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"
55 #include "support/textutils.h"
59 using namespace lyx::support;
64 // Helper class for deciding what should be ignored
69 : ignoreFamily_(false), ignoreSeries_(false),
70 ignoreShape_(false), ignoreUnderline_(false),
71 ignoreMarkUp_(false), ignoreStrikeOut_(false),
72 ignoreSectioning_(false), ignoreFrontMatter_(false),
73 ignoreColor_(false), ignoreLanguage_(false) {}
75 bool getFamily() { return ignoreFamily_; };
77 bool getSeries() { return ignoreSeries_; };
79 bool getShape() { return ignoreShape_; };
81 bool getUnderline() { return ignoreUnderline_; };
83 bool getMarkUp() { return ignoreMarkUp_; };
85 bool getStrikeOut() { return ignoreStrikeOut_; };
87 bool getSectioning() { return ignoreSectioning_; };
89 bool getFrontMatter() { return ignoreFrontMatter_; };
91 bool getColor() { return ignoreColor_; };
93 bool getLanguage() { return ignoreLanguage_; };
95 void setIgnoreFormat(string type, bool value);
105 bool ignoreUnderline_;
109 bool ignoreStrikeOut_;
111 bool ignoreSectioning_;
113 bool ignoreFrontMatter_;
117 bool ignoreLanguage_;
121 void IgnoreFormats::setIgnoreFormat(string type, bool value)
123 if (type == "color") {
124 ignoreColor_ = value;
126 else if (type == "language") {
127 ignoreLanguage_ = value;
129 else if (type == "sectioning") {
130 ignoreSectioning_ = value;
131 ignoreFrontMatter_ = value;
133 else if (type == "font") {
134 ignoreSeries_ = value;
135 ignoreShape_ = value;
136 ignoreFamily_ = value;
138 else if (type == "series") {
139 ignoreSeries_ = value;
141 else if (type == "shape") {
142 ignoreShape_ = value;
144 else if (type == "family") {
145 ignoreFamily_ = value;
147 else if (type == "markup") {
148 ignoreMarkUp_ = value;
150 else if (type == "underline") {
151 ignoreUnderline_ = value;
153 else if (type == "strike") {
154 ignoreStrikeOut_ = value;
158 // The global variable that can be changed from outside
159 IgnoreFormats ignoreFormats;
162 void setIgnoreFormat(string type, bool value)
164 ignoreFormats.setIgnoreFormat(type, value);
170 bool parse_bool(docstring & howto)
175 howto = split(howto, var, ' ');
180 class MatchString : public binary_function<Paragraph, pos_type, int>
183 MatchString(docstring const & str, bool cs, bool mw)
184 : str(str), case_sens(cs), whole_words(mw)
187 // returns true if the specified string is at the specified position
188 // del specifies whether deleted strings in ct mode will be considered
189 int operator()(Paragraph const & par, pos_type pos, bool del = true) const
191 return par.find(str, case_sens, whole_words, pos, del);
199 // match whole words only
204 int findForward(DocIterator & cur, MatchString const & match,
205 bool find_del = true)
207 for (; cur; cur.forwardChar())
208 if (cur.inTexted()) {
209 int len = match(cur.paragraph(), cur.pos(), find_del);
217 int findBackwards(DocIterator & cur, MatchString const & match,
218 bool find_del = true)
222 if (cur.inTexted()) {
223 int len = match(cur.paragraph(), cur.pos(), find_del);
232 bool searchAllowed(docstring const & str)
235 frontend::Alert::error(_("Search error"), _("Search string is empty"));
242 bool findOne(BufferView * bv, docstring const & searchstr,
243 bool case_sens, bool whole, bool forward,
244 bool find_del = true, bool check_wrap = false)
246 if (!searchAllowed(searchstr))
249 DocIterator cur = forward
250 ? bv->cursor().selectionEnd()
251 : bv->cursor().selectionBegin();
253 MatchString const match(searchstr, case_sens, whole);
255 int match_len = forward
256 ? findForward(cur, match, find_del)
257 : findBackwards(cur, match, find_del);
260 bv->putSelectionAt(cur, match_len, !forward);
261 else if (check_wrap) {
262 DocIterator cur_orig(bv->cursor());
265 q = _("End of file reached while searching forward.\n"
266 "Continue searching from the beginning?");
268 q = _("Beginning of file reached while searching backward.\n"
269 "Continue searching from the end?");
270 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
271 q, 0, 1, _("&Yes"), _("&No"));
272 if (wrap_answer == 0) {
274 bv->cursor().clear();
275 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
277 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
278 bv->cursor().backwardPos();
280 bv->clearSelection();
281 if (findOne(bv, searchstr, case_sens, whole, forward, find_del, false))
284 bv->cursor().setCursor(cur_orig);
288 return match_len > 0;
292 int replaceAll(BufferView * bv,
293 docstring const & searchstr, docstring const & replacestr,
294 bool case_sens, bool whole)
296 Buffer & buf = bv->buffer();
298 if (!searchAllowed(searchstr) || buf.isReadonly())
301 DocIterator cur_orig(bv->cursor());
303 MatchString const match(searchstr, case_sens, whole);
306 int const rsize = replacestr.size();
307 int const ssize = searchstr.size();
310 cur.setCursor(doc_iterator_begin(&buf));
311 int match_len = findForward(cur, match, false);
312 while (match_len > 0) {
313 // Backup current cursor position and font.
314 pos_type const pos = cur.pos();
315 Font const font = cur.paragraph().getFontSettings(buf.params(), pos);
317 int striked = ssize -
318 cur.paragraph().eraseChars(pos, pos + match_len,
319 buf.params().track_changes);
320 cur.paragraph().insert(pos, replacestr, font,
321 Change(buf.params().track_changes
323 : Change::UNCHANGED));
324 for (int i = 0; i < rsize + striked; ++i)
327 match_len = findForward(cur, match, false);
330 bv->putSelectionAt(doc_iterator_begin(&buf), 0, false);
332 cur_orig.fixIfBroken();
333 bv->setCursor(cur_orig);
339 // the idea here is that we are going to replace the string that
340 // is selected IF it is the search string.
341 // if there is a selection, but it is not the search string, then
342 // we basically ignore it. (FIXME We ought to replace only within
344 // if there is no selection, then:
345 // (i) if some search string has been provided, then we find it.
346 // (think of how the dialog works when you hit "replace" the
348 // (ii) if no search string has been provided, then we treat the
349 // word the cursor is in as the search string. (why? i have no
350 // idea.) but this only works in text?
352 // returns the number of replacements made (one, if any) and
353 // whether anything at all was done.
354 pair<bool, int> replaceOne(BufferView * bv, docstring searchstr,
355 docstring const & replacestr, bool case_sens,
356 bool whole, bool forward, bool findnext)
358 Cursor & cur = bv->cursor();
360 if (!cur.selection()) {
361 // no selection, non-empty search string: find it
362 if (!searchstr.empty()) {
363 found = findOne(bv, searchstr, case_sens, whole, forward, true, findnext);
364 return make_pair(found, 0);
366 // empty search string
369 return make_pair(false, 0);
370 // select current word and treat it as the search string.
371 // This causes a minor bug as undo will restore this selection,
372 // which the user did not create (#8986).
373 cur.innerText()->selectWord(cur, WHOLE_WORD);
374 searchstr = cur.selectionAsString(false);
377 // if we still don't have a search string, report the error
379 if (!searchAllowed(searchstr))
380 return make_pair(false, 0);
382 bool have_selection = cur.selection();
383 docstring const selected = cur.selectionAsString(false);
386 ? searchstr == selected
387 : compare_no_case(searchstr, selected) == 0;
389 // no selection or current selection is not search word:
390 // just find the search word
391 if (!have_selection || !match) {
392 found = findOne(bv, searchstr, case_sens, whole, forward, true, findnext);
393 return make_pair(found, 0);
396 // we're now actually ready to replace. if the buffer is
397 // read-only, we can't, though.
398 if (bv->buffer().isReadonly())
399 return make_pair(false, 0);
401 cap::replaceSelectionWithString(cur, replacestr);
403 cur.pos() += replacestr.length();
404 LASSERT(cur.pos() <= cur.lastpos(),
405 cur.pos() = cur.lastpos());
408 findOne(bv, searchstr, case_sens, whole, forward, false, findnext);
410 return make_pair(true, 1);
416 docstring const find2string(docstring const & search,
417 bool casesensitive, bool matchword, bool forward)
421 << int(casesensitive) << ' '
422 << int(matchword) << ' '
428 docstring const replace2string(docstring const & replace,
429 docstring const & search,
430 bool casesensitive, bool matchword,
431 bool all, bool forward, bool findnext)
434 ss << replace << '\n'
436 << int(casesensitive) << ' '
437 << int(matchword) << ' '
439 << int(forward) << ' '
445 bool lyxfind(BufferView * bv, FuncRequest const & ev)
447 if (!bv || ev.action() != LFUN_WORD_FIND)
450 //lyxerr << "find called, cmd: " << ev << endl;
452 // data is of the form
454 // <casesensitive> <matchword> <forward>"
456 docstring howto = split(ev.argument(), search, '\n');
458 bool casesensitive = parse_bool(howto);
459 bool matchword = parse_bool(howto);
460 bool forward = parse_bool(howto);
462 return findOne(bv, search, casesensitive, matchword, forward, true, true);
466 bool lyxreplace(BufferView * bv,
467 FuncRequest const & ev, bool has_deleted)
469 if (!bv || ev.action() != LFUN_WORD_REPLACE)
472 // data is of the form
475 // <casesensitive> <matchword> <all> <forward> <findnext>"
478 docstring howto = split(ev.argument(), rplc, '\n');
479 howto = split(howto, search, '\n');
481 bool casesensitive = parse_bool(howto);
482 bool matchword = parse_bool(howto);
483 bool all = parse_bool(howto);
484 bool forward = parse_bool(howto);
485 bool findnext = howto.empty() ? true : parse_bool(howto);
490 int replace_count = 0;
492 replace_count = replaceAll(bv, search, rplc, casesensitive, matchword);
493 update = replace_count > 0;
496 replaceOne(bv, search, rplc, casesensitive, matchword, forward, findnext);
498 replace_count = rv.second;
501 Buffer const & buf = bv->buffer();
503 // emit message signal.
504 buf.message(_("String not found."));
506 if (replace_count == 0) {
507 buf.message(_("String found."));
508 } else if (replace_count == 1) {
509 buf.message(_("String has been replaced."));
511 docstring const str =
512 bformat(_("%1$d strings have been replaced."), replace_count);
516 } else if (findnext) {
517 // if we have deleted characters, we do not replace at all, but
518 // rather search for the next occurence
519 if (findOne(bv, search, casesensitive, matchword, forward, true, findnext))
522 bv->message(_("String not found."));
528 bool findNextChange(BufferView * bv, Cursor & cur, bool const check_wrap)
530 for (; cur; cur.forwardPos())
531 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
535 DocIterator cur_orig(bv->cursor());
536 docstring q = _("End of file reached while searching forward.\n"
537 "Continue searching from the beginning?");
538 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
539 q, 0, 1, _("&Yes"), _("&No"));
540 if (wrap_answer == 0) {
541 bv->cursor().clear();
542 bv->cursor().push_back(CursorSlice(bv->buffer().inset()));
543 bv->clearSelection();
544 cur.setCursor(bv->cursor().selectionBegin());
545 if (findNextChange(bv, cur, false))
548 bv->cursor().setCursor(cur_orig);
555 bool findPreviousChange(BufferView * bv, Cursor & cur, bool const check_wrap)
557 for (cur.backwardPos(); cur; cur.backwardPos()) {
558 if (cur.inTexted() && cur.paragraph().isChanged(cur.pos()))
563 DocIterator cur_orig(bv->cursor());
564 docstring q = _("Beginning of file reached while searching backward.\n"
565 "Continue searching from the end?");
566 int wrap_answer = frontend::Alert::prompt(_("Wrap search?"),
567 q, 0, 1, _("&Yes"), _("&No"));
568 if (wrap_answer == 0) {
569 bv->cursor().setCursor(doc_iterator_end(&bv->buffer()));
570 bv->cursor().backwardPos();
571 bv->clearSelection();
572 cur.setCursor(bv->cursor().selectionBegin());
573 if (findPreviousChange(bv, cur, false))
576 bv->cursor().setCursor(cur_orig);
583 bool selectChange(Cursor & cur, bool forward)
585 if (!cur.inTexted() || !cur.paragraph().isChanged(cur.pos()))
587 Change ch = cur.paragraph().lookupChange(cur.pos());
589 CursorSlice tip1 = cur.top();
590 for (; tip1.pit() < tip1.lastpit() || tip1.pos() < tip1.lastpos(); tip1.forwardPos()) {
591 Change ch2 = tip1.paragraph().lookupChange(tip1.pos());
592 if (!ch2.isSimilarTo(ch))
595 CursorSlice tip2 = cur.top();
596 for (; tip2.pit() > 0 || tip2.pos() > 0;) {
598 Change ch2 = tip2.paragraph().lookupChange(tip2.pos());
599 if (!ch2.isSimilarTo(ch)) {
600 // take a step forward to correctly set the selection
608 cur.bv().mouseSetCursor(cur, false);
610 cur.bv().mouseSetCursor(cur, true);
618 bool findChange(BufferView * bv, bool forward)
621 cur.setCursor(forward ? bv->cursor().selectionEnd()
622 : bv->cursor().selectionBegin());
623 forward ? findNextChange(bv, cur, true) : findPreviousChange(bv, cur, true);
624 return selectChange(cur, forward);
629 bool findNextChange(BufferView * bv)
631 return findChange(bv, true);
635 bool findPreviousChange(BufferView * bv)
637 return findChange(bv, false);
644 typedef vector<pair<string, string> > Escapes;
646 /// A map of symbols and their escaped equivalent needed within a regex.
647 /// @note Beware of order
648 Escapes const & get_regexp_escapes()
650 typedef std::pair<std::string, std::string> P;
652 static Escapes escape_map;
653 if (escape_map.empty()) {
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(".", "_x_."));
664 escape_map.push_back(P("\\", "(?:\\\\|\\\\backslash)"));
665 escape_map.push_back(P("~", "(?:\\\\textasciitilde|\\\\sim)"));
666 escape_map.push_back(P("^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\textasciicircum|\\\\mathcircumflex)"));
667 escape_map.push_back(P("_x_", "\\"));
672 /// A map of lyx escaped strings and their unescaped equivalent.
673 Escapes const & get_lyx_unescapes()
675 typedef std::pair<std::string, std::string> P;
677 static Escapes escape_map;
678 if (escape_map.empty()) {
679 escape_map.push_back(P("\\%", "%"));
680 escape_map.push_back(P("\\mathcircumflex ", "^"));
681 escape_map.push_back(P("\\mathcircumflex", "^"));
682 escape_map.push_back(P("\\backslash ", "\\"));
683 escape_map.push_back(P("\\backslash", "\\"));
684 escape_map.push_back(P("\\\\{", "_x_<"));
685 escape_map.push_back(P("\\\\}", "_x_>"));
686 escape_map.push_back(P("\\sim ", "~"));
687 escape_map.push_back(P("\\sim", "~"));
692 /// A map of escapes turning a regexp matching text to one matching latex.
693 Escapes const & get_regexp_latex_escapes()
695 typedef std::pair<std::string, std::string> P;
697 static Escapes escape_map;
698 if (escape_map.empty()) {
699 escape_map.push_back(P("\\\\", "(?:\\\\\\\\|\\\\backslash|\\\\textbackslash\\{\\}|\\\\textbackslash)"));
700 escape_map.push_back(P("(<?!\\\\\\\\textbackslash)\\{", "\\\\\\{"));
701 escape_map.push_back(P("(<?!\\\\\\\\textbackslash\\\\\\{)\\}", "\\\\\\}"));
702 escape_map.push_back(P("\\[", "\\{\\[\\}"));
703 escape_map.push_back(P("\\]", "\\{\\]\\}"));
704 escape_map.push_back(P("\\^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\textasciicircum|\\\\mathcircumflex)"));
705 escape_map.push_back(P("%", "\\\\\\%"));
706 escape_map.push_back(P("#", "\\\\#"));
711 /** @todo Probably the maps need to be migrated to regexps, in order to distinguish if
712 ** the found occurrence were escaped.
714 string apply_escapes(string s, Escapes const & escape_map)
716 LYXERR(Debug::FIND, "Escaping: '" << s << "'");
717 Escapes::const_iterator it;
718 for (it = escape_map.begin(); it != escape_map.end(); ++it) {
719 // LYXERR(Debug::FIND, "Escaping " << it->first << " as " << it->second);
720 unsigned int pos = 0;
721 while (pos < s.length() && (pos = s.find(it->first, pos)) < s.length()) {
722 s.replace(pos, it->first.length(), it->second);
723 LYXERR(Debug::FIND, "After escape: " << s);
724 pos += it->second.length();
725 // LYXERR(Debug::FIND, "pos: " << pos);
728 LYXERR(Debug::FIND, "Escaped : '" << s << "'");
733 /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
734 /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
735 /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
736 string escape_for_regex(string s, bool match_latex)
739 while (pos < s.size()) {
740 size_t new_pos = s.find("\\regexp{", pos);
741 if (new_pos == string::npos)
745 LYXERR(Debug::FIND, "new_pos: " << new_pos);
746 t = apply_escapes(s.substr(pos, new_pos - pos), get_lyx_unescapes());
747 LYXERR(Debug::FIND, "t [lyx]: " << t);
748 t = apply_escapes(t, get_regexp_escapes());
749 LYXERR(Debug::FIND, "t [rxp]: " << t);
750 s.replace(pos, new_pos - pos, t);
751 new_pos = pos + t.size();
752 LYXERR(Debug::FIND, "Regexp after escaping: " << s);
753 LYXERR(Debug::FIND, "new_pos: " << new_pos);
754 if (new_pos == s.size())
757 // Might fail if \\endregexp{} is preceeded by unexpected stuff (weird escapes)
758 size_t end_pos = s.find("\\endregexp{}}", new_pos + 8);
759 LYXERR(Debug::FIND, "end_pos: " << end_pos);
760 t = s.substr(new_pos + 8, end_pos - (new_pos + 8));
761 LYXERR(Debug::FIND, "t in regexp : " << t);
762 t = apply_escapes(t, get_lyx_unescapes());
763 LYXERR(Debug::FIND, "t in regexp [lyx]: " << t);
765 t = apply_escapes(t, get_regexp_latex_escapes());
766 LYXERR(Debug::FIND, "t in regexp [ltx]: " << t);
768 if (end_pos == s.size()) {
769 s.replace(new_pos, end_pos - new_pos, t);
770 LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s);
773 s.replace(new_pos, end_pos + 13 - new_pos, t);
774 LYXERR(Debug::FIND, "Regexp after \\regexp{...\\endregexp{}} removal: " << s);
775 pos = new_pos + t.size();
776 LYXERR(Debug::FIND, "pos: " << pos);
782 /// Wrapper for lyx::regex_replace with simpler interface
783 bool regex_replace(string const & s, string & t, string const & searchstr,
784 string const & replacestr)
786 lyx::regex e(searchstr, regex_constants::ECMAScript);
788 ostream_iterator<char, char> it(oss);
789 lyx::regex_replace(it, s.begin(), s.end(), e, replacestr);
790 // tolerate t and s be references to the same variable
791 bool rv = (s != oss.str());
797 /** Checks if supplied string segment is well-formed from the standpoint of matching open-closed braces.
799 ** Verify that closed braces exactly match open braces. This avoids that, for example,
800 ** \frac{.*}{x} matches \frac{x+\frac{y}{x}}{z} with .* being 'x+\frac{y'.
803 ** Number of open braces that must remain open at the end for the verification to succeed.
805 bool braces_match(string::const_iterator const & beg,
806 string::const_iterator const & end,
810 string::const_iterator it = beg;
811 LYXERR(Debug::FIND, "Checking " << unmatched << " unmatched braces in '" << string(beg, end) << "'");
812 for (; it != end; ++it) {
813 // Skip escaped braces in the count
818 } else if (*it == '{') {
820 } else if (*it == '}') {
821 if (open_pars == 0) {
822 LYXERR(Debug::FIND, "Found unmatched closed brace");
828 if (open_pars != unmatched) {
829 LYXERR(Debug::FIND, "Found " << open_pars
830 << " instead of " << unmatched
831 << " unmatched open braces at the end of count");
834 LYXERR(Debug::FIND, "Braces match as expected");
844 MatchResult(): match_len(0),match2end(0), pos(0) {};
847 /** The class performing a match between a position in the document and the FindAdvOptions.
849 class MatchStringAdv {
851 MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt);
853 /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv
854 ** constructor as opt.search, under the opt.* options settings.
857 ** If set, then match is searched only against beginning of text starting at cur.
858 ** If unset, then match is searched anywhere in text starting at cur.
861 ** The length of the matching text, or zero if no match was found.
863 MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
868 /// first buffer on which search was started
869 lyx::Buffer * const p_first_buf;
871 FindAndReplaceOptions const & opt;
874 /// Auxiliary find method (does not account for opt.matchword)
875 MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
877 /** Normalize a stringified or latexified LyX paragraph.
881 ** <li>if search is not casesensitive, then lowercase the string;
882 ** <li>remove any newline at begin or end of the string;
883 ** <li>replace any newline in the middle of the string with a simple space;
884 ** <li>remove stale empty styles and environments, like \emph{} and \textbf{}.
887 ** @todo Normalization should also expand macros, if the corresponding
888 ** search option was checked.
890 string normalize(docstring const & s, bool hack_braces) const;
891 // normalized string to search
892 string par_as_string;
893 // regular expression to use for searching
895 // same as regexp, but prefixed with a ".*?"
897 // leading format material as string
898 string lead_as_string;
899 // par_as_string after removal of lead_as_string
900 string par_as_string_nolead;
901 // unmatched open braces in the search string/regexp
903 // number of (.*?) subexpressions added at end of search regexp for closing
904 // environments, math mode, styles, etc...
906 // Are we searching with regular expressions ?
911 static docstring buffer_to_latex(Buffer & buffer)
913 OutputParams runparams(&buffer.params().encoding());
914 odocstringstream ods;
916 runparams.nice = true;
917 runparams.flavor = OutputParams::LATEX;
918 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
919 // No side effect of file copying and image conversion
920 runparams.dryrun = true;
921 runparams.for_search = true;
922 pit_type const endpit = buffer.paragraphs().size();
923 for (pit_type pit = 0; pit != endpit; ++pit) {
924 TeXOnePar(buffer, buffer.text(), pit, os, runparams);
925 LYXERR(Debug::FIND, "searchString up to here: " << ods.str());
931 static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt)
934 if (!opt.ignoreformat) {
935 str = buffer_to_latex(buffer);
937 OutputParams runparams(&buffer.params().encoding());
938 runparams.nice = true;
939 runparams.flavor = OutputParams::LATEX;
940 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
941 runparams.dryrun = true;
942 runparams.for_search = true;
943 for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
944 Paragraph const & par = buffer.paragraphs().at(pit);
945 LYXERR(Debug::FIND, "Adding to search string: '"
946 << par.asString(pos_type(0), par.size(),
947 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
950 str += par.asString(pos_type(0), par.size(),
951 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
959 /// Return separation pos between the leading material and the rest
960 static size_t identifyLeading(string const & s)
963 // @TODO Support \item[text]
964 // Kornel: Added textsl, textsf, textit, texttt and noun
965 // + allow to search for colored text too
966 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)\\*?)\\{", "")
967 || regex_replace(t, t, REGEX_BOS "\\$", "")
968 || regex_replace(t, t, REGEX_BOS "\\\\\\[ ", "")
969 || regex_replace(t, t, REGEX_BOS " ?\\\\item\\{[a-z]+\\}", "")
970 || regex_replace(t, t, REGEX_BOS "\\\\begin\\{[a-zA-Z_]*\\*?\\} ", ""))
972 LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'");
977 * Given a latexified string, retrieve some handled features
978 * The features of the regex will later be compared with the features
979 * of the searched text. If the regex features are not a
980 * subset of the analized, then, in not format ignoring search
981 * we can early stop the search in the relevant inset.
983 typedef map<string, bool> Features;
985 static Features identifyFeatures(string const & s)
987 static regex const feature("\\\\(([a-zA-Z]+(\\{([a-z]+\\*?)\\}|\\*)?))\\{");
988 static regex const valid("^("
990 "(footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge|"
991 "emph|noun|text(bf|md|sl|sf|it|tt)|"
992 "(textcolor|foreignlanguage|item|listitem|latexenvironment)\\{[a-z]+\\*?\\})|"
993 "(u|uu)line|(s|x)out|uwave|"
994 "(sub|extra)?title|author|subject|publishers|dedication|(upper|lower)titleback|lyx(right)?address)|"
995 "((sub)?(((sub)?section)|paragraph)|part|chapter|lyxslide)\\*?)$");
1000 for (sregex_iterator it(s.begin(), s.end(), feature), end; it != end; ++it) {
1003 if (sub.str(1).compare("regexp") == 0) {
1007 string token = sub.str(1);
1009 if (regex_match(token, sub2, valid)) {
1017 if (sub.str(1).compare("endregexp") == 0) {
1027 * defines values features of a key "\\[a-z]+{"
1032 /* Char type with content discarded
1033 * like \hspace{1cm} */
1035 /* Char, like \backslash */
1037 /* \part, \section*, ... */
1039 /* title, author etc */
1041 /* \foreignlanguage{ngerman}, ... */
1044 * to discard language in content */
1047 /* \begin{eqnarray}...\end{eqnarray}, ... $...$ */
1049 /* fonts, colors, markups, ... */
1051 /* footnotesize, ... large, ...
1052 * Ignore all of them */
1055 /* inputencoding, ...
1056 * Discard also content, because they do not help in search */
1059 * like remove, but also all arguments */
1061 /* item, listitem */
1063 /* tex, latex, ... like isChar */
1065 /* like \lettrine[lines=5]{}{} */
1067 /* End of arguments marker for lettrine,
1068 * so that they can be ignored */
1078 parenthesiscount(1),
1082 KeyInfo(KeyType type, int parcount, bool disable)
1088 parenthesiscount(parcount),
1097 int parenthesiscount;
1099 bool used; /* by pattern */
1104 Border(int l=0, int u=0) : low(l), upper(u) {};
1109 #define MAXOPENED 30
1111 bool isPatternString;
1113 explicit Intervall(bool isPattern) :
1114 isPatternString(isPattern),
1116 actualdeptindex(0) { depts[0] = 0; closes[0] = 0;};
1119 int depts[MAXOPENED];
1120 int closes[MAXOPENED];
1121 int actualdeptindex;
1122 Border borders[2*MAXOPENED];
1123 int previousNotIgnored(int);
1124 int nextNotIgnored(int);
1125 void handleOpenP(int i);
1126 void handleCloseP(int i, bool closingAllowed);
1127 void resetOpenedP(int openPos);
1128 void addIntervall(int upper);
1129 void addIntervall(int low, int upper); /* if explicit */
1130 void setForDefaultLang(KeyInfo &defLang);
1131 int findclosing(int start, int end, char up, char down, int repeat);
1132 void handleParentheses(int lastpos, bool closingAllowed);
1134 int isOpeningPar(int pos);
1136 void output(ostringstream &os, int lastpos);
1137 // string show(int lastpos);
1140 int Intervall::isOpeningPar(int pos)
1142 if ((pos < 0) || (size_t(pos) >= par.size()))
1144 if (par[pos] != '{')
1146 if (size_t(pos) + 2 >= par.size())
1148 if (par[pos+2] != '}')
1150 if (par[pos+1] == '[' || par[pos+1] == ']')
1155 void Intervall::setForDefaultLang(KeyInfo &defLang)
1157 // Enable the use of first token again
1158 if (ignoreidx >= 0) {
1159 int value = defLang._tokenstart + defLang._tokensize;
1161 if (borders[0].low < value)
1162 borders[0].low = value;
1163 if (borders[0].upper < value)
1164 borders[0].upper = value;
1169 static void checkDepthIndex(int val)
1171 static int maxdepthidx = MAXOPENED-2;
1172 if (val > maxdepthidx) {
1174 LYXERR0("maxdepthidx now " << val);
1178 static void checkIgnoreIdx(int val)
1180 static int maxignoreidx = 2*MAXOPENED - 4;
1181 if (val > maxignoreidx) {
1183 LYXERR0("maxignoreidx now " << val);
1188 * Expand the region of ignored parts of the input latex string
1189 * The region is only relevant in output()
1191 void Intervall::addIntervall(int low, int upper)
1194 if (low == upper) return;
1195 for (idx = ignoreidx+1; idx > 0; --idx) {
1196 if (low > borders[idx-1].upper) {
1200 Border br(low, upper);
1201 if (idx > ignoreidx) {
1204 checkIgnoreIdx(ignoreidx);
1208 // Expand only if one of the new bound is inside the interwall
1209 // We know here that br.low > borders[idx-1].upper
1210 if (br.upper < borders[idx].low) {
1211 // We have to insert at this pos
1212 for (int i = ignoreidx+1; i > idx; --i) {
1213 borders[i] = borders[i-1];
1217 checkIgnoreIdx(ignoreidx);
1220 // Here we know, that we are overlapping
1221 if (br.low > borders[idx].low)
1222 br.low = borders[idx].low;
1223 // check what has to be concatenated
1225 for (int i = idx; i <= ignoreidx; i++) {
1226 if (br.upper >= borders[i].low) {
1228 if (br.upper < borders[i].upper)
1229 br.upper = borders[i].upper;
1235 // count should be >= 1 here
1238 for (int i = idx + count; i <= ignoreidx; i++) {
1239 borders[i-count+1] = borders[i];
1241 ignoreidx -= count - 1;
1247 void Intervall::handleOpenP(int i)
1250 depts[actualdeptindex] = i+1;
1251 closes[actualdeptindex] = -1;
1252 checkDepthIndex(actualdeptindex);
1255 void Intervall::handleCloseP(int i, bool closingAllowed)
1257 if (actualdeptindex <= 0) {
1258 if (! closingAllowed)
1259 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1260 // if we are at the very end
1261 addIntervall(i, i+1);
1264 closes[actualdeptindex] = i+1;
1269 void Intervall::resetOpenedP(int openPos)
1271 // Used as initializer for foreignlanguage entry
1272 actualdeptindex = 1;
1273 depts[1] = openPos+1;
1277 int Intervall::previousNotIgnored(int start)
1279 int idx = 0; /* int intervalls */
1280 for (idx = ignoreidx; idx >= 0; --idx) {
1281 if (start > borders[idx].upper)
1283 if (start >= borders[idx].low)
1284 start = borders[idx].low-1;
1289 int Intervall::nextNotIgnored(int start)
1291 int idx = 0; /* int intervalls */
1292 for (idx = 0; idx <= ignoreidx; idx++) {
1293 if (start < borders[idx].low)
1295 if (start < borders[idx].upper)
1296 start = borders[idx].upper;
1301 typedef map<string, KeyInfo> KeysMap;
1302 typedef vector< KeyInfo> Entries;
1303 static KeysMap keys = map<string, KeyInfo>();
1310 void buildKeys(bool);
1311 void buildEntries(bool);
1312 void makeKey(const string &, KeyInfo, bool isPatternString);
1313 void processRegion(int start, int region_end); /* remove {} parts */
1314 void removeHead(KeyInfo&, int count=0);
1317 LatexInfo(string par, bool isPatternString) : entidx(-1), interval(isPatternString) {
1319 interval.hasTitle = false;
1320 interval.titleValue = "";
1321 buildKeys(isPatternString);
1322 entries = vector<KeyInfo>();
1323 buildEntries(isPatternString);
1327 if (entries.empty()) {
1330 if (entries[0].keytype == KeyInfo::isTitle) {
1331 if (! entries[0].disabled) {
1332 interval.hasTitle = true;
1333 interval.titleValue = entries[0].head;
1336 interval.hasTitle = false;
1337 interval.titleValue = "";
1339 removeHead(entries[0]);
1340 if (entries.size() > 1)
1349 if (int(entries.size()) > entidx) {
1356 bool setNextKey(int idx) {
1357 if ((idx == entidx) && (entidx >= 0)) {
1364 int find(int start, KeyInfo::KeyType keytype) {
1368 while (tmpIdx < int(entries.size())) {
1369 if (entries[tmpIdx].keytype == keytype)
1375 int process(ostringstream &os, KeyInfo &actual);
1376 int dispatch(ostringstream &os, int previousStart, KeyInfo &actual);
1377 // string show(int lastpos) { return interval.show(lastpos);};
1378 int nextNotIgnored(int start) { return interval.nextNotIgnored(start);};
1379 KeyInfo &getKeyInfo(int keyinfo) {
1380 static KeyInfo invalidInfo = KeyInfo();
1381 if ((keyinfo < 0) || ( keyinfo >= int(entries.size())))
1384 return entries[keyinfo];
1386 void setForDefaultLang(KeyInfo &defLang) {interval.setForDefaultLang(defLang);};
1387 void addIntervall(int low, int up) { interval.addIntervall(low, up); };
1391 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
1396 for (int i = start; i < end; i += 1 + skip) {
1400 if (c == '\\') skip = 1;
1404 else if (c == down) {
1406 if ((repeat <= 0) || (par[i+1] != up))
1424 vector<MathEntry> entries;
1429 void insert(string wait, size_t start, size_t end) {
1430 MathEntry m = MathEntry();
1432 m.mathStart = start;
1434 m.mathSize = end - start;
1435 entries.push_back(m);
1437 bool empty() { return entries.empty(); };
1438 size_t getEndPos() {
1439 if (entries.empty() || (actualIdx >= entries.size())) {
1442 return entries[actualIdx].mathEnd;
1444 size_t getStartPos() {
1445 if (entries.empty() || (actualIdx >= entries.size())) {
1446 return 100000; /* definitely enough? */
1448 return entries[actualIdx].mathStart;
1450 size_t getFirstPos() {
1452 return getStartPos();
1455 if (entries.empty() || (actualIdx >= entries.size())) {
1458 return entries[actualIdx].mathSize;
1460 void incrEntry() { actualIdx++; };
1463 void LatexInfo::buildEntries(bool isPatternString)
1465 static regex const rmath("\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|alignat)\\*?)\\}");
1466 static regex const rkeys("\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?))");
1467 static bool disableLanguageOverride = false;
1468 smatch sub, submath;
1469 bool evaluatingRegexp = false;
1471 bool evaluatingMath = false;
1472 bool evaluatingCode = false;
1474 bool evaluatingOptional = false;
1475 size_t optionalEnd = 0;
1478 bool math_end_waiting = false;
1479 size_t math_pos = 10000;
1482 for (sregex_iterator itmath(interval.par.begin(), interval.par.end(), rmath), end; itmath != end; ++itmath) {
1484 if (math_end_waiting) {
1485 size_t pos = submath.position(size_t(0));
1486 if ((math_end == "$") &&
1487 (submath.str(0) == "$") &&
1488 (interval.par[pos-1] != '\\')) {
1489 mi.insert("$", math_pos, pos + 1);
1490 math_end_waiting = false;
1492 else if ((math_end == "\\]") &&
1493 (submath.str(0) == "\\]")) {
1494 mi.insert("\\]", math_pos, pos + 2);
1495 math_end_waiting = false;
1497 else if ((submath.str(1).compare("end") == 0) &&
1498 (submath.str(2).compare(math_end) == 0)) {
1499 mi.insert(math_end, math_pos, pos + submath.str(0).length());
1500 math_end_waiting = false;
1506 if (submath.str(1).compare("begin") == 0) {
1507 math_end_waiting = true;
1508 math_end = submath.str(2);
1509 math_pos = submath.position(size_t(0));
1511 else if (submath.str(0).compare("\\[") == 0) {
1512 math_end_waiting = true;
1514 math_pos = submath.position(size_t(0));
1516 else if (submath.str(0) == "$") {
1517 size_t pos = submath.position(size_t(0));
1518 if ((pos == 0) || (interval.par[pos-1] != '\\')) {
1519 math_end_waiting = true;
1526 // Ignore language if there is math somewhere in pattern-string
1527 if (isPatternString) {
1530 keys["foreignlanguage"].disabled = true;
1531 disableLanguageOverride = true;
1534 disableLanguageOverride = false;
1537 if (disableLanguageOverride) {
1538 keys["foreignlanguage"].disabled = true;
1541 math_pos = mi.getFirstPos();
1542 for (sregex_iterator it(interval.par.begin(), interval.par.end(), rkeys), end; it != end; ++it) {
1544 string key = sub.str(3);
1546 if (sub.str(0)[0] == '\\')
1547 key = sub.str(0)[1];
1551 size_t k_pos = sub.position(size_t(0));
1552 if ((k_pos > 0) && (interval.par[k_pos - 1] == '\\')) {
1553 // Escaped '$', ignoring
1559 if (evaluatingRegexp) {
1560 if (sub.str(1).compare("endregexp") == 0) {
1561 evaluatingRegexp = false;
1562 // found._tokenstart already set
1563 found._dataEnd = sub.position(size_t(0)) + 13;
1564 found._dataStart = found._dataEnd;
1565 found._tokensize = found._dataEnd - found._tokenstart;
1566 found.parenthesiscount = 0;
1567 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1573 if (evaluatingMath) {
1574 if (size_t(sub.position(size_t(0))) < mi.getEndPos())
1576 evaluatingMath = false;
1578 math_pos = mi.getStartPos();
1580 if (keys.find(key) == keys.end()) {
1581 found = KeyInfo(KeyInfo::isStandard, 0, true);
1582 if (isPatternString) {
1583 found.keytype = KeyInfo::isChar;
1584 found.disabled = false;
1591 if (key.compare("regexp") == 0) {
1592 evaluatingRegexp = true;
1593 found._tokenstart = sub.position(size_t(0));
1594 found._tokensize = 0;
1598 // Handle the other params of key
1599 if (found.keytype == KeyInfo::isIgnored)
1601 else if (found.keytype == KeyInfo::isMath) {
1602 if (size_t(sub.position(size_t(0))) == math_pos) {
1604 found._tokenstart = sub.position(size_t(0));
1605 found._tokensize = mi.getSize();
1606 found._dataEnd = found._tokenstart + found._tokensize;
1607 found._dataStart = found._dataEnd;
1608 found.parenthesiscount = 0;
1609 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1610 evaluatingMath = true;
1613 // begin|end of unknown env, discard
1614 // First handle tables
1615 // longtable|tabular
1616 bool discardComment;
1618 found.keytype = KeyInfo::doRemove;
1619 if ((sub.str(5).compare("longtable") == 0) ||
1620 (sub.str(5).compare("tabular") == 0)) {
1621 discardComment = true; /* '%' */
1624 discardComment = false;
1625 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
1627 string token = sub.str(5);
1628 if (regex_match(token, sub2, removeArgs)) {
1629 found.keytype = KeyInfo::removeWithArg;
1632 // discard spaces before pos(0)
1633 int pos = sub.position(size_t(0));
1635 for (count = 0; pos - count > 0; count++) {
1636 char c = interval.par[pos-count-1];
1637 if (discardComment) {
1638 if ((c != ' ') && (c != '%'))
1644 found._tokenstart = pos - count;
1645 if (sub.str(1).compare(0, 5, "begin") == 0) {
1646 size_t pos1 = pos + sub.str(0).length();
1647 if (sub.str(5).compare("cjk") == 0) {
1648 pos1 = interval.findclosing(pos1+1, interval.par.length()) + 1;
1649 if ((interval.par[pos1] == '{') && (interval.par[pos1+1] == '}'))
1651 found.keytype = KeyInfo::isMain;
1652 found._dataStart = pos1;
1653 found._dataEnd = interval.par.length();
1654 found.disabled = keys["foreignlanguage"].disabled;
1655 found.used = keys["foreignlanguage"].used;
1656 found._tokensize = pos1 - found._tokenstart;
1657 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1660 // Swallow possible optional params
1661 while (interval.par[pos1] == '[') {
1662 pos1 = interval.findclosing(pos1+1, interval.par.length(), '[', ']')+1;
1664 // Swallow also the eventual parameter
1665 if (interval.par[pos1] == '{') {
1666 found._dataEnd = interval.findclosing(pos1+1, interval.par.length()) + 1;
1669 found._dataEnd = pos1;
1671 found._dataStart = found._dataEnd;
1672 found._tokensize = count + found._dataEnd - pos;
1673 found.parenthesiscount = 0;
1674 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1675 found.disabled = true;
1679 // Handle "\end{...}"
1680 found._dataStart = pos + sub.str(0).length();
1681 found._dataEnd = found._dataStart;
1682 found._tokensize = count + found._dataEnd - pos;
1683 found.parenthesiscount = 0;
1684 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1685 found.disabled = true;
1689 else if (found.keytype != KeyInfo::isRegex) {
1690 found._tokenstart = sub.position(size_t(0));
1691 if (found.parenthesiscount == 0) {
1692 // Probably to be discarded
1693 size_t following_pos = sub.position(size_t(0)) + sub.str(3).length() + 1;
1694 char following = interval.par[following_pos];
1695 if (following == ' ')
1696 found.head = "\\" + sub.str(3) + " ";
1697 else if (following == '=') {
1698 // like \uldepth=1000pt
1699 found.head = sub.str(0);
1702 found.head = "\\" + key;
1703 found._tokensize = found.head.length();
1704 found._dataEnd = found._tokenstart + found._tokensize;
1705 found._dataStart = found._dataEnd;
1708 int params = found._tokenstart + key.length() + 1;
1709 if (evaluatingOptional) {
1710 if (size_t(found._tokenstart) > optionalEnd) {
1711 evaluatingOptional = false;
1714 found.disabled = true;
1717 int optend = params;
1718 while (interval.par[optend] == '[') {
1719 // discard optional parameters
1720 optend = interval.findclosing(optend+1, interval.par.length(), '[', ']') + 1;
1722 if (optend > params) {
1723 key += interval.par.substr(params, optend-params);
1724 evaluatingOptional = true;
1725 optionalEnd = optend;
1727 string token = sub.str(5);
1728 int closings = found.parenthesiscount;
1729 if (found.parenthesiscount == 1) {
1730 found.head = "\\" + key + "{";
1732 else if (found.parenthesiscount > 1) {
1734 found.head = sub.str(0) + "{";
1735 closings = found.parenthesiscount - 1;
1738 found.head = "\\" + key + "{";
1741 found._tokensize = found.head.length();
1742 found._dataStart = found._tokenstart + found.head.length();
1743 if (interval.par.substr(found._dataStart-1, 15).compare("\\endarguments{}") == 0) {
1744 found._dataStart += 15;
1746 size_t endpos = interval.findclosing(found._dataStart, interval.par.length(), '{', '}', closings);
1747 if (found.keytype == KeyInfo::isList) {
1748 // Check if it really is list env
1749 static regex const listre("^([a-z]+)$");
1751 if (!regex_match(token, sub2, listre)) {
1752 // Change the key of this entry. It is not in a list/item environment
1753 found.keytype = KeyInfo::endArguments;
1756 if (found.keytype == KeyInfo::noMain) {
1757 evaluatingCode = true;
1759 codeStart = found._dataStart;
1761 else if (evaluatingCode) {
1762 if (size_t(found._dataStart) > codeEnd)
1763 evaluatingCode = false;
1764 else if (found.keytype == KeyInfo::isMain) {
1765 // Disable this key, treate it as standard
1766 found.keytype = KeyInfo::isStandard;
1767 found.disabled = true;
1768 if ((codeEnd == interval.par.length()) &&
1769 (found._tokenstart == codeStart)) {
1770 // trickery, because the code inset starts
1771 // with \selectlanguage ...
1773 if (entries.size() > 1) {
1774 entries[entries.size()-1]._dataEnd = codeEnd;
1779 if ((endpos == interval.par.length()) &&
1780 (found.keytype == KeyInfo::doRemove)) {
1781 // Missing closing => error in latex-input?
1782 // therefore do not delete remaining data
1783 found._dataStart -= 1;
1784 found._dataEnd = found._dataStart;
1787 found._dataEnd = endpos;
1789 if (isPatternString) {
1790 keys[key].used = true;
1793 entries.push_back(found);
1797 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
1799 stringstream s(keysstring);
1801 const char delim = '|';
1802 while (getline(s, key, delim)) {
1803 KeyInfo keyII(keyI);
1804 if (isPatternString) {
1807 else if ( !keys[key].used)
1808 keyII.disabled = true;
1813 void LatexInfo::buildKeys(bool isPatternString)
1816 static bool keysBuilt = false;
1817 if (keysBuilt && !isPatternString) return;
1819 // Known standard keys with 1 parameter.
1820 // Split is done, if not at start of region
1821 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
1822 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
1823 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
1824 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
1825 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
1826 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
1828 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
1829 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1830 makeKey("section*|subsection*|subsubsection*|paragraph*",
1831 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1832 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1833 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
1835 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
1837 // Split is done, if not at start of region
1838 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
1839 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
1841 // Split is done always.
1842 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
1846 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1847 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1848 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1849 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1851 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1852 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1853 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1855 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1856 // Custom space/skip, remove the content (== length value)
1857 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
1858 // Found in fr/UserGuide.lyx
1859 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1861 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1862 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1863 // Known macros to remove (including their parameter)
1865 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
1866 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
1867 // handle like standard keys with 1 parameter.
1868 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
1870 // Macros to remove, but let the parameter survive
1872 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1874 // Remove language spec from content of these insets
1875 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
1877 // Same effect as previous, parameter will survive (because there is no one anyway)
1879 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1880 // Remove table decorations
1881 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
1882 // Discard shape-header.
1883 // For footnote or shortcut too, because of lang settings
1884 // and wrong handling if used 'KeyInfo::noMain'
1885 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1886 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1887 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1888 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1889 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1890 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
1891 // like ('tiny{}' or '\tiny ' ... )
1892 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
1894 // Survives, like known character
1895 makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1896 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
1898 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1899 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1900 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1902 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1903 // Remove RTL/LTR marker
1904 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1905 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
1906 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
1907 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
1908 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
1909 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
1910 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1911 if (isPatternString) {
1912 // Allow the first searched string to rebuild the keys too
1916 // no need to rebuild again
1922 * Keep the list of actual opened parentheses actual
1923 * (e.g. depth == 4 means there are 4 '{' not processed yet)
1925 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
1928 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
1932 if (c == '\\') skip = 1;
1933 else if (c == '{') {
1936 else if (c == '}') {
1937 handleCloseP(i, closingAllowed);
1943 string Intervall::show(int lastpos)
1945 int idx = 0; /* int intervalls */
1948 for (idx = 0; idx <= ignoreidx; idx++) {
1949 while (i < lastpos) {
1951 if (i <= borders[idx].low) {
1952 if (borders[idx].low > lastpos)
1953 printsize = lastpos - i;
1955 printsize = borders[idx].low - i;
1956 s += par.substr(i, printsize);
1958 if (i >= borders[idx].low)
1959 i = borders[idx].upper;
1962 i = borders[idx].upper;
1968 s += par.substr(i, lastpos-i);
1974 void Intervall::output(ostringstream &os, int lastpos)
1976 // get number of chars to output
1977 int idx = 0; /* int intervalls */
1980 string startTitle = titleValue;
1981 for (idx = 0; idx <= ignoreidx; idx++) {
1983 if (i <= borders[idx].low) {
1985 if (borders[idx].low > lastpos)
1986 printsize = lastpos - i;
1988 printsize = borders[idx].low - i;
1989 if (printsize > 0) {
1990 os << startTitle << par.substr(i, printsize);
1992 printed += printsize;
1995 handleParentheses(i, false);
1996 if (i >= borders[idx].low)
1997 i = borders[idx].upper;
2000 i = borders[idx].upper;
2007 os << startTitle << par.substr(i, lastpos-i);
2008 printed += lastpos-i;
2010 handleParentheses(lastpos, false);
2011 for (int i = actualdeptindex; i > 0; --i) {
2014 if (hasTitle && (printed > 0))
2016 if (! isPatternString)
2018 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2021 void LatexInfo::processRegion(int start, int region_end)
2023 while (start < region_end) { /* Let {[} and {]} survive */
2024 int cnt = interval.isOpeningPar(start);
2026 // Closing is allowed past the region
2027 int closing = interval.findclosing(start+1, interval.par.length());
2028 interval.addIntervall(start, start+1);
2029 interval.addIntervall(closing, closing+1);
2033 start = interval.nextNotIgnored(start+1);
2037 void LatexInfo::removeHead(KeyInfo &actual, int count)
2039 if (actual.parenthesiscount == 0) {
2040 // "{\tiny{} ...}" ==> "{{} ...}"
2041 interval.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2044 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2045 interval.addIntervall(actual._tokenstart - count, actual._dataStart);
2046 interval.addIntervall(actual._dataEnd, actual._dataEnd+1);
2050 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2053 switch (actual.keytype)
2055 case KeyInfo::isTitle: {
2057 nextKeyIdx = getNextKey();
2060 case KeyInfo::cleanToStart: {
2061 actual._dataEnd = actual._dataStart;
2062 nextKeyIdx = getNextKey();
2063 // Search for end of arguments
2064 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2066 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2067 entries[i].disabled = true;
2069 actual._dataEnd = entries[tmpIdx]._dataEnd;
2071 while (interval.par[actual._dataEnd] == ' ')
2073 interval.addIntervall(0, actual._dataEnd+1);
2074 interval.actualdeptindex = 0;
2075 interval.depts[0] = actual._dataEnd+1;
2076 interval.closes[0] = -1;
2079 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2080 if (actual.disabled)
2081 interval.addIntervall(actual._tokenstart, actual._dataEnd);
2083 interval.addIntervall(actual._dataStart, actual._dataEnd);
2086 case KeyInfo::isChar: {
2087 nextKeyIdx = getNextKey();
2090 case KeyInfo::isSize: {
2091 if (actual.disabled || (interval.par[actual._dataStart] != '{') || (interval.par[actual._dataStart-1] == ' ')) {
2092 processRegion(actual._dataEnd, actual._dataEnd+1); /* remove possibly following {} */
2093 interval.addIntervall(actual._tokenstart, actual._dataEnd+1);
2094 nextKeyIdx = getNextKey();
2096 // Here _dataStart points to '{', so correct it
2097 actual._dataStart += 1;
2098 actual._tokensize += 1;
2099 actual.parenthesiscount = 1;
2100 if (interval.par[actual._dataStart] == '}') {
2101 // Determine the end if used like '{\tiny{}...}'
2102 actual._dataEnd = interval.findclosing(actual._dataStart+1, interval.par.length()) + 1;
2103 interval.addIntervall(actual._dataStart, actual._dataStart+1);
2106 // Determine the end if used like '\tiny{...}'
2107 actual._dataEnd = interval.findclosing(actