)
3751 : find_buf_name(_find_buf_name), casesensitive(_casesensitive), matchword(_matchword),
3752 forward(_forward), expandmacros(_expandmacros), ignoreformat(_ignoreformat),
3753 repl_buf_name(_repl_buf_name), keep_case(_keep_case), scope(_scope), restr(_restr), replace_all(_replace_all)
3761 /** Check if 'len' letters following cursor are all non-lowercase */
3762 static bool allNonLowercase(Cursor const & cur, int len)
3764 pos_type beg_pos = cur.selectionBegin().pos();
3765 pos_type end_pos = cur.selectionBegin().pos() + len;
3766 if (len > cur.lastpos() + 1 - beg_pos) {
3767 LYXERR(Debug::FIND, "This should not happen, more debug needed");
3768 len = cur.lastpos() + 1 - beg_pos;
3769 end_pos = beg_pos + len;
3771 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
3772 if (isLowerCase(cur.paragraph().getChar(pos)))
3778 /** Check if first letter is upper case and second one is lower case */
3779 static bool firstUppercase(Cursor const & cur)
3782 pos_type pos = cur.selectionBegin().pos();
3783 if (pos >= cur.lastpos() - 1) {
3784 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
3787 ch1 = cur.paragraph().getChar(pos);
3788 ch2 = cur.paragraph().getChar(pos + 1);
3789 bool result = isUpperCase(ch1) && isLowerCase(ch2);
3790 LYXERR(Debug::FIND, "firstUppercase(): "
3791 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
3792 << ch2 << "(" << char(ch2) << ")"
3793 << ", result=" << result << ", cur=" << cur);
3798 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
3800 ** \fixme What to do with possible further paragraphs in replace buffer ?
3802 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
3804 ParagraphList::iterator pit = buffer.paragraphs().begin();
3805 LASSERT(!pit->empty(), /**/);
3806 pos_type right = pos_type(1);
3807 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
3808 right = pit->size();
3809 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
3815 static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
3817 Cursor & cur = bv->cursor();
3818 if (opt.repl_buf_name.empty()
3819 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
3820 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3823 DocIterator sel_beg = cur.selectionBegin();
3824 DocIterator sel_end = cur.selectionEnd();
3825 if (&sel_beg.inset() != &sel_end.inset()
3826 || sel_beg.pit() != sel_end.pit()
3827 || sel_beg.idx() != sel_end.idx())
3829 int sel_len = sel_end.pos() - sel_beg.pos();
3830 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
3831 << ", sel_len: " << sel_len << endl);
3834 LASSERT(sel_len > 0, return 0);
3836 if (!matchAdv(sel_beg, sel_len).match_len)
3839 // Build a copy of the replace buffer, adapted to the KeepCase option
3840 Buffer const & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
3842 repl_buffer_orig.write(oss);
3843 string lyx = oss.str();
3844 Buffer repl_buffer("", false);
3845 repl_buffer.setUnnamed(true);
3846 LASSERT(repl_buffer.readString(lyx), return 0);
3847 if (opt.keep_case && sel_len >= 2) {
3848 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
3849 if (cur.inTexted()) {
3850 if (firstUppercase(cur))
3851 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
3852 else if (allNonLowercase(cur, sel_len))
3853 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
3856 cap::cutSelection(cur, false);
3857 if (cur.inTexted()) {
3858 repl_buffer.changeLanguage(
3859 repl_buffer.language(),
3860 cur.getFont().language());
3861 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
3862 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
3863 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
3864 repl_buffer.params().documentClassPtr(),
3865 bv->buffer().errorList("Paste"));
3866 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
3867 sel_len = repl_buffer.paragraphs().begin()->size();
3868 } else if (cur.inMathed()) {
3869 odocstringstream ods;
3871 // OutputParams runparams(&repl_buffer.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 runparams.dryrun = true;
3877 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
3878 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
3879 docstring repl_latex = ods.str();
3880 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
3882 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
3883 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
3884 repl_latex = from_utf8(s);
3885 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
3886 MathData ar(cur.buffer());
3887 asArray(repl_latex, ar, Parse::NORMAL);
3889 sel_len = ar.size();
3890 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3892 if (cur.pos() >= sel_len)
3893 cur.pos() -= sel_len;
3896 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3897 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
3898 bv->processUpdateFlags(Update::Force);
3903 /// Perform a FindAdv operation.
3904 bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt)
3909 // e.g., when invoking word-findadv from mini-buffer wither with
3910 // wrong options syntax or before ever opening advanced F&R pane
3911 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3915 MatchStringAdv matchAdv(bv->buffer(), opt);
3916 #if (QT_VERSION >= 0x050000)
3917 if (!matchAdv.regexIsValid) {
3918 bv->message(lyx::from_utf8(matchAdv.regexError));
3922 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
3924 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
3925 num_replaced += findAdvReplace(bv, opt, matchAdv);
3928 match_len = findForwardAdv(cur, matchAdv);
3930 match_len = findBackwardsAdv(cur, matchAdv);
3932 // This may only be raised by lyx::regex()
3933 bv->message(_("Invalid regular expression!"));
3937 if (match_len == 0) {
3938 if (num_replaced > 0) {
3939 switch (num_replaced)
3942 bv->message(_("One match has been replaced."));
3945 bv->message(_("Two matches have been replaced."));
3948 bv->message(bformat(_("%1$d matches have been replaced."), num_replaced));
3954 bv->message(_("Match not found."));
3959 if (num_replaced > 0)
3960 bv->message(_("Match has been replaced."));
3962 bv->message(_("Match found."));
3964 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << match_len);
3965 bv->putSelectionAt(cur, match_len, !opt.forward);
3971 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
3973 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
3974 << opt.casesensitive << ' '
3975 << opt.matchword << ' '
3976 << opt.forward << ' '
3977 << opt.expandmacros << ' '
3978 << opt.ignoreformat << ' '
3979 << opt.replace_all << ' '
3980 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
3981 << opt.keep_case << ' '
3982 << int(opt.scope) << ' '
3985 LYXERR(Debug::FIND, "built: " << os.str());
3991 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
3993 LYXERR(Debug::FIND, "parsing");
3997 while (line != "EOSS") {
4001 if (is.eof()) // Tolerate malformed request
4005 LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
4006 opt.find_buf_name = from_utf8(s);
4007 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.replace_all;
4008 is.get(); // Waste space before replace string
4011 while (line != "EOSS") {
4015 if (is.eof()) // Tolerate malformed request
4019 LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
4020 opt.repl_buf_name = from_utf8(s);
4021 is >> opt.keep_case;
4024 opt.scope = FindAndReplaceOptions::SearchScope(i);
4026 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
4028 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
4029 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
4030 << opt.scope << ' ' << opt.restr);