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"
56 using namespace lyx::support;
61 // Helper class for deciding what should be ignored
65 IgnoreFormats() = default;
67 bool getFamily() const { return ignoreFamily_; }
69 bool getSeries() const { return ignoreSeries_; }
71 bool getShape() const { return ignoreShape_; }
73 bool getUnderline() const { return ignoreUnderline_; }
75 bool getMarkUp() const { return ignoreMarkUp_; }
77 bool getStrikeOut() const { return ignoreStrikeOut_; }
79 bool getSectioning() const { return ignoreSectioning_; }
81 bool getFrontMatter() const { return ignoreFrontMatter_; }
83 bool getColor() const { return ignoreColor_; }
85 bool getLanguage() const { return ignoreLanguage_; }
87 void setIgnoreFormat(string const & type, bool value);
91 bool ignoreFamily_ = false;
93 bool ignoreSeries_ = false;
95 bool ignoreShape_ = false;
97 bool ignoreUnderline_ = false;
99 bool ignoreMarkUp_ = false;
101 bool ignoreStrikeOut_ = false;
103 bool ignoreSectioning_ = false;
105 bool ignoreFrontMatter_ = false;
107 bool ignoreColor_ = false;
109 bool ignoreLanguage_ = false;
113 void IgnoreFormats::setIgnoreFormat(string const & type, bool value)
115 if (type == "color") {
116 ignoreColor_ = value;
118 else if (type == "language") {
119 ignoreLanguage_ = value;
121 else if (type == "sectioning") {
122 ignoreSectioning_ = value;
123 ignoreFrontMatter_ = value;
125 else if (type == "font") {
126 ignoreSeries_ = value;
127 ignoreShape_ = value;
128 ignoreFamily_ = value;
130 else if (type == "series") {
131 ignoreSeries_ = value;
133 else if (type == "shape") {
134 ignoreShape_ = value;
136 else if (type == "family") {
137 ignoreFamily_ = value;
139 else if (type == "markup") {
140 ignoreMarkUp_ = value;
142 else if (type == "underline") {
143 ignoreUnderline_ = value;
145 else if (type == "strike") {
146 ignoreStrikeOut_ = value;
150 // The global variable that can be changed from outside
151 IgnoreFormats ignoreFormats;
154 void setIgnoreFormat(string const & type, bool value)
156 ignoreFormats.setIgnoreFormat(type, value);
162 bool parse_bool(docstring & howto)
167 howto = split(howto, var, ' ');
175 MatchString(docstring const & s, bool cs, bool mw)
176 : str(s), case_sens(cs), whole_words(mw)
179 // returns true if the specified string is at the specified position
180 // del specifies whether deleted strings in ct mode will be considered
181 int operator()(Paragraph const & par, pos_type pos, bool del = true) const
183 return par.find(str, case_sens, whole_words, pos, del);
191 // match whole words only
196 int findForward(DocIterator & cur, MatchString const & match,
197 bool find_del = true)
199 for (; cur; cur.forwardChar())
200 if (cur.inTexted()) {
201 int len = match(cur.paragraph(), cur.pos(), find_del);
209 int findBackwards(DocIterator & cur, MatchString const & match,
210 bool find_del = true)
214 if (cur.inTexted()) {
215 int len = match(cur.paragraph(), cur.pos(), find_del);
224 bool searchAllowed(docstring const & str)
227 frontend::Alert::error(_("Search error"), _("Search string is empty"));
234 bool findOne(BufferView * bv, docstring const & searchstr,
235 bool case_sens, bool whole, bool forward,
236 bool find_del = true, bool check_wrap = false)
238 if (!searchAllowed(searchstr))
241 DocIterator cur = forward
242 ? bv->cursor().selectionEnd()
243 : bv->cursor().selectionBegin();
245 MatchString const match(searchstr, case_sens, whole);
247 int match_len = forward
248 ? findForward(cur, match, find_del)
249 : findBackwards(cur, match, find_del);
252 bv->putSelectionAt(cur, match_len, !forward);
253 else if (check_wrap) {
254 DocIterator cur_orig(bv->cursor());
257 q = _("End of file reached while searching forward.\n"
258 "Continue searching from the beginning?");
260 q = _("Beginning of file reached while searching backward.\n"
261 "Continue searching from the end?");
262 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
263 q, 0, 1, _("&Yes"), _("&No"));
264 if (wrap_answer == 0) {
266 bv->cursor().clear();
267 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
269 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
270 bv->cursor().backwardPos();
272 bv->clearSelection();
273 if (findOne(bv, searchstr, case_sens, whole, forward, find_del, false))
276 bv->cursor().setCursor(cur_orig);
280 return match_len > 0;
284 int replaceAll(BufferView * bv,
285 docstring const & searchstr, docstring const & replacestr,
286 bool case_sens, bool whole)
288 Buffer & buf = bv->buffer();
290 if (!searchAllowed(searchstr) || buf.isReadonly())
293 DocIterator cur_orig(bv->cursor());
295 MatchString const match(searchstr, case_sens, whole);
298 int const rsize = replacestr.size();
299 int const ssize = searchstr.size();
302 cur.setCursor(doc_iterator_begin(&buf));
303 int match_len = findForward(cur, match, false);
304 while (match_len > 0) {
305 // Backup current cursor position and font.
306 pos_type const pos = cur.pos();
307 Font const font = cur.paragraph().getFontSettings(buf.params(), pos);
309 int striked = ssize -
310 cur.paragraph().eraseChars(pos, pos + match_len,
311 buf.params().track_changes);
312 cur.paragraph().insert(pos, replacestr, font,
313 Change(buf.params().track_changes
315 : Change::UNCHANGED));
316 for (int i = 0; i < rsize + striked; ++i)
319 match_len = findForward(cur, match, false);
322 bv->putSelectionAt(doc_iterator_begin(&buf), 0, false);
324 cur_orig.fixIfBroken();
325 bv->setCursor(cur_orig);
331 // the idea here is that we are going to replace the string that
332 // is selected IF it is the search string.
333 // if there is a selection, but it is not the search string, then
334 // we basically ignore it. (FIXME We ought to replace only within
336 // if there is no selection, then:
337 // (i) if some search string has been provided, then we find it.
338 // (think of how the dialog works when you hit "replace" the
340 // (ii) if no search string has been provided, then we treat the
341 // word the cursor is in as the search string. (why? i have no
342 // idea.) but this only works in text?
344 // returns the number of replacements made (one, if any) and
345 // whether anything at all was done.
346 pair<bool, int> replaceOne(BufferView * bv, docstring searchstr,
347 docstring const & replacestr, bool case_sens,
348 bool whole, bool forward, bool findnext)
350 Cursor & cur = bv->cursor();
351 if (!cur.selection()) {
352 // no selection, non-empty search string: find it
353 if (!searchstr.empty()) {
354 bool const found = findOne(bv, searchstr, case_sens, whole, forward, true, findnext);
355 return make_pair(found, 0);
357 // empty search string
360 return make_pair(false, 0);
361 // select current word and treat it as the search string.
362 // This causes a minor bug as undo will restore this selection,
363 // which the user did not create (#8986).
364 cur.innerText()->selectWord(cur, WHOLE_WORD);
365 searchstr = cur.selectionAsString(false);
368 // if we still don't have a search string, report the error
370 if (!searchAllowed(searchstr))
371 return make_pair(false, 0);
373 bool have_selection = cur.selection();
374 docstring const selected = cur.selectionAsString(false);
377 ? searchstr == selected
378 : compare_no_case(searchstr, selected) == 0;
380 // no selection or current selection is not search word:
381 // just find the search word
382 if (!have_selection || !match) {
383 bool const found = findOne(bv, searchstr, case_sens, whole, forward, true, findnext);
384 return make_pair(found, 0);
387 // we're now actually ready to replace. if the buffer is
388 // read-only, we can't, though.
389 if (bv->buffer().isReadonly())
390 return make_pair(false, 0);
392 cap::replaceSelectionWithString(cur, replacestr);
394 cur.pos() += replacestr.length();
395 LASSERT(cur.pos() <= cur.lastpos(),
396 cur.pos() = cur.lastpos());
399 findOne(bv, searchstr, case_sens, whole, forward, false, findnext);
401 return make_pair(true, 1);
407 docstring const find2string(docstring const & search,
408 bool casesensitive, bool matchword, bool forward)
412 << int(casesensitive) << ' '
413 << int(matchword) << ' '
419 docstring const replace2string(docstring const & replace,
420 docstring const & search,
421 bool casesensitive, bool matchword,
422 bool all, bool forward, bool findnext)
425 ss << replace << '\n'
427 << int(casesensitive) << ' '
428 << int(matchword) << ' '
430 << int(forward) << ' '
436 bool lyxfind(BufferView * bv, FuncRequest const & ev)
438 if (!bv || ev.action() != LFUN_WORD_FIND)
441 //lyxerr << "find called, cmd: " << ev << endl;
443 // data is of the form
445 // <casesensitive> <matchword> <forward>"
447 docstring howto = split(ev.argument(), search, '\n');
449 bool casesensitive = parse_bool(howto);
450 bool matchword = parse_bool(howto);
451 bool forward = parse_bool(howto);
453 return findOne(bv, search, casesensitive, matchword, forward, true, true);
457 bool lyxreplace(BufferView * bv,
458 FuncRequest const & ev, bool has_deleted)
460 if (!bv || ev.action() != LFUN_WORD_REPLACE)
463 // data is of the form
466 // <casesensitive> <matchword> <all> <forward> <findnext>"
469 docstring howto = split(ev.argument(), rplc, '\n');
470 howto = split(howto, search, '\n');
472 bool casesensitive = parse_bool(howto);
473 bool matchword = parse_bool(howto);
474 bool all = parse_bool(howto);
475 bool forward = parse_bool(howto);
476 bool findnext = howto.empty() ? true : parse_bool(howto);
481 int replace_count = 0;
483 replace_count = replaceAll(bv, search, rplc, casesensitive, matchword);
484 update = replace_count > 0;
487 replaceOne(bv, search, rplc, casesensitive, matchword, forward, findnext);
489 replace_count = rv.second;
492 Buffer const & buf = bv->buffer();
494 // emit message signal.
495 buf.message(_("String not found."));
497 if (replace_count == 0) {
498 buf.message(_("String found."));
499 } else if (replace_count == 1) {
500 buf.message(_("String has been replaced."));
502 docstring const str =
503 bformat(_("%1$d strings have been replaced."), replace_count);
507 } else if (findnext) {
508 // if we have deleted characters, we do not replace at all, but
509 // rather search for the next occurence
510 if (findOne(bv, search, casesensitive, matchword, forward, true, findnext))
513 bv->message(_("String not found."));
519 bool findNextChange(BufferView * bv, Cursor & cur, bool const check_wrap)
521 for (; cur; cur.forwardPos())
522 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
526 DocIterator cur_orig(bv->cursor());
527 docstring q = _("End of file reached while searching forward.\n"
528 "Continue searching from the beginning?");
529 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
530 q, 0, 1, _("&Yes"), _("&No"));
531 if (wrap_answer == 0) {
532 bv->cursor().clear();
533 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
534 bv->clearSelection();
535 cur.setCursor(bv->cursor().selectionBegin());
536 if (findNextChange(bv, cur, false))
539 bv->cursor().setCursor(cur_orig);
546 bool findPreviousChange(BufferView * bv, Cursor & cur, bool const check_wrap)
548 for (cur.backwardPos(); cur; cur.backwardPos()) {
549 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
554 DocIterator cur_orig(bv->cursor());
555 docstring q = _("Beginning of file reached while searching backward.\n"
556 "Continue searching from the end?");
557 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
558 q, 0, 1, _("&Yes"), _("&No"));
559 if (wrap_answer == 0) {
560 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
561 bv->cursor().backwardPos();
562 bv->clearSelection();
563 cur.setCursor(bv->cursor().selectionBegin());
564 if (findPreviousChange(bv, cur, false))
567 bv->cursor().setCursor(cur_orig);
574 bool selectChange(Cursor & cur, bool forward)
576 if (!cur.inTexted() || !cur.paragraph().isChanged(cur.pos()))
578 Change ch = cur.paragraph().lookupChange(cur.pos());
580 CursorSlice tip1 = cur.top();
581 for (; tip1.pit() < tip1.lastpit() || tip1.pos() < tip1.lastpos(); tip1.forwardPos()) {
582 Change ch2 = tip1.paragraph().lookupChange(tip1.pos());
583 if (!ch2.isSimilarTo(ch))
586 CursorSlice tip2 = cur.top();
587 for (; tip2.pit() > 0 || tip2.pos() > 0;) {
589 Change ch2 = tip2.paragraph().lookupChange(tip2.pos());
590 if (!ch2.isSimilarTo(ch)) {
591 // take a step forward to correctly set the selection
599 cur.bv().mouseSetCursor(cur, false);
601 cur.bv().mouseSetCursor(cur, true);
609 bool findChange(BufferView * bv, bool forward)
612 cur.setCursor(forward ? bv->cursor().selectionEnd()
613 : bv->cursor().selectionBegin());
614 forward ? findNextChange(bv, cur, true) : findPreviousChange(bv, cur, true);
615 return selectChange(cur, forward);
620 bool findNextChange(BufferView * bv)
622 return findChange(bv, true);
626 bool findPreviousChange(BufferView * bv)
628 return findChange(bv, false);
635 typedef vector<pair<string, string> > Escapes;
637 /// A map of symbols and their escaped equivalent needed within a regex.
638 /// @note Beware of order
639 Escapes const & get_regexp_escapes()
641 typedef std::pair<std::string, std::string> P;
643 static Escapes escape_map;
644 if (escape_map.empty()) {
645 escape_map.push_back(P("$", "_x_$"));
646 escape_map.push_back(P("{", "_x_{"));
647 escape_map.push_back(P("}", "_x_}"));
648 escape_map.push_back(P("[", "_x_["));
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("\\", "(?:\\\\|\\\\backslash)"));
656 escape_map.push_back(P("~", "(?:\\\\textasciitilde|\\\\sim)"));
657 escape_map.push_back(P("^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\textasciicircum|\\\\mathcircumflex)"));
658 escape_map.push_back(P("_x_", "\\"));
663 /// A map of lyx escaped strings and their unescaped equivalent.
664 Escapes const & get_lyx_unescapes()
666 typedef std::pair<std::string, std::string> P;
668 static Escapes escape_map;
669 if (escape_map.empty()) {
670 escape_map.push_back(P("\\%", "%"));
671 escape_map.push_back(P("\\{", "{"));
672 escape_map.push_back(P("\\}", "}"));
673 escape_map.push_back(P("\\mathcircumflex ", "^"));
674 escape_map.push_back(P("\\mathcircumflex", "^"));
675 escape_map.push_back(P("\\backslash ", "\\"));
676 escape_map.push_back(P("\\backslash", "\\"));
677 escape_map.push_back(P("\\sim ", "~"));
678 escape_map.push_back(P("\\sim", "~"));
683 /// A map of escapes turning a regexp matching text to one matching latex.
684 Escapes const & get_regexp_latex_escapes()
686 typedef std::pair<std::string, std::string> P;
688 static Escapes escape_map;
689 if (escape_map.empty()) {
690 escape_map.push_back(P("\\\\", "(?:\\\\\\\\|\\\\backslash|\\\\textbackslash\\{\\}|\\\\textbackslash)"));
691 escape_map.push_back(P("(<?!\\\\\\\\textbackslash)\\{", "\\\\\\{"));
692 escape_map.push_back(P("(<?!\\\\\\\\textbackslash\\\\\\{)\\}", "\\\\\\}"));
693 escape_map.push_back(P("\\[", "\\{\\[\\}"));
694 escape_map.push_back(P("\\]", "\\{\\]\\}"));
695 escape_map.push_back(P("\\^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\textasciicircum|\\\\mathcircumflex)"));
696 escape_map.push_back(P("%", "\\\\\\%"));
697 escape_map.push_back(P("#", "\\\\#"));
702 /** @todo Probably the maps need to be migrated to regexps, in order to distinguish if
703 ** the found occurrence were escaped.
705 string apply_escapes(string s, Escapes const & escape_map)
707 LYXERR(Debug::FIND, "Escaping: '" << s << "'");
708 Escapes::const_iterator it;
709 for (it = escape_map.begin(); it != escape_map.end(); ++it) {
710 // LYXERR(Debug::FIND, "Escaping " << it->first << " as " << it->second);
711 unsigned int pos = 0;
712 while (pos < s.length() && (pos = s.find(it->first, pos)) < s.length()) {
713 s.replace(pos, it->first.length(), it->second);
714 LYXERR(Debug::FIND, "After escape: " << s);
715 pos += it->second.length();
716 // LYXERR(Debug::FIND, "pos: " << pos);
719 LYXERR(Debug::FIND, "Escaped : '" << s << "'");
724 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
725 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
726 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
727 string escape_for_regex(string s, bool match_latex)
730 while (pos < s.size()) {
731 size_t new_pos = s.find("\\regexp{", pos);
732 if (new_pos == string::npos)
736 LYXERR(Debug::FIND, "new_pos: " << new_pos);
737 t = apply_escapes(s.substr(pos, new_pos - pos), get_lyx_unescapes());
738 LYXERR(Debug::FIND, "t [lyx]: " << t);
739 t = apply_escapes(t, get_regexp_escapes());
740 LYXERR(Debug::FIND, "t [rxp]: " << t);
741 s.replace(pos, new_pos - pos, t);
742 new_pos = pos + t.size();
743 LYXERR(Debug::FIND, "Regexp after escaping: " << s);
744 LYXERR(Debug::FIND, "new_pos: " << new_pos);
745 if (new_pos == s.size())
748 // Might fail if \\endregexp{} is preceeded by unexpected stuff (weird escapes)
749 size_t end_pos = s.find("\\endregexp{}}", new_pos + 8);
750 LYXERR(Debug::FIND, "end_pos: " << end_pos);
751 t = s.substr(new_pos + 8, end_pos - (new_pos + 8));
752 LYXERR(Debug::FIND, "t in regexp : " << t);
753 t = apply_escapes(t, get_lyx_unescapes());
754 LYXERR(Debug::FIND, "t in regexp after unescapes [lyx]: " << t);
756 t = apply_escapes(t, get_regexp_latex_escapes());
757 LYXERR(Debug::FIND, "t in regexp after latex_escapes [ltx]: " << t);
759 if (end_pos == s.size()) {
760 s.replace(new_pos, end_pos - new_pos, t);
761 LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s);
764 s.replace(new_pos, end_pos + 13 - new_pos, t);
765 LYXERR(Debug::FIND, "Regexp after \\regexp{...\\endregexp{}} removal: " << s);
766 pos = new_pos + t.size();
767 LYXERR(Debug::FIND, "pos: " << pos);
773 /// Wrapper for lyx::regex_replace with simpler interface
774 bool regex_replace(string const & s, string & t, string const & searchstr,
775 string const & replacestr)
777 regex e(searchstr, regex_constants::ECMAScript);
779 ostream_iterator<char, char> it(oss);
780 regex_replace(it, s.begin(), s.end(), e, replacestr);
781 // tolerate t and s be references to the same variable
782 bool rv = (s != oss.str());
788 /** Checks if supplied string segment is well-formed from the standpoint of matching open-closed braces.
790 ** Verify that closed braces exactly match open braces. This avoids that, for example,
791 ** \frac{.*}{x} matches \frac{x+\frac{y}{x}}{z} with .* being 'x+\frac{y'.
794 ** Number of open braces that must remain open at the end for the verification to succeed.
796 bool braces_match(string::const_iterator const & beg,
797 string::const_iterator const & end,
801 string::const_iterator it = beg;
802 LYXERR(Debug::FIND, "Checking " << unmatched << " unmatched braces in '" << string(beg, end) << "'");
803 for (; it != end; ++it) {
804 // Skip escaped braces in the count
809 } else if (*it == '{') {
811 } else if (*it == '}') {
812 if (open_pars == 0) {
813 LYXERR(Debug::FIND, "Found unmatched closed brace");
819 if (open_pars != unmatched) {
820 LYXERR(Debug::FIND, "Found " << open_pars
821 << " instead of " << unmatched
822 << " unmatched open braces at the end of count");
825 LYXERR(Debug::FIND, "Braces match as expected");
835 MatchResult(): match_len(0),match2end(0), pos(0) {};
838 /** The class performing a match between a position in the document and the FindAdvOptions.
840 class MatchStringAdv {
842 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt);
844 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
845 ** constructor as opt.search, under the opt.* options settings.
848 ** If set, then match is searched only against beginning of text starting at cur.
849 ** If unset, then match is searched anywhere in text starting at cur.
852 ** The length of the matching text, or zero if no match was found.
854 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
859 /// first buffer on which search was started
860 lyx::Buffer * const p_first_buf;
862 FindAndReplaceOptions const & opt;
865 /// Auxiliary find method (does not account for opt.matchword)
866 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
868 /** Normalize a stringified or latexified LyX paragraph.
872 ** <li>if search is not casesensitive, then lowercase the string;
873 ** <li>remove any newline at begin or end of the string;
874 ** <li>replace any newline in the middle of the string with a simple space;
875 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
878 ** @todo Normalization should also expand macros, if the corresponding
879 ** search option was checked.
881 string normalize(docstring const & s, bool hack_braces) const;
882 // normalized string to search
883 string par_as_string;
884 // regular expression to use for searching
886 // same as regexp, but prefixed with a ".*?"
888 // leading format material as string
889 string lead_as_string;
890 // par_as_string after removal of lead_as_string
891 string par_as_string_nolead;
892 // unmatched open braces in the search string/regexp
894 // number of (.*?) subexpressions added at end of search regexp for closing
895 // environments, math mode, styles, etc...
897 // Are we searching with regular expressions ?
902 static docstring buffer_to_latex(Buffer & buffer)
904 //OutputParams runparams(&buffer.params().encoding());
905 OutputParams runparams(encodings.fromLyXName("utf8"));
906 odocstringstream ods;
908 runparams.nice = true;
909 runparams.flavor = Flavor::XeTeX;
910 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
911 // No side effect of file copying and image conversion
912 runparams.dryrun = true;
913 runparams.for_search = true;
914 pit_type const endpit = buffer.paragraphs().size();
915 for (pit_type pit = 0; pit != endpit; ++pit) {
916 TeXOnePar(buffer, buffer.text(), pit, os, runparams);
917 LYXERR(Debug::FIND, "searchString up to here: " << ods.str());
923 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
926 if (!opt.ignoreformat) {
927 str = buffer_to_latex(buffer);
929 // OutputParams runparams(&buffer.params().encoding());
930 OutputParams runparams(encodings.fromLyXName("utf8"));
931 runparams.nice = true;
932 runparams.flavor = Flavor::XeTeX;
933 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
934 runparams.dryrun = true;
935 runparams.for_search = true;
936 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
937 Paragraph const & par = buffer.paragraphs().at(pit);
938 LYXERR(Debug::FIND, "Adding to search string: '"
939 << par.asString(pos_type(0), par.size(),
940 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
943 str += par.asString(pos_type(0), par.size(),
944 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
947 // Even in ignore-format we have to remove "\text{}, \lyxmathsym{}" parts
948 string t = to_utf8(str);
949 while (regex_replace(t, t, "\\\\(text|lyxmathsym)\\{([^\\}]*)\\}", "$2"));
956 /// Return separation pos between the leading material and the rest
957 static size_t identifyLeading(string const & s)
960 // @TODO Support \item[text]
961 // Kornel: Added textsl, textsf, textit, texttt and noun
962 // + allow to search for colored text too
963 while (regex_replace(t, t, "^\\\\(("
964 "(author|title|subtitle|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|"
965 "lyxaddress|lyxrightaddress|"
966 "footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
967 "emph|noun|minisec|text(bf|md|sl|sf|it|tt))|"
968 "((textcolor|foreignlanguage|latexenvironment)\\{[a-z]+\\*?\\})|"
969 "(u|uu)line|(s|x)out|uwave)|((sub)?(((sub)?section)|paragraph)|part|chapter)\\*?)\\{", "")
970 || regex_replace(t, t, "^\\$", "")
971 || regex_replace(t, t, "^\\\\\\[", "")
972 || regex_replace(t, t, "^ ?\\\\item\\{[a-z]+\\}", "")
973 || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\}", ""))
975 LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
980 * Given a latexified string, retrieve some handled features
981 * The features of the regex will later be compared with the features
982 * of the searched text. If the regex features are not a
983 * subset of the analized, then, in not format ignoring search
984 * we can early stop the search in the relevant inset.
986 typedef map<string, bool> Features;
988 static Features identifyFeatures(string const & s)
990 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
991 static regex const valid("^("
993 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
994 "emph|noun|text(bf|md|sl|sf|it|tt)|"
995 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
996 "(u|uu)line|(s|x)out|uwave|"
997 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
998 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
1003 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1006 if (sub.str(1).compare("regexp") == 0) {
1010 string token = sub.str(1);
1012 if (regex_match(token, sub2, valid)) {
1020 if (sub.str(1).compare("endregexp") == 0) {
1030 * defines values features of a key "\\[a-z]+{"
1035 /* Char type with content discarded
1036 * like \hspace{1cm} */
1038 /* Char, like \backslash */
1040 /* \part, \section*, ... */
1042 /* title, author etc */
1044 /* \foreignlanguage{ngerman}, ... */
1047 * to discard language in content */
1050 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1052 /* fonts, colors, markups, ... */
1054 /* footnotesize, ... large, ...
1055 * Ignore all of them */
1058 /* inputencoding, ...
1059 * Discard also content, because they do not help in search */
1062 * like remove, but also all arguments */
1064 /* item, listitem */
1066 /* tex, latex, ... like isChar */
1068 /* like \lettrine[lines=5]{}{} */
1070 // like isStandard, but always remove head
1072 /* End of arguments marker for lettrine,
1073 * so that they can be ignored */
1076 KeyInfo() = default;
1077 KeyInfo(KeyType type, int parcount, bool disable)
1079 parenthesiscount(parcount),
1080 disabled(disable) {}
1081 KeyType keytype = invalid;
1083 int _tokensize = -1;
1084 int _tokenstart = -1;
1085 int _dataStart = -1;
1087 int parenthesiscount = 1;
1088 bool disabled = false;
1089 bool used = false; /* by pattern */
1094 Border(int l=0, int u=0) : low(l), upper(u) {};
1099 #define MAXOPENED 30
1101 bool isPatternString_;
1103 explicit Intervall(bool isPattern, string const & p) :
1104 isPatternString_(isPattern), par(p), ignoreidx(-1), actualdeptindex(0),
1105 hasTitle(false), langcount(0)
1113 static vector<Border> borders;
1114 int depts[MAXOPENED];
1115 int closes[MAXOPENED];
1116 int actualdeptindex;
1117 int previousNotIgnored(int) const;
1118 int nextNotIgnored(int) const;
1119 void handleOpenP(int i);
1120 void handleCloseP(int i, bool closingAllowed);
1121 void resetOpenedP(int openPos);
1122 void addIntervall(int upper);
1123 void addIntervall(int low, int upper); /* if explicit */
1124 void removeAccents();
1125 void setForDefaultLang(KeyInfo const & defLang) const;
1126 int findclosing(int start, int end, char up, char down, int repeat);
1127 void handleParentheses(int lastpos, bool closingAllowed);
1129 int langcount; // Number of disabled language specs up to current position in actual interval
1130 int isOpeningPar(int pos) const;
1132 void output(ostringstream &os, int lastpos);
1133 // string show(int lastpos);
1136 vector<Border> Intervall::borders = vector<Border>(30);
1138 int Intervall::isOpeningPar(int pos) const
1140 if ((pos < 0) || (size_t(pos) >= par.size()))
1142 if (par[pos] != '{')
1144 if (size_t(pos) + 2 >= par.size())
1146 if (par[pos+2] != '}')
1148 if (par[pos+1] == '[' || par[pos+1] == ']')
1153 void Intervall::setForDefaultLang(KeyInfo const & defLang) const
1155 // Enable the use of first token again
1156 if (ignoreidx >= 0) {
1157 int value = defLang._tokenstart + defLang._tokensize;
1163 if (borders[borderidx].low < value)
1164 borders[borderidx].low = value;
1165 if (borders[borderidx].upper < value)
1166 borders[borderidx].upper = value;
1171 static void checkDepthIndex(int val)
1173 static int maxdepthidx = MAXOPENED-2;
1174 static int lastmaxdepth = 0;
1175 if (val > lastmaxdepth) {
1176 LYXERR(Debug::INFO, "Depth reached " << val);
1179 if (val > maxdepthidx) {
1181 LYXERR(Debug::INFO, "maxdepthidx now " << val);
1186 // Not needed, because borders are now dynamically expanded
1187 static void checkIgnoreIdx(int val)
1189 static int lastmaxignore = -1;
1190 if ((lastmaxignore < val) && (size_t(val+1) >= borders.size())) {
1191 LYXERR(Debug::INFO, "IgnoreIdx reached " << val);
1192 lastmaxignore = val;
1198 * Expand the region of ignored parts of the input latex string
1199 * The region is only relevant in output()
1201 void Intervall::addIntervall(int low, int upper)
1204 if (low == upper) return;
1205 for (idx = ignoreidx+1; idx > 0; --idx) {
1206 if (low > borders[idx-1].upper) {
1210 Border br(low, upper);
1211 if (idx > ignoreidx) {
1212 if (borders.size() <= size_t(idx)) {
1213 borders.push_back(br);
1219 // checkIgnoreIdx(ignoreidx);
1223 // Expand only if one of the new bound is inside the interwall
1224 // We know here that br.low > borders[idx-1].upper
1225 if (br.upper < borders[idx].low) {
1226 // We have to insert at this pos
1227 if (size_t(ignoreidx+1) >= borders.size()) {
1228 borders.push_back(borders[ignoreidx]);
1231 borders[ignoreidx+1] = borders[ignoreidx];
1233 for (int i = ignoreidx; i > idx; --i) {
1234 borders[i] = borders[i-1];
1238 // checkIgnoreIdx(ignoreidx);
1241 // Here we know, that we are overlapping
1242 if (br.low > borders[idx].low)
1243 br.low = borders[idx].low;
1244 // check what has to be concatenated
1246 for (int i = idx; i <= ignoreidx; i++) {
1247 if (br.upper >= borders[i].low) {
1249 if (br.upper < borders[i].upper)
1250 br.upper = borders[i].upper;
1256 // count should be >= 1 here
1259 for (int i = idx + count; i <= ignoreidx; i++) {
1260 borders[i-count+1] = borders[i];
1262 ignoreidx -= count - 1;
1268 typedef map<string, string> AccentsMap;
1269 static AccentsMap accents = map<string, string>();
1271 static void buildaccent(string n, string param, string values)
1275 const char delim = '|';
1276 while (getline(s, name, delim)) {
1278 for (char c : param) {
1279 string key = name + "{" + c + "}";
1280 // get the corresponding utf8-value
1281 if ((values[start] & 0xc0) != 0xc0) {
1282 // should not happen, utf8 encoding starts at least with 11xxxxxx
1283 // but value for '\dot{i}' is 'i', which is ascii
1284 if ((values[start] & 0x80) == 0) {
1286 accents[key] = values.substr(start, 1);
1287 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1292 for (int j = 1; ;j++) {
1293 if (start + j >= values.size()) {
1294 accents[key] = values.substr(start, j);
1295 start = values.size() - 1;
1298 else if ((values[start+j] & 0xc0) != 0x80) {
1299 // This is the first byte of following utf8 char
1300 accents[key] = values.substr(start, j);
1302 // LYXERR(Debug::INFO, "" << key << "=" << accents[key]);
1311 static string getutf8(unsigned uchar)
1314 string ret = string();
1316 if (uchar <= 0x7f) {
1317 c[maxc-1] = uchar & 0x7f;
1320 unsigned char rest = 0x40;
1321 unsigned char first = 0x80;
1323 if (uchar >= (1 << 11)) {
1324 // needed at least 3 bytes
1325 c[4] = (uchar & 0x0f) | 0x30;
1329 for (int i = start; i >=0; --i) {
1331 c[i] = first + uchar;
1334 c[i] = 0x80 | (uchar & 0x3f);
1341 for (int i = 0; i < maxc; i++) {
1342 if (c[i] == 0) continue;
1348 static void buildAccentsMap()
1350 accents["imath"] = "ı";
1352 accents["jmath"] = "ȷ";
1353 accents["cdot"] = "·";
1354 accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
1355 accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
1356 accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
1357 accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
1358 accents["negmedspace"] = getutf8(0xf0005);
1359 accents["thickspace"] = getutf8(0xf0006);
1360 accents["negthickspace"] = getutf8(0xf0007);
1361 accents["lyx"] = getutf8(0xf0010); // Used logos
1362 accents["LyX"] = getutf8(0xf0010);
1363 accents["tex"] = getutf8(0xf0011);
1364 accents["TeX"] = getutf8(0xf0011);
1365 accents["latex"] = getutf8(0xf0012);
1366 accents["LaTeX"] = getutf8(0xf0012);
1367 accents["latexe"] = getutf8(0xf0013);
1368 accents["LaTeXe"] = getutf8(0xf0013);
1369 accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash
1370 accents["backslash LyX"] = getutf8(0xf0010);
1371 accents["backslash tex"] = getutf8(0xf0011);
1372 accents["backslash TeX"] = getutf8(0xf0011);
1373 accents["backslash latex"] = getutf8(0xf0012);
1374 accents["backslash LaTeX"] = getutf8(0xf0012);
1375 accents["backslash latexe"] = getutf8(0xf0013);
1376 accents["backslash LaTeXe"] = getutf8(0xf0013);
1377 accents["ddot{\\imath}"] = "ï";
1378 buildaccent("ddot", "aAeEhHiIioOtuUwWxXyY",
1379 "äÄëËḧḦïÏïöÖẗüÜẅẄẍẌÿŸ"); // umlaut
1380 buildaccent("dot|.", "aAbBcCdDeEfFGghHIimMnNoOpPrRsStTwWxXyYzZ",
1381 "ȧȦḃḂċĊḋḊėĖḟḞĠġḣḢİİṁṀṅṄȯȮṗṖṙṘṡṠṫṪẇẆẋẊẏẎżŻ"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'İ'
1382 accents["acute{\\imath}"] = "í";
1383 buildaccent("acute", "aAcCeEgGkKlLmMoOnNpPrRsSuUwWyYzZiI",
1384 "áÁćĆéÉǵǴḱḰĺĹḿḾóÓńŃṕṔŕŔśŚúÚẃẂýÝźŹíÍ");
1385 buildaccent("dacute|H|h", "oOuU", "őŐűŰ"); // double acute
1386 buildaccent("mathring|r", "aAuUwy",
1388 accents["check{\\imath}"] = "ǐ";
1389 accents["check{\\jmath}"] = "ǰ";
1390 buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
1391 "čČďĎǎǍěĚǐǏǒǑǔǓǧǦǩǨȟȞľĽňŇřŘšŠŤťžŽ"); // caron
1392 accents["hat{\\imath}"] = "î";
1393 accents["hat{\\jmath}"] = "ĵ";
1394 buildaccent("hat|^", "aAcCeEgGhHiIjJoOsSuUwWyYzZ",
1395 "âÂĉĈêÊĝĜĥĤîÎĵĴôÔŝŜûÛŵŴŷŶẑẐ"); // circ
1396 accents["bar{\\imath}"] = "ī";
1397 buildaccent("bar|=", "aAeEiIoOuUyY",
1398 "āĀēĒīĪōŌūŪȳȲ"); // macron
1399 accents["tilde{\\imath}"] = "ĩ";
1400 buildaccent("tilde", "aAeEiInNoOuUvVyY",
1401 "ãÃẽẼĩĨñÑõÕũŨṽṼỹỸ"); // tilde
1402 accents["breve{\\imath}"] = "ĭ";
1403 buildaccent("breve|u", "aAeEgGiIoOuU",
1404 "ăĂĕĔğĞĭĬŏŎŭŬ"); // breve
1405 accents["grave{\\imath}"] = "ì";
1406 buildaccent("grave|`", "aAeEiIoOuUnNwWyY",
1407 "àÀèÈìÌòÒùÙǹǸẁẀỳỲ"); // grave
1408 buildaccent("subdot|d", "BbDdHhKkLlMmNnRrSsTtVvWwZzAaEeIiOoUuYy",
1409 "ḄḅḌḍḤḥḲḳḶḷṂṃṆṇṚṛṢṣṬṭṾṿẈẉẒẓẠạẸẹỊịỌọỤụỴỵ"); // dot below
1410 buildaccent("ogonek|k", "AaEeIiUuOo",
1411 "ĄąĘęĮįŲųǪǫ"); // ogonek
1412 buildaccent("cedilla|c", "CcGgKkLlNnRrSsTtEeDdHh",
1413 "ÇçĢģĶķĻļŅņŖŗŞşŢţȨȩḐḑḨḩ"); // cedilla
1414 buildaccent("subring|textsubring", "Aa",
1416 buildaccent("subhat|textsubcircum", "DdEeLlNnTtUu",
1417 "ḒḓḘḙḼḽṊṋṰṱṶṷ"); // subcircum
1418 buildaccent("subtilde|textsubtilde", "EeIiUu",
1419 "ḚḛḬḭṴṵ"); // subtilde
1420 accents["dgrave{\\imath}"] = "ȉ";
1421 accents["textdoublegrave{\\i}"] = "ȉ";
1422 buildaccent("dgrave|textdoublegrave", "AaEeIiOoRrUu",
1423 "ȀȁȄȅȈȉȌȍȐȑȔȕ"); // double grave
1424 accents["rcap{\\imath}"] = "ȉ";
1425 accents["textroundcap{\\i}"] = "ȉ";
1426 buildaccent("rcap|textroundcap", "AaEeIiOoRrUu",
1427 "ȂȃȆȇȊȋȎȏȒȓȖȗ"); // inverted breve
1428 buildaccent("slashed", "oO",
1433 * Created accents in math or regexp environment
1434 * are macros, but we need the utf8 equivalent
1436 void Intervall::removeAccents()
1438 if (accents.empty())
1440 static regex const accre("\\\\(([\\S]|grave|breve|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot|ogonek|"
1441 "cedilla|subring|textsubring|subhat|textsubcircum|subtilde|textsubtilde|dgrave|textdoublegrave|rcap|textroundcap|slashed)\\{[^\\{\\}]+\\}"
1442 "|((i|imath|jmath|cdot|[a-z]+space)|((backslash )?([lL]y[xX]|[tT]e[xX]|[lL]a[tT]e[xX]e?)))(?![a-zA-Z]))");
1444 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
1446 string key = sub.str(1);
1447 if (accents.find(key) != accents.end()) {
1448 string val = accents[key];
1449 size_t pos = sub.position(size_t(0));
1450 for (size_t i = 0; i < val.size(); i++) {
1451 par[pos+i] = val[i];
1453 // Remove possibly following space too
1454 if (par[pos+sub.str(0).size()] == ' ')
1455 addIntervall(pos+val.size(), pos + sub.str(0).size()+1);
1457 addIntervall(pos+val.size(), pos + sub.str(0).size());
1458 for (size_t i = pos+val.size(); i < pos + sub.str(0).size(); i++) {
1459 // remove traces of any remaining chars
1464 LYXERR(Debug::INFO, "Not added accent for \"" << key << "\"");
1469 void Intervall::handleOpenP(int i)
1472 depts[actualdeptindex] = i+1;
1473 closes[actualdeptindex] = -1;
1474 checkDepthIndex(actualdeptindex);
1477 void Intervall::handleCloseP(int i, bool closingAllowed)
1479 if (actualdeptindex <= 0) {
1480 if (! closingAllowed)
1481 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1482 // if we are at the very end
1483 addIntervall(i, i+1);
1486 closes[actualdeptindex] = i+1;
1491 void Intervall::resetOpenedP(int openPos)
1493 // Used as initializer for foreignlanguage entry
1494 actualdeptindex = 1;
1495 depts[1] = openPos+1;
1499 int Intervall::previousNotIgnored(int start) const
1501 int idx = 0; /* int intervalls */
1502 for (idx = ignoreidx; idx >= 0; --idx) {
1503 if (start > borders[idx].upper)
1505 if (start >= borders[idx].low)
1506 start = borders[idx].low-1;
1511 int Intervall::nextNotIgnored(int start) const
1513 int idx = 0; /* int intervalls */
1514 for (idx = 0; idx <= ignoreidx; idx++) {
1515 if (start < borders[idx].low)
1517 if (start < borders[idx].upper)
1518 start = borders[idx].upper;
1523 typedef map<string, KeyInfo> KeysMap;
1524 typedef vector< KeyInfo> Entries;
1525 static KeysMap keys = map<string, KeyInfo>();
1531 Intervall interval_;
1532 void buildKeys(bool);
1533 void buildEntries(bool);
1534 void makeKey(const string &, KeyInfo, bool isPatternString);
1535 void processRegion(int start, int region_end); /* remove {} parts */
1536 void removeHead(KeyInfo const &, int count=0);
1539 LatexInfo(string const & par, bool isPatternString)
1540 : entidx_(-1), interval_(isPatternString, par)
1542 buildKeys(isPatternString);
1543 entries_ = vector<KeyInfo>();
1544 buildEntries(isPatternString);
1548 if (entries_.empty()) {
1551 if (entries_[0].keytype == KeyInfo::isTitle) {
1552 interval_.hasTitle = true;
1553 if (! entries_[0].disabled) {
1554 interval_.titleValue = entries_[0].head;
1557 interval_.titleValue = "";
1559 removeHead(entries_[0]);
1560 if (entries_.size() > 1)
1569 if (int(entries_.size()) > entidx_) {
1576 bool setNextKey(int idx) {
1577 if ((idx == entidx_) && (entidx_ >= 0)) {
1584 int find(int start, KeyInfo::KeyType keytype) const {
1588 while (tmpIdx < int(entries_.size())) {
1589 if (entries_[tmpIdx].keytype == keytype)
1595 int process(ostringstream & os, KeyInfo const & actual);
1596 int dispatch(ostringstream & os, int previousStart, KeyInfo & actual);
1597 // string show(int lastpos) { return interval.show(lastpos);};
1598 int nextNotIgnored(int start) { return interval_.nextNotIgnored(start);};
1599 KeyInfo &getKeyInfo(int keyinfo) {
1600 static KeyInfo invalidInfo = KeyInfo();
1601 if ((keyinfo < 0) || ( keyinfo >= int(entries_.size())))
1604 return entries_[keyinfo];
1606 void setForDefaultLang(KeyInfo const & defLang) {interval_.setForDefaultLang(defLang);};
1607 void addIntervall(int low, int up) { interval_.addIntervall(low, up); };
1611 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
1615 for (int i = start; i < end; i += 1 + skip) {
1619 if (c == '\\') skip = 1;
1623 else if (c == down) {
1626 if ((repeat <= 0) || (par[i+1] != up))
1644 vector<MathEntry> entries_;
1649 void insert(string const & wait, size_t start, size_t end) {
1650 MathEntry m = MathEntry();
1652 m.mathStart = start;
1654 m.mathSize = end - start;
1655 entries_.push_back(m);
1657 bool empty() const { return entries_.empty(); };
1658 size_t getEndPos() const {
1659 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1662 return entries_[actualIdx_].mathEnd;
1664 size_t getStartPos() const {
1665 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1666 return 100000; /* definitely enough? */
1668 return entries_[actualIdx_].mathStart;
1670 size_t getFirstPos() {
1672 return getStartPos();
1674 size_t getSize() const {
1675 if (entries_.empty() || (actualIdx_ >= entries_.size())) {
1678 return entries_[actualIdx_].mathSize;
1680 void incrEntry() { actualIdx_++; };
1683 void LatexInfo::buildEntries(bool isPatternString)
1685 static regex const rmath("\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|alignat)\\*?)\\}");
1686 static regex const rkeys("\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?))");
1687 static bool disableLanguageOverride = false;
1688 smatch sub, submath;
1689 bool evaluatingRegexp = false;
1691 bool evaluatingMath = false;
1692 bool evaluatingCode = false;
1694 bool evaluatingOptional = false;
1695 size_t optionalEnd = 0;
1698 bool math_end_waiting = false;
1699 size_t math_pos = 10000;
1702 interval_.removeAccents();
1704 for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) {
1706 if (math_end_waiting) {
1707 size_t pos = submath.position(size_t(0));
1708 if ((math_end == "$") &&
1709 (submath.str(0) == "$") &&
1710 (interval_.par[pos-1] != '\\')) {
1711 mi.insert("$", math_pos, pos + 1);
1712 math_end_waiting = false;
1714 else if ((math_end == "\\]") &&
1715 (submath.str(0) == "\\]")) {
1716 mi.insert("\\]", math_pos, pos + 2);
1717 math_end_waiting = false;
1719 else if ((submath.str(1).compare("end") == 0) &&
1720 (submath.str(2).compare(math_end) == 0)) {
1721 mi.insert(math_end, math_pos, pos + submath.str(0).length());
1722 math_end_waiting = false;
1728 if (submath.str(1).compare("begin") == 0) {
1729 math_end_waiting = true;
1730 math_end = submath.str(2);
1731 math_pos = submath.position(size_t(0));
1733 else if (submath.str(0).compare("\\[") == 0) {
1734 math_end_waiting = true;
1736 math_pos = submath.position(size_t(0));
1738 else if (submath.str(0) == "$") {
1739 size_t pos = submath.position(size_t(0));
1740 if ((pos == 0) || (interval_.par[pos-1] != '\\')) {
1741 math_end_waiting = true;
1748 // Ignore language if there is math somewhere in pattern-string
1749 if (isPatternString) {
1752 keys["foreignlanguage"].disabled = true;
1753 disableLanguageOverride = true;
1756 disableLanguageOverride = false;
1759 if (disableLanguageOverride) {
1760 keys["foreignlanguage"].disabled = true;
1763 math_pos = mi.getFirstPos();
1764 for (sregex_iterator it(interval_.par.begin(), interval_.par.end(), rkeys), end; it != end; ++it) {
1766 string key = sub.str(3);
1768 if (sub.str(0)[0] == '\\')
1769 key = sub.str(0)[1];
1773 size_t k_pos = sub.position(size_t(0));
1774 if ((k_pos > 0) && (interval_.par[k_pos - 1] == '\\')) {
1775 // Escaped '$', ignoring
1781 if (keys.find(key) != keys.end()) {
1782 if (keys[key].keytype == KeyInfo::headRemove) {
1783 KeyInfo found1 = keys[key];
1784 found1.disabled = true;
1785 found1.head = "\\" + key + "{";
1786 found1._tokenstart = sub.position(size_t(0));
1787 found1._tokensize = found1.head.length();
1788 found1._dataStart = found1._tokenstart + found1.head.length();
1789 int endpos = interval_.findclosing(found1._dataStart, interval_.par.length(), '{', '}', 1);
1790 found1._dataEnd = endpos;
1795 if (evaluatingRegexp) {
1796 if (sub.str(1).compare("endregexp") == 0) {
1797 evaluatingRegexp = false;
1798 // found._tokenstart already set
1799 found._dataEnd = sub.position(size_t(0)) + 13;
1800 found._dataStart = found._dataEnd;
1801 found._tokensize = found._dataEnd - found._tokenstart;
1802 found.parenthesiscount = 0;
1803 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1810 if (evaluatingMath) {
1811 if (size_t(sub.position(size_t(0))) < mi.getEndPos())
1813 evaluatingMath = false;
1815 math_pos = mi.getStartPos();
1817 if (keys.find(key) == keys.end()) {
1818 found = KeyInfo(KeyInfo::isStandard, 0, true);
1819 if (isPatternString) {
1820 found.keytype = KeyInfo::isChar;
1821 found.disabled = false;
1828 if (key.compare("regexp") == 0) {
1829 evaluatingRegexp = true;
1830 found._tokenstart = sub.position(size_t(0));
1831 found._tokensize = 0;
1835 // Handle the other params of key
1836 if (found.keytype == KeyInfo::isIgnored)
1838 else if (found.keytype == KeyInfo::isMath) {
1839 if (size_t(sub.position(size_t(0))) == math_pos) {
1841 found._tokenstart = sub.position(size_t(0));
1842 found._tokensize = mi.getSize();
1843 found._dataEnd = found._tokenstart + found._tokensize;
1844 found._dataStart = found._dataEnd;
1845 found.parenthesiscount = 0;
1846 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1847 evaluatingMath = true;
1850 // begin|end of unknown env, discard
1851 // First handle tables
1852 // longtable|tabular
1853 bool discardComment;
1855 found.keytype = KeyInfo::doRemove;
1856 if ((sub.str(5).compare("longtable") == 0) ||
1857 (sub.str(5).compare("tabular") == 0)) {
1858 discardComment = true; /* '%' */
1861 discardComment = false;
1862 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
1864 string token = sub.str(5);
1865 if (regex_match(token, sub2, removeArgs)) {
1866 found.keytype = KeyInfo::removeWithArg;
1869 // discard spaces before pos(0)
1870 int pos = sub.position(size_t(0));
1872 for (count = 0; pos - count > 0; count++) {
1873 char c = interval_.par[pos-count-1];
1874 if (discardComment) {
1875 if ((c != ' ') && (c != '%'))
1881 found._tokenstart = pos - count;
1882 if (sub.str(1).compare(0, 5, "begin") == 0) {
1883 size_t pos1 = pos + sub.str(0).length();
1884 if (sub.str(5).compare("cjk") == 0) {
1885 pos1 = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
1886 if ((interval_.par[pos1] == '{') && (interval_.par[pos1+1] == '}'))
1888 found.keytype = KeyInfo::isMain;
1889 found._dataStart = pos1;
1890 found._dataEnd = interval_.par.length();
1891 found.disabled = keys["foreignlanguage"].disabled;
1892 found.used = keys["foreignlanguage"].used;
1893 found._tokensize = pos1 - found._tokenstart;
1894 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1897 // Swallow possible optional params
1898 while (interval_.par[pos1] == '[') {
1899 pos1 = interval_.findclosing(pos1+1, interval_.par.length(), '[', ']')+1;
1901 // Swallow also the eventual parameter
1902 if (interval_.par[pos1] == '{') {
1903 found._dataEnd = interval_.findclosing(pos1+1, interval_.par.length()) + 1;
1906 found._dataEnd = pos1;
1908 found._dataStart = found._dataEnd;
1909 found._tokensize = count + found._dataEnd - pos;
1910 found.parenthesiscount = 0;
1911 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1912 found.disabled = true;
1916 // Handle "\end{...}"
1917 found._dataStart = pos + sub.str(0).length();
1918 found._dataEnd = found._dataStart;
1919 found._tokensize = count + found._dataEnd - pos;
1920 found.parenthesiscount = 0;
1921 found.head = interval_.par.substr(found._tokenstart, found._tokensize);
1922 found.disabled = true;
1926 else if (found.keytype != KeyInfo::isRegex) {
1927 found._tokenstart = sub.position(size_t(0));
1928 if (found.parenthesiscount == 0) {
1929 // Probably to be discarded
1930 size_t following_pos = sub.position(size_t(0)) + sub.str(3).length() + 1;
1931 char following = interval_.par[following_pos];
1932 if (following == ' ')
1933 found.head = "\\" + sub.str(3) + " ";
1934 else if (following == '=') {
1935 // like \uldepth=1000pt
1936 found.head = sub.str(0);
1939 found.head = "\\" + key;
1940 found._tokensize = found.head.length();
1941 found._dataEnd = found._tokenstart + found._tokensize;
1942 found._dataStart = found._dataEnd;
1945 int params = found._tokenstart + key.length() + 1;
1946 if (evaluatingOptional) {
1947 if (size_t(found._tokenstart) > optionalEnd) {
1948 evaluatingOptional = false;
1951 found.disabled = true;
1954 int optend = params;
1955 while (interval_.par[optend] == '[') {
1956 // discard optional parameters
1957 optend = interval_.findclosing(optend+1, interval_.par.length(), '[', ']') + 1;
1959 if (optend > params) {
1960 key += interval_.par.substr(params, optend-params);
1961 evaluatingOptional = true;
1962 optionalEnd = optend;
1964 string token = sub.str(5);
1965 int closings = found.parenthesiscount;
1966 if (found.parenthesiscount == 1) {
1967 found.head = "\\" + key + "{";
1969 else if (found.parenthesiscount > 1) {
1971 found.head = sub.str(0) + "{";
1972 closings = found.parenthesiscount - 1;
1975 found.head = "\\" + key + "{";
1978 found._tokensize = found.head.length();
1979 found._dataStart = found._tokenstart + found.head.length();
1980 if (found.keytype == KeyInfo::doRemove) {
1981 int endpar = 2 + interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
1982 found._dataStart = endpar;
1983 found._tokensize = found._dataStart - found._tokenstart;
1986 if (interval_.par.substr(found._dataStart-1, 15).compare("\\endarguments{}") == 0) {
1987 found._dataStart += 15;
1991 endpos = found._dataStart - 1;
1993 endpos = interval_.findclosing(found._dataStart, interval_.par.length(), '{', '}', closings);
1994 if (found.keytype == KeyInfo::isList) {
1995 // Check if it really is list env
1996 static regex const listre("^([a-z]+)$");
1998 if (!regex_match(token, sub2, listre)) {
1999 // Change the key of this entry. It is not in a list/item environment
2000 found.keytype = KeyInfo::endArguments;
2003 if (found.keytype == KeyInfo::noMain) {
2004 evaluatingCode = true;
2006 codeStart = found._dataStart;
2008 else if (evaluatingCode) {
2009 if (size_t(found._dataStart) > codeEnd)
2010 evaluatingCode = false;
2011 else if (found.keytype == KeyInfo::isMain) {
2012 // Disable this key, treate it as standard
2013 found.keytype = KeyInfo::isStandard;
2014 found.disabled = true;
2015 if ((codeEnd == interval_.par.length()) &&
2016 (found._tokenstart == codeStart)) {
2017 // trickery, because the code inset starts
2018 // with \selectlanguage ...
2020 if (entries_.size() > 1) {
2021 entries_[entries_.size()-1]._dataEnd = codeEnd;
2026 if ((endpos == interval_.par.length()) &&
2027 (found.keytype == KeyInfo::doRemove)) {
2028 // Missing closing => error in latex-input?
2029 // therefore do not delete remaining data
2030 found._dataStart -= 1;
2031 found._dataEnd = found._dataStart;
2034 found._dataEnd = endpos;
2036 if (isPatternString) {
2037 keys[key].used = true;
2040 entries_.push_back(found);
2044 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
2046 stringstream s(keysstring);
2048 const char delim = '|';
2049 while (getline(s, key, delim)) {
2050 KeyInfo keyII(keyI);
2051 if (isPatternString) {
2054 else if ( !keys[key].used)
2055 keyII.disabled = true;
2060 void LatexInfo::buildKeys(bool isPatternString)
2063 static bool keysBuilt = false;
2064 if (keysBuilt && !isPatternString) return;
2066 // Keys to ignore in any case
2067 makeKey("text|textcyrillic|lyxmathsym", KeyInfo(KeyInfo::headRemove, 1, true), true);
2068 // Known standard keys with 1 parameter.
2069 // Split is done, if not at start of region
2070 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
2071 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
2072 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
2073 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
2074 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
2075 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
2077 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
2078 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2079 makeKey("section*|subsection*|subsubsection*|paragraph*",
2080 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2081 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
2082 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
2084 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
2086 // Split is done, if not at start of region
2087 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
2088 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
2090 // Split is done always.
2091 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
2095 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2096 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2097 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2098 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2100 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2101 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2102 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2103 makeKey("thickspace|medspace|thinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2105 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2106 // Custom space/skip, remove the content (== length value)
2107 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
2108 // Found in fr/UserGuide.lyx
2109 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2111 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2112 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2113 // Known macros to remove (including their parameter)
2115 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
2116 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
2117 // handle like standard keys with 1 parameter.
2118 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
2120 // Ignore deleted text
2121 makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString);
2122 // but preserve added text
2123 makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString);
2125 // Macros to remove, but let the parameter survive
2127 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2129 // Remove language spec from content of these insets
2130 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
2132 // Same effect as previous, parameter will survive (because there is no one anyway)
2134 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2135 // Remove table decorations
2136 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
2137 // Discard shape-header.
2138 // For footnote or shortcut too, because of lang settings
2139 // and wrong handling if used 'KeyInfo::noMain'
2140 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2141 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2142 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2143 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2144 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
2145 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
2146 // like ('tiny{}' or '\tiny ' ... )
2147 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
2149 // Survives, like known character
2150 // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
2151 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
2153 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2154 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2155 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
2157 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2158 // Remove RTL/LTR marker
2159 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2160 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
2161 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
2162 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
2163 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
2164 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
2165 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
2166 if (isPatternString) {
2167 // Allow the first searched string to rebuild the keys too
2171 // no need to rebuild again
2177 * Keep the list of actual opened parentheses actual
2178 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2180 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2183 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2187 if (c == '\\') skip = 1;
2188 else if (c == '{') {
2191 else if (c == '}') {
2192 handleCloseP(i, closingAllowed);
2198 string Intervall::show(int lastpos)
2200 int idx = 0; /* int intervalls */
2203 for (idx = 0; idx <= ignoreidx; idx++) {
2204 while (i < lastpos) {
2206 if (i <= borders[idx].low) {
2207 if (borders[idx].low > lastpos)
2208 printsize = lastpos - i;
2210 printsize = borders[idx].low - i;
2211 s += par.substr(i, printsize);
2213 if (i >= borders[idx].low)
2214 i = borders[idx].upper;
2217 i = borders[idx].upper;
2223 s += par.substr(i, lastpos-i);
2229 void Intervall::output(ostringstream &os, int lastpos)
2231 // get number of chars to output
2232 int idx = 0; /* int intervalls */
2235 string startTitle = titleValue;
2236 for (idx = 0; idx <= ignoreidx; idx++) {
2238 if (i <= borders[idx].low) {
2240 if (borders[idx].low > lastpos)
2241 printsize = lastpos - i;
2243 printsize = borders[idx].low - i;
2244 if (printsize > 0) {
2245 os << startTitle << par.substr(i, printsize);
2247 printed += printsize;
2250 handleParentheses(i, false);
2251 if (i >= borders[idx].low)
2252 i = borders[idx].upper;
2255 i = borders[idx].upper;
2262 os << startTitle << par.substr(i, lastpos-i);
2263 printed += lastpos-i;
2265 handleParentheses(lastpos, false);
2267 if (keys["foreignlanguage"].disabled)
2268 startindex = actualdeptindex-langcount;
2270 startindex = actualdeptindex;
2271 for (int i = startindex; i > 0; --i) {
2274 if (hasTitle && (printed > 0))
2276 if (! isPatternString_)
2278 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2281 void LatexInfo::processRegion(int start, int region_end)
2283 while (start < region_end) { /* Let {[} and {]} survive */
2284 int cnt = interval_.isOpeningPar(start);
2286 // Closing is allowed past the region
2287 int closing = interval_.findclosing(start+1, interval_.par.length());
2288 interval_.addIntervall(start, start+1);
2289 interval_.addIntervall(closing, closing+1);
2293 start = interval_.nextNotIgnored(start+1);
2297 void LatexInfo::removeHead(KeyInfo const & actual, int count)
2299 if (actual.parenthesiscount == 0) {
2300 // "{\tiny{} ...}" ==> "{{} ...}"
2301 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2304 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2305 interval_.addIntervall(actual._tokenstart - count, actual._dataStart);
2306 interval_.addIntervall(actual._dataEnd, actual._dataEnd+1);
2310 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2313 switch (actual.keytype)
2315 case KeyInfo::isTitle: {
2317 nextKeyIdx = getNextKey();
2320 case KeyInfo::cleanToStart: {
2321 actual._dataEnd = actual._dataStart;
2322 nextKeyIdx = getNextKey();
2323 // Search for end of arguments
2324 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2326 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2327 entries_[i].disabled = true;
2329 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2331 while (interval_.par[actual._dataEnd] == ' ')
2333 interval_.addIntervall(0, actual._dataEnd+1);
2334 interval_.actualdeptindex = 0;
2335 interval_.depts[0] = actual._dataEnd+1;
2336 interval_.closes[0] = -1;
2339 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2340 if (actual.disabled)
2341 interval_.addIntervall(actual._tokenstart, actual._dataEnd);
2343 interval_.addIntervall(actual._dataStart, actual._dataEnd);
2346 case KeyInfo::isChar: {
2347 nextKeyIdx = getNextKey();
2350 case KeyInfo::isSize: {
2351 if (actual.disabled || (interval_.par[actual._dataStart] != '{') || (interval_.par[actual._dataStart-1] == ' ')) {
2352 processRegion(actual._dataEnd, actual._dataEnd+1); /* remove possibly following {} */
2353 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2354 nextKeyIdx = getNextKey();
2356 // Here _dataStart points to '{', so correct it
2357 actual._dataStart += 1;
2358 actual._tokensize += 1;
2359 actual.parenthesiscount = 1;
2360 if (interval_.par[actual._dataStart] == '}') {
2361 // Determine the end if used like '{\tiny{}...}'
2362 actual._dataEnd = interval_.findclosing(actual._dataStart+1, interval_.par.length()) + 1;
2363 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2366 // Determine the end if used like '\tiny{...}'
2367 actual._dataEnd = interval_.findclosing(actual._dataStart, interval_.par.length()) + 1;
2369 // Split on this key if not at start
2370 int start = interval_.nextNotIgnored(previousStart);
2371 if (start < actual._tokenstart) {
2372 interval_.output(os, actual._tokenstart);
2373 interval_.addIntervall(start, actual._tokenstart);
2375 // discard entry if at end of actual
2376 nextKeyIdx = process(os, actual);
2380 case KeyInfo::endArguments: {
2381 // Remove trailing '{}' too
2382 actual._dataStart += 1;
2383 actual._dataEnd += 1;
2384 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2385 nextKeyIdx = getNextKey();
2388 case KeyInfo::noMain:
2390 case KeyInfo::isStandard: {
2391 if (actual.disabled) {
2393 processRegion(actual._dataStart, actual._dataStart+1);
2394 nextKeyIdx = getNextKey();
2396 // Split on this key if not at datastart of calling entry
2397 int start = interval_.nextNotIgnored(previousStart);
2398 if (start < actual._tokenstart) {
2399 interval_.output(os, actual._tokenstart);
2400 interval_.addIntervall(start, actual._tokenstart);
2402 // discard entry if at end of actual
2403 nextKeyIdx = process(os, actual);
2407 case KeyInfo::removeWithArg: {
2408 nextKeyIdx = getNextKey();
2409 // Search for end of arguments
2410 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2412 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2413 entries_[i].disabled = true;
2415 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2417 interval_.addIntervall(actual._tokenstart, actual._dataEnd+1);
2420 case KeyInfo::doRemove: {
2421 // Remove the key with all parameters and following spaces
2423 for (pos = actual._dataEnd+1; pos < interval_.par.length(); pos++) {
2424 if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
2427 // Remove also enclosing parentheses [] and {}
2430 while (actual._tokenstart > numpars) {
2431 if (pos+numpars >= interval_.par.size())
2433 else if (interval_.par[pos+numpars] == ']' && interval_.par[actual._tokenstart-numpars-1] == '[')
2435 else if (interval_.par[pos+numpars] == '}' && interval_.par[actual._tokenstart-numpars-1] == '{')
2441 if (interval_.par[pos+numpars] == ' ')
2445 interval_.addIntervall(actual._tokenstart-numpars, pos+numpars+spaces);
2446 nextKeyIdx = getNextKey();
2449 case KeyInfo::isList: {
2450 // Discard space before _tokenstart
2452 for (count = 0; count < actual._tokenstart; count++) {
2453 if (interval_.par[actual._tokenstart-count-1] != ' ')
2456 nextKeyIdx = getNextKey();
2457 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2459 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2461 // How else can we catch this one?
2462 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2463 entries_[i].disabled = true;
2465 actual._dataEnd = entries_[tmpIdx]._dataEnd;
2467 else if (nextKeyIdx > 0) {
2468 // Ignore any lang entries inside data region
2469 for (int i = nextKeyIdx; i < int(entries_.size()) && entries_[i]._tokenstart < actual._dataEnd; i++) {
2470 if (entries_[i].keytype == KeyInfo::isMain)
2471 entries_[i].disabled = true;
2474 if (actual.disabled) {
2475 interval_.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2478 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2480 if (interval_.par[actual._dataEnd+1] == '[') {
2481 int posdown = interval_.findclosing(actual._dataEnd+2, interval_.par.length(), '[', ']');
2482 if ((interval_.par[actual._dataEnd+2] == '{') &&
2483 (interval_.par[posdown-1] == '}')) {
2484 interval_.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
2485 interval_.addIntervall(posdown-1, posdown+1);
2488 interval_.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
2489 interval_.addIntervall(posdown, posdown+1);
2491 int blk = interval_.nextNotIgnored(actual._dataEnd+1);
2492 if (blk > posdown) {
2493 // Discard at most 1 space after empty item
2495 for (count = 0; count < 1; count++) {
2496 if (interval_.par[blk+count] != ' ')
2500 interval_.addIntervall(blk, blk+count);
2505 case KeyInfo::isSectioning: {
2506 // Discard spaces before _tokenstart
2508 int val = actual._tokenstart;
2509 for (count = 0; count < actual._tokenstart;) {
2510 val = interval_.previousNotIgnored(val-1);
2511 if (val < 0 || interval_.par[val] != ' ')
2514 count = actual._tokenstart - val;
2517 if (actual.disabled) {
2518 removeHead(actual, count);
2519 nextKeyIdx = getNextKey();
2521 interval_.addIntervall(actual._tokenstart-count, actual._tokenstart);
2522 nextKeyIdx = process(os, actual);
2526 case KeyInfo::isMath: {
2527 // Same as regex, use the content unchanged
2528 nextKeyIdx = getNextKey();
2531 case KeyInfo::isRegex: {
2532 // DO NOT SPLIT ON REGEX
2534 nextKeyIdx = getNextKey();
2537 case KeyInfo::isIgnored: {
2538 // Treat like a character for now
2539 nextKeyIdx = getNextKey();
2542 case KeyInfo::isMain: {
2543 if (interval_.par.substr(actual._dataStart, 2) == "% ")
2544 interval_.addIntervall(actual._dataStart, actual._dataStart+2);
2545 if (actual._tokenstart > 0) {
2546 int prev = interval_.previousNotIgnored(actual._tokenstart - 1);
2547 if ((prev >= 0) && interval_.par[prev] == '%')
2548 interval_.addIntervall(prev, prev+1);
2550 if (actual.disabled) {
2552 interval_.langcount++;
2553 if ((interval_.par.substr(actual._dataStart, 3) == " \\[") ||
2554 (interval_.par.substr(actual._dataStart, 8) == " \\begin{")) {
2555 // Discard also the space before math-equation
2556 interval_.addIntervall(actual._dataStart, actual._dataStart+1);
2558 nextKeyIdx = getNextKey();
2559 // interval.resetOpenedP(actual._dataStart-1);
2562 if (actual._tokenstart < 26) {
2563 // for the first (and maybe dummy) language
2564 interval_.setForDefaultLang(actual);
2566 interval_.resetOpenedP(actual._dataStart-1);
2570 case KeyInfo::invalid:
2571 case KeyInfo::headRemove:
2572 // These two cases cannot happen, already handled
2575 // LYXERR(Debug::INFO, "Unhandled keytype");
2576 nextKeyIdx = getNextKey();
2583 int LatexInfo::process(ostringstream & os, KeyInfo const & actual )
2585 int end = interval_.nextNotIgnored(actual._dataEnd);
2586 int oldStart = actual._dataStart;
2587 int nextKeyIdx = getNextKey();
2589 if ((nextKeyIdx < 0) ||
2590 (entries_[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
2591 (entries_[nextKeyIdx].keytype == KeyInfo::invalid)) {
2592 if (oldStart <= end) {
2593 processRegion(oldStart, end);
2598 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
2600 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
2601 (void) dispatch(os, actual._dataStart, nextKey);
2602 end = nextKey._tokenstart;
2605 processRegion(oldStart, nextKey._tokenstart);
2606 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
2608 oldStart = nextKey._dataEnd+1;
2610 // now nextKey is either invalid or is outside of actual._dataEnd
2611 // output the remaining and discard myself
2612 if (oldStart <= end) {
2613 processRegion(oldStart, end);
2615 if (interval_.par.size() > (size_t) end && interval_.par[end] == '}') {
2617 // This is the normal case.
2618 // But if using the firstlanguage, the closing may be missing
2620 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
2622 if (actual._dataEnd < end)
2623 output_end = interval_.nextNotIgnored(actual._dataEnd);
2625 output_end = interval_.nextNotIgnored(end);
2626 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
2627 interval_.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
2629 // Remove possible empty data
2630 int dstart = interval_.nextNotIgnored(actual._dataStart);
2631 while (interval_.isOpeningPar(dstart) == 1) {
2632 interval_.addIntervall(dstart, dstart+1);
2633 int dend = interval_.findclosing(dstart+1, output_end);
2634 interval_.addIntervall(dend, dend+1);
2635 dstart = interval_.nextNotIgnored(dstart+1);
2637 if (dstart < output_end)
2638 interval_.output(os, output_end);
2639 interval_.addIntervall(actual._tokenstart, end);
2643 string splitOnKnownMacros(string par, bool isPatternString)
2646 LatexInfo li(par, isPatternString);
2647 // LYXERR(Debug::INFO, "Berfore split: " << par);
2648 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
2650 DummyKey._tokensize = 0;
2651 DummyKey._dataStart = 0;
2652 DummyKey._dataEnd = par.length();
2653 DummyKey.disabled = true;
2654 int firstkeyIdx = li.getFirstKey();
2656 if (firstkeyIdx >= 0) {
2657 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
2658 DummyKey._tokenstart = firstKey._tokenstart;
2660 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
2661 // Use dummy firstKey
2662 firstKey = DummyKey;
2663 (void) li.setNextKey(firstkeyIdx);
2666 if (par.substr(firstKey._dataStart, 2) == "% ")
2667 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
2669 nextkeyIdx = li.process(os, firstKey);
2670 while (nextkeyIdx >= 0) {
2671 // Check for a possible gap between the last
2672 // entry and this one
2673 int datastart = li.nextNotIgnored(firstKey._dataStart);
2674 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
2675 if ((nextKey._tokenstart > datastart)) {
2677 firstKey._dataStart = datastart;
2678 firstKey._dataEnd = par.length();
2679 (void) li.setNextKey(nextkeyIdx);
2680 // Fake the last opened parenthesis
2681 li.setForDefaultLang(firstKey);
2682 nextkeyIdx = li.process(os, firstKey);
2685 if (nextKey.keytype != KeyInfo::isMain) {
2686 firstKey._dataStart = datastart;
2687 firstKey._dataEnd = nextKey._dataEnd+1;
2688 (void) li.setNextKey(nextkeyIdx);
2689 li.setForDefaultLang(firstKey);
2690 nextkeyIdx = li.process(os, firstKey);
2693 nextkeyIdx = li.process(os, nextKey);
2697 // Handle the remaining
2698 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
2699 firstKey._dataEnd = par.length();
2701 if ((firstKey._dataStart < firstKey._dataEnd) &&
2702 (par[firstKey._dataStart] != '}')) {
2703 li.setForDefaultLang(firstKey);
2704 (void) li.process(os, firstKey);
2708 // return string definitelly impossible to match
2709 s = "\\foreignlanguage{ignore}{ }";
2713 s = par; /* no known macros found */
2714 // LYXERR(Debug::INFO, "After split: " << s);
2719 * Try to unify the language specs in the latexified text.
2720 * Resulting modified string is set to "", if
2721 * the searched tex does not contain all the features in the search pattern
2723 static string correctlanguagesetting(string par, bool isPatternString, bool withformat)
2725 static Features regex_f;
2726 static int missed = 0;
2727 static bool regex_with_format = false;
2729 int parlen = par.length();
2731 while ((parlen > 0) && (par[parlen-1] == '\n')) {
2734 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
2735 // Happens to be there in case of description or labeling environment
2740 // Split the latex input into pieces which
2741 // can be digested by our search engine
2742 LYXERR(Debug::FIND, "input: \"" << par << "\"");
2743 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
2744 LYXERR(Debug::FIND, "After split: \"" << result << "\"");
2747 result = par.substr(0, parlen);
2748 if (isPatternString) {
2751 regex_f = identifyFeatures(result);
2752 string features = "";
2753 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2754 string a = it->first;
2755 regex_with_format = true;
2756 features += " " + a;
2757 // LYXERR(Debug::INFO, "Identified regex format:" << a);
2759 LYXERR(Debug::FIND, "Identified Features" << features);
2762 } else if (regex_with_format) {
2763 Features info = identifyFeatures(result);
2764 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2765 string a = it->first;
2766 bool b = it->second;
2767 if (b && ! info[a]) {
2769 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
2775 // LYXERR(Debug::INFO, "No regex formats");
2781 // Remove trailing closure of math, macros and environments, so to catch parts of them.
2782 static int identifyClosing(string & t)
2784 int open_braces = 0;
2786 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
2787 if (regex_replace(t, t, "(.*[^\\\\])\\$$", "$1"))
2789 if (regex_replace(t, t, "(.*[^\\\\])\\\\\\]$", "$1"))
2791 if (regex_replace(t, t, "(.*[^\\\\])\\\\end\\{[a-zA-Z_]*\\*?\\}$", "$1"))
2793 if (regex_replace(t, t, "(.*[^\\\\])\\}$", "$1")) {
2802 static int num_replaced = 0;
2803 static bool previous_single_replace = true;
2805 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt)
2806 : p_buf(&buf), p_first_buf(&buf), opt(opt)
2808 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
2809 docstring const & ds = stringifySearchBuffer(find_buf, opt);
2810 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
2811 if (opt.replace_all && previous_single_replace) {
2812 previous_single_replace = false;
2815 else if (!opt.replace_all) {
2816 num_replaced = 0; // count number of replaced strings
2817 previous_single_replace = true;
2819 // When using regexp, braces are hacked already by escape_for_regex()
2820 par_as_string = normalize(ds, !use_regexp);
2822 close_wildcards = 0;
2824 size_t lead_size = 0;
2825 // correct the language settings
2826 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat);
2827 if (opt.ignoreformat) {
2829 // if par_as_string_nolead were emty,
2830 // the following call to findAux will always *find* the string
2831 // in the checked data, and thus always using the slow
2832 // examining of the current text part.
2833 par_as_string_nolead = par_as_string;
2836 lead_size = identifyLeading(par_as_string);
2837 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
2838 lead_as_string = par_as_string.substr(0, lead_size);
2839 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
2843 open_braces = identifyClosing(par_as_string);
2844 identifyClosing(par_as_string_nolead);
2845 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2846 LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'");
2848 string lead_as_regexp;
2849 if (lead_size > 0) {
2850 // @todo No need to search for \regexp{} insets in leading material
2851 lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat);
2852 par_as_string = par_as_string_nolead;
2853 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
2854 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2856 LYXERR(Debug::FIND, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
2857 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
2858 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
2859 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2861 // Insert .* before trailing '\$' ('$' has been escaped by escape_for_regex)
2862 regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\$)\\'", "$1(.*?)$2")
2863 // Insert .* before trailing '\\\]' ('\]' has been escaped by escape_for_regex)
2864 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])( \\\\\\\\\\\\\\])\\'", "$1(.*?)$2")
2865 // Insert .* before trailing '\\end\{...}' ('\end{...}' has been escaped by escape_for_regex)
2866 || regex_replace(par_as_string, par_as_string,
2867 "(.*[^\\\\])( \\\\\\\\end\\\\\\{[a-zA-Z_]*)(\\\\\\*)?(\\\\\\})\\'", "$1(.*?)$2$3$4")
2868 // Insert .* before trailing '\}' ('}' has been escaped by escape_for_regex)
2869 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\})\\'", "$1(.*?)$2")
2873 if (!opt.ignoreformat) {
2874 // Remove extra '\}' at end if not part of \{\.\}
2875 size_t lng = par_as_string.size();
2877 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
2879 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
2888 if (lng < par_as_string.size())
2889 par_as_string = par_as_string.substr(0,lng);
2892 regex_replace(par_as_string, par_as_string, "\\\\\\.", "_xxbdotxx_");
2893 // handle '.' -> '[^]', replace later as '[^\}\{\\]'
2894 regex_replace(par_as_string, par_as_string, "\\.", "[^]");
2895 // replace '[^...]' with '[^...\}\{\\]'
2896 regex_replace(par_as_string, par_as_string, "\\[\\^([^\\\\\\]]*)\\]", "_xxbrlxx_$1\\}\\{\\\\_xxbrrxx_");
2897 regex_replace(par_as_string, par_as_string, "_xxbrlxx_", "[^");
2898 regex_replace(par_as_string, par_as_string, "_xxbrrxx_", "]");
2900 regex_replace(par_as_string, par_as_string, "_xxbdotxx_", "\\.");
2903 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2904 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2905 LYXERR(Debug::FIND, "Close .*? : " << close_wildcards);
2906 LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
2908 // If entered regexp must match at begin of searched string buffer
2909 // Kornel: Added parentheses to use $1 for size of the leading string
2913 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
2914 // Unfortunately is '\1', '\2', etc not working for strings with extra format
2915 // so the convert has no effect in that case
2916 for (int i = 8; i > 0; --i) {
2917 string orig = "\\\\" + std::to_string(i);
2918 string dest = "\\" + std::to_string(i+1);
2919 while (regex_replace(par_as_string, par_as_string, orig, dest));
2921 regexp_str = "(" + lead_as_regexp + ")" + par_as_string;
2922 regexp2_str = "(" + lead_as_regexp + ").*?" + par_as_string;
2924 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
2925 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
2926 if (! opt.casesensitive) {
2927 regexp = regex(regexp_str, std::regex_constants::icase);
2928 regexp2 = regex(regexp2_str, std::regex_constants::icase);
2931 regexp = regex(regexp_str);
2932 regexp2 = regex(regexp2_str);
2938 // Count number of characters in string
2942 // \\[a-zA-Z]+ ==> 1
2943 static int computeSize(string s, int len)
2949 for (int i = 0; i < len; i += skip, count++) {
2952 if (isalpha(s[i+1])) {
2953 for (int j = 2; i+j < len; j++) {
2954 if (! isalpha(s[i+j])) {
2957 else if ((s[i+j] == '{') && s[i+j+1] == '}')
2959 else if ((s[i+j] == '{') && (i + j + 1 >= len))
2967 else if (s[i] == '{') {
2973 else if (s[i] == '-') {
2974 if (s[i+1] == '-') {
2990 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
2995 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
2998 docstring docstr = stringifyFromForSearch(opt, cur, len);
3000 if (use_regexp || opt.casesensitive)
3001 str = normalize(docstr, true);
3003 str = normalize(lowercase(docstr), true);
3004 if (!opt.ignoreformat) {
3005 str = correctlanguagesetting(str, false, !opt.ignoreformat);
3008 mres.match_len = -1;
3011 LYXERR(Debug::FIND, "Matching against '" << lyx::to_utf8(docstr) << "'");
3012 LYXERR(Debug::FIND, "After normalization: '" << str << "'");
3015 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
3016 regex const *p_regexp;
3017 regex_constants::match_flag_type flags;
3019 flags = regex_constants::match_continuous;
3022 flags = regex_constants::match_default;
3023 p_regexp = ®exp2;
3025 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
3026 if (re_it == sregex_iterator())
3028 match_results<string::const_iterator> const & m = *re_it;
3030 // Check braces on segments that matched all (.*?) subexpressions,
3031 // except the last "padding" one inserted by lyx.
3032 for (size_t i = 1; i < m.size() - 1; ++i)
3033 if (!braces_match(m[i].first, m[i].second, open_braces))
3036 // Exclude from the returned match length any length
3037 // due to close wildcards added at end of regexp
3038 // and also the length of the leading (e.g. '\emph{}')
3040 // Whole found string, including the leading: m[0].second - m[0].first
3041 // Size of the leading string: m[1].second - m[1].first
3042 int leadingsize = 0;
3044 leadingsize = m[1].second - m[1].first;
3046 for (size_t i = 0; i < m.size(); i++) {
3047 LYXERR(Debug::FIND, "Match " << i << " is " << m[i].second - m[i].first << " long");
3049 if (close_wildcards == 0)
3050 result = m[0].second - m[0].first;
3053 result = m[m.size() - close_wildcards].first - m[0].first;
3055 size_t pos = m.position(size_t(0));
3056 // Ignore last closing characters
3057 while (result > 0) {
3058 if (str[pos+result-1] == '}')
3063 if (result > leadingsize)
3064 result -= leadingsize;
3067 mres.match_len = computeSize(str.substr(pos+leadingsize,result), result);
3068 mres.match2end = str.size() - pos - leadingsize;
3069 mres.pos = pos+leadingsize;
3073 // else !use_regexp: but all code paths above return
3074 LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='"
3075 << par_as_string << "', str='" << str << "'");
3076 LYXERR(Debug::FIND, "Searching in normal mode: lead_as_string='"
3077 << lead_as_string << "', par_as_string_nolead='"
3078 << par_as_string_nolead << "'");
3081 LYXERR(Debug::FIND, "size=" << par_as_string.size()
3082 << ", substr='" << str.substr(0, par_as_string.size()) << "'");
3083 if (str.substr(0, par_as_string.size()) == par_as_string) {
3084 mres.match_len = par_as_string.size();
3085 mres.match2end = str.size();
3090 // Start the search _after_ the leading part
3091 size_t pos = str.find(par_as_string_nolead, lead_as_string.size());
3092 if (pos != string::npos) {
3093 mres.match_len = par_as_string.size();
3094 mres.match2end = str.size() - pos;
3103 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
3105 MatchResult mres = findAux(cur, len, at_begin);
3106 int res = mres.match_len;
3108 "res=" << res << ", at_begin=" << at_begin
3109 << ", matchword=" << opt.matchword
3110 << ", inTexted=" << cur.inTexted());
3111 if (res == 0 || !at_begin || !opt.matchword || !cur.inTexted())
3113 if ((len > 0) && (res < len)) {
3117 Paragraph const & par = cur.paragraph();
3118 bool ws_left = (cur.pos() > 0)
3119 ? par.isWordSeparator(cur.pos() - 1)
3125 ws_right = (cur.pos() + len < par.size())
3126 ? par.isWordSeparator(cur.pos() + len)
3130 "cur.pos()=" << cur.pos() << ", res=" << res
3131 << ", separ: " << ws_left << ", " << ws_right
3134 if (ws_left && ws_right) {
3135 // Check for word separators inside the found 'word'
3136 for (int i = 0; i < len; i++) {
3137 if (par.isWordSeparator(cur.pos() + i)) {
3149 string MatchStringAdv::normalize(docstring const & s, bool hack_braces) const
3152 t = lyx::to_utf8(s);
3153 // Remove \n at begin
3154 while (!t.empty() && t[0] == '\n')
3157 while (!t.empty() && t[t.size() - 1] == '\n')
3158 t = t.substr(0, t.size() - 1);
3160 // Handle all other '\n'
3161 while ((pos = t.find("\n")) != string::npos) {
3162 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
3164 if (isAlnumASCII(t[pos+1])) {
3165 t.replace(pos-2, 3, " ");
3168 t.replace(pos-2, 3, "");
3171 else if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
3172 // '\n' adjacent to non-alpha-numerics, discard
3173 t.replace(pos, 1, "");
3176 // Replace all other \n with spaces
3177 t.replace(pos, 1, " ");
3180 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
3181 // Kornel: Added textsl, textsf, textit, texttt and noun
3182 // + allow to seach for colored text too
3183 LYXERR(Debug::FIND, "Removing stale empty \\emph{}, \\textbf{}, \\*section{} macros from: " << t);
3184 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
3185 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3186 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
3187 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
3189 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
3190 // FIXME - check what preceeds the brace
3192 if (opt.ignoreformat)
3193 while (regex_replace(t, t, "\\{", "_x_<")
3194 || regex_replace(t, t, "\\}", "_x_>"))
3195 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
3197 while (regex_replace(t, t, "\\\\\\{", "_x_<")
3198 || regex_replace(t, t, "\\\\\\}", "_x_>"))
3199 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
3206 docstring stringifyFromCursor(DocIterator const & cur, int len)
3208 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
3209 if (cur.inTexted()) {
3210 Paragraph const & par = cur.paragraph();
3211 // TODO what about searching beyond/across paragraph breaks ?
3212 // TODO Try adding a AS_STR_INSERTS as last arg
3213 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
3214 int(par.size()) : cur.pos() + len;
3215 // OutputParams runparams(&cur.buffer()->params().encoding());
3216 OutputParams runparams(encodings.fromLyXName("utf8"));
3217 runparams.nice = true;
3218 runparams.flavor = Flavor::XeTeX;
3219 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
3220 // No side effect of file copying and image conversion
3221 runparams.dryrun = true;
3222 runparams.for_search = true;
3223 LYXERR(Debug::FIND, "Stringifying with cur: "
3224 << cur << ", from pos: " << cur.pos() << ", end: " << end);
3225 return par.asString(cur.pos(), end,
3226 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
3228 } else if (cur.inMathed()) {
3229 CursorSlice cs = cur.top();
3230 MathData md = cs.cell();
3231 MathData::const_iterator it_end =
3232 (( len == -1 || cs.pos() + len > int(md.size()))
3234 : md.begin() + cs.pos() + len );
3236 for (MathData::const_iterator it = md.begin() + cs.pos();
3239 docstring s = asString(md2);
3240 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
3243 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3248 /** Computes the LaTeX export of buf starting from cur and ending len positions
3249 * after cur, if len is positive, or at the paragraph or innermost inset end
3252 docstring latexifyFromCursor(DocIterator const & cur, int len)
3254 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
3255 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
3256 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
3257 Buffer const & buf = *cur.buffer();
3259 odocstringstream ods;
3261 //OutputParams runparams(&buf.params().encoding());
3262 OutputParams runparams(encodings.fromLyXName("utf8"));
3263 runparams.nice = false;
3264 runparams.flavor = Flavor::XeTeX;
3265 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3266 // No side effect of file copying and image conversion
3267 runparams.dryrun = true;
3268 runparams.for_search = true;
3270 if (cur.inTexted()) {
3271 // @TODO what about searching beyond/across paragraph breaks ?
3272 pos_type endpos = cur.paragraph().size();
3273 if (len != -1 && endpos > cur.pos() + len)
3274 endpos = cur.pos() + len;
3275 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
3276 string(), cur.pos(), endpos);
3277 string s = lyx::to_utf8(ods.str());
3278 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
3279 return(lyx::from_utf8(s));
3280 } else if (cur.inMathed()) {
3281 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
3282 for (int s = cur.depth() - 1; s >= 0; --s) {
3283 CursorSlice const & cs = cur[s];
3284 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
3286 cs.asInsetMath()->asHullInset()->header_write(ws);
3291 CursorSlice const & cs = cur.top();
3292 MathData md = cs.cell();
3293 MathData::const_iterator it_end =
3294 ((len == -1 || cs.pos() + len > int(md.size()))
3296 : md.begin() + cs.pos() + len);
3298 for (MathData::const_iterator it = md.begin() + cs.pos();
3302 ods << asString(md2);
3303 // Retrieve the math environment type, and add '$' or '$]'
3304 // or others (\end{equation}) accordingly
3305 for (int s = cur.depth() - 1; s >= 0; --s) {
3306 CursorSlice const & cs2 = cur[s];
3307 InsetMath * inset = cs2.asInsetMath();
3308 if (inset && inset->asHullInset()) {
3310 inset->asHullInset()->footer_write(ws);
3314 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3316 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3322 /** Finalize an advanced find operation, advancing the cursor to the innermost
3323 ** position that matches, plus computing the length of the matching text to
3326 int findAdvFinalize(DocIterator & cur, MatchStringAdv const & match)
3328 // Search the foremost position that matches (avoids find of entire math
3329 // inset when match at start of it)
3331 DocIterator old_cur(cur.buffer());
3333 LYXERR(Debug::FIND, "Forwarding one step (searching for innermost match)");
3337 } while (cur && cur.depth() > d && match(cur).match_len > 0);
3339 int max_match = match(cur).match_len; /* match valid only if not searching whole words */
3340 if (max_match <= 0) return 0;
3341 LYXERR(Debug::FIND, "Ok");
3343 // Compute the match length
3345 if (cur.pos() + len > cur.lastpos())
3347 if (match.opt.matchword) {
3348 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3349 while (cur.pos() + len <= cur.lastpos() && match(cur, len).match_len <= 0) {
3351 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3353 // Length of matched text (different from len param)
3354 int old_match = match(cur, len).match_len;
3358 // Greedy behaviour while matching regexps
3359 while ((new_match = match(cur, len + 1).match_len) > old_match) {
3361 old_match = new_match;
3362 LYXERR(Debug::FIND, "verifying match with len = " << len);
3369 int maxl = cur.lastpos() - cur.pos();
3370 // Greedy behaviour while matching regexps
3371 while (maxl > minl) {
3372 int actual_match = match(cur, len).match_len;
3373 if (actual_match >= max_match) {
3374 // actual_match > max_match _can_ happen,
3375 // if the search area splits
3376 // some following word so that the regex
3377 // (e.g. 'r.*r\b' matches 'r' from the middle of the
3379 // This means, the len value is too big
3381 len = (int)((maxl + minl)/2);
3384 // (actual_match < max_match)
3386 len = (int)((maxl + minl)/2);
3390 // Search for real start of matched characters
3395 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
3396 if (cur.depth() < old_cur.depth()) {
3398 LYXERR(Debug::INFO, "cur.depth() < old_cur.depth(), this should never happen");
3401 if (cur.pos() != old_cur.pos()) {
3402 // OK, forwarded 1 pos in actual inset
3403 actual_match = match(cur, len-1).match_len;
3404 if (actual_match == max_match) {
3405 // Ha, got it! The shorter selection has the same match length
3410 // OK, the shorter selection matches less chars, revert to previous value
3416 LYXERR(Debug::INFO, "cur.pos() == old_cur.pos(), this should never happen");
3417 actual_match = match(cur, len).match_len;
3418 if (actual_match == max_match)
3428 int findForwardAdv(DocIterator & cur, MatchStringAdv const & match)
3432 while (!theApp()->longOperationCancelled() && cur) {
3433 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
3434 MatchResult mres = match(cur, -1, false);
3435 int match_len = mres.match_len;
3436 LYXERR(Debug::FIND, "match_len: " << match_len);
3437 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
3438 LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
3441 if (match_len > 0) {
3442 // Try to find the begin of searched string
3443 int increment = mres.pos/2;
3444 while (mres.pos > 5 && (increment > 5)) {
3445 DocIterator old_cur = cur;
3446 for (int i = 0; i < increment && cur; cur.forwardPos(), i++) {
3448 if (! cur || (cur.pit() > old_cur.pit())) {
3449 // Are we outside of the paragraph?
3450 // This can happen if moving past some UTF8-encoded chars
3455 MatchResult mres2 = match(cur, -1, false);
3456 if ((mres2.match2end < mres.match2end) ||
3457 (mres2.match_len < mres.match_len)) {
3464 if (increment > mres.pos/2)
3465 increment = mres.pos/2;
3469 int match_len_zero_count = 0;
3470 for (int i = 0; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) {
3472 int remaining_len = match(cur, -1, false).match_len;
3473 if (remaining_len <= 0) {
3474 // Apparently the searched string is not in the remaining part
3481 LYXERR(Debug::FIND, "Advancing cur: " << cur);
3482 int match_len3 = match(cur, 1).match_len;
3485 int match_len2 = match(cur).match_len;
3486 LYXERR(Debug::FIND, "match_len2: " << match_len2);
3487 if (match_len2 > 0) {
3488 // Sometimes in finalize we understand it wasn't a match
3489 // and we need to continue the outest loop
3490 int len = findAdvFinalize(cur, match);
3495 if (match_len2 >= 0) {
3496 if (match_len2 == 0)
3497 match_len_zero_count++;
3499 match_len_zero_count = 0;
3502 if (++match_len_zero_count > 3) {
3503 LYXERR(Debug::FIND, "match_len2_zero_count: " << match_len_zero_count << ", match_len was " << match_len);
3511 if (match_len >= 0 && cur.pit() < cur.lastpit()) {
3512 LYXERR(Debug::FIND, "Advancing par: cur=" << cur);
3515 // This should exit nested insets, if any, or otherwise undefine the currsor.
3516 cur.pos() = cur.lastpos();
3517 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
3525 /// Find the most backward consecutive match within same paragraph while searching backwards.
3526 int findMostBackwards(DocIterator & cur, MatchStringAdv const & match)
3528 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3529 DocIterator tmp_cur = cur;
3530 int len = findAdvFinalize(tmp_cur, match);
3531 Inset & inset = cur.inset();
3532 for (; cur != cur_begin; cur.backwardPos()) {
3533 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
3534 DocIterator new_cur = cur;
3535 new_cur.backwardPos();
3536 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
3538 int new_len = findAdvFinalize(new_cur, match);
3543 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
3549 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
3553 // Backup of original position
3554 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3555 if (cur == cur_begin)
3558 DocIterator cur_orig(cur);
3559 bool pit_changed = false;
3562 bool found_match = (match(cur, -1, false).match_len > 0);
3566 cur.pos() = cur.lastpos();
3568 cur.pos() = cur_orig.pos();
3569 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
3570 DocIterator cur_prev_iter;
3572 found_match = (match(cur).match_len > 0);
3573 LYXERR(Debug::FIND, "findBackAdv3: found_match="
3574 << found_match << ", cur: " << cur);
3576 return findMostBackwards(cur, match);
3578 // Stop if begin of document reached
3579 if (cur == cur_begin)
3581 cur_prev_iter = cur;
3585 if (cur == cur_begin)
3592 } while (!theApp()->longOperationCancelled());
3600 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
3601 DocIterator const & cur, int len)
3603 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
3605 if (!opt.ignoreformat)
3606 return latexifyFromCursor(cur, len);
3608 return stringifyFromCursor(cur, len);
3612 FindAndReplaceOptions::FindAndReplaceOptions(
3613 docstring const & _find_buf_name, bool _casesensitive,
3614 bool _matchword, bool _forward, bool _expandmacros, bool _ignoreformat,
3615 docstring const & _repl_buf_name, bool _keep_case,
3616 SearchScope _scope, SearchRestriction _restr, bool _replace_all)
3617 : find_buf_name(_find_buf_name), casesensitive(_casesensitive), matchword(_matchword),
3618 forward(_forward), expandmacros(_expandmacros), ignoreformat(_ignoreformat),
3619 repl_buf_name(_repl_buf_name), keep_case(_keep_case), scope(_scope), restr(_restr), replace_all(_replace_all)
3627 /** Check if 'len' letters following cursor are all non-lowercase */
3628 static bool allNonLowercase(Cursor const & cur, int len)
3630 pos_type beg_pos = cur.selectionBegin().pos();
3631 pos_type end_pos = cur.selectionBegin().pos() + len;
3632 if (len > cur.lastpos() + 1 - beg_pos) {
3633 LYXERR(Debug::FIND, "This should not happen, more debug needed");
3634 len = cur.lastpos() + 1 - beg_pos;
3635 end_pos = beg_pos + len;
3637 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
3638 if (isLowerCase(cur.paragraph().getChar(pos)))
3644 /** Check if first letter is upper case and second one is lower case */
3645 static bool firstUppercase(Cursor const & cur)
3648 pos_type pos = cur.selectionBegin().pos();
3649 if (pos >= cur.lastpos() - 1) {
3650 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
3653 ch1 = cur.paragraph().getChar(pos);
3654 ch2 = cur.paragraph().getChar(pos + 1);
3655 bool result = isUpperCase(ch1) && isLowerCase(ch2);
3656 LYXERR(Debug::FIND, "firstUppercase(): "
3657 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
3658 << ch2 << "(" << char(ch2) << ")"
3659 << ", result=" << result << ", cur=" << cur);
3664 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
3666 ** \fixme What to do with possible further paragraphs in replace buffer ?
3668 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
3670 ParagraphList::iterator pit = buffer.paragraphs().begin();
3671 LASSERT(!pit->empty(), /**/);
3672 pos_type right = pos_type(1);
3673 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
3674 right = pit->size();
3675 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
3681 static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
3683 Cursor & cur = bv->cursor();
3684 if (opt.repl_buf_name.empty()
3685 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
3686 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3689 DocIterator sel_beg = cur.selectionBegin();
3690 DocIterator sel_end = cur.selectionEnd();
3691 if (&sel_beg.inset() != &sel_end.inset()
3692 || sel_beg.pit() != sel_end.pit()
3693 || sel_beg.idx() != sel_end.idx())
3695 int sel_len = sel_end.pos() - sel_beg.pos();
3696 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
3697 << ", sel_len: " << sel_len << endl);
3700 LASSERT(sel_len > 0, return 0);
3702 if (!matchAdv(sel_beg, sel_len).match_len)
3705 // Build a copy of the replace buffer, adapted to the KeepCase option
3706 Buffer const & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
3708 repl_buffer_orig.write(oss);
3709 string lyx = oss.str();
3710 Buffer repl_buffer("", false);
3711 repl_buffer.setUnnamed(true);
3712 LASSERT(repl_buffer.readString(lyx), return 0);
3713 if (opt.keep_case && sel_len >= 2) {
3714 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
3715 if (cur.inTexted()) {
3716 if (firstUppercase(cur))
3717 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
3718 else if (allNonLowercase(cur, sel_len))
3719 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
3722 cap::cutSelection(cur, false);
3723 if (cur.inTexted()) {
3724 repl_buffer.changeLanguage(
3725 repl_buffer.language(),
3726 cur.getFont().language());
3727 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
3728 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
3729 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
3730 repl_buffer.params().documentClassPtr(),
3731 bv->buffer().errorList("Paste"));
3732 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
3733 sel_len = repl_buffer.paragraphs().begin()->size();
3734 } else if (cur.inMathed()) {
3735 odocstringstream ods;
3737 // OutputParams runparams(&repl_buffer.params().encoding());
3738 OutputParams runparams(encodings.fromLyXName("utf8"));
3739 runparams.nice = false;
3740 runparams.flavor = Flavor::XeTeX;
3741 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3742 runparams.dryrun = true;
3743 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
3744 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
3745 docstring repl_latex = ods.str();
3746 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
3748 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
3749 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
3750 repl_latex = from_utf8(s);
3751 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
3752 MathData ar(cur.buffer());
3753 asArray(repl_latex, ar, Parse::NORMAL);
3755 sel_len = ar.size();
3756 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3758 if (cur.pos() >= sel_len)
3759 cur.pos() -= sel_len;
3762 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3763 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
3764 bv->processUpdateFlags(Update::Force);
3769 /// Perform a FindAdv operation.
3770 bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt)
3775 // e.g., when invoking word-findadv from mini-buffer wither with
3776 // wrong options syntax or before ever opening advanced F&R pane
3777 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3781 MatchStringAdv matchAdv(bv->buffer(), opt);
3782 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
3784 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
3785 num_replaced += findAdvReplace(bv, opt, matchAdv);
3788 match_len = findForwardAdv(cur, matchAdv);
3790 match_len = findBackwardsAdv(cur, matchAdv);
3792 // This may only be raised by lyx::regex()
3793 bv->message(_("Invalid regular expression!"));
3797 if (match_len == 0) {
3798 if (num_replaced > 0) {
3799 switch (num_replaced)
3802 bv->message(_("One match has been replaced."));
3805 bv->message(_("Two matches have been replaced."));
3808 bv->message(bformat(_("%1$d matches have been replaced."), num_replaced));
3814 bv->message(_("Match not found."));
3819 if (num_replaced > 0)
3820 bv->message(_("Match has been replaced."));
3822 bv->message(_("Match found."));
3824 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << match_len);
3825 bv->putSelectionAt(cur, match_len, !opt.forward);
3831 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
3833 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
3834 << opt.casesensitive << ' '
3835 << opt.matchword << ' '
3836 << opt.forward << ' '
3837 << opt.expandmacros << ' '
3838 << opt.ignoreformat << ' '
3839 << opt.replace_all << ' '
3840 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
3841 << opt.keep_case << ' '
3842 << int(opt.scope) << ' '
3845 LYXERR(Debug::FIND, "built: " << os.str());
3851 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
3853 LYXERR(Debug::FIND, "parsing");
3857 while (line != "EOSS") {
3861 if (is.eof()) // Tolerate malformed request
3865 LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
3866 opt.find_buf_name = from_utf8(s);
3867 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.replace_all;
3868 is.get(); // Waste space before replace string
3871 while (line != "EOSS") {
3875 if (is.eof()) // Tolerate malformed request
3879 LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
3880 opt.repl_buf_name = from_utf8(s);
3881 is >> opt.keep_case;
3884 opt.scope = FindAndReplaceOptions::SearchScope(i);
3886 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
3888 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
3889 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
3890 << opt.scope << ' ' << opt.restr);