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)
742 LYXERR(Debug::FIND, "new_pos: " << new_pos);
743 string t = apply_escapes(s.substr(pos, new_pos - pos), get_lyx_unescapes());
744 LYXERR(Debug::FIND, "t [lyx]: " << t);
745 t = apply_escapes(t, get_regexp_escapes());
746 LYXERR(Debug::FIND, "t [rxp]: " << t);
747 s.replace(pos, new_pos - pos, t);
748 new_pos = pos + t.size();
749 LYXERR(Debug::FIND, "Regexp after escaping: " << s);
750 LYXERR(Debug::FIND, "new_pos: " << new_pos);
751 if (new_pos == s.size())
753 // Might fail if \\endregexp{} is preceeded by unexpected stuff (weird escapes)
754 size_t end_pos = s.find("\\endregexp{}}", new_pos + 8);
755 LYXERR(Debug::FIND, "end_pos: " << end_pos);
756 t = s.substr(new_pos + 8, end_pos - (new_pos + 8));
757 LYXERR(Debug::FIND, "t in regexp : " << t);
758 t = apply_escapes(t, get_lyx_unescapes());
759 LYXERR(Debug::FIND, "t in regexp [lyx]: " << t);
761 t = apply_escapes(t, get_regexp_latex_escapes());
762 LYXERR(Debug::FIND, "t in regexp [ltx]: " << t);
764 if (end_pos == s.size()) {
765 s.replace(new_pos, end_pos - new_pos, t);
766 LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s);
769 s.replace(new_pos, end_pos + 13 - new_pos, t);
770 LYXERR(Debug::FIND, "Regexp after \\regexp{...\\endregexp{}} removal: " << s);
771 pos = new_pos + t.size();
772 LYXERR(Debug::FIND, "pos: " << pos);
778 /// Wrapper for lyx::regex_replace with simpler interface
779 bool regex_replace(string const & s, string & t, string const & searchstr,
780 string const & replacestr)
782 lyx::regex e(searchstr, regex_constants::ECMAScript);
784 ostream_iterator<char, char> it(oss);
785 lyx::regex_replace(it, s.begin(), s.end(), e, replacestr);
786 // tolerate t and s be references to the same variable
787 bool rv = (s != oss.str());
793 /** Checks if supplied string segment is well-formed from the standpoint of matching open-closed braces.
795 ** Verify that closed braces exactly match open braces. This avoids that, for example,
796 ** \frac{.*}{x} matches \frac{x+\frac{y}{x}}{z} with .* being 'x+\frac{y'.
799 ** Number of open braces that must remain open at the end for the verification to succeed.
801 bool braces_match(string::const_iterator const & beg,
802 string::const_iterator const & end,
806 string::const_iterator it = beg;
807 LYXERR(Debug::FIND, "Checking " << unmatched << " unmatched braces in '" << string(beg, end) << "'");
808 for (; it != end; ++it) {
809 // Skip escaped braces in the count
814 } else if (*it == '{') {
816 } else if (*it == '}') {
817 if (open_pars == 0) {
818 LYXERR(Debug::FIND, "Found unmatched closed brace");
824 if (open_pars != unmatched) {
825 LYXERR(Debug::FIND, "Found " << open_pars
826 << " instead of " << unmatched
827 << " unmatched open braces at the end of count");
830 LYXERR(Debug::FIND, "Braces match as expected");
835 /** The class performing a match between a position in the document and the FindAdvOptions.
837 class MatchStringAdv {
839 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt);
841 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
842 ** constructor as opt.search, under the opt.* options settings.
845 ** If set, then match is searched only against beginning of text starting at cur.
846 ** If unset, then match is searched anywhere in text starting at cur.
849 ** The length of the matching text, or zero if no match was found.
851 int operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
856 /// first buffer on which search was started
857 lyx::Buffer * const p_first_buf;
859 FindAndReplaceOptions const & opt;
862 /// Auxiliary find method (does not account for opt.matchword)
863 int findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
865 /** Normalize a stringified or latexified LyX paragraph.
869 ** <li>if search is not casesensitive, then lowercase the string;
870 ** <li>remove any newline at begin or end of the string;
871 ** <li>replace any newline in the middle of the string with a simple space;
872 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
875 ** @todo Normalization should also expand macros, if the corresponding
876 ** search option was checked.
878 string normalize(docstring const & s, bool hack_braces) const;
879 // normalized string to search
880 string par_as_string;
881 // regular expression to use for searching
883 // same as regexp, but prefixed with a ".*?"
885 // leading format material as string
886 string lead_as_string;
887 // par_as_string after removal of lead_as_string
888 string par_as_string_nolead;
889 // unmatched open braces in the search string/regexp
891 // number of (.*?) subexpressions added at end of search regexp for closing
892 // environments, math mode, styles, etc...
894 // Are we searching with regular expressions ?
899 static docstring buffer_to_latex(Buffer & buffer)
901 OutputParams runparams(&buffer.params().encoding());
902 odocstringstream ods;
904 runparams.nice = true;
905 runparams.flavor = OutputParams::LATEX;
906 runparams.linelen = 100000; //lyxrc.plaintext_linelen;
907 // No side effect of file copying and image conversion
908 runparams.dryrun = true;
909 runparams.for_search = true;
910 pit_type const endpit = buffer.paragraphs().size();
911 for (pit_type pit = 0; pit != endpit; ++pit) {
912 TeXOnePar(buffer, buffer.text(), pit, os, runparams);
913 LYXERR(Debug::FIND, "searchString up to here: " << ods.str());
919 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
922 if (!opt.ignoreformat) {
923 str = buffer_to_latex(buffer);
925 OutputParams runparams(&buffer.params().encoding());
926 runparams.nice = true;
927 runparams.flavor = OutputParams::LATEX;
928 runparams.linelen = 100000; //lyxrc.plaintext_linelen;
929 runparams.dryrun = true;
930 runparams.for_search = true;
931 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
932 Paragraph const & par = buffer.paragraphs().at(pit);
933 LYXERR(Debug::FIND, "Adding to search string: '"
934 << par.asString(pos_type(0), par.size(),
935 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
938 str += par.asString(pos_type(0), par.size(),
939 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
947 /// Return separation pos between the leading material and the rest
948 static size_t identifyLeading(string const & s)
951 // @TODO Support \item[text]
952 // Kornel: Added textsl, textsf, textit, texttt and noun
953 // + allow to search for colored text too
954 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)\\*?)\\{", "")
955 || regex_replace(t, t, REGEX_BOS "\\$", "")
956 || regex_replace(t, t, REGEX_BOS "\\\\\\[ ", "")
957 || regex_replace(t, t, REGEX_BOS " ?\\\\item\\{[a-z]+\\}", "")
958 || regex_replace(t, t, REGEX_BOS "\\\\begin\\{[a-zA-Z_]*\\*?\\} ", ""))
960 LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
965 * Given a latexified string, retrieve some handled features
966 * The features of the regex will later be compared with the features
967 * of the searched text. If the regex features are not a
968 * subset of the analized, then, in not format ignoring search
969 * we can early stop the search in the relevant inset.
971 typedef map<string, bool> Features;
973 static Features identifyFeatures(string const & s)
975 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
976 static regex const valid("^("
978 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
979 "emph|noun|text(bf|md|sl|sf|it|tt)|"
980 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
981 "(u|uu)line|(s|x)out|uwave|"
982 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
983 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
988 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
991 if (sub.str(1).compare("regexp") == 0) {
995 string token = sub.str(1);
997 if (regex_match(token, sub2, valid)) {
1005 if (sub.str(1).compare("endregexp") == 0) {
1015 * defines values features of a key "\\[a-z]+{"
1020 /* Char type with content discarded
1021 * like \hspace{1cm} */
1023 /* Char, like \backslash */
1025 /* \part, \section*, ... */
1027 /* title, author etc */
1029 /* \foreignlanguage{ngerman}, ... */
1032 * to discard language in content */
1035 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1037 /* fonts, colors, markups, ... */
1039 /* footnotesize, ... large, ...
1040 * Ignore all of them */
1043 /* inputencoding, ...
