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
12 * Full author contact details are available in file CREDITS.
20 #include "buffer_funcs.h"
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"
32 #include "ParIterator.h"
36 #include "frontends/Application.h"
37 #include "frontends/alert.h"
39 #include "mathed/InsetMath.h"
40 #include "mathed/InsetMathGrid.h"
41 #include "mathed/InsetMathHull.h"
42 #include "mathed/MathData.h"
43 #include "mathed/MathStream.h"
44 #include "mathed/MathSupport.h"
46 #include "support/convert.h"
47 #include "support/debug.h"
48 #include "support/docstream.h"
49 #include "support/FileName.h"
50 #include "support/gettext.h"
51 #include "support/lassert.h"
52 #include "support/lstrings.h"
54 #include "support/regex.h"
58 using namespace lyx::support;
63 // Helper class for deciding what should be ignored
68 : ignoreFamily_(false), ignoreSeries_(false),
69 ignoreShape_(false), ignoreUnderline_(false),
70 ignoreMarkUp_(false), ignoreStrikeOut_(false),
71 ignoreSectioning_(false), ignoreFrontMatter_(false),
72 ignoreColor_(false), ignoreLanguage_(false) {}
74 bool getFamily() { return ignoreFamily_; };
76 bool getSeries() { return ignoreSeries_; };
78 bool getShape() { return ignoreShape_; };
80 bool getUnderline() { return ignoreUnderline_; };
82 bool getMarkUp() { return ignoreMarkUp_; };
84 bool getStrikeOut() { return ignoreStrikeOut_; };
86 bool getSectioning() { return ignoreSectioning_; };
88 bool getFrontMatter() { return ignoreFrontMatter_; };
90 bool getColor() { return ignoreColor_; };
92 bool getLanguage() { return ignoreLanguage_; };
94 void setIgnoreFormat(string type, bool value);
104 bool ignoreUnderline_;
108 bool ignoreStrikeOut_;
110 bool ignoreSectioning_;
112 bool ignoreFrontMatter_;
116 bool ignoreLanguage_;
120 void IgnoreFormats::setIgnoreFormat(string type, bool value)
122 if (type == "color") {
123 ignoreColor_ = value;
125 else if (type == "language") {
126 ignoreLanguage_ = value;
128 else if (type == "sectioning") {
129 ignoreSectioning_ = value;
130 ignoreFrontMatter_ = value;
132 else if (type == "font") {
133 ignoreSeries_ = value;
134 ignoreShape_ = value;
135 ignoreFamily_ = value;
137 else if (type == "series") {
138 ignoreSeries_ = value;
140 else if (type == "shape") {
141 ignoreShape_ = value;
143 else if (type == "family") {
144 ignoreFamily_ = value;
146 else if (type == "markup") {
147 ignoreMarkUp_ = value;
149 else if (type == "underline") {
150 ignoreUnderline_ = value;
152 else if (type == "strike") {
153 ignoreStrikeOut_ = value;
157 // The global variable that can be changed from outside
158 IgnoreFormats ignoreFormats;
161 void setIgnoreFormat(string type, bool value)
163 ignoreFormats.setIgnoreFormat(type, value);
169 bool parse_bool(docstring & howto)
174 howto = split(howto, var, ' ');
179 class MatchString : public binary_function<Paragraph, pos_type, int>
182 MatchString(docstring const & str, bool cs, bool mw)
183 : str(str), case_sens(cs), whole_words(mw)
186 // returns true if the specified string is at the specified position
187 // del specifies whether deleted strings in ct mode will be considered
188 int operator()(Paragraph const & par, pos_type pos, bool del = true) const
190 return par.find(str, case_sens, whole_words, pos, del);
198 // match whole words only
203 int findForward(DocIterator & cur, MatchString const & match,
204 bool find_del = true)
206 for (; cur; cur.forwardChar())
207 if (cur.inTexted()) {
208 int len = match(cur.paragraph(), cur.pos(), find_del);
216 int findBackwards(DocIterator & cur, MatchString const & match,
217 bool find_del = true)
221 if (cur.inTexted()) {
222 int len = match(cur.paragraph(), cur.pos(), find_del);
231 bool searchAllowed(docstring const & str)
234 frontend::Alert::error(_("Search error"), _("Search string is empty"));
241 bool findOne(BufferView * bv, docstring const & searchstr,
242 bool case_sens, bool whole, bool forward,
243 bool find_del = true, bool check_wrap = false)
245 if (!searchAllowed(searchstr))
248 DocIterator cur = forward
249 ? bv->cursor().selectionEnd()
250 : bv->cursor().selectionBegin();
252 MatchString const match(searchstr, case_sens, whole);
254 int match_len = forward
255 ? findForward(cur, match, find_del)
256 : findBackwards(cur, match, find_del);
259 bv->putSelectionAt(cur, match_len, !forward);
260 else if (check_wrap) {
261 DocIterator cur_orig(bv->cursor());
264 q = _("End of file reached while searching forward.\n"
265 "Continue searching from the beginning?");
267 q = _("Beginning of file reached while searching backward.\n"
268 "Continue searching from the end?");
269 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
270 q, 0, 1, _("&Yes"), _("&No"));
271 if (wrap_answer == 0) {
273 bv->cursor().clear();
274 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
276 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
277 bv->cursor().backwardPos();
279 bv->clearSelection();
280 if (findOne(bv, searchstr, case_sens, whole, forward, find_del, false))
283 bv->cursor().setCursor(cur_orig);
287 return match_len > 0;
291 int replaceAll(BufferView * bv,
292 docstring const & searchstr, docstring const & replacestr,
293 bool case_sens, bool whole)
295 Buffer & buf = bv->buffer();
297 if (!searchAllowed(searchstr) || buf.isReadonly())
300 DocIterator cur_orig(bv->cursor());
302 MatchString const match(searchstr, case_sens, whole);
305 int const rsize = replacestr.size();
306 int const ssize = searchstr.size();
309 cur.setCursor(doc_iterator_begin(&buf));
310 int match_len = findForward(cur, match, false);
311 while (match_len > 0) {
312 // Backup current cursor position and font.
313 pos_type const pos = cur.pos();
314 Font const font = cur.paragraph().getFontSettings(buf.params(), pos);
316 int striked = ssize -
317 cur.paragraph().eraseChars(pos, pos + match_len,
318 buf.params().track_changes);
319 cur.paragraph().insert(pos, replacestr, font,
320 Change(buf.params().track_changes
322 : Change::UNCHANGED));
323 for (int i = 0; i < rsize + striked; ++i)
326 match_len = findForward(cur, match, false);
329 bv->putSelectionAt(doc_iterator_begin(&buf), 0, false);
331 cur_orig.fixIfBroken();
332 bv->setCursor(cur_orig);
338 // the idea here is that we are going to replace the string that
339 // is selected IF it is the search string.
340 // if there is a selection, but it is not the search string, then
341 // we basically ignore it. (FIXME We ought to replace only within
343 // if there is no selection, then:
344 // (i) if some search string has been provided, then we find it.
345 // (think of how the dialog works when you hit "replace" the
347 // (ii) if no search string has been provided, then we treat the
348 // word the cursor is in as the search string. (why? i have no
349 // idea.) but this only works in text?
351 // returns the number of replacements made (one, if any) and
352 // whether anything at all was done.
353 pair<bool, int> replaceOne(BufferView * bv, docstring searchstr,
354 docstring const & replacestr, bool case_sens,
355 bool whole, bool forward, bool findnext)
357 Cursor & cur = bv->cursor();
359 if (!cur.selection()) {
360 // no selection, non-empty search string: find it
361 if (!searchstr.empty()) {
362 found = findOne(bv, searchstr, case_sens, whole, forward, true, findnext);
363 return make_pair(found, 0);
365 // empty search string
368 return make_pair(false, 0);
369 // select current word and treat it as the search string.
370 // This causes a minor bug as undo will restore this selection,
371 // which the user did not create (#8986).
372 cur.innerText()->selectWord(cur, WHOLE_WORD);
373 searchstr = cur.selectionAsString(false);
376 // if we still don't have a search string, report the error
378 if (!searchAllowed(searchstr))
379 return make_pair(false, 0);
381 bool have_selection = cur.selection();
382 docstring const selected = cur.selectionAsString(false);
385 ? searchstr == selected
386 : compare_no_case(searchstr, selected) == 0;
388 // no selection or current selection is not search word:
389 // just find the search word
390 if (!have_selection || !match) {
391 found = findOne(bv, searchstr, case_sens, whole, forward, true, findnext);
392 return make_pair(found, 0);
395 // we're now actually ready to replace. if the buffer is
396 // read-only, we can't, though.
397 if (bv->buffer().isReadonly())
398 return make_pair(false, 0);
400 cap::replaceSelectionWithString(cur, replacestr);
402 cur.pos() += replacestr.length();
403 LASSERT(cur.pos() <= cur.lastpos(),
404 cur.pos() = cur.lastpos());
407 findOne(bv, searchstr, case_sens, whole, forward, false, findnext);
409 return make_pair(true, 1);
415 docstring const find2string(docstring const & search,
416 bool casesensitive, bool matchword, bool forward)
420 << int(casesensitive) << ' '
421 << int(matchword) << ' '
427 docstring const replace2string(docstring const & replace,
428 docstring const & search,
429 bool casesensitive, bool matchword,
430 bool all, bool forward, bool findnext)
433 ss << replace << '\n'
435 << int(casesensitive) << ' '
436 << int(matchword) << ' '
438 << int(forward) << ' '
444 bool lyxfind(BufferView * bv, FuncRequest const & ev)
446 if (!bv || ev.action() != LFUN_WORD_FIND)
449 //lyxerr << "find called, cmd: " << ev << endl;
451 // data is of the form
453 // <casesensitive> <matchword> <forward>"
455 docstring howto = split(ev.argument(), search, '\n');
457 bool casesensitive = parse_bool(howto);
458 bool matchword = parse_bool(howto);
459 bool forward = parse_bool(howto);
461 return findOne(bv, search, casesensitive, matchword, forward, true, true);
465 bool lyxreplace(BufferView * bv,
466 FuncRequest const & ev, bool has_deleted)
468 if (!bv || ev.action() != LFUN_WORD_REPLACE)
471 // data is of the form
474 // <casesensitive> <matchword> <all> <forward> <findnext>"
477 docstring howto = split(ev.argument(), rplc, '\n');
478 howto = split(howto, search, '\n');
480 bool casesensitive = parse_bool(howto);
481 bool matchword = parse_bool(howto);
482 bool all = parse_bool(howto);
483 bool forward = parse_bool(howto);
484 bool findnext = howto.empty() ? true : parse_bool(howto);
489 int replace_count = 0;
491 replace_count = replaceAll(bv, search, rplc, casesensitive, matchword);
492 update = replace_count > 0;
495 replaceOne(bv, search, rplc, casesensitive, matchword, forward, findnext);
497 replace_count = rv.second;
500 Buffer const & buf = bv->buffer();
502 // emit message signal.
503 buf.message(_("String not found."));
505 if (replace_count == 0) {
506 buf.message(_("String found."));
507 } else if (replace_count == 1) {
508 buf.message(_("String has been replaced."));
510 docstring const str =
511 bformat(_("%1$d strings have been replaced."), replace_count);
515 } else if (findnext) {
516 // if we have deleted characters, we do not replace at all, but
517 // rather search for the next occurence
518 if (findOne(bv, search, casesensitive, matchword, forward, true, findnext))
521 bv->message(_("String not found."));
527 bool findNextChange(BufferView * bv, Cursor & cur, bool const check_wrap)
529 for (; cur; cur.forwardPos())
530 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
534 DocIterator cur_orig(bv->cursor());
535 docstring q = _("End of file reached while searching forward.\n"
536 "Continue searching from the beginning?");
537 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
538 q, 0, 1, _("&Yes"), _("&No"));
539 if (wrap_answer == 0) {
540 bv->cursor().clear();
541 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
542 bv->clearSelection();
543 cur.setCursor(bv->cursor().selectionBegin());
544 if (findNextChange(bv, cur, false))
547 bv->cursor().setCursor(cur_orig);
554 bool findPreviousChange(BufferView * bv, Cursor & cur, bool const check_wrap)
556 for (cur.backwardPos(); cur; cur.backwardPos()) {
557 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
562 DocIterator cur_orig(bv->cursor());
563 docstring q = _("Beginning of file reached while searching backward.\n"
564 "Continue searching from the end?");
565 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
566 q, 0, 1, _("&Yes"), _("&No"));
567 if (wrap_answer == 0) {
568 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
569 bv->cursor().backwardPos();
570 bv->clearSelection();
571 cur.setCursor(bv->cursor().selectionBegin());
572 if (findPreviousChange(bv, cur, false))
575 bv->cursor().setCursor(cur_orig);
582 bool selectChange(Cursor & cur, bool forward)
584 if (!cur.inTexted() || !cur.paragraph().isChanged(cur.pos()))
586 Change ch = cur.paragraph().lookupChange(cur.pos());
588 CursorSlice tip1 = cur.top();
589 for (; tip1.pit() < tip1.lastpit() || tip1.pos() < tip1.lastpos(); tip1.forwardPos()) {
590 Change ch2 = tip1.paragraph().lookupChange(tip1.pos());
591 if (!ch2.isSimilarTo(ch))
594 CursorSlice tip2 = cur.top();
595 for (; tip2.pit() > 0 || tip2.pos() > 0;) {
597 Change ch2 = tip2.paragraph().lookupChange(tip2.pos());
598 if (!ch2.isSimilarTo(ch)) {
599 // take a step forward to correctly set the selection
607 cur.bv().mouseSetCursor(cur, false);
609 cur.bv().mouseSetCursor(cur, true);
617 bool findChange(BufferView * bv, bool forward)
620 cur.setCursor(forward ? bv->cursor().selectionEnd()
621 : bv->cursor().selectionBegin());
622 forward ? findNextChange(bv, cur, true) : findPreviousChange(bv, cur, true);
623 return selectChange(cur, forward);
628 bool findNextChange(BufferView * bv)
630 return findChange(bv, true);
634 bool findPreviousChange(BufferView * bv)
636 return findChange(bv, false);
643 typedef vector<pair<string, string> > Escapes;
645 /// A map of symbols and their escaped equivalent needed within a regex.
646 /// @note Beware of order
647 Escapes const & get_regexp_escapes()
649 typedef std::pair<std::string, std::string> P;
651 static Escapes escape_map;
652 if (escape_map.empty()) {
653 escape_map.push_back(P("$", "_x_$"));
654 escape_map.push_back(P("{", "_x_{"));
655 escape_map.push_back(P("}", "_x_}"));
656 escape_map.push_back(P("[", "_x_["));
657 escape_map.push_back(P("]", "_x_]"));
658 escape_map.push_back(P("(", "_x_("));
659 escape_map.push_back(P(")", "_x_)"));
660 escape_map.push_back(P("+", "_x_+"));
661 escape_map.push_back(P("*", "_x_*"));
662 escape_map.push_back(P(".", "_x_."));
663 escape_map.push_back(P("\\", "(?:\\\\|\\\\backslash)"));
664 escape_map.push_back(P("~", "(?:\\\\textasciitilde|\\\\sim)"));
665 escape_map.push_back(P("^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\textasciicircum|\\\\mathcircumflex)"));
666 escape_map.push_back(P("_x_", "\\"));
671 /// A map of lyx escaped strings and their unescaped equivalent.
672 Escapes const & get_lyx_unescapes()
674 typedef std::pair<std::string, std::string> P;
676 static Escapes escape_map;
677 if (escape_map.empty()) {
678 escape_map.push_back(P("\\%", "%"));
679 escape_map.push_back(P("\\mathcircumflex ", "^"));
680 escape_map.push_back(P("\\mathcircumflex", "^"));
681 escape_map.push_back(P("\\backslash ", "\\"));
682 escape_map.push_back(P("\\backslash", "\\"));
683 escape_map.push_back(P("\\\\{", "_x_<"));
684 escape_map.push_back(P("\\\\}", "_x_>"));
685 escape_map.push_back(P("\\sim ", "~"));
686 escape_map.push_back(P("\\sim", "~"));
691 /// A map of escapes turning a regexp matching text to one matching latex.
692 Escapes const & get_regexp_latex_escapes()
694 typedef std::pair<std::string, std::string> P;
696 static Escapes escape_map;
697 if (escape_map.empty()) {
698 escape_map.push_back(P("\\\\", "(?:\\\\\\\\|\\\\backslash|\\\\textbackslash\\{\\}|\\\\textbackslash)"));
699 escape_map.push_back(P("(<?!\\\\\\\\textbackslash)\\{", "\\\\\\{"));
700 escape_map.push_back(P("(<?!\\\\\\\\textbackslash\\\\\\{)\\}", "\\\\\\}"));
701 escape_map.push_back(P("\\[", "\\{\\[\\}"));
702 escape_map.push_back(P("\\]", "\\{\\]\\}"));
703 escape_map.push_back(P("\\^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\textasciicircum|\\\\mathcircumflex)"));
704 escape_map.push_back(P("%", "\\\\\\%"));
705 escape_map.push_back(P("#", "\\\\#"));
710 /** @todo Probably the maps need to be migrated to regexps, in order to distinguish if
711 ** the found occurrence were escaped.
713 string apply_escapes(string s, Escapes const & escape_map)
715 LYXERR(Debug::FIND, "Escaping: '" << s << "'");
716 Escapes::const_iterator it;
717 for (it = escape_map.begin(); it != escape_map.end(); ++it) {
718 // LYXERR(Debug::FIND, "Escaping " << it->first << " as " << it->second);
719 unsigned int pos = 0;
720 while (pos < s.length() && (pos = s.find(it->first, pos)) < s.length()) {
721 s.replace(pos, it->first.length(), it->second);
722 LYXERR(Debug::FIND, "After escape: " << s);
723 pos += it->second.length();
724 // LYXERR(Debug::FIND, "pos: " << pos);
727 LYXERR(Debug::FIND, "Escaped : '" << s << "'");
732 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
733 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
734 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
735 string escape_for_regex(string s, bool match_latex)
738 while (pos < s.size()) {
739 size_t new_pos = s.find("\\regexp{", pos);
740 if (new_pos == string::npos)
744 LYXERR(Debug::FIND, "new_pos: " << new_pos);
745 t = apply_escapes(s.substr(pos, new_pos - pos), get_lyx_unescapes());
746 LYXERR(Debug::FIND, "t [lyx]: " << t);
747 t = apply_escapes(t, get_regexp_escapes());
748 LYXERR(Debug::FIND, "t [rxp]: " << t);
749 s.replace(pos, new_pos - pos, t);
750 new_pos = pos + t.size();
751 LYXERR(Debug::FIND, "Regexp after escaping: " << s);
752 LYXERR(Debug::FIND, "new_pos: " << new_pos);
753 if (new_pos == s.size())
756 // Might fail if \\endregexp{} is preceeded by unexpected stuff (weird escapes)
757 size_t end_pos = s.find("\\endregexp{}}", new_pos + 8);
758 LYXERR(Debug::FIND, "end_pos: " << end_pos);
759 t = s.substr(new_pos + 8, end_pos - (new_pos + 8));
760 LYXERR(Debug::FIND, "t in regexp : " << t);
761 t = apply_escapes(t, get_lyx_unescapes());
762 LYXERR(Debug::FIND, "t in regexp [lyx]: " << t);
764 t = apply_escapes(t, get_regexp_latex_escapes());
765 LYXERR(Debug::FIND, "t in regexp [ltx]: " << t);
767 if (end_pos == s.size()) {
768 s.replace(new_pos, end_pos - new_pos, t);
769 LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s);
772 s.replace(new_pos, end_pos + 13 - new_pos, t);
773 LYXERR(Debug::FIND, "Regexp after \\regexp{...\\endregexp{}} removal: " << s);
774 pos = new_pos + t.size();
775 LYXERR(Debug::FIND, "pos: " << pos);
781 /// Wrapper for lyx::regex_replace with simpler interface
782 bool regex_replace(string const & s, string & t, string const & searchstr,
783 string const & replacestr)
785 lyx::regex e(searchstr, regex_constants::ECMAScript);
787 ostream_iterator<char, char> it(oss);
788 lyx::regex_replace(it, s.begin(), s.end(), e, replacestr);
789 // tolerate t and s be references to the same variable
790 bool rv = (s != oss.str());
796 /** Checks if supplied string segment is well-formed from the standpoint of matching open-closed braces.
798 ** Verify that closed braces exactly match open braces. This avoids that, for example,
799 ** \frac{.*}{x} matches \frac{x+\frac{y}{x}}{z} with .* being 'x+\frac{y'.
802 ** Number of open braces that must remain open at the end for the verification to succeed.
804 bool braces_match(string::const_iterator const & beg,
805 string::const_iterator const & end,
809 string::const_iterator it = beg;
810 LYXERR(Debug::FIND, "Checking " << unmatched << " unmatched braces in '" << string(beg, end) << "'");
811 for (; it != end; ++it) {
812 // Skip escaped braces in the count
817 } else if (*it == '{') {
819 } else if (*it == '}') {
820 if (open_pars == 0) {
821 LYXERR(Debug::FIND, "Found unmatched closed brace");
827 if (open_pars != unmatched) {
828 LYXERR(Debug::FIND, "Found " << open_pars
829 << " instead of " << unmatched
830 << " unmatched open braces at the end of count");
833 LYXERR(Debug::FIND, "Braces match as expected");
843 MatchResult(): match_len(0),match2end(0), pos(0) {};
846 /** The class performing a match between a position in the document and the FindAdvOptions.
848 class MatchStringAdv {
850 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt);
852 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
853 ** constructor as opt.search, under the opt.* options settings.
856 ** If set, then match is searched only against beginning of text starting at cur.
857 ** If unset, then match is searched anywhere in text starting at cur.
860 ** The length of the matching text, or zero if no match was found.
862 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
867 /// first buffer on which search was started
868 lyx::Buffer * const p_first_buf;
870 FindAndReplaceOptions const & opt;
873 /// Auxiliary find method (does not account for opt.matchword)
874 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
876 /** Normalize a stringified or latexified LyX paragraph.
880 ** <li>if search is not casesensitive, then lowercase the string;
881 ** <li>remove any newline at begin or end of the string;
882 ** <li>replace any newline in the middle of the string with a simple space;
883 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
886 ** @todo Normalization should also expand macros, if the corresponding
887 ** search option was checked.
889 string normalize(docstring const & s, bool hack_braces) const;
890 // normalized string to search
891 string par_as_string;
892 // regular expression to use for searching
894 // same as regexp, but prefixed with a ".*?"
896 // leading format material as string
897 string lead_as_string;
898 // par_as_string after removal of lead_as_string
899 string par_as_string_nolead;
900 // unmatched open braces in the search string/regexp
902 // number of (.*?) subexpressions added at end of search regexp for closing
903 // environments, math mode, styles, etc...
905 // Are we searching with regular expressions ?
910 static docstring buffer_to_latex(Buffer & buffer)
912 OutputParams runparams(&buffer.params().encoding());
913 odocstringstream ods;
915 runparams.nice = true;
916 runparams.flavor = OutputParams::LATEX;
917 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
918 // No side effect of file copying and image conversion
919 runparams.dryrun = true;
920 runparams.for_search = true;
921 pit_type const endpit = buffer.paragraphs().size();
922 for (pit_type pit = 0; pit != endpit; ++pit) {
923 TeXOnePar(buffer, buffer.text(), pit, os, runparams);
924 LYXERR(Debug::FIND, "searchString up to here: " << ods.str());
930 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
933 if (!opt.ignoreformat) {
934 str = buffer_to_latex(buffer);
936 OutputParams runparams(&buffer.params().encoding());
937 runparams.nice = true;
938 runparams.flavor = OutputParams::LATEX;
939 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
940 runparams.dryrun = true;
941 runparams.for_search = true;
942 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
943 Paragraph const & par = buffer.paragraphs().at(pit);
944 LYXERR(Debug::FIND, "Adding to search string: '"
945 << par.asString(pos_type(0), par.size(),
946 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
949 str += par.asString(pos_type(0), par.size(),
950 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
958 /// Return separation pos between the leading material and the rest
959 static size_t identifyLeading(string const & s)
962 // @TODO Support \item[text]
963 // Kornel: Added textsl, textsf, textit, texttt and noun
964 // + allow to search for colored text too
965 while (regex_replace(t, t, REGEX_BOS "\\\\(((footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|emph|noun|minisec|text(bf|md|sl|sf|it|tt))|((textcolor|foreignlanguage|latexenvironment)\\{[a-z]+\\*?\\})|(u|uu)line|(s|x)out|uwave)|((sub)?(((sub)?section)|paragraph)|part|chapter)\\*?)\\{", "")
966 || regex_replace(t, t, REGEX_BOS "\\$", "")
967 || regex_replace(t, t, REGEX_BOS "\\\\\\[ ", "")
968 || regex_replace(t, t, REGEX_BOS " ?\\\\item\\{[a-z]+\\}", "")
969 || regex_replace(t, t, REGEX_BOS "\\\\begin\\{[a-zA-Z_]*\\*?\\} ", ""))
971 LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
976 * Given a latexified string, retrieve some handled features
977 * The features of the regex will later be compared with the features
978 * of the searched text. If the regex features are not a
979 * subset of the analized, then, in not format ignoring search
980 * we can early stop the search in the relevant inset.
982 typedef map<string, bool> Features;
984 static Features identifyFeatures(string const & s)
986 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
987 static regex const valid("^("
989 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
990 "emph|noun|text(bf|md|sl|sf|it|tt)|"
991 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
992 "(u|uu)line|(s|x)out|uwave|"
993 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
994 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
999 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1002 if (sub.str(1).compare("regexp") == 0) {
1006 string token = sub.str(1);
1008 if (regex_match(token, sub2, valid)) {
1016 if (sub.str(1).compare("endregexp") == 0) {
1026 * defines values features of a key "\\[a-z]+{"
1031 /* Char type with content discarded
1032 * like \hspace{1cm} */
1034 /* Char, like \backslash */
1036 /* \part, \section*, ... */
1038 /* title, author etc */
1040 /* \foreignlanguage{ngerman}, ... */
1043 * to discard language in content */
1046 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1048 /* fonts, colors, markups, ... */
1050 /* footnotesize, ... large, ...
1051 * Ignore all of them */
1054 /* inputencoding, ...
1055 * Discard also content, because they do not help in search */
1058 * like remove, but also all arguments */
1060 /* item, listitem */
1062 /* tex, latex, ... like isChar */
1064 /* like \lettrine[lines=5]{}{} */
1066 /* End of arguments marker for lettrine,
1067 * so that they can be ignored */
1077 parenthesiscount(1),
1081 KeyInfo(KeyType type, int parcount, bool disable)
1087 parenthesiscount(parcount),
1096 int parenthesiscount;
1098 bool used; /* by pattern */
1103 Border(int l=0, int u=0) : low(l), upper(u) {};
1108 #define MAXOPENED 30
1110 bool isPatternString;
1112 explicit Intervall(bool isPattern) :
1113 isPatternString(isPattern),
1115 actualdeptindex(0) { depts[0] = 0; closes[0] = 0;};
1118 int depts[MAXOPENED];
1119 int closes[MAXOPENED];
1120 int actualdeptindex;
1121 Border borders[2*MAXOPENED];
1122 int previousNotIgnored(int);
1123 int nextNotIgnored(int);
1124 void handleOpenP(int i);
1125 void handleCloseP(int i, bool closingAllowed);
1126 void resetOpenedP(int openPos);
1127 void addIntervall(int upper);
1128 void addIntervall(int low, int upper); /* if explicit */
1129 void setForDefaultLang(KeyInfo &defLang);
1130 int findclosing(int start, int end, char up, char down, int repeat);
1131 void handleParentheses(int lastpos, bool closingAllowed);
1133 int isOpeningPar(int pos);
1135 void output(ostringstream &os, int lastpos);
1136 // string show(int lastpos);
1139 int Intervall::isOpeningPar(int pos)
1141 if ((pos < 0) || (size_t(pos) >= par.size()))
1143 if (par[pos] != '{')
1145 if (size_t(pos) + 2 >= par.size())
1147 if (par[pos+2] != '}')
1149 if (par[pos+1] == '[' || par[pos+1] == ']')
1154 void Intervall::setForDefaultLang(KeyInfo &defLang)
1156 // Enable the use of first token again
1157 if (ignoreidx >= 0) {
1158 int value = defLang._tokenstart + defLang._tokensize;
1160 if (borders[0].low < value)
1161 borders[0].low = value;
1162 if (borders[0].upper < value)
1163 borders[0].upper = value;
1168 static void checkDepthIndex(int val)
1170 static int maxdepthidx = MAXOPENED-2;
1171 if (val > maxdepthidx) {
1173 LYXERR0("maxdepthidx now " << val);
1177 static void checkIgnoreIdx(int val)
1179 static int maxignoreidx = 2*MAXOPENED - 4;
1180 if (val > maxignoreidx) {
1182 LYXERR0("maxignoreidx now " << val);
1187 * Expand the region of ignored parts of the input latex string
1188 * The region is only relevant in output()
1190 void Intervall::addIntervall(int low, int upper)
1193 if (low == upper) return;
1194 for (idx = ignoreidx+1; idx > 0; --idx) {
1195 if (low > borders[idx-1].upper) {
1199 Border br(low, upper);
1200 if (idx > ignoreidx) {
1203 checkIgnoreIdx(ignoreidx);
1207 // Expand only if one of the new bound is inside the interwall
1208 // We know here that br.low > borders[idx-1].upper
1209 if (br.upper < borders[idx].low) {
1210 // We have to insert at this pos
1211 for (int i = ignoreidx+1; i > idx; --i) {
1212 borders[i] = borders[i-1];
1216 checkIgnoreIdx(ignoreidx);
1219 // Here we know, that we are overlapping
1220 if (br.low > borders[idx].low)
1221 br.low = borders[idx].low;
1222 // check what has to be concatenated
1224 for (int i = idx; i <= ignoreidx; i++) {
1225 if (br.upper >= borders[i].low) {
1227 if (br.upper < borders[i].upper)
1228 br.upper = borders[i].upper;
1234 // count should be >= 1 here
1237 for (int i = idx + count; i <= ignoreidx; i++) {
1238 borders[i-count+1] = borders[i];
1240 ignoreidx -= count - 1;
1246 void Intervall::handleOpenP(int i)
1249 depts[actualdeptindex] = i+1;
1250 closes[actualdeptindex] = -1;
1251 checkDepthIndex(actualdeptindex);
1254 void Intervall::handleCloseP(int i, bool closingAllowed)
1256 if (actualdeptindex <= 0) {
1257 if (! closingAllowed)
1258 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1259 // if we are at the very end
1260 addIntervall(i, i+1);
1263 closes[actualdeptindex] = i+1;
1268 void Intervall::resetOpenedP(int openPos)
1270 // Used as initializer for foreignlanguage entry
1271 actualdeptindex = 1;
1272 depts[1] = openPos+1;
1276 int Intervall::previousNotIgnored(int start)
1278 int idx = 0; /* int intervalls */
1279 for (idx = ignoreidx; idx >= 0; --idx) {
1280 if (start > borders[idx].upper)
1282 if (start >= borders[idx].low)
1283 start = borders[idx].low-1;
1288 int Intervall::nextNotIgnored(int start)
1290 int idx = 0; /* int intervalls */
1291 for (idx = 0; idx <= ignoreidx; idx++) {
1292 if (start < borders[idx].low)
1294 if (start < borders[idx].upper)
1295 start = borders[idx].upper;
1300 typedef map<string, KeyInfo> KeysMap;
1301 typedef vector< KeyInfo> Entries;
1302 static KeysMap keys = map<string, KeyInfo>();
1309 void buildKeys(bool);
1310 void buildEntries(bool);
1311 void makeKey(const string &, KeyInfo, bool isPatternString);
1312 void processRegion(int start, int region_end); /* remove {} parts */
1313 void removeHead(KeyInfo&, int count=0);
1316 LatexInfo(string par, bool isPatternString) : entidx(-1), interval(isPatternString) {
1318 interval.hasTitle = false;
1319 interval.titleValue = "";
1320 buildKeys(isPatternString);
1321 entries = vector<KeyInfo>();
1322 buildEntries(isPatternString);
1326 if (entries.empty()) {
1329 if (entries[0].keytype == KeyInfo::isTitle) {
1330 if (! entries[0].disabled) {
1331 interval.hasTitle = true;
1332 interval.titleValue = entries[0].head;
1335 interval.hasTitle = false;
1336 interval.titleValue = "";
1338 removeHead(entries[0]);
1339 if (entries.size() > 1)
1348 if (int(entries.size()) > entidx) {
1355 bool setNextKey(int idx) {
1356 if ((idx == entidx) && (entidx >= 0)) {
1363 int find(int start, KeyInfo::KeyType keytype) {
1367 while (tmpIdx < int(entries.size())) {
1368 if (entries[tmpIdx].keytype == keytype)
1374 int process(ostringstream &os, KeyInfo &actual);
1375 int dispatch(ostringstream &os, int previousStart, KeyInfo &actual);
1376 // string show(int lastpos) { return interval.show(lastpos);};
1377 int nextNotIgnored(int start) { return interval.nextNotIgnored(start);};
1378 KeyInfo &getKeyInfo(int keyinfo) {
1379 static KeyInfo invalidInfo = KeyInfo();
1380 if ((keyinfo < 0) || ( keyinfo >= int(entries.size())))
1383 return entries[keyinfo];
1385 void setForDefaultLang(KeyInfo &defLang) {interval.setForDefaultLang(defLang);};
1386 void addIntervall(int low, int up) { interval.addIntervall(low, up); };
1390 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
1395 for (int i = start; i < end; i += 1 + skip) {
1399 if (c == '\\') skip = 1;
1403 else if (c == down) {
1405 if ((repeat <= 0) || (par[i+1] != up))
1423 vector<MathEntry> entries;
1428 void insert(string wait, size_t start, size_t end) {
1429 MathEntry m = MathEntry();
1431 m.mathStart = start;
1433 m.mathSize = end - start;
1434 entries.push_back(m);
1436 bool empty() { return entries.empty(); };
1437 size_t getEndPos() {
1438 if (entries.empty() || (actualIdx >= entries.size())) {
1441 return entries[actualIdx].mathEnd;
1443 size_t getStartPos() {
1444 if (entries.empty() || (actualIdx >= entries.size())) {
1445 return 100000; /* definitely enough? */
1447 return entries[actualIdx].mathStart;
1449 size_t getFirstPos() {
1451 return getStartPos();
1454 if (entries.empty() || (actualIdx >= entries.size())) {
1457 return entries[actualIdx].mathSize;
1459 void incrEntry() { actualIdx++; };
1462 void LatexInfo::buildEntries(bool isPatternString)
1464 static regex const rmath("\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|alignat)\\*?)\\}");
1465 static regex const rkeys("\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?))");
1466 static bool disableLanguageOverride = false;
1467 smatch sub, submath;
1468 bool evaluatingRegexp = false;
1470 bool evaluatingMath = false;
1471 bool evaluatingCode = false;
1473 bool evaluatingOptional = false;
1474 size_t optionalEnd = 0;
1477 bool math_end_waiting = false;
1478 size_t math_pos = 10000;
1481 for (sregex_iterator itmath(interval.par.begin(), interval.par.end(), rmath), end; itmath != end; ++itmath) {
1483 if (math_end_waiting) {
1484 size_t pos = submath.position(size_t(0));
1485 if ((math_end == "$") &&
1486 (submath.str(0) == "$") &&
1487 (interval.par[pos-1] != '\\')) {
1488 mi.insert("$", math_pos, pos + 1);
1489 math_end_waiting = false;
1491 else if ((math_end == "\\]") &&
1492 (submath.str(0) == "\\]")) {
1493 mi.insert("\\]", math_pos, pos + 2);
1494 math_end_waiting = false;
1496 else if ((submath.str(1).compare("end") == 0) &&
1497 (submath.str(2).compare(math_end) == 0)) {
1498 mi.insert(math_end, math_pos, pos + submath.str(0).length());
1499 math_end_waiting = false;
1505 if (submath.str(1).compare("begin") == 0) {
1506 math_end_waiting = true;
1507 math_end = submath.str(2);
1508 math_pos = submath.position(size_t(0));
1510 else if (submath.str(0).compare("\\[") == 0) {
1511 math_end_waiting = true;
1513 math_pos = submath.position(size_t(0));
1515 else if (submath.str(0) == "$") {
1516 size_t pos = submath.position(size_t(0));
1517 if ((pos == 0) || (interval.par[pos-1] != '\\')) {
1518 math_end_waiting = true;
1525 // Ignore language if there is math somewhere in pattern-string
1526 if (isPatternString) {
1529 keys["foreignlanguage"].disabled = true;
1530 disableLanguageOverride = true;
1533 disableLanguageOverride = false;
1536 if (disableLanguageOverride) {
1537 keys["foreignlanguage"].disabled = true;
1540 math_pos = mi.getFirstPos();
1541 for (sregex_iterator it(interval.par.begin(), interval.par.end(), rkeys), end; it != end; ++it) {
1543 string key = sub.str(3);
1545 if (sub.str(0)[0] == '\\')
1546 key = sub.str(0)[1];
1550 size_t k_pos = sub.position(size_t(0));
1551 if ((k_pos > 0) && (interval.par[k_pos - 1] == '\\')) {
1552 // Escaped '$', ignoring
1558 if (evaluatingRegexp) {
1559 if (sub.str(1).compare("endregexp") == 0) {
1560 evaluatingRegexp = false;
1561 // found._tokenstart already set
1562 found._dataEnd = sub.position(size_t(0)) + 13;
1563 found._dataStart = found._dataEnd;
1564 found._tokensize = found._dataEnd - found._tokenstart;
1565 found.parenthesiscount = 0;
1566 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1572 if (evaluatingMath) {
1573 if (size_t(sub.position(size_t(0))) < mi.getEndPos())
1575 evaluatingMath = false;
1577 math_pos = mi.getStartPos();
1579 if (keys.find(key) == keys.end()) {
1580 found = KeyInfo(KeyInfo::isStandard, 0, true);
1581 if (isPatternString) {
1582 found.keytype = KeyInfo::isChar;
1583 found.disabled = false;
1590 if (key.compare("regexp") == 0) {
1591 evaluatingRegexp = true;
1592 found._tokenstart = sub.position(size_t(0));
1593 found._tokensize = 0;
1597 // Handle the other params of key
1598 if (found.keytype == KeyInfo::isIgnored)
1600 else if (found.keytype == KeyInfo::isMath) {
1601 if (size_t(sub.position(size_t(0))) == math_pos) {
1603 found._tokenstart = sub.position(size_t(0));
1604 found._tokensize = mi.getSize();
1605 found._dataEnd = found._tokenstart + found._tokensize;
1606 found._dataStart = found._dataEnd;
1607 found.parenthesiscount = 0;
1608 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1609 evaluatingMath = true;
1612 // begin|end of unknown env, discard
1613 // First handle tables
1614 // longtable|tabular
1615 bool discardComment;
1617 found.keytype = KeyInfo::doRemove;
1618 if ((sub.str(5).compare("longtable") == 0) ||
1619 (sub.str(5).compare("tabular") == 0)) {
1620 discardComment = true; /* '%' */
1623 discardComment = false;
1624 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
1626 string token = sub.str(5);
1627 if (regex_match(token, sub2, removeArgs)) {
1628 found.keytype = KeyInfo::removeWithArg;
1631 // discard spaces before pos(0)
1632 int pos = sub.position(size_t(0));
1634 for (count = 0; pos - count > 0; count++) {
1635 char c = interval.par[pos-count-1];
1636 if (discardComment) {
1637 if ((c != ' ') && (c != '%'))
1643 found._tokenstart = pos - count;
1644 if (sub.str(1).compare(0, 5, "begin") == 0) {
1645 size_t pos1 = pos + sub.str(0).length();
1646 if (sub.str(5).compare("cjk") == 0) {
1647 pos1 = interval.findclosing(pos1+1, interval.par.length()) + 1;
1648 if ((interval.par[pos1] == '{') && (interval.par[pos1+1] == '}'))
1650 found.keytype = KeyInfo::isMain;
1651 found._dataStart = pos1;
1652 found._dataEnd = interval.par.length();
1653 found.disabled = keys["foreignlanguage"].disabled;
1654 found.used = keys["foreignlanguage"].used;
1655 found._tokensize = pos1 - found._tokenstart;
1656 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1659 // Swallow possible optional params
1660 while (interval.par[pos1] == '[') {
1661 pos1 = interval.findclosing(pos1+1, interval.par.length(), '[', ']')+1;
1663 // Swallow also the eventual parameter
1664 if (interval.par[pos1] == '{') {
1665 found._dataEnd = interval.findclosing(pos1+1, interval.par.length()) + 1;
1668 found._dataEnd = pos1;
1670 found._dataStart = found._dataEnd;
1671 found._tokensize = count + found._dataEnd - pos;
1672 found.parenthesiscount = 0;
1673 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1674 found.disabled = true;
1678 // Handle "\end{...}"
1679 found._dataStart = pos + sub.str(0).length();
1680 found._dataEnd = found._dataStart;
1681 found._tokensize = count + found._dataEnd - pos;
1682 found.parenthesiscount = 0;
1683 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1684 found.disabled = true;
1688 else if (found.keytype != KeyInfo::isRegex) {
1689 found._tokenstart = sub.position(size_t(0));
1690 if (found.parenthesiscount == 0) {
1691 // Probably to be discarded
1692 size_t following_pos = sub.position(size_t(0)) + sub.str(3).length() + 1;
1693 char following = interval.par[following_pos];
1694 if (following == ' ')
1695 found.head = "\\" + sub.str(3) + " ";
1696 else if (following == '=') {
1697 // like \uldepth=1000pt
1698 found.head = sub.str(0);
1701 found.head = "\\" + key;
1702 found._tokensize = found.head.length();
1703 found._dataEnd = found._tokenstart + found._tokensize;
1704 found._dataStart = found._dataEnd;
1707 int params = found._tokenstart + key.length() + 1;
1708 if (evaluatingOptional) {
1709 if (size_t(found._tokenstart) > optionalEnd) {
1710 evaluatingOptional = false;
1713 found.disabled = true;
1716 int optend = params;
1717 while (interval.par[optend] == '[') {
1718 // discard optional parameters
1719 optend = interval.findclosing(optend+1, interval.par.length(), '[', ']') + 1;
1721 if (optend > params) {
1722 key += interval.par.substr(params, optend-params);
1723 evaluatingOptional = true;
1724 optionalEnd = optend;
1726 string token = sub.str(5);
1727 int closings = found.parenthesiscount;
1728 if (found.parenthesiscount == 1) {
1729 found.head = "\\" + key + "{";
1731 else if (found.parenthesiscount > 1) {
1733 found.head = sub.str(0) + "{";
1734 closings = found.parenthesiscount - 1;
1737 found.head = "\\" + key + "{";
1740 found._tokensize = found.head.length();
1741 found._dataStart = found._tokenstart + found.head.length();
1742 if (interval.par.substr(found._dataStart-1, 15).compare("\\endarguments{}") == 0) {
1743 found._dataStart += 15;
1745 size_t endpos = interval.findclosing(found._dataStart, interval.par.length(), '{', '}', closings);
1746 if (found.keytype == KeyInfo::isList) {
1747 // Check if it really is list env
1748 static regex const listre("^([a-z]+)$");
1750 if (!regex_match(token, sub2, listre)) {
1751 // Change the key of this entry. It is not in a list/item environment
1752 found.keytype = KeyInfo::endArguments;
1755 if (found.keytype == KeyInfo::noMain) {
1756 evaluatingCode = true;
1758 codeStart = found._dataStart;
1760 else if (evaluatingCode) {
1761 if (size_t(found._dataStart) > codeEnd)
1762 evaluatingCode = false;
1763 else if (found.keytype == KeyInfo::isMain) {
1764 // Disable this key, treate it as standard
1765 found.keytype = KeyInfo::isStandard;
1766 found.disabled = true;
1767 if ((codeEnd == interval.par.length()) &&
1768 (found._tokenstart == codeStart)) {
1769 // trickery, because the code inset starts
1770 // with \selectlanguage ...
1772 if (entries.size() > 1) {
1773 entries[entries.size()-1]._dataEnd = codeEnd;
1778 if ((endpos == interval.par.length()) &&
1779 (found.keytype == KeyInfo::doRemove)) {
1780 // Missing closing => error in latex-input?
1781 // therefore do not delete remaining data
1782 found._dataStart -= 1;
1783 found._dataEnd = found._dataStart;
1786 found._dataEnd = endpos;
1788 if (isPatternString) {
1789 keys[key].used = true;
1792 entries.push_back(found);
1796 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
1798 stringstream s(keysstring);
1800 const char delim = '|';
1801 while (getline(s, key, delim)) {
1802 KeyInfo keyII(keyI);
1803 if (isPatternString) {
1806 else if ( !keys[key].used)
1807 keyII.disabled = true;
1812 void LatexInfo::buildKeys(bool isPatternString)
1815 static bool keysBuilt = false;
1816 if (keysBuilt && !isPatternString) return;
1818 // Known standard keys with 1 parameter.
1819 // Split is done, if not at start of region
1820 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
1821 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
1822 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
1823 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
1824 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
1825 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
1827 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
1828 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1829 makeKey("section*|subsection*|subsubsection*|paragraph*",
1830 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1831 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1832 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
1834 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
1836 // Split is done, if not at start of region
1837 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
1838 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
1840 // Split is done always.
1841 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
1845 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1846 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1847 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1848 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1850 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1851 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1852 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1854 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1855 // Custom space/skip, remove the content (== length value)
1856 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
1857 // Found in fr/UserGuide.lyx
1858 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1860 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1861 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1862 // Known macros to remove (including their parameter)
1864 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
1865 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
1866 // handle like standard keys with 1 parameter.
1867 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
1869 // Macros to remove, but let the parameter survive
1871 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1873 // Remove language spec from content of these insets
1874 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
1876 // Same effect as previous, parameter will survive (because there is no one anyway)
1878 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1879 // Remove table decorations
1880 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
1881 // Discard shape-header.
1882 // For footnote or shortcut too, because of lang settings
1883 // and wrong handling if used 'KeyInfo::noMain'
1884 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1885 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1886 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1887 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1888 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1889 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
1890 // like ('tiny{}' or '\tiny ' ... )
1891 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
1893 // Survives, like known character
1894 makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1895 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
1897 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1898 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1899 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1901 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1902 // Remove RTL/LTR marker
1903 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1904 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
1905 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
1906 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
1907 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
1908 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
1909 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1910 if (isPatternString) {
1911 // Allow the first searched string to rebuild the keys too
1915 // no need to rebuild again
1921 * Keep the list of actual opened parentheses actual
1922 * (e.g. depth == 4 means there are 4 '{' not processed yet)
1924 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
1927 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
1931 if (c == '\\') skip = 1;
1932 else if (c == '{') {
1935 else if (c == '}') {
1936 handleCloseP(i, closingAllowed);
1942 string Intervall::show(int lastpos)
1944 int idx = 0; /* int intervalls */
1947 for (idx = 0; idx <= ignoreidx; idx++) {
1948 while (i < lastpos) {
1950 if (i <= borders[idx].low) {
1951 if (borders[idx].low > lastpos)
1952 printsize = lastpos - i;
1954 printsize = borders[idx].low - i;
1955 s += par.substr(i, printsize);
1957 if (i >= borders[idx].low)
1958 i = borders[idx].upper;
1961 i = borders[idx].upper;
1967 s += par.substr(i, lastpos-i);
1973 void Intervall::output(ostringstream &os, int lastpos)
1975 // get number of chars to output
1976 int idx = 0; /* int intervalls */
1979 string startTitle = titleValue;
1980 for (idx = 0; idx <= ignoreidx; idx++) {
1982 if (i <= borders[idx].low) {
1984 if (borders[idx].low > lastpos)
1985 printsize = lastpos - i;
1987 printsize = borders[idx].low - i;
1988 if (printsize > 0) {
1989 os << startTitle << par.substr(i, printsize);
1991 printed += printsize;
1994 handleParentheses(i, false);
1995 if (i >= borders[idx].low)
1996 i = borders[idx].upper;
1999 i = borders[idx].upper;
2006 os << startTitle << par.substr(i, lastpos-i);
2007 printed += lastpos-i;
2009 handleParentheses(lastpos, false);
2010 for (int i = actualdeptindex; i > 0; --i) {
2013 if (hasTitle && (printed > 0))
2015 if (! isPatternString)
2017 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2020 void LatexInfo::processRegion(int start, int region_end)
2022 while (start < region_end) { /* Let {[} and {]} survive */
2023 int cnt = interval.isOpeningPar(start);
2025 // Closing is allowed past the region
2026 int closing = interval.findclosing(start+1, interval.par.length());
2027 interval.addIntervall(start, start+1);
2028 interval.addIntervall(closing, closing+1);
2032 start = interval.nextNotIgnored(start+1);
2036 void LatexInfo::removeHead(KeyInfo &actual, int count)
2038 if (actual.parenthesiscount == 0) {
2039 // "{\tiny{} ...}" ==> "{{} ...}"
2040 interval.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2043 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2044 interval.addIntervall(actual._tokenstart - count, actual._dataStart);
2045 interval.addIntervall(actual._dataEnd, actual._dataEnd+1);
2049 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2052 switch (actual.keytype)
2054 case KeyInfo::isTitle: {
2056 nextKeyIdx = getNextKey();
2059 case KeyInfo::cleanToStart: {
2060 actual._dataEnd = actual._dataStart;
2061 nextKeyIdx = getNextKey();
2062 // Search for end of arguments
2063 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2065 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2066 entries[i].disabled = true;
2068 actual._dataEnd = entries[tmpIdx]._dataEnd;
2070 while (interval.par[actual._dataEnd] == ' ')
2072 interval.addIntervall(0, actual._dataEnd+1);
2073 interval.actualdeptindex = 0;
2074 interval.depts[0] = actual._dataEnd+1;
2075 interval.closes[0] = -1;
2078 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2079 if (actual.disabled)
2080 interval.addIntervall(actual._tokenstart, actual._dataEnd);
2082 interval.addIntervall(actual._dataStart, actual._dataEnd);
2085 case KeyInfo::isChar: {
2086 nextKeyIdx = getNextKey();
2089 case KeyInfo::isSize: {
2090 if (actual.disabled || (interval.par[actual._dataStart] != '{') || (interval.par[actual._dataStart-1] == ' ')) {
2091 processRegion(actual._dataEnd, actual._dataEnd+1); /* remove possibly following {} */
2092 interval.addIntervall(actual._tokenstart, actual._dataEnd+1);
2093 nextKeyIdx = getNextKey();
2095 // Here _dataStart points to '{', so correct it
2096 actual._dataStart += 1;
2097 actual._tokensize += 1;
2098 actual.parenthesiscount = 1;
2099 if (interval.par[actual._dataStart] == '}') {
2100 // Determine the end if used like '{\tiny{}...}'
2101 actual._dataEnd = interval.findclosing(actual._dataStart+1, interval.par.length()) + 1;
2102 interval.addIntervall(actual._dataStart, actual._dataStart+1);
2105 // Determine the end if used like '\tiny{...}'
2106 actual._dataEnd = interval.findclosing(actual._dataStart, interval.par.length()) + 1;
2108 // Split on this key if not at start
2109 int start = interval.nextNotIgnored(previousStart);
2110 if (start < actual._tokenstart) {
2111 interval.output(os, actual._tokenstart);
2112 interval.addIntervall(start, actual._tokenstart);
2114 // discard entry if at end of actual
2115 nextKeyIdx = process(os, actual);
2119 case KeyInfo::endArguments:
2120 // Remove trailing '{}' too
2121 actual._dataStart += 2;
2122 actual._dataEnd += 2;
2123 interval.addIntervall(actual._tokenstart, actual._dataEnd);
2124 nextKeyIdx = getNextKey();
2126 case KeyInfo::noMain:
2128 case KeyInfo::isStandard: {
2129 if (actual.disabled) {
2131 processRegion(actual._dataStart, actual._dataStart+1);
2132 nextKeyIdx = getNextKey();
2134 // Split on this key if not at datastart of calling entry
2135 int start = interval.nextNotIgnored(previousStart);
2136 if (start < actual._tokenstart) {
2137 interval.output(os, actual._tokenstart);
2138 interval.addIntervall(start, actual._tokenstart);
2140 // discard entry if at end of actual
2141 nextKeyIdx = process(os, actual);
2145 case KeyInfo::removeWithArg: {
2146 nextKeyIdx = getNextKey();
2147 // Search for end of arguments
2148 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2150 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2151 entries[i].disabled = true;
2153 actual._dataEnd = entries[tmpIdx]._dataEnd;
2155 interval.addIntervall(actual._tokenstart, actual._dataEnd+1);
2158 case KeyInfo::doRemove: {
2159 // Remove the key with all parameters and following spaces
2161 for (pos = actual._dataEnd+1; pos < interval.par.length(); pos++) {
2162 if ((interval.par[pos] != ' ') && (interval.par[pos] != '%'))
2165 interval.addIntervall(actual._tokenstart, pos);
2166 nextKeyIdx = getNextKey();
2169 case KeyInfo::isList: {
2170 // Discard space before _tokenstart
2172 for (count = 0; count < actual._tokenstart; count++) {
2173 if (interval.par[actual._tokenstart-count-1] != ' ')
2176 nextKeyIdx = getNextKey();
2177 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2179 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2181 // How else can we catch this one?
2182 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2183 entries[i].disabled = true;
2185 actual._dataEnd = entries[tmpIdx]._dataEnd;
2187 else if (nextKeyIdx > 0) {
2188 // Ignore any lang entries inside data region
2189 for (int i = nextKeyIdx; i < int(entries.size()) && entries[i]._tokenstart < actual._dataEnd; i++) {
2190 if (entries[i].keytype == KeyInfo::isMain)
2191 entries[i].disabled = true;
2194 if (actual.disabled) {
2195 interval.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2198 interval.addIntervall(actual._tokenstart-count, actual._tokenstart);
2200 if (interval.par[actual._dataEnd+1] == '[') {
2201 int posdown = interval.findclosing(actual._dataEnd+2, interval.par.length(), '[', ']');
2202 if ((interval.par[actual._dataEnd+2] == '{') &&
2203 (interval.par[posdown-1] == '}')) {
2204 interval.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
2205 interval.addIntervall(posdown-1, posdown+1);
2208 interval.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
2209 interval.addIntervall(posdown, posdown+1);
2211 int blk = interval.nextNotIgnored(actual._dataEnd+1);
2212 if (blk > posdown) {
2213 // Discard at most 1 space after empty item
2215 for (count = 0; count < 1; count++) {
2216 if (interval.par[blk+count] != ' ')
2220 interval.addIntervall(blk, blk+count);
2225 case KeyInfo::isSectioning: {
2226 // Discard spaces before _tokenstart
2228 int val = actual._tokenstart;
2229 for (count = 0; count < actual._tokenstart;) {
2230 val = interval.previousNotIgnored(val-1);
2231 if (interval.par[val] != ' ')
2234 count = actual._tokenstart - val;
2237 if (actual.disabled) {
2238 removeHead(actual, count);
2239 nextKeyIdx = getNextKey();
2241 interval.addIntervall(actual._tokenstart-count, actual._tokenstart);
2242 nextKeyIdx = process(os, actual);
2246 case KeyInfo::isMath: {
2247 // Same as regex, use the content unchanged
2248 nextKeyIdx = getNextKey();
2251 case KeyInfo::isRegex: {
2252 // DO NOT SPLIT ON REGEX
2254 nextKeyIdx = getNextKey();
2257 case KeyInfo::isIgnored: {
2258 // Treat like a character for now
2259 nextKeyIdx = getNextKey();
2262 case KeyInfo::isMain: {
2263 if (interval.par.substr(actual._dataStart, 2) == "% ")
2264 interval.addIntervall(actual._dataStart, actual._dataStart+2);
2265 if (actual._tokenstart > 0) {
2266 int prev = interval.previousNotIgnored(actual._tokenstart - 1);
2267 if ((prev >= 0) && interval.par[prev] == '%')
2268 interval.addIntervall(prev, prev+1);
2270 if (actual.disabled) {
2272 if ((interval.par.substr(actual._dataStart, 3) == " \\[") ||
2273 (interval.par.substr(actual._dataStart, 8) == " \\begin{")) {
2274 // Discard also the space before math-equation
2275 interval.addIntervall(actual._dataStart, actual._dataStart+1);
2277 nextKeyIdx = getNextKey();
2278 // interval.resetOpenedP(actual._dataStart-1);
2281 if (actual._tokenstart < 26) {
2282 // for the first (and maybe dummy) language
2283 interval.setForDefaultLang(actual);
2285 interval.resetOpenedP(actual._dataStart-1);
2289 case KeyInfo::invalid:
2290 // This cannot happen, already handled
2293 // LYXERR0("Unhandled keytype");
2294 nextKeyIdx = getNextKey();
2301 int LatexInfo::process(ostringstream &os, KeyInfo &actual )
2303 int end = interval.nextNotIgnored(actual._dataEnd);
2304 int oldStart = actual._dataStart;
2305 int nextKeyIdx = getNextKey();
2307 if ((nextKeyIdx < 0) ||
2308 (entries[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
2309 (entries[nextKeyIdx].keytype == KeyInfo::invalid)) {
2310 if (oldStart <= end) {
2311 processRegion(oldStart, end);
2316 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
2318 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
2319 (void) dispatch(os, actual._dataStart, nextKey);
2320 end = nextKey._tokenstart;
2323 processRegion(oldStart, nextKey._tokenstart);
2324 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
2326 oldStart = nextKey._dataEnd+1;
2328 // now nextKey is either invalid or is outside of actual._dataEnd
2329 // output the remaining and discard myself
2330 if (oldStart <= end) {
2331 processRegion(oldStart, end);
2333 if (interval.par[end] == '}') {
2335 // This is the normal case.
2336 // But if using the firstlanguage, the closing may be missing
2338 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
2340 if (actual._dataEnd < end)
2341 output_end = interval.nextNotIgnored(actual._dataEnd);
2343 output_end = interval.nextNotIgnored(end);
2344 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
2345 interval.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
2347 // Remove possible empty data
2348 int dstart = interval.nextNotIgnored(actual._dataStart);
2349 while (interval.isOpeningPar(dstart) == 1) {
2350 interval.addIntervall(dstart, dstart+1);
2351 int dend = interval.findclosing(dstart+1, output_end);
2352 interval.addIntervall(dend, dend+1);
2353 dstart = interval.nextNotIgnored(dstart+1);
2355 if (dstart < output_end)
2356 interval.output(os, output_end);
2357 interval.addIntervall(actual._tokenstart, end);
2361 string splitOnKnownMacros(string par, bool isPatternString)
2364 LatexInfo li(par, isPatternString);
2365 // LYXERR0("Berfore split: " << par);
2366 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
2368 DummyKey._tokensize = 0;
2369 DummyKey._dataStart = 0;
2370 DummyKey._dataEnd = par.length();
2371 DummyKey.disabled = true;
2372 int firstkeyIdx = li.getFirstKey();
2374 if (firstkeyIdx >= 0) {
2375 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
2376 DummyKey._tokenstart = firstKey._tokenstart;
2378 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
2379 // Use dummy firstKey
2380 firstKey = DummyKey;
2381 (void) li.setNextKey(firstkeyIdx);
2384 if (par.substr(firstKey._dataStart, 2) == "% ")
2385 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
2387 nextkeyIdx = li.process(os, firstKey);
2388 while (nextkeyIdx >= 0) {
2389 // Check for a possible gap between the last
2390 // entry and this one
2391 int datastart = li.nextNotIgnored(firstKey._dataStart);
2392 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
2393 if ((nextKey._tokenstart > datastart)) {
2395 firstKey._dataStart = datastart;
2396 firstKey._dataEnd = par.length();
2397 (void) li.setNextKey(nextkeyIdx);
2398 // Fake the last opened parenthesis
2399 li.setForDefaultLang(firstKey);
2400 nextkeyIdx = li.process(os, firstKey);
2403 if (nextKey.keytype != KeyInfo::isMain) {
2404 firstKey._dataStart = datastart;
2405 firstKey._dataEnd = nextKey._dataEnd+1;
2406 (void) li.setNextKey(nextkeyIdx);
2407 li.setForDefaultLang(firstKey);
2408 nextkeyIdx = li.process(os, firstKey);
2411 nextkeyIdx = li.process(os, nextKey);
2415 // Handle the remaining
2416 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
2417 firstKey._dataEnd = par.length();
2419 if ((firstKey._dataStart < firstKey._dataEnd) &&
2420 (par[firstKey._dataStart] != '}')) {
2421 li.setForDefaultLang(firstKey);
2422 (void) li.process(os, firstKey);
2426 // return string definitelly impossible to match
2427 s = "\\foreignlanguage{ignore}{ }";
2431 s = par; /* no known macros found */
2432 // LYXERR0("After split: " << s);
2437 * Try to unify the language specs in the latexified text.
2438 * Resulting modified string is set to "", if
2439 * the searched tex does not contain all the features in the search pattern
2441 static string correctlanguagesetting(string par, bool isPatternString, bool withformat)
2443 static Features regex_f;
2444 static int missed = 0;
2445 static bool regex_with_format = false;
2447 int parlen = par.length();
2449 while ((parlen > 0) && (par[parlen-1] == '\n')) {
2452 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
2453 // Happens to be there in case of description or labeling environment
2458 // Split the latex input into pieces which
2459 // can be digested by our search engine
2460 LYXERR(Debug::FIND, "input: \"" << par << "\"");
2461 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
2462 LYXERR(Debug::FIND, "After split: \"" << result << "\"");
2465 result = par.substr(0, parlen);
2466 if (isPatternString) {
2469 regex_f = identifyFeatures(result);
2470 string features = "";
2471 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2472 string a = it->first;
2473 regex_with_format = true;
2474 features += " " + a;
2475 // LYXERR0("Identified regex format:" << a);
2477 LYXERR(Debug::FIND, "Identified Features" << features);
2480 } else if (regex_with_format) {
2481 Features info = identifyFeatures(result);
2482 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2483 string a = it->first;
2484 bool b = it->second;
2485 if (b && ! info[a]) {
2487 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
2493 // LYXERR0("No regex formats");
2499 // Remove trailing closure of math, macros and environments, so to catch parts of them.
2500 static int identifyClosing(string & t)
2502 int open_braces = 0;
2504 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
2505 if (regex_replace(t, t, "(.*[^\\\\])\\$" REGEX_EOS, "$1"))
2507 if (regex_replace(t, t, "(.*[^\\\\]) \\\\\\]" REGEX_EOS, "$1"))
2509 if (regex_replace(t, t, "(.*[^\\\\]) \\\\end\\{[a-zA-Z_]*\\*?\\}" REGEX_EOS, "$1"))
2511 if (regex_replace(t, t, "(.*[^\\\\])\\}" REGEX_EOS, "$1")) {
2521 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt)
2522 : p_buf(&buf), p_first_buf(&buf), opt(opt)
2524 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
2525 docstring const & ds = stringifySearchBuffer(find_buf, opt);
2526 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
2527 // When using regexp, braces are hacked already by escape_for_regex()
2528 par_as_string = normalize(ds, !use_regexp);
2530 close_wildcards = 0;
2532 size_t lead_size = 0;
2533 // correct the language settings
2534 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat);
2535 if (opt.ignoreformat) {
2537 // if par_as_string_nolead were emty,
2538 // the following call to findAux will always *find* the string
2539 // in the checked data, and thus always using the slow
2540 // examining of the current text part.
2541 par_as_string_nolead = par_as_string;
2544 lead_size = identifyLeading(par_as_string);
2545 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
2546 lead_as_string = par_as_string.substr(0, lead_size);
2547 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
2551 open_braces = identifyClosing(par_as_string);
2552 identifyClosing(par_as_string_nolead);
2553 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2554 LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'");
2556 string lead_as_regexp;
2557 if (lead_size > 0) {
2558 // @todo No need to search for \regexp{} insets in leading material
2559 lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat);
2560 par_as_string = par_as_string_nolead;
2561 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
2562 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2564 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
2565 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
2566 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2568 // Insert .* before trailing '\$' ('$' has been escaped by escape_for_regex)
2569 regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\$)\\'", "$1(.*?)$2")
2570 // Insert .* before trailing '\\\]' ('\]' has been escaped by escape_for_regex)
2571 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])( \\\\\\\\\\\\\\])\\'", "$1(.*?)$2")
2572 // Insert .* before trailing '\\end\{...}' ('\end{...}' has been escaped by escape_for_regex)
2573 || regex_replace(par_as_string, par_as_string,
2574 "(.*[^\\\\])( \\\\\\\\end\\\\\\{[a-zA-Z_]*)(\\\\\\*)?(\\\\\\})\\'", "$1(.*?)$2$3$4")
2575 // Insert .* before trailing '\}' ('}' has been escaped by escape_for_regex)
2576 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\})\\'", "$1(.*?)$2")
2580 if (!opt.ignoreformat) {
2581 // Remove extra '\}' at end if not part of \{\.\}
2582 size_t lng = par_as_string.size();
2584 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
2586 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
2595 if (lng < par_as_string.size())
2596 par_as_string = par_as_string.substr(0,lng);
2599 regex_replace(par_as_string, par_as_string, "\\\\\\.", "_xxbdotxx_");
2600 // handle '.' -> '[^]', replace later as '[^\}\{\\]'
2601 regex_replace(par_as_string, par_as_string, "\\.", "[^]");
2602 // replace '[^...]' with '[^...\}\{\\]'
2603 regex_replace(par_as_string, par_as_string, "\\[\\^([^\\\\\\]]*)\\]", "_xxbrlxx_$1\\}\\{\\\\_xxbrrxx_");
2604 regex_replace(par_as_string, par_as_string, "_xxbrlxx_", "[^");
2605 regex_replace(par_as_string, par_as_string, "_xxbrrxx_", "]");
2607 regex_replace(par_as_string, par_as_string, "_xxbdotxx_", "\\.");
2610 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2611 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2612 LYXERR(Debug::FIND, "Close .*? : " << close_wildcards);
2613 LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
2615 // If entered regexp must match at begin of searched string buffer
2616 // Kornel: Added parentheses to use $1 for size of the leading string
2620 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
2621 // Unfortunately is '\1', '\2', etc not working for strings with extra format
2622 // so the convert has no effect in that case
2623 for (int i = 8; i > 0; --i) {
2624 string orig = "\\\\" + std::to_string(i);
2625 string dest = "\\" + std::to_string(i+1);
2626 while (regex_replace(par_as_string, par_as_string, orig, dest));
2628 regexp_str = "(" + lead_as_regexp + ")" + par_as_string;
2629 regexp2_str = "(" + lead_as_regexp + ").*?" + par_as_string;
2631 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
2632 regexp = lyx::regex(regexp_str);
2634 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
2635 regexp2 = lyx::regex(regexp2_str);
2640 // Count number of characters in string
2644 // \\[a-zA-Z]+ ==> 1
2645 static int computeSize(string s, int len)
2651 for (int i = 0; i < len; i += skip, count++) {
2654 if (isalpha(s[i+1])) {
2655 for (int j = 2; i+j < len; j++) {
2656 if (! isalpha(s[i+j])) {
2659 else if ((s[i+j] == '{') && s[i+j+1] == '}')
2661 else if ((s[i+j] == '{') && (i + j + 1 >= len))
2669 else if (s[i] == '{') {
2675 else if (s[i] == '-') {
2676 if (s[i+1] == '-') {
2692 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
2697 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
2700 docstring docstr = stringifyFromForSearch(opt, cur, len);
2701 string str = normalize(docstr, true);
2702 if (!opt.ignoreformat) {
2703 str = correctlanguagesetting(str, false, !opt.ignoreformat);
2706 mres.match_len = -1;
2709 LYXERR(Debug::FIND, "Matching against '" << lyx::to_utf8(docstr) << "'");
2710 LYXERR(Debug::FIND, "After normalization: '" << str << "'");
2713 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
2714 regex const *p_regexp;
2715 regex_constants::match_flag_type flags;
2717 flags = regex_constants::match_continuous;
2720 flags = regex_constants::match_default;
2721 p_regexp = ®exp2;
2723 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
2724 if (re_it == sregex_iterator())
2726 match_results<string::const_iterator> const & m = *re_it;
2728 if (0) { // Kornel Benko: DO NOT CHECKK
2729 // Check braces on the segment that matched the entire regexp expression,
2730 // plus the last subexpression, if a (.*?) was inserted in the constructor.
2731 if (!braces_match(m[0].first, m[0].second, open_braces))
2735 // Check braces on segments that matched all (.*?) subexpressions,
2736 // except the last "padding" one inserted by lyx.
2737 for (size_t i = 1; i < m.size() - 1; ++i)
2738 if (!braces_match(m[i].first, m[i].second, open_braces))
2741 // Exclude from the returned match length any length
2742 // due to close wildcards added at end of regexp
2743 // and also the length of the leading (e.g. '\emph{')
2745 // Whole found string, including the leading: m[0].second - m[0].first
2746 // Size of the leading string: m[1].second - m[1].first
2747 int leadingsize = 0;
2749 leadingsize = m[1].second - m[1].first;
2751 for (size_t i = 0; i < m.size(); i++) {
2752 LYXERR(Debug::FIND, "Match " << i << " is " << m[i].second - m[i].first << " long");
2754 if (close_wildcards == 0)
2755 result = m[0].second - m[0].first;
2758 result = m[m.size() - close_wildcards].first - m[0].first;
2760 size_t pos = m.position(size_t(0));
2761 // Ignore last closing characters
2762 while (result > 0) {
2763 if (str[pos+result-1] == '}')
2768 if (result > leadingsize)
2769 result -= leadingsize;
2772 mres.match_len = computeSize(str.substr(pos+leadingsize,result), result);
2773 mres.match2end = str.size() - pos - leadingsize;
2774 mres.pos = pos+leadingsize;
2778 // else !use_regexp: but all code paths above return
2779 LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='"
2780 << par_as_string << "', str='" << str << "'");
2781 LYXERR(Debug::FIND, "Searching in normal mode: lead_as_string='"
2782 << lead_as_string << "', par_as_string_nolead='"
2783 << par_as_string_nolead << "'");
2786 LYXERR(Debug::FIND, "size=" << par_as_string.size()
2787 << ", substr='" << str.substr(0, par_as_string.size()) << "'");
2788 if (str.substr(0, par_as_string.size()) == par_as_string) {
2789 mres.match_len = par_as_string.size();
2790 mres.match2end = str.size();
2795 size_t pos = str.find(par_as_string_nolead);
2796 if (pos != string::npos) {
2797 mres.match_len = par_as_string.size();
2798 mres.match2end = str.size() - pos;
2807 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
2809 MatchResult mres = findAux(cur, len, at_begin);
2810 int res = mres.match_len;
2812 "res=" << res << ", at_begin=" << at_begin
2813 << ", matchword=" << opt.matchword
2814 << ", inTexted=" << cur.inTexted());
2815 if (res == 0 || !at_begin || !opt.matchword || !cur.inTexted())
2817 if ((len > 0) && (res < len)) {
2821 Paragraph const & par = cur.paragraph();
2822 bool ws_left = (cur.pos() > 0)
2823 ? par.isWordSeparator(cur.pos() - 1)
2825 bool ws_right = (cur.pos() + len < par.size())
2826 ? par.isWordSeparator(cur.pos() + len)
2829 "cur.pos()=" << cur.pos() << ", res=" << res
2830 << ", separ: " << ws_left << ", " << ws_right
2833 if (ws_left && ws_right) {
2834 // Check for word separators inside the found 'word'
2835 for (int i = 0; i < len; i++) {
2836 if (par.isWordSeparator(cur.pos() + i)) {
2848 string MatchStringAdv::normalize(docstring const & s, bool hack_braces) const
2851 if (! opt.casesensitive)
2852 t = lyx::to_utf8(lowercase(s));
2854 t = lyx::to_utf8(s);
2855 // Remove \n at begin
2856 while (!t.empty() && t[0] == '\n')
2859 while (!t.empty() && t[t.size() - 1] == '\n')
2860 t = t.substr(0, t.size() - 1);
2862 // Handle all other '\n'
2863 while ((pos = t.find("\n")) != string::npos) {
2864 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
2866 if (std::isalnum(t[pos+1])) {
2867 t.replace(pos-2, 3, " ");
2870 t.replace(pos-2, 3, "");
2873 else if (!std::isalnum(t[pos+1]) || !std::isalnum(t[pos-1])) {
2874 // '\n' adjacent to non-alpha-numerics, discard
2875 t.replace(pos, 1, "");
2878 // Replace all other \n with spaces
2879 t.replace(pos, 1, " ");
2882 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
2883 // Kornel: Added textsl, textsf, textit, texttt and noun
2884 // + allow to seach for colored text too
2885 LYXERR(Debug::FIND, "Removing stale empty \\emph{}, \\textbf{}, \\*section{} macros from: " << t);
2886 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
2887 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
2888 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
2889 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
2891 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
2892 // FIXME - check what preceeds the brace
2894 if (opt.ignoreformat)
2895 while (regex_replace(t, t, "\\{", "_x_<")
2896 || regex_replace(t, t, "\\}", "_x_>"))
2897 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
2899 while (regex_replace(t, t, "\\\\\\{", "_x_<")
2900 || regex_replace(t, t, "\\\\\\}", "_x_>"))
2901 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
2908 docstring stringifyFromCursor(DocIterator const & cur, int len)
2910 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
2911 if (cur.inTexted()) {
2912 Paragraph const & par = cur.paragraph();
2913 // TODO what about searching beyond/across paragraph breaks ?
2914 // TODO Try adding a AS_STR_INSERTS as last arg
2915 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
2916 int(par.size()) : cur.pos() + len;
2917 OutputParams runparams(&cur.buffer()->params().encoding());
2918 runparams.nice = true;
2919 runparams.flavor = OutputParams::LATEX;
2920 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
2921 // No side effect of file copying and image conversion
2922 runparams.dryrun = true;
2923 LYXERR(Debug::FIND, "Stringifying with cur: "
2924 << cur << ", from pos: " << cur.pos() << ", end: " << end);
2925 docstring result = par.asString(cur.pos(), end,
2926 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
2929 } else if (cur.inMathed()) {
2930 CursorSlice cs = cur.top();
2931 MathData md = cs.cell();
2932 MathData::const_iterator it_end =
2933 (( len == -1 || cs.pos() + len > int(md.size()))
2935 : md.begin() + cs.pos() + len );
2937 for (MathData::const_iterator it = md.begin() + cs.pos();
2940 docstring s = asString(md2);
2941 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
2944 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
2949 /** Computes the LaTeX export of buf starting from cur and ending len positions
2950 * after cur, if len is positive, or at the paragraph or innermost inset end
2953 docstring latexifyFromCursor(DocIterator const & cur, int len)
2955 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
2956 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
2957 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
2958 Buffer const & buf = *cur.buffer();
2960 odocstringstream ods;
2962 OutputParams runparams(&buf.params().encoding());
2963 runparams.nice = false;
2964 runparams.flavor = OutputParams::LATEX;
2965 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
2966 // No side effect of file copying and image conversion
2967 runparams.dryrun = true;
2968 runparams.for_search = true;
2970 if (cur.inTexted()) {
2971 // @TODO what about searching beyond/across paragraph breaks ?
2972 pos_type endpos = cur.paragraph().size();
2973 if (len != -1 && endpos > cur.pos() + len)
2974 endpos = cur.pos() + len;
2975 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
2976 string(), cur.pos(), endpos);
2977 string s = lyx::to_utf8(ods.str());
2978 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
2979 return(lyx::from_utf8(s));
2980 } else if (cur.inMathed()) {
2981 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
2982 for (int s = cur.depth() - 1; s >= 0; --s) {
2983 CursorSlice const & cs = cur[s];
2984 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
2986 cs.asInsetMath()->asHullInset()->header_write(ws);
2991 CursorSlice const & cs = cur.top();
2992 MathData md = cs.cell();
2993 MathData::const_iterator it_end =
2994 ((len == -1 || cs.pos() + len > int(md.size()))
2996 : md.begin() + cs.pos() + len);
2998 for (MathData::const_iterator it = md.begin() + cs.pos();
3002 ods << asString(md2);
3003 // Retrieve the math environment type, and add '$' or '$]'
3004 // or others (\end{equation}) accordingly
3005 for (int s = cur.depth() - 1; s >= 0; --s) {
3006 CursorSlice const & cs2 = cur[s];
3007 InsetMath * inset = cs2.asInsetMath();
3008 if (inset && inset->asHullInset()) {
3010 inset->asHullInset()->footer_write(ws);
3014 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3016 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3022 /** Finalize an advanced find operation, advancing the cursor to the innermost
3023 ** position that matches, plus computing the length of the matching text to
3026 int findAdvFinalize(DocIterator & cur, MatchStringAdv const & match)
3028 // Search the foremost position that matches (avoids find of entire math
3029 // inset when match at start of it)
3031 DocIterator old_cur(cur.buffer());
3033 LYXERR(Debug::FIND, "Forwarding one step (searching for innermost match)");
3037 } while (cur && cur.depth() > d && match(cur).match_len > 0);
3039 int max_match = match(cur).match_len; /* match valid only if not searching whole words */
3040 if (max_match <= 0) return 0;
3041 LYXERR(Debug::FIND, "Ok");
3043 // Compute the match length
3045 if (cur.pos() + len > cur.lastpos())
3047 if (match.opt.matchword) {
3048 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3049 while (cur.pos() + len <= cur.lastpos() && match(cur, len).match_len <= 0) {
3051 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3053 // Length of matched text (different from len param)
3054 int old_match = match(cur, len).match_len;
3058 // Greedy behaviour while matching regexps
3059 while ((new_match = match(cur, len + 1).match_len) > old_match) {
3061 old_match = new_match;
3062 LYXERR(Debug::FIND, "verifying match with len = " << len);
3069 int maxl = cur.lastpos() - cur.pos();
3070 // Greedy behaviour while matching regexps
3071 while (maxl > minl) {
3072 int actual_match = match(cur, len).match_len;
3073 if (actual_match >= max_match) {
3074 // actual_match > max_match _can_ happen,
3075 // if the search area splits
3076 // some following word so that the regex
3077 // (e.g. 'r.*r\b' matches 'r' from the middle of the
3079 // This means, the len value is too big
3081 len = (int)((maxl + minl)/2);
3084 // (actual_match < max_match)
3086 len = (int)((maxl + minl)/2);
3090 // Search for real start of matched characters
3095 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
3096 if (cur.depth() < old_cur.depth()) {
3098 LYXERR0("cur.depth() < old_cur.depth(), this should never happen");
3101 if (cur.pos() != old_cur.pos()) {
3102 // OK, forwarded 1 pos in actual inset
3103 actual_match = match(cur, len-1).match_len;
3104 if (actual_match == max_match) {
3105 // Ha, got it! The shorter selection has the same match length
3110 // OK, the shorter selection matches less chars, revert to previous value
3116 LYXERR0("cur.pos() == old_cur.pos(), this should never happen");
3117 actual_match = match(cur, len).match_len;
3118 if (actual_match == max_match)
3128 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
3132 while (!theApp()->longOperationCancelled() && cur) {
3133 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
3134 MatchResult mres = match(cur, -1, false);
3135 int match_len = mres.match_len;
3136 LYXERR(Debug::FIND, "match_len: " << match_len);
3137 if ((mres.pos > 10000) || (mres.match2end > 10000) || (match_len > 10000)) {
3138 LYXERR0("BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
3141 if (match_len > 0) {
3142 // Try to find the begin of searched string
3143 int increment = mres.pos/2;
3144 while (mres.pos > 5 && (increment > 5)) {
3145 DocIterator old_cur = cur;
3146 for (int i = 0; i < increment && cur; cur.forwardPos(), i++) {
3148 while (cur && cur.depth() != old_cur.depth())
3157 MatchResult mres2 = match(cur, -1, false);
3158 if ((mres2.match2end < mres.match2end) ||
3159 (mres2.match_len < mres.match_len)) {
3166 if (increment > mres.pos/2)
3167 increment = mres.pos/2;
3171 int match_len_zero_count = 0;
3172 for (int i = 0; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) {
3174 int remaining_len = match(cur, -1, false).match_len;
3175 if (remaining_len <= 0) {
3176 // Apparently the searched string is not in the remaining part
3183 LYXERR(Debug::FIND, "Advancing cur: " << cur);
3184 int match_len3 = match(cur, 1).match_len;
3187 int match_len2 = match(cur).match_len;
3188 LYXERR(Debug::FIND, "match_len2: " << match_len2);
3189 if (match_len2 > 0) {
3190 // Sometimes in finalize we understand it wasn't a match
3191 // and we need to continue the outest loop
3192 int len = findAdvFinalize(cur, match);
3197 if (match_len2 >= 0) {
3198 if (match_len2 == 0)
3199 match_len_zero_count++;
3201 match_len_zero_count = 0;
3204 if (++match_len_zero_count > 3) {
3205 LYXERR(Debug::FIND, "match_len2_zero_count: " << match_len_zero_count << ", match_len was " << match_len);
3206 match_len_zero_count = 0;
3214 if (match_len >= 0 && cur.pit() < cur.lastpit()) {
3215 LYXERR(Debug::FIND, "Advancing par: cur=" << cur);
3218 // This should exit nested insets, if any, or otherwise undefine the currsor.
3219 cur.pos() = cur.lastpos();
3220 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
3228 /// Find the most backward consecutive match within same paragraph while searching backwards.
3229 int findMostBackwards(DocIterator & cur, MatchStringAdv const & match)
3231 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3232 DocIterator tmp_cur = cur;
3233 int len = findAdvFinalize(tmp_cur, match);
3234 Inset & inset = cur.inset();
3235 for (; cur != cur_begin; cur.backwardPos()) {
3236 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
3237 DocIterator new_cur = cur;
3238 new_cur.backwardPos();
3239 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
3241 int new_len = findAdvFinalize(new_cur, match);
3246 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
3252 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
3256 // Backup of original position
3257 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3258 if (cur == cur_begin)
3261 DocIterator cur_orig(cur);
3262 bool pit_changed = false;
3265 bool found_match = (match(cur, -1, false).match_len > 0);
3269 cur.pos() = cur.lastpos();
3271 cur.pos() = cur_orig.pos();
3272 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
3273 DocIterator cur_prev_iter;
3275 found_match = (match(cur).match_len > 0);
3276 LYXERR(Debug::FIND, "findBackAdv3: found_match="
3277 << found_match << ", cur: " << cur);
3279 return findMostBackwards(cur, match);
3281 // Stop if begin of document reached
3282 if (cur == cur_begin)
3284 cur_prev_iter = cur;
3288 if (cur == cur_begin)
3295 } while (!theApp()->longOperationCancelled());
3303 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
3304 DocIterator const & cur, int len)
3306 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
3308 if (!opt.ignoreformat)
3309 return latexifyFromCursor(cur, len);
3311 return stringifyFromCursor(cur, len);
3315 FindAndReplaceOptions::FindAndReplaceOptions(
3316 docstring const & find_buf_name, bool casesensitive,
3317 bool matchword, bool forward, bool expandmacros, bool ignoreformat,
3318 docstring const & repl_buf_name, bool keep_case,
3319 SearchScope scope, SearchRestriction restr)
3320 : find_buf_name(find_buf_name), casesensitive(casesensitive), matchword(matchword),
3321 forward(forward), expandmacros(expandmacros), ignoreformat(ignoreformat),
3322 repl_buf_name(repl_buf_name), keep_case(keep_case), scope(scope), restr(restr)
3330 /** Check if 'len' letters following cursor are all non-lowercase */
3331 static bool allNonLowercase(Cursor const & cur, int len)
3333 pos_type beg_pos = cur.selectionBegin().pos();
3334 pos_type end_pos = cur.selectionBegin().pos() + len;
3335 if (len > cur.lastpos() + 1 - beg_pos) {
3336 LYXERR(Debug::FIND, "This should not happen, more debug needed");
3337 len = cur.lastpos() + 1 - beg_pos;
3338 end_pos = beg_pos + len;
3340 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
3341 if (isLowerCase(cur.paragraph().getChar(pos)))
3347 /** Check if first letter is upper case and second one is lower case */
3348 static bool firstUppercase(Cursor const & cur)
3351 pos_type pos = cur.selectionBegin().pos();
3352 if (pos >= cur.lastpos() - 1) {
3353 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
3356 ch1 = cur.paragraph().getChar(pos);
3357 ch2 = cur.paragraph().getChar(pos + 1);
3358 bool result = isUpperCase(ch1) && isLowerCase(ch2);
3359 LYXERR(Debug::FIND, "firstUppercase(): "
3360 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
3361 << ch2 << "(" << char(ch2) << ")"
3362 << ", result=" << result << ", cur=" << cur);
3367 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
3369 ** \fixme What to do with possible further paragraphs in replace buffer ?
3371 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
3373 ParagraphList::iterator pit = buffer.paragraphs().begin();
3374 LASSERT(pit->size() >= 1, /**/);
3375 pos_type right = pos_type(1);
3376 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
3377 right = pit->size();
3378 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
3384 static void findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
3386 Cursor & cur = bv->cursor();
3387 if (opt.repl_buf_name == docstring()
3388 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
3389 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3392 DocIterator sel_beg = cur.selectionBegin();
3393 DocIterator sel_end = cur.selectionEnd();
3394 if (&sel_beg.inset() != &sel_end.inset()
3395 || sel_beg.pit() != sel_end.pit()
3396 || sel_beg.idx() != sel_end.idx())
3398 int sel_len = sel_end.pos() - sel_beg.pos();
3399 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
3400 << ", sel_len: " << sel_len << endl);
3403 LASSERT(sel_len > 0, return);
3405 if (!matchAdv(sel_beg, sel_len).match_len)
3408 // Build a copy of the replace buffer, adapted to the KeepCase option
3409 Buffer & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
3411 repl_buffer_orig.write(oss);
3412 string lyx = oss.str();
3413 Buffer repl_buffer("", false);
3414 repl_buffer.setUnnamed(true);
3415 LASSERT(repl_buffer.readString(lyx), return);
3416 if (opt.keep_case && sel_len >= 2) {
3417 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
3418 if (cur.inTexted()) {
3419 if (firstUppercase(cur))
3420 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
3421 else if (allNonLowercase(cur, sel_len))
3422 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
3425 cap::cutSelection(cur, false);
3426 if (cur.inTexted()) {
3427 repl_buffer.changeLanguage(
3428 repl_buffer.language(),
3429 cur.getFont().language());
3430 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
3431 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
3432 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
3433 repl_buffer.params().documentClassPtr(),
3434 bv->buffer().errorList("Paste"));
3435 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
3436 sel_len = repl_buffer.paragraphs().begin()->size();
3437 } else if (cur.inMathed()) {
3438 odocstringstream ods;
3440 OutputParams runparams(&repl_buffer.params().encoding());
3441 runparams.nice = false;
3442 runparams.flavor = OutputParams::LATEX;
3443 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3444 runparams.dryrun = true;
3445 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
3446 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
3447 docstring repl_latex = ods.str();
3448 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
3450 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
3451 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
3452 repl_latex = from_utf8(s);
3453 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
3454 MathData ar(cur.buffer());
3455 asArray(repl_latex, ar, Parse::NORMAL);
3457 sel_len = ar.size();
3458 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3460 if (cur.pos() >= sel_len)
3461 cur.pos() -= sel_len;
3464 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3465 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
3466 bv->processUpdateFlags(Update::Force);
3470 /// Perform a FindAdv operation.
3471 bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt)
3476 // e.g., when invoking word-findadv from mini-buffer wither with
3477 // wrong options syntax or before ever opening advanced F&R pane
3478 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3482 MatchStringAdv matchAdv(bv->buffer(), opt);
3483 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
3485 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
3486 findAdvReplace(bv, opt, matchAdv);
3489 match_len = findForwardAdv(cur, matchAdv);
3491 match_len = findBackwardsAdv(cur, matchAdv);
3493 // This may only be raised by lyx::regex()
3494 bv->message(_("Invalid regular expression!"));
3498 if (match_len == 0) {
3499 bv->message(_("Match not found!"));
3503 bv->message(_("Match found!"));
3505 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << match_len);
3506 bv->putSelectionAt(cur, match_len, !opt.forward);
3512 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
3514 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
3515 << opt.casesensitive << ' '
3516 << opt.matchword << ' '
3517 << opt.forward << ' '
3518 << opt.expandmacros << ' '
3519 << opt.ignoreformat << ' '
3520 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
3521 << opt.keep_case << ' '
3522 << int(opt.scope) << ' '
3525 LYXERR(Debug::FIND, "built: " << os.str());
3531 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
3533 LYXERR(Debug::FIND, "parsing");
3537 while (line != "EOSS") {
3541 if (is.eof()) // Tolerate malformed request
3545 LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
3546 opt.find_buf_name = from_utf8(s);
3547 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat;
3548 is.get(); // Waste space before replace string
3551 while (line != "EOSS") {
3555 if (is.eof()) // Tolerate malformed request
3559 LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
3560 opt.repl_buf_name = from_utf8(s);
3561 is >> opt.keep_case;
3564 opt.scope = FindAndReplaceOptions::SearchScope(i);
3566 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
3568 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
3569 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
3570 << opt.scope << ' ' << opt.restr);