1044 * Discard also content, because they do not help in search */
1047 * like remove, but also all arguments */
1049 /* item, listitem */
1051 /* tex, latex, ... like isChar */
1053 /* like \lettrine[lines=5]{}{} */
1055 /* End of arguments marker for lettrine,
1056 * so that they can be ignored */
1066 parenthesiscount(1),
1070 KeyInfo(KeyType type, int parcount, bool disable)
1076 parenthesiscount(parcount),
1085 int parenthesiscount;
1087 bool used; /* by pattern */
1092 Border(int l=0, int u=0) : low(l), upper(u) {};
1097 #define MAXOPENED 30
1099 bool isPatternString;
1101 explicit Intervall(bool isPattern) :
1102 isPatternString(isPattern),
1104 actualdeptindex(0) { depts[0] = 0; closes[0] = 0;};
1107 int depts[MAXOPENED];
1108 int closes[MAXOPENED];
1109 int actualdeptindex;
1110 Border borders[2*MAXOPENED];
1111 int previousNotIgnored(int);
1112 int nextNotIgnored(int);
1113 void handleOpenP(int i);
1114 void handleCloseP(int i, bool closingAllowed);
1115 void resetOpenedP(int openPos);
1116 void addIntervall(int upper);
1117 void addIntervall(int low, int upper); /* if explicit */
1118 void setForDefaultLang(KeyInfo &defLang);
1119 int findclosing(int start, int end, char up, char down, int repeat);
1120 void handleParentheses(int lastpos, bool closingAllowed);
1123 void output(ostringstream &os, int lastpos);
1124 // string show(int lastpos);
1127 void Intervall::setForDefaultLang(KeyInfo &defLang)
1129 // Enable the use of first token again
1130 if (ignoreidx >= 0) {
1131 int value = defLang._tokenstart + defLang._tokensize;
1133 if (borders[0].low < value)
1134 borders[0].low = value;
1135 if (borders[0].upper < value)
1136 borders[0].upper = value;
1141 static void checkDepthIndex(int val)
1143 static int maxdepthidx = MAXOPENED-2;
1144 if (val > maxdepthidx) {
1146 LYXERR0("maxdepthidx now " << val);
1150 static void checkIgnoreIdx(int val)
1152 static int maxignoreidx = 2*MAXOPENED - 4;
1153 if (val > maxignoreidx) {
1155 LYXERR0("maxignoreidx now " << val);
1160 * Expand the region of ignored parts of the input latex string
1161 * The region is only relevant in output()
1163 void Intervall::addIntervall(int low, int upper)
1166 if (low == upper) return;
1167 for (idx = ignoreidx+1; idx > 0; --idx) {
1168 if (low > borders[idx-1].upper) {
1172 Border br(low, upper);
1173 if (idx > ignoreidx) {
1176 checkIgnoreIdx(ignoreidx);
1180 // Expand only if one of the new bound is inside the interwall
1181 // We know here that br.low > borders[idx-1].upper
1182 if (br.upper < borders[idx].low) {
1183 // We have to insert at this pos
1184 for (int i = ignoreidx+1; i > idx; --i) {
1185 borders[i] = borders[i-1];
1189 checkIgnoreIdx(ignoreidx);
1192 // Here we know, that we are overlapping
1193 if (br.low > borders[idx].low)
1194 br.low = borders[idx].low;
1195 // check what has to be concatenated
1197 for (int i = idx; i <= ignoreidx; i++) {
1198 if (br.upper >= borders[i].low) {
1200 if (br.upper < borders[i].upper)
1201 br.upper = borders[i].upper;
1207 // count should be >= 1 here
1210 for (int i = idx + count; i <= ignoreidx; i++) {
1211 borders[i-count+1] = borders[i];
1213 ignoreidx -= count - 1;
1219 void Intervall::handleOpenP(int i)
1222 depts[actualdeptindex] = i+1;
1223 closes[actualdeptindex] = -1;
1224 checkDepthIndex(actualdeptindex);
1227 void Intervall::handleCloseP(int i, bool closingAllowed)
1229 if (actualdeptindex <= 0) {
1230 if (! closingAllowed)
1231 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1232 // if we are at the very end
1233 addIntervall(i, i+1);
1236 closes[actualdeptindex] = i+1;
1241 void Intervall::resetOpenedP(int openPos)
1243 // Used as initializer for foreignlanguage entry
1244 actualdeptindex = 1;
1245 depts[1] = openPos+1;
1249 int Intervall::previousNotIgnored(int start)
1251 int idx = 0; /* int intervalls */
1252 for (idx = ignoreidx; idx >= 0; --idx) {
1253 if (start > borders[idx].upper)
1255 if (start >= borders[idx].low)
1256 start = borders[idx].low-1;
1261 int Intervall::nextNotIgnored(int start)
1263 int idx = 0; /* int intervalls */
1264 for (idx = 0; idx <= ignoreidx; idx++) {
1265 if (start < borders[idx].low)
1267 if (start < borders[idx].upper)
1268 start = borders[idx].upper;
1273 typedef map<string, KeyInfo> KeysMap;
1274 typedef vector< KeyInfo> Entries;
1275 static KeysMap keys = map<string, KeyInfo>();
1282 void buildKeys(bool);
1283 void buildEntries(bool);
1284 void makeKey(const string &, KeyInfo, bool isPatternString);
1285 void processRegion(int start, int region_end); /* remove {} parts */
1286 void removeHead(KeyInfo&, int count=0);
1289 LatexInfo(string par, bool isPatternString) : entidx(-1), interval(isPatternString) {
1291 interval.hasTitle = false;
1292 interval.titleValue = "";
1293 buildKeys(isPatternString);
1294 entries = vector<KeyInfo>();
1295 buildEntries(isPatternString);
1299 if (entries.empty()) {
1302 if (entries[0].keytype == KeyInfo::isTitle) {
1303 if (! entries[0].disabled) {
1304 interval.hasTitle = true;
1305 interval.titleValue = entries[0].head;
1308 interval.hasTitle = false;
1309 interval.titleValue = "";
1311 removeHead(entries[0]);
1312 if (entries.size() > 1)
1321 if (int(entries.size()) > entidx) {
1328 bool setNextKey(int idx) {
1329 if ((idx == entidx) && (entidx >= 0)) {
1336 int find(int start, KeyInfo::KeyType keytype) {
1340 while (tmpIdx < int(entries.size())) {
1341 if (entries[tmpIdx].keytype == keytype)
1347 int process(ostringstream &os, KeyInfo &actual);
1348 int dispatch(ostringstream &os, int previousStart, KeyInfo &actual);
1349 // string show(int lastpos) { return interval.show(lastpos);};
1350 int nextNotIgnored(int start) { return interval.nextNotIgnored(start);};
1351 KeyInfo &getKeyInfo(int keyinfo) {
1352 static KeyInfo invalidInfo = KeyInfo();
1353 if ((keyinfo < 0) || ( keyinfo >= int(entries.size())))
1356 return entries[keyinfo];
1358 void setForDefaultLang(KeyInfo &defLang) {interval.setForDefaultLang(defLang);};
1359 void addIntervall(int low, int up) { interval.addIntervall(low, up); };
1363 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
1368 for (int i = start; i < end; i += 1 + skip) {
1372 if (c == '\\') skip = 1;
1376 else if (c == down) {
1378 if ((repeat <= 0) || (par[i+1] != up))
1396 vector<MathEntry> entries;
1401 void insert(string wait, size_t start, size_t end) {
1402 MathEntry m = MathEntry();
1404 m.mathStart = start;
1406 m.mathSize = end - start;
1407 entries.push_back(m);
1409 bool empty() { return entries.empty(); };
1410 size_t getEndPos() {
1411 if (entries.empty() || (actualIdx >= entries.size())) {
1414 return entries[actualIdx].mathEnd;
1416 size_t getStartPos() {
1417 if (entries.empty() || (actualIdx >= entries.size())) {
1418 return 100000; /* definitely enough? */
1420 return entries[actualIdx].mathStart;
1422 size_t getFirstPos() {
1424 return getStartPos();
1427 if (entries.empty() || (actualIdx >= entries.size())) {
1430 return entries[actualIdx].mathSize;
1432 void incrEntry() { actualIdx++; };
1435 void LatexInfo::buildEntries(bool isPatternString)
1437 static regex const rmath("\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|alignat)\\*?)\\}");
1438 static regex const rkeys("\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?))");
1439 static bool disableLanguageOverride = false;
1440 smatch sub, submath;
1441 bool evaluatingRegexp = false;
1443 bool evaluatingMath = false;
1444 bool evaluatingCode = false;
1446 bool evaluatingOptional = false;
1447 size_t optionalEnd = 0;
1450 bool math_end_waiting = false;
1451 size_t math_pos = 10000;
1454 for (sregex_iterator itmath(interval.par.begin(), interval.par.end(), rmath), end; itmath != end; ++itmath) {
1456 if (math_end_waiting) {
1457 size_t pos = submath.position(size_t(0));
1458 if ((math_end == "$") &&
1459 (submath.str(0) == "$") &&
1460 (interval.par[pos-1] != '\\')) {
1461 mi.insert("$", math_pos, pos + 1);
1462 math_end_waiting = false;
1464 else if ((math_end == "\\]") &&
1465 (submath.str(0) == "\\]")) {
1466 mi.insert("\\]", math_pos, pos + 2);
1467 math_end_waiting = false;
1469 else if ((submath.str(1).compare("end") == 0) &&
1470 (submath.str(2).compare(math_end) == 0)) {
1471 mi.insert(math_end, math_pos, pos + submath.str(0).length());
1472 math_end_waiting = false;
1478 if (submath.str(1).compare("begin") == 0) {
1479 math_end_waiting = true;
1480 math_end = submath.str(2);
1481 math_pos = submath.position(size_t(0));
1483 else if (submath.str(0).compare("\\[") == 0) {
1484 math_end_waiting = true;
1486 math_pos = submath.position(size_t(0));
1488 else if (submath.str(0) == "$") {
1489 size_t pos = submath.position(size_t(0));
1490 if ((pos == 0) || (interval.par[pos-1] != '\\')) {
1491 math_end_waiting = true;
1498 // Ignore language if there is math somewhere in pattern-string
1499 if (isPatternString) {
1502 keys["foreignlanguage"].disabled = true;
1503 disableLanguageOverride = true;
1506 disableLanguageOverride = false;
1509 if (disableLanguageOverride) {
1510 keys["foreignlanguage"].disabled = true;
1513 math_pos = mi.getFirstPos();
1514 for (sregex_iterator it(interval.par.begin(), interval.par.end(), rkeys), end; it != end; ++it) {
1516 string key = sub.str(3);
1518 if (sub.str(0)[0] == '\\')
1519 key = sub.str(0)[1];
1523 size_t k_pos = sub.position(size_t(0));
1524 if ((k_pos > 0) && (interval.par[k_pos - 1] == '\\')) {
1525 // Escaped '$', ignoring
1531 if (evaluatingRegexp) {
1532 if (sub.str(1).compare("endregexp") == 0) {
1533 evaluatingRegexp = false;
1534 // found._tokenstart already set
1535 found._dataEnd = sub.position(size_t(0)) + 13;
1536 found._dataStart = found._dataEnd;
1537 found._tokensize = found._dataEnd - found._tokenstart;
1538 found.parenthesiscount = 0;
1542 if (evaluatingMath) {
1543 if (size_t(sub.position(size_t(0))) < mi.getEndPos())
1545 evaluatingMath = false;
1547 math_pos = mi.getStartPos();
1549 if (keys.find(key) == keys.end()) {
1550 found = KeyInfo(KeyInfo::isStandard, 0, true);
1551 if (isPatternString) {
1552 found.keytype = KeyInfo::isChar;
1553 found.disabled = false;
1560 if (key.compare("regexp") == 0) {
1561 evaluatingRegexp = true;
1562 found._tokenstart = sub.position(size_t(0));
1563 found._tokensize = 0;
1567 // Handle the other params of key
1568 if (found.keytype == KeyInfo::isIgnored)
1570 else if (found.keytype == KeyInfo::isMath) {
1571 if (size_t(sub.position(size_t(0))) == math_pos) {
1573 found._tokenstart = sub.position(size_t(0));
1574 found._tokensize = mi.getSize();
1575 found._dataEnd = found._tokenstart + found._tokensize;
1576 found._dataStart = found._dataEnd;
1577 found.parenthesiscount = 0;
1578 evaluatingMath = true;
1581 // begin|end of unknown env, discard
1582 // First handle tables
1583 // longtable|tabular
1584 bool discardComment;
1586 found.keytype = KeyInfo::doRemove;
1587 if ((sub.str(5).compare("longtable") == 0) ||
1588 (sub.str(5).compare("tabular") == 0)) {
1589 discardComment = true; /* '%' */
1592 discardComment = false;
1593 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
1595 string token = sub.str(5);
1596 if (regex_match(token, sub2, removeArgs)) {
1597 found.keytype = KeyInfo::removeWithArg;
1600 // discard spaces before pos(0)
1601 int pos = sub.position(size_t(0));
1603 for (count = 0; pos - count > 0; count++) {
1604 char c = interval.par[pos-count-1];
1605 if (discardComment) {
1606 if ((c != ' ') && (c != '%'))
1612 found._tokenstart = pos - count;
1613 if (sub.str(1).compare(0, 5, "begin") == 0) {
1614 size_t pos1 = pos + sub.str(0).length();
1615 if (sub.str(5).compare("cjk") == 0) {
1616 pos1 = interval.findclosing(pos1+1, interval.par.length()) + 1;
1617 if ((interval.par[pos1] == '{') && (interval.par[pos1+1] == '}'))
1619 found.keytype = KeyInfo::isMain;
1620 found._dataStart = pos1;
1621 found._dataEnd = interval.par.length();
1622 found.disabled = keys["foreignlanguage"].disabled;
1623 found.used = keys["foreignlanguage"].used;
1624 found._tokensize = pos1 - found._tokenstart;
1625 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1628 while (interval.par[pos1] == '[') {
1629 pos1 = interval.findclosing(pos1+1, interval.par.length(), '[', ']')+1;
1631 if (interval.par[pos1] == '{') {
1632 found._dataEnd = interval.findclosing(pos1+1, interval.par.length()) + 1;
1635 found._dataEnd = pos1;
1637 found._dataStart = found._dataEnd;
1638 found._tokensize = count + found._dataEnd - pos;
1639 found.parenthesiscount = 0;
1640 found.disabled = true;
1644 // Handle "\end{...}"
1645 found._dataStart = pos + sub.str(0).length();
1646 found._dataEnd = found._dataStart;
1647 found._tokensize = count + found._dataEnd - pos;
1648 found.parenthesiscount = 0;
1649 found.disabled = true;
1653 else if (found.keytype != KeyInfo::isRegex) {
1654 found._tokenstart = sub.position(size_t(0));
1655 if (found.parenthesiscount == 0) {
1656 // Probably to be discarded
1657 size_t following_pos = sub.position(size_t(0)) + sub.str(3).length() + 1;
1658 char following = interval.par[following_pos];
1659 if (following == ' ')
1660 found.head = "\\" + sub.str(3) + " ";
1661 else if (following == '=') {
1662 // like \uldepth=1000pt
1663 found.head = sub.str(0);
1666 found.head = "\\" + key;
1667 found._tokensize = found.head.length();
1668 found._dataEnd = found._tokenstart + found._tokensize;
1669 found._dataStart = found._dataEnd;
1672 int params = found._tokenstart + key.length() + 1;
1673 if (evaluatingOptional) {
1674 if (size_t(found._tokenstart) > optionalEnd) {
1675 evaluatingOptional = false;
1678 found.disabled = true;
1681 int optend = params;
1682 while (interval.par[optend] == '[') {
1683 // discard optional parameters
1684 optend = interval.findclosing(optend+1, interval.par.length(), '[', ']') + 1;
1686 if (optend > params) {
1687 key += interval.par.substr(params, optend-params);
1688 evaluatingOptional = true;
1689 optionalEnd = optend;
1691 string token = sub.str(5);
1692 int closings = found.parenthesiscount;
1693 if (found.parenthesiscount == 1) {
1694 found.head = "\\" + key + "{";
1696 else if (found.parenthesiscount > 1) {
1698 found.head = sub.str(0) + "{";
1699 closings = found.parenthesiscount - 1;
1702 found.head = "\\" + key + "{";
1705 found._tokensize = found.head.length();
1706 found._dataStart = found._tokenstart + found.head.length();
1707 if (interval.par.substr(found._dataStart-1, 15).compare("\\endarguments{}") == 0) {
1708 found._dataStart += 15;
1710 size_t endpos = interval.findclosing(found._dataStart, interval.par.length(), '{', '}', closings);
1711 if (found.keytype == KeyInfo::isList) {
1712 // Check if it really is list env
1713 static regex const listre("^([a-z]+)$");
1715 if (!regex_match(token, sub2, listre)) {
1716 // Change the key of this entry. It is not in a list/item environment
1717 found.keytype = KeyInfo::endArguments;
1720 if (found.keytype == KeyInfo::noMain) {
1721 evaluatingCode = true;
1723 codeStart = found._dataStart;
1725 else if (evaluatingCode) {
1726 if (size_t(found._dataStart) > codeEnd)
1727 evaluatingCode = false;
1728 else if (found.keytype == KeyInfo::isMain) {
1729 // Disable this key, treate it as standard
1730 found.keytype = KeyInfo::isStandard;
1731 found.disabled = true;
1732 if ((codeEnd == interval.par.length()) &&
1733 (found._tokenstart == codeStart)) {
1734 // trickery, because the code inset starts
1735 // with \selectlanguage ...
1737 if (entries.size() > 1) {
1738 entries[entries.size()-1]._dataEnd = codeEnd;
1743 if ((endpos == interval.par.length()) &&
1744 (found.keytype == KeyInfo::doRemove)) {
1745 // Missing closing => error in latex-input?
1746 // therefore do not delete remaining data
1747 found._dataStart -= 1;
1748 found._dataEnd = found._dataStart;
1751 found._dataEnd = endpos;
1753 if (isPatternString) {
1754 keys[key].used = true;
1757 entries.push_back(found);
1761 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
1763 stringstream s(keysstring);
1765 const char delim = '|';
1766 while (getline(s, key, delim)) {
1767 KeyInfo keyII(keyI);
1768 if (isPatternString) {
1771 else if ( !keys[key].used)
1772 keyII.disabled = true;
1777 void LatexInfo::buildKeys(bool isPatternString)
1780 static bool keysBuilt = false;
1781 if (keysBuilt && !isPatternString) return;
1783 // Known standard keys with 1 parameter.
1784 // Split is done, if not at start of region
1785 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
1786 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
1787 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
1788 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
1789 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
1790 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
1792 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
1793 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1794 makeKey("section*|subsection*|subsubsection*|paragraph*",
1795 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1796 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1797 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
1799 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
1801 // Split is done, if not at start of region
1802 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
1803 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
1805 // Split is done always.
1806 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
1810 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1811 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1812 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1813 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1815 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1816 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1817 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1819 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1820 // Custom space/skip, remove the content (== length value)
1821 makeKey("vspace|hspace|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
1822 // Found in fr/UserGuide.lyx
1823 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1825 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1826 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1827 // Known macros to remove (including their parameter)
1829 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
1830 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
1831 // handle like standard keys with 1 parameter.
1832 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
1834 // Macros to remove, but let the parameter survive
1836 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1838 // Remove language spec from content of these insets
1839 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
1841 // Same effect as previous, parameter will survive (because there is no one anyway)
1843 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1844 // Remove table decorations
1845 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
1846 // Discard shape-header.
1847 // For footnote or shortcut too, because of lang settings
1848 // and wrong handling if used 'KeyInfo::noMain'
1849 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1850 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1851 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1852 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1853 makeKey("hphantom|footnote|shortcut|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1854 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
1855 // like ('tiny{}' or '\tiny ' ... )
1856 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
1858 // Survives, like known character
1859 makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1860 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
1862 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1863 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1864 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1866 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1867 // Remove RTL/LTR marker
1868 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1869 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
1870 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
1871 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
1872 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
1873 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
1874 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1875 if (isPatternString) {
1876 // Allow the first searched string to rebuild the keys too
1880 // no need to rebuild again
1886 * Keep the list of actual opened parentheses actual
1887 * (e.g. depth == 4 means there are 4 '{' not processed yet)
1889 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
1892 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
1896 if (c == '\\') skip = 1;
1897 else if (c == '{') {
1900 else if (c == '}') {
1901 handleCloseP(i, closingAllowed);
1907 string Intervall::show(int lastpos)
1909 int idx = 0; /* int intervalls */
1912 for (idx = 0; idx <= ignoreidx; idx++) {
1913 while (i < lastpos) {
1915 if (i <= borders[idx].low) {
1916 if (borders[idx].low > lastpos)
1917 printsize = lastpos - i;
1919 printsize = borders[idx].low - i;
1920 s += par.substr(i, printsize);
1922 if (i >= borders[idx].low)
1923 i = borders[idx].upper;
1926 i = borders[idx].upper;
1932 s += par.substr(i, lastpos-i);
1938 void Intervall::output(ostringstream &os, int lastpos)
1940 // get number of chars to output
1941 int idx = 0; /* int intervalls */
1944 string startTitle = titleValue;
1945 for (idx = 0; idx <= ignoreidx; idx++) {
1947 if (i <= borders[idx].low) {
1949 if (borders[idx].low > lastpos)
1950 printsize = lastpos - i;
1952 printsize = borders[idx].low - i;
1953 if (printsize > 0) {
1954 os << startTitle << par.substr(i, printsize);
1956 printed += printsize;
1959 handleParentheses(i, false);
1960 if (i >= borders[idx].low)
1961 i = borders[idx].upper;
1964 i = borders[idx].upper;
1971 os << startTitle << par.substr(i, lastpos-i);
1972 printed += lastpos-i;
1974 handleParentheses(lastpos, false);
1975 for (int i = actualdeptindex; i > 0; --i) {
1978 if (hasTitle && (printed > 0))
1980 if (! isPatternString)
1982 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
1985 void LatexInfo::processRegion(int start, int region_end)
1987 while (start < region_end) { /* Let {[} and {]} survive */
1988 if ((interval.par[start] == '{') &&
1989 (interval.par[start+1] != ']') &&
1990 (interval.par[start+1] != '[')) {
1991 // Closing is allowed past the region
1992 int closing = interval.findclosing(start+1, interval.par.length());
1993 interval.addIntervall(start, start+1);
1994 interval.addIntervall(closing, closing+1);
1996 start = interval.nextNotIgnored(start+1);
2000 void LatexInfo::removeHead(KeyInfo &actual, int count)
2002 if (actual.parenthesiscount == 0) {
2003 // "{\tiny{} ...}" ==> "{{} ...}"
2004 interval.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2007 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2008 interval.addIntervall(actual._tokenstart - count, actual._dataStart);
2009 interval.addIntervall(actual._dataEnd, actual._dataEnd+1);
2013 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2016 switch (actual.keytype)
2018 case KeyInfo::isTitle: {
2020 nextKeyIdx = getNextKey();
2023 case KeyInfo::cleanToStart: {
2024 actual._dataEnd = actual._dataStart;
2025 nextKeyIdx = getNextKey();
2026 // Search for end of arguments
2027 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2029 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2030 entries[i].disabled = true;
2032 actual._dataEnd = entries[tmpIdx]._dataEnd;
2034 while (interval.par[actual._dataEnd] == ' ')
2036 interval.addIntervall(0, actual._dataEnd+1);
2037 interval.actualdeptindex = 0;
2038 interval.depts[0] = actual._dataEnd+1;
2039 interval.closes[0] = -1;
2042 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2043 if (actual.disabled)
2044 interval.addIntervall(actual._tokenstart, actual._dataEnd);
2046 interval.addIntervall(actual._dataStart, actual._dataEnd);
2049 case KeyInfo::isChar: {
2050 nextKeyIdx = getNextKey();
2053 case KeyInfo::isSize: {
2054 if (actual.disabled || (interval.par[actual._dataStart] != '{') || (interval.par[actual._dataStart-1] == ' ')) {
2055 processRegion(actual._dataEnd, actual._dataEnd+1); /* remove possibly following {} */
2056 interval.addIntervall(actual._tokenstart, actual._dataEnd+1);
2057 nextKeyIdx = getNextKey();
2059 // Here _dataStart points to '{', so correct it
2060 actual._dataStart += 1;
2061 actual._tokensize += 1;
2062 actual.parenthesiscount = 1;
2063 if (interval.par[actual._dataStart] == '}') {
2064 // Determine the end if used like '{\tiny{}...}'
2065 actual._dataEnd = interval.findclosing(actual._dataStart+1, interval.par.length()) + 1;
2066 interval.addIntervall(actual._dataStart, actual._dataStart+1);
2069 // Determine the end if used like '\tiny{...}'
2070 actual._dataEnd = interval.findclosing(actual._dataStart, interval.par.length()) + 1;
2072 // Split on this key if not at start
2073 int start = interval.nextNotIgnored(previousStart);
2074 if (start < actual._tokenstart) {
2075 interval.output(os, actual._tokenstart);
2076 interval.addIntervall(start, actual._tokenstart);
2078 // discard entry if at end of actual
2079 nextKeyIdx = process(os, actual);
2083 case KeyInfo::endArguments:
2084 // Remove trailing '{}' too
2085 actual._dataStart += 2;
2086 actual._dataEnd += 2;
2087 interval.addIntervall(actual._tokenstart, actual._dataEnd);
2088 nextKeyIdx = getNextKey();
2090 case KeyInfo::noMain:
2092 case KeyInfo::isStandard: {
2093 if (actual.disabled) {
2095 processRegion(actual._dataStart, actual._dataStart+1);
2096 nextKeyIdx = getNextKey();
2098 // Split on this key if not at datastart of calling entry
2099 int start = interval.nextNotIgnored(previousStart);
2100 if (start < actual._tokenstart) {
2101 interval.output(os, actual._tokenstart);
2102 interval.addIntervall(start, actual._tokenstart);
2104 // discard entry if at end of actual
2105 nextKeyIdx = process(os, actual);
2109 case KeyInfo::removeWithArg: {
2110 nextKeyIdx = getNextKey();
2111 // Search for end of arguments
2112 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2114 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2115 entries[i].disabled = true;
2117 actual._dataEnd = entries[tmpIdx]._dataEnd;
2119 interval.addIntervall(actual._tokenstart, actual._dataEnd+1);
2122 case KeyInfo::doRemove: {
2123 // Remove the key with all parameters and following spaces
2125 for (pos = actual._dataEnd+1; pos < interval.par.length(); pos++) {
2126 if ((interval.par[pos] != ' ') && (interval.par[pos] != '%'))
2129 interval.addIntervall(actual._tokenstart, pos);
2130 nextKeyIdx = getNextKey();
2133 case KeyInfo::isList: {
2134 // Discard space before _tokenstart
2136 for (count = 0; count < actual._tokenstart; count++) {
2137 if (interval.par[actual._tokenstart-count-1] != ' ')
2140 nextKeyIdx = getNextKey();
2141 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2143 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2145 // How else can we catch this one?
2146 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2147 entries[i].disabled = true;
2149 actual._dataEnd = entries[tmpIdx]._dataEnd;
2151 else if (nextKeyIdx > 0) {
2152 // Ignore any lang entries inside data region
2153 for (int i = nextKeyIdx; i < int(entries.size()) && entries[i]._tokenstart < actual._dataEnd; i++) {
2154 if (entries[i].keytype == KeyInfo::isMain)
2155 entries[i].disabled = true;
2158 if (actual.disabled) {
2159 interval.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2162 interval.addIntervall(actual._tokenstart-count, actual._tokenstart);
2164 if (interval.par[actual._dataEnd+1] == '[') {
2165 int posdown = interval.findclosing(actual._dataEnd+2, interval.par.length(), '[', ']');
2166 if ((interval.par[actual._dataEnd+2] == '{') &&
2167 (interval.par[posdown-1] == '}')) {
2168 interval.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
2169 interval.addIntervall(posdown-1, posdown+1);
2172 interval.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
2173 interval.addIntervall(posdown, posdown+1);
2175 int blk = interval.nextNotIgnored(actual._dataEnd+1);
2176 if (blk > posdown) {
2177 // Discard at most 1 space after empty item
2179 for (count = 0; count < 1; count++) {
2180 if (interval.par[blk+count] != ' ')
2184 interval.addIntervall(blk, blk+count);
2189 case KeyInfo::isSectioning: {
2190 // Discard spaces before _tokenstart
2192 int val = actual._tokenstart;
2193 for (count = 0; count < actual._tokenstart;) {
2194 val = interval.previousNotIgnored(val-1);
2195 if (interval.par[val] != ' ')
2198 count = actual._tokenstart - val;
2201 if (actual.disabled) {
2202 removeHead(actual, count);
2203 nextKeyIdx = getNextKey();
2205 interval.addIntervall(actual._tokenstart-count, actual._tokenstart);
2206 nextKeyIdx = process(os, actual);
2210 case KeyInfo::isMath: {
2211 // Same as regex, use the content unchanged
2212 nextKeyIdx = getNextKey();
2215 case KeyInfo::isRegex: {
2216 // DO NOT SPLIT ON REGEX
2218 nextKeyIdx = getNextKey();
2221 case KeyInfo::isIgnored: {
2222 // Treat like a character for now
2223 nextKeyIdx = getNextKey();
2226 case KeyInfo::isMain: {
2227 if (interval.par.substr(actual._dataStart, 2) == "% ")
2228 interval.addIntervall(actual._dataStart, actual._dataStart+2);
2229 if (actual._tokenstart > 0) {
2230 int prev = interval.previousNotIgnored(actual._tokenstart - 1);
2231 if ((prev >= 0) && interval.par[prev] == '%')
2232 interval.addIntervall(prev, prev+1);
2234 if (actual.disabled) {
2236 if ((interval.par.substr(actual._dataStart, 3) == " \\[") ||
2237 (interval.par.substr(actual._dataStart, 8) == " \\begin{")) {
2238 // Discard also the space before math-equation
2239 interval.addIntervall(actual._dataStart, actual._dataStart+1);
2241 nextKeyIdx = getNextKey();
2242 // interval.resetOpenedP(actual._dataStart-1);
2245 if (actual._tokenstart < 26) {
2246 // for the first (and maybe dummy) language
2247 interval.setForDefaultLang(actual);
2249 interval.resetOpenedP(actual._dataStart-1);
2253 case KeyInfo::invalid:
2254 // This cannot happen, already handled
2257 // LYXERR0("Unhandled keytype");
2258 nextKeyIdx = getNextKey();
2265 int LatexInfo::process(ostringstream &os, KeyInfo &actual )
2267 int end = interval.nextNotIgnored(actual._dataEnd);
2268 int oldStart = actual._dataStart;
2269 int nextKeyIdx = getNextKey();
2271 if ((nextKeyIdx < 0) ||
2272 (entries[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
2273 (entries[nextKeyIdx].keytype == KeyInfo::invalid)) {
2274 if (oldStart <= end) {
2275 processRegion(oldStart, end);
2280 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
2282 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
2283 (void) dispatch(os, actual._dataStart, nextKey);
2284 end = nextKey._tokenstart;
2287 processRegion(oldStart, nextKey._tokenstart);
2288 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
2290 oldStart = nextKey._dataEnd+1;
2292 // now nextKey is either invalid or is outside of actual._dataEnd
2293 // output the remaining and discard myself
2294 if (oldStart <= end) {
2295 processRegion(oldStart, end);
2297 if (interval.par[end] == '}') {
2299 // This is the normal case.
2300 // But if using the firstlanguage, the closing may be missing
2302 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
2304 if (actual._dataEnd < end)
2305 output_end = interval.nextNotIgnored(actual._dataEnd);
2307 output_end = interval.nextNotIgnored(end);
2308 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
2309 interval.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
2311 // Remove possible empty data
2312 int dstart = interval.nextNotIgnored(actual._dataStart);
2313 while ((dstart < output_end) && (interval.par[dstart] == '{')) {
2314 interval.addIntervall(dstart, dstart+1);
2315 int dend = interval.findclosing(dstart+1, output_end);
2316 interval.addIntervall(dend, dend+1);
2317 dstart = interval.nextNotIgnored(dstart+1);
2319 if (dstart < output_end)
2320 interval.output(os, output_end);
2321 interval.addIntervall(actual._tokenstart, end);
2325 string splitOnKnownMacros(string par, bool isPatternString)
2328 LatexInfo li(par, isPatternString);
2329 // LYXERR0("Berfore split: " << par);
2330 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
2332 DummyKey._tokensize = 0;
2333 DummyKey._dataStart = 0;
2334 DummyKey._dataEnd = par.length();
2335 DummyKey.disabled = true;
2336 int firstkeyIdx = li.getFirstKey();
2338 if (firstkeyIdx >= 0) {
2339 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
2340 DummyKey._tokenstart = firstKey._tokenstart;
2342 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
2343 // Use dummy firstKey
2344 firstKey = DummyKey;
2345 (void) li.setNextKey(firstkeyIdx);
2348 if (par.substr(firstKey._dataStart, 2) == "% ")
2349 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
2351 nextkeyIdx = li.process(os, firstKey);
2352 while (nextkeyIdx >= 0) {
2353 // Check for a possible gap between the last
2354 // entry and this one
2355 int datastart = li.nextNotIgnored(firstKey._dataStart);
2356 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
2357 if ((nextKey._tokenstart > datastart)) {
2359 firstKey._dataStart = datastart;
2360 firstKey._dataEnd = par.length();
2361 (void) li.setNextKey(nextkeyIdx);
2362 // Fake the last opened parenthesis
2363 li.setForDefaultLang(firstKey);
2364 nextkeyIdx = li.process(os, firstKey);
2367 if (nextKey.keytype != KeyInfo::isMain) {
2368 firstKey._dataStart = datastart;
2369 firstKey._dataEnd = nextKey._dataEnd+1;
2370 (void) li.setNextKey(nextkeyIdx);
2371 li.setForDefaultLang(firstKey);
2372 nextkeyIdx = li.process(os, firstKey);
2375 nextkeyIdx = li.process(os, nextKey);
2379 // Handle the remaining
2380 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
2381 firstKey._dataEnd = par.length();
2383 if ((firstKey._dataStart < firstKey._dataEnd) &&
2384 (par[firstKey._dataStart] != '}')) {
2385 li.setForDefaultLang(firstKey);
2386 (void) li.process(os, firstKey);
2390 // return string definitelly impossible to match
2391 s = "\\foreignlanguage{ignore}{ }";
2395 s = par; /* no known macros found */
2396 // LYXERR0("After split: " << s);
2401 * Try to unify the language specs in the latexified text.
2402 * Resulting modified string is set to "", if
2403 * the searched tex does not contain all the features in the search pattern
2405 static string correctlanguagesetting(string par, bool isPatternString, bool withformat)
2407 static Features regex_f;
2408 static int missed = 0;
2409 static bool regex_with_format = false;
2411 int parlen = par.length();
2413 while ((parlen > 0) && (par[parlen-1] == '\n')) {
2416 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
2417 // Happens to be there in case of description or labeling environment
2422 // Split the latex input into pieces which
2423 // can be digested by our search engine
2424 LYXERR(Debug::FIND, "input: \"" << par << "\"");
2425 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
2426 LYXERR(Debug::FIND, "After split: \"" << result << "\"");
2429 result = par.substr(0, parlen);
2430 if (isPatternString) {
2433 regex_f = identifyFeatures(result);
2434 string features = "";
2435 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2436 string a = it->first;
2437 regex_with_format = true;
2438 features += " " + a;
2439 // LYXERR0("Identified regex format:" << a);
2441 LYXERR(Debug::FIND, "Identified Features" << features);
2444 } else if (regex_with_format) {
2445 Features info = identifyFeatures(result);
2446 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2447 string a = it->first;
2448 bool b = it->second;
2449 if (b && ! info[a]) {
2451 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
2457 // LYXERR0("No regex formats");
2463 // Remove trailing closure of math, macros and environments, so to catch parts of them.
2464 static int identifyClosing(string & t)
2466 int open_braces = 0;
2468 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
2469 if (regex_replace(t, t, "(.*[^\\\\])\\$" REGEX_EOS, "$1"))
2471 if (regex_replace(t, t, "(.*[^\\\\]) \\\\\\]" REGEX_EOS, "$1"))
2473 if (regex_replace(t, t, "(.*[^\\\\]) \\\\end\\{[a-zA-Z_]*\\*?\\}" REGEX_EOS, "$1"))
2475 if (regex_replace(t, t, "(.*[^\\\\])\\}" REGEX_EOS, "$1")) {
2485 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt)
2486 : p_buf(&buf), p_first_buf(&buf), opt(opt)
2488 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
2489 docstring const & ds = stringifySearchBuffer(find_buf, opt);
2490 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
2491 // When using regexp, braces are hacked already by escape_for_regex()
2492 par_as_string = normalize(ds, !use_regexp);
2494 close_wildcards = 0;
2496 size_t lead_size = 0;
2497 // correct the language settings
2498 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat);
2499 if (opt.ignoreformat) {
2501 // if par_as_string_nolead were emty,
2502 // the following call to findAux will always *find* the string
2503 // in the checked data, and thus always using the slow
2504 // examining of the current text part.
2505 par_as_string_nolead = par_as_string;
2508 lead_size = identifyLeading(par_as_string);
2509 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
2510 lead_as_string = par_as_string.substr(0, lead_size);
2511 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
2515 open_braces = identifyClosing(par_as_string);
2516 identifyClosing(par_as_string_nolead);
2517 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2518 LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'");
2520 string lead_as_regexp;
2521 if (lead_size > 0) {
2522 // @todo No need to search for \regexp{} insets in leading material
2523 lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat);
2524 par_as_string = par_as_string_nolead;
2525 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
2526 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2528 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
2529 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
2530 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2532 // Insert .* before trailing '\$' ('$' has been escaped by escape_for_regex)
2533 regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\$)\\'", "$1(.*?)$2")
2534 // Insert .* before trailing '\\\]' ('\]' has been escaped by escape_for_regex)
2535 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])( \\\\\\\\\\\\\\])\\'", "$1(.*?)$2")
2536 // Insert .* before trailing '\\end\{...}' ('\end{...}' has been escaped by escape_for_regex)
2537 || regex_replace(par_as_string, par_as_string,
2538 "(.*[^\\\\])( \\\\\\\\end\\\\\\{[a-zA-Z_]*)(\\\\\\*)?(\\\\\\})\\'", "$1(.*?)$2$3$4")
2539 // Insert .* before trailing '\}' ('}' has been escaped by escape_for_regex)
2540 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\})\\'", "$1(.*?)$2")
2544 if (!opt.ignoreformat) {
2545 // Remove extra '\}' at end
2546 while ( regex_replace(par_as_string, par_as_string, "(.*)\\\\}$", "$1")) {
2551 regex_replace(par_as_string, par_as_string, "\\\\\\.", "_xxbdotxx_");
2552 // handle '.' -> '[^]', replace later as '[^\}\{\\]'
2553 regex_replace(par_as_string, par_as_string, "\\.", "[^]");
2554 // replace '[^...]' with '[^...\}\{\\]'
2555 regex_replace(par_as_string, par_as_string, "\\[\\^([^\\\\\\]]*)\\]", "_xxbrlxx_$1\\}\\{\\\\_xxbrrxx_");
2556 regex_replace(par_as_string, par_as_string, "_xxbrlxx_", "[^");
2557 regex_replace(par_as_string, par_as_string, "_xxbrrxx_", "]");
2559 regex_replace(par_as_string, par_as_string, "_xxbdotxx_", "\\.");
2562 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2563 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2564 LYXERR(Debug::FIND, "Close .*? : " << close_wildcards);
2565 LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
2567 // If entered regexp must match at begin of searched string buffer
2568 // Kornel: Added parentheses to use $1 for size of the leading string
2572 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
2573 // Unfortunately is '\1', '\2', etc not working for strings with extra format
2574 // so the convert has no effect in that case
2575 for (int i = 8; i > 0; --i) {
2576 string orig = "\\\\" + std::to_string(i);
2577 string dest = "\\" + std::to_string(i+1);
2578 while (regex_replace(par_as_string, par_as_string, orig, dest));
2580 regexp_str = "(" + lead_as_regexp + ")" + par_as_string;
2581 regexp2_str = "(" + lead_as_regexp + ").*?" + par_as_string;
2583 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
2584 regexp = lyx::regex(regexp_str);
2586 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
2587 regexp2 = lyx::regex(regexp2_str);
2592 // Count number of characters in string
2596 // \\[a-zA-Z]+ ==> 1
2597 static int computeSize(string s, int len)
2603 for (int i = 0; i < len; i += skip, count++) {
2606 if (isalpha(s[i+1])) {
2607 for (int j = 2; i+j < len; j++) {
2608 if (! isalpha(s[i+j])) {
2611 else if ((s[i+j] == '{') && s[i+j+1] == '}')
2613 else if ((s[i+j] == '{') && (i + j + 1 >= len))
2621 else if (s[i] == '{') {
2627 else if (s[i] == '-') {
2628 if (s[i+1] == '-') {
2644 int MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
2647 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
2650 docstring docstr = stringifyFromForSearch(opt, cur, len);
2651 string str = normalize(docstr, true);
2652 if (!opt.ignoreformat) {
2653 str = correctlanguagesetting(str, false, !opt.ignoreformat);
2655 if (str.empty()) return(-1);
2656 LYXERR(Debug::FIND, "Matching against '" << lyx::to_utf8(docstr) << "'");
2657 LYXERR(Debug::FIND, "After normalization: '" << str << "'");
2660 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
2661 regex const *p_regexp;
2662 regex_constants::match_flag_type flags;
2664 flags = regex_constants::match_continuous;
2667 flags = regex_constants::match_default;
2668 p_regexp = ®exp2;
2670 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
2671 if (re_it == sregex_iterator())
2673 match_results<string::const_iterator> const & m = *re_it;
2675 if (0) { // Kornel Benko: DO NOT CHECKK
2676 // Check braces on the segment that matched the entire regexp expression,
2677 // plus the last subexpression, if a (.*?) was inserted in the constructor.
2678 if (!braces_match(m[0].first, m[0].second, open_braces))
2682 // Check braces on segments that matched all (.*?) subexpressions,
2683 // except the last "padding" one inserted by lyx.
2684 for (size_t i = 1; i < m.size() - 1; ++i)
2685 if (!braces_match(m[i].first, m[i].second, open_braces))
2688 // Exclude from the returned match length any length
2689 // due to close wildcards added at end of regexp
2690 // and also the length of the leading (e.g. '\emph{')
2692 // Whole found string, including the leading: m[0].second - m[0].first
2693 // Size of the leading string: m[1].second - m[1].first
2694 int leadingsize = 0;
2696 leadingsize = m[1].second - m[1].first;
2698 for (size_t i = 0; i < m.size(); i++) {
2699 LYXERR(Debug::FIND, "Match " << i << " is " << m[i].second - m[i].first << " long");
2701 if (close_wildcards == 0)
2702 result = m[0].second - m[0].first;
2705 result = m[m.size() - close_wildcards].first - m[0].first;
2707 size_t pos = m.position(size_t(0));
2708 // Ignore last closing characters
2709 while (result > 0) {
2710 if (str[pos+result-1] == '}')
2715 if (result > leadingsize)
2716 result -= leadingsize;
2719 return computeSize(str.substr(pos+leadingsize,result), result);
2722 // else !use_regexp: but all code paths above return
2723 LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='"
2724 << par_as_string << "', str='" << str << "'");
2725 LYXERR(Debug::FIND, "Searching in normal mode: lead_as_string='"
2726 << lead_as_string << "', par_as_string_nolead='"
2727 << par_as_string_nolead << "'");
2730 LYXERR(Debug::FIND, "size=" << par_as_string.size()
2731 << ", substr='" << str.substr(0, par_as_string.size()) << "'");
2732 if (str.substr(0, par_as_string.size()) == par_as_string)
2733 return par_as_string.size();
2735 size_t pos = str.find(par_as_string_nolead);
2736 if (pos != string::npos)
2737 return par_as_string.size();
2743 int MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
2745 int res = findAux(cur, len, at_begin);
2747 "res=" << res << ", at_begin=" << at_begin
2748 << ", matchword=" << opt.matchword
2749 << ", inTexted=" << cur.inTexted());
2750 if (res == 0 || !at_begin || !opt.matchword || !cur.inTexted())
2752 if ((len > 0) && (res < len))
2754 Paragraph const & par = cur.paragraph();
2755 bool ws_left = (cur.pos() > 0)
2756 ? par.isWordSeparator(cur.pos() - 1)
2758 bool ws_right = (cur.pos() + len < par.size())
2759 ? par.isWordSeparator(cur.pos() + len)
2762 "cur.pos()=" << cur.pos() << ", res=" << res
2763 << ", separ: " << ws_left << ", " << ws_right
2766 if (ws_left && ws_right) {
2767 // Check for word separators inside the found 'word'
2768 for (int i = 0; i < len; i++) {
2769 if (par.isWordSeparator(cur.pos() + i))
2778 string MatchStringAdv::normalize(docstring const & s, bool hack_braces) const
2781 if (! opt.casesensitive)
2782 t = lyx::to_utf8(lowercase(s));
2784 t = lyx::to_utf8(s);
2785 // Remove \n at begin
2786 while (!t.empty() && t[0] == '\n')
2789 while (!t.empty() && t[t.size() - 1] == '\n')
2790 t = t.substr(0, t.size() - 1);
2792 // Handle all other '\n'
2793 while ((pos = t.find("\n")) != string::npos) {
2794 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
2796 if (std::isalnum(t[pos+1])) {
2797 t.replace(pos-2, 3, " ");
2800 t.replace(pos-2, 3, "");
2803 else if (!std::isalnum(t[pos+1]) || !std::isalnum(t[pos-1])) {
2804 // '\n' adjacent to non-alpha-numerics, discard
2805 t.replace(pos, 1, "");
2808 // Replace all other \n with spaces
2809 t.replace(pos, 1, " ");
2812 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
2813 // Kornel: Added textsl, textsf, textit, texttt and noun
2814 // + allow to seach for colored text too
2815 LYXERR(Debug::FIND, "Removing stale empty \\emph{}, \\textbf{}, \\*section{} macros from: " << t);
2816 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
2817 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
2818 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
2819 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
2821 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
2822 // FIXME - check what preceeds the brace
2824 if (opt.ignoreformat)
2825 while (regex_replace(t, t, "\\{", "_x_<")
2826 || regex_replace(t, t, "\\}", "_x_>"))
2827 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
2829 while (regex_replace(t, t, "\\\\\\{", "_x_<")
2830 || regex_replace(t, t, "\\\\\\}", "_x_>"))
2831 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
2838 docstring stringifyFromCursor(DocIterator const & cur, int len)
2840 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
2841 if (cur.inTexted()) {
2842 Paragraph const & par = cur.paragraph();
2843 // TODO what about searching beyond/across paragraph breaks ?
2844 // TODO Try adding a AS_STR_INSERTS as last arg
2845 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
2846 int(par.size()) : cur.pos() + len;
2847 OutputParams runparams(&cur.buffer()->params().encoding());
2848 runparams.nice = true;
2849 runparams.flavor = OutputParams::LATEX;
2850 runparams.linelen = 100000; //lyxrc.plaintext_linelen;
2851 // No side effect of file copying and image conversion
2852 runparams.dryrun = true;
2853 LYXERR(Debug::FIND, "Stringifying with cur: "
2854 << cur << ", from pos: " << cur.pos() << ", end: " << end);
2855 return par.asString(cur.pos(), end,
2856 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
2858 } else if (cur.inMathed()) {
2859 CursorSlice cs = cur.top();
2860 MathData md = cs.cell();
2861 MathData::const_iterator it_end =
2862 (( len == -1 || cs.pos() + len > int(md.size()))
2864 : md.begin() + cs.pos() + len );
2866 for (MathData::const_iterator it = md.begin() + cs.pos();
2869 docstring s = asString(md2);
2870 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
2873 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
2878 /** Computes the LaTeX export of buf starting from cur and ending len positions
2879 * after cur, if len is positive, or at the paragraph or innermost inset end
2882 docstring latexifyFromCursor(DocIterator const & cur, int len)
2884 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
2885 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
2886 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
2887 Buffer const & buf = *cur.buffer();
2889 odocstringstream ods;
2891 OutputParams runparams(&buf.params().encoding());
2892 runparams.nice = false;
2893 runparams.flavor = OutputParams::LATEX;
2894 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
2895 // No side effect of file copying and image conversion
2896 runparams.dryrun = true;
2897 runparams.for_search = true;
2899 if (cur.inTexted()) {
2900 // @TODO what about searching beyond/across paragraph breaks ?
2901 pos_type endpos = cur.paragraph().size();
2902 if (len != -1 && endpos > cur.pos() + len)
2903 endpos = cur.pos() + len;
2904 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
2905 string(), cur.pos(), endpos);
2906 string s = lyx::to_utf8(ods.str());
2907 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
2908 return(lyx::from_utf8(s));
2909 } else if (cur.inMathed()) {
2910 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
2911 for (int s = cur.depth() - 1; s >= 0; --s) {
2912 CursorSlice const & cs = cur[s];
2913 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
2915 cs.asInsetMath()->asHullInset()->header_write(ws);
2920 CursorSlice const & cs = cur.top();
2921 MathData md = cs.cell();
2922 MathData::const_iterator it_end =
2923 ((len == -1 || cs.pos() + len > int(md.size()))
2925 : md.begin() + cs.pos() + len);
2927 for (MathData::const_iterator it = md.begin() + cs.pos();
2931 ods << asString(md2);
2932 // Retrieve the math environment type, and add '$' or '$]'
2933 // or others (\end{equation}) accordingly
2934 for (int s = cur.depth() - 1; s >= 0; --s) {
2935 CursorSlice const & cs2 = cur[s];
2936 InsetMath * inset = cs2.asInsetMath();
2937 if (inset && inset->asHullInset()) {
2939 inset->asHullInset()->footer_write(ws);
2943 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
2945 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
2951 /** Finalize an advanced find operation, advancing the cursor to the innermost
2952 ** position that matches, plus computing the length of the matching text to
2955 int findAdvFinalize(DocIterator & cur, MatchStringAdv const & match)
2957 // Search the foremost position that matches (avoids find of entire math
2958 // inset when match at start of it)
2960 DocIterator old_cur(cur.buffer());
2962 LYXERR(Debug::FIND, "Forwarding one step (searching for innermost match)");
2966 } while (cur && cur.depth() > d && match(cur) > 0);
2968 int max_match = match(cur); /* match valid only if not searching whole words */
2969 if (max_match <= 0) return 0;
2970 LYXERR(Debug::FIND, "Ok");
2972 // Compute the match length
2974 if (cur.pos() + len > cur.lastpos())
2976 if (match.opt.matchword) {
2977 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
2978 while (cur.pos() + len <= cur.lastpos() && match(cur, len) <= 0) {
2980 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
2982 // Length of matched text (different from len param)
2983 int old_match = match(cur, len);
2987 // Greedy behaviour while matching regexps
2988 while ((new_match = match(cur, len + 1)) > old_match) {
2990 old_match = new_match;
2991 LYXERR(Debug::FIND, "verifying match with len = " << len);
2998 int maxl = cur.lastpos() - cur.pos();
2999 // Greedy behaviour while matching regexps
3000 while (maxl > minl) {
3001 int actual_match = match(cur, len);
3002 if (actual_match >= max_match) {
3003 // actual_match > max_match _can_ happen,
3004 // if the search area splits
3005 // some following word so that the regex
3006 // (e.g. 'r.*r\b' matches 'r' from the middle of the
3008 // This means, the len value is too big
3010 len = (int)((maxl + minl)/2);
3013 // (actual_match < max_match)
3015 len = (int)((maxl + minl)/2);
3019 // Search for real start of matched characters
3024 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
3025 if (cur.depth() < old_cur.depth()) {
3027 LYXERR0("cur.depth() < old_cur.depth(), this should never happen");
3030 if (cur.pos() != old_cur.pos()) {
3031 // OK, forwarded 1 pos in actual inset
3032 actual_match = match(cur, len-1);
3033 if (actual_match == max_match) {
3034 // Ha, got it! The shorter selection has the same match length
3039 // OK, the shorter selection matches less chars, revert to previous value
3045 LYXERR0("cur.pos() == old_cur.pos(), this should never happen");
3046 actual_match = match(cur, len);
3047 if (actual_match == max_match)
3057 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
3061 while (!theApp()->longOperationCancelled() && cur) {
3062 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
3063 int match_len = match(cur, -1, false);
3064 LYXERR(Debug::FIND, "match_len: " << match_len);
3065 if (match_len > 0) {
3066 int match_len_zero_count = 0;
3067 for (; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) {
3068 LYXERR(Debug::FIND, "Advancing cur: " << cur);
3069 int match_len3 = match(cur, 1);
3072 int match_len2 = match(cur);
3073 LYXERR(Debug::FIND, "match_len2: " << match_len2);
3074 if (match_len2 > 0) {
3075 // Sometimes in finalize we understand it wasn't a match
3076 // and we need to continue the outest loop
3077 int len = findAdvFinalize(cur, match);
3082 if (match_len2 >= 0) {
3083 if (match_len2 == 0)
3084 match_len_zero_count++;
3086 match_len_zero_count = 0;
3089 if (++match_len_zero_count > 3) {
3090 LYXERR(Debug::FIND, "match_len2_zero_count: " << match_len_zero_count << ", match_len was " << match_len);
3091 match_len_zero_count = 0;
3099 if (match_len >= 0 && cur.pit() < cur.lastpit()) {
3100 LYXERR(Debug::FIND, "Advancing par: cur=" << cur);
3103 // This should exit nested insets, if any, or otherwise undefine the currsor.
3104 cur.pos() = cur.lastpos();
3105 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
3113 /// Find the most backward consecutive match within same paragraph while searching backwards.
3114 int findMostBackwards(DocIterator & cur, MatchStringAdv const & match)
3116 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3117 DocIterator tmp_cur = cur;
3118 int len = findAdvFinalize(tmp_cur, match);
3119 Inset & inset = cur.inset();
3120 for (; cur != cur_begin; cur.backwardPos()) {
3121 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
3122 DocIterator new_cur = cur;
3123 new_cur.backwardPos();
3124 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur))
3126 int new_len = findAdvFinalize(new_cur, match);
3131 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
3137 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
3141 // Backup of original position
3142 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3143 if (cur == cur_begin)
3146 DocIterator cur_orig(cur);
3147 bool pit_changed = false;
3150 bool found_match = match(cur, -1, false);
3154 cur.pos() = cur.lastpos();
3156 cur.pos() = cur_orig.pos();
3157 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
3158 DocIterator cur_prev_iter;
3160 found_match = match(cur);
3161 LYXERR(Debug::FIND, "findBackAdv3: found_match="
3162 << found_match << ", cur: " << cur);
3164 return findMostBackwards(cur, match);
3166 // Stop if begin of document reached
3167 if (cur == cur_begin)
3169 cur_prev_iter = cur;
3173 if (cur == cur_begin)
3180 } while (!theApp()->longOperationCancelled());
3188 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
3189 DocIterator const & cur, int len)
3191 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
3193 if (!opt.ignoreformat)
3194 return latexifyFromCursor(cur, len);
3196 return stringifyFromCursor(cur, len);
3200 FindAndReplaceOptions::FindAndReplaceOptions(
3201 docstring const & find_buf_name, bool casesensitive,
3202 bool matchword, bool forward, bool expandmacros, bool ignoreformat,
3203 docstring const & repl_buf_name, bool keep_case,
3204 SearchScope scope, SearchRestriction restr)
3205 : find_buf_name(find_buf_name), casesensitive(casesensitive), matchword(matchword),
3206 forward(forward), expandmacros(expandmacros), ignoreformat(ignoreformat),
3207 repl_buf_name(repl_buf_name), keep_case(keep_case), scope(scope), restr(restr)
3215 /** Check if 'len' letters following cursor are all non-lowercase */
3216 static bool allNonLowercase(Cursor const & cur, int len)
3218 pos_type beg_pos = cur.selectionBegin().pos();
3219 pos_type end_pos = cur.selectionBegin().pos() + len;
3220 if (len > cur.lastpos() + 1 - beg_pos) {
3221 LYXERR(Debug::FIND, "This should not happen, more debug needed");
3222 len = cur.lastpos() + 1 - beg_pos;
3223 end_pos = beg_pos + len;
3225 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
3226 if (isLowerCase(cur.paragraph().getChar(pos)))
3232 /** Check if first letter is upper case and second one is lower case */
3233 static bool firstUppercase(Cursor const & cur)
3236 pos_type pos = cur.selectionBegin().pos();
3237 if (pos >= cur.lastpos() - 1) {
3238 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
3241 ch1 = cur.paragraph().getChar(pos);
3242 ch2 = cur.paragraph().getChar(pos + 1);
3243 bool result = isUpperCase(ch1) && isLowerCase(ch2);
3244 LYXERR(Debug::FIND, "firstUppercase(): "
3245 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
3246 << ch2 << "(" << char(ch2) << ")"
3247 << ", result=" << result << ", cur=" << cur);
3252 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
3254 ** \fixme What to do with possible further paragraphs in replace buffer ?
3256 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
3258 ParagraphList::iterator pit = buffer.paragraphs().begin();
3259 LASSERT(pit->size() >= 1, /**/);
3260 pos_type right = pos_type(1);
3261 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
3262 right = pit->size();
3263 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
3269 static void findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
3271 Cursor & cur = bv->cursor();
3272 if (opt.repl_buf_name == docstring()
3273 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
3274 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3277 DocIterator sel_beg = cur.selectionBegin();
3278 DocIterator sel_end = cur.selectionEnd();
3279 if (&sel_beg.inset() != &sel_end.inset()
3280 || sel_beg.pit() != sel_end.pit()
3281 || sel_beg.idx() != sel_end.idx())
3283 int sel_len = sel_end.pos() - sel_beg.pos();
3284 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
3285 << ", sel_len: " << sel_len << endl);
3288 LASSERT(sel_len > 0, return);
3290 if (!matchAdv(sel_beg, sel_len))
3293 // Build a copy of the replace buffer, adapted to the KeepCase option
3294 Buffer & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
3296 repl_buffer_orig.write(oss);
3297 string lyx = oss.str();
3298 Buffer repl_buffer("", false);
3299 repl_buffer.setUnnamed(true);
3300 LASSERT(repl_buffer.readString(lyx), return);
3301 if (opt.keep_case && sel_len >= 2) {
3302 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
3303 if (cur.inTexted()) {
3304 if (firstUppercase(cur))
3305 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
3306 else if (allNonLowercase(cur, sel_len))
3307 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
3310 cap::cutSelection(cur, false);
3311 if (cur.inTexted()) {
3312 repl_buffer.changeLanguage(
3313 repl_buffer.language(),
3314 cur.getFont().language());
3315 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
3316 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
3317 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
3318 repl_buffer.params().documentClassPtr(),
3319 bv->buffer().errorList("Paste"));
3320 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
3321 sel_len = repl_buffer.paragraphs().begin()->size();
3322 } else if (cur.inMathed()) {
3323 odocstringstream ods;
3325 OutputParams runparams(&repl_buffer.params().encoding());
3326 runparams.nice = false;
3327 runparams.flavor = OutputParams::LATEX;
3328 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3329 runparams.dryrun = true;
3330 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
3331 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
3332 docstring repl_latex = ods.str();
3333 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
3335 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
3336 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
3337 repl_latex = from_utf8(s);
3338 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
3339 MathData ar(cur.buffer());
3340 asArray(repl_latex, ar, Parse::NORMAL);
3342 sel_len = ar.size();
3343 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3345 if (cur.pos() >= sel_len)
3346 cur.pos() -= sel_len;
3349 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3350 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
3351 bv->processUpdateFlags(Update::Force);
3355 /// Perform a FindAdv operation.
3356 bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt)
3361 // e.g., when invoking word-findadv from mini-buffer wither with
3362 // wrong options syntax or before ever opening advanced F&R pane
3363 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3367 MatchStringAdv matchAdv(bv->buffer(), opt);
3368 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
3370 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
3371 findAdvReplace(bv, opt, matchAdv);
3374 match_len = findForwardAdv(cur, matchAdv);
3376 match_len = findBackwardsAdv(cur, matchAdv);
3378 // This may only be raised by lyx::regex()
3379 bv->message(_("Invalid regular expression!"));
3383 if (match_len == 0) {
3384 bv->message(_("Match not found!"));
3388 bv->message(_("Match found!"));
3390 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << match_len);
3391 bv->putSelectionAt(cur, match_len, !opt.forward);
3397 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
3399 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
3400 << opt.casesensitive << ' '
3401 << opt.matchword << ' '
3402 << opt.forward << ' '
3403 << opt.expandmacros << ' '
3404 << opt.ignoreformat << ' '
3405 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
3406 << opt.keep_case << ' '
3407 << int(opt.scope) << ' '
3410 LYXERR(Debug::FIND, "built: " << os.str());
3416 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
3418 LYXERR(Debug::FIND, "parsing");
3422 while (line != "EOSS") {
3426 if (is.eof()) // Tolerate malformed request
3430 LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
3431 opt.find_buf_name = from_utf8(s);
3432 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat;
3433 is.get(); // Waste space before replace string
3436 while (line != "EOSS") {
3440 if (is.eof()) // Tolerate malformed request
3444 LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
3445 opt.repl_buf_name = from_utf8(s);
3446 is >> opt.keep_case;
3449 opt.scope = FindAndReplaceOptions::SearchScope(i);
3451 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
3453 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
3454 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
3455 << opt.scope << ' ' << opt.restr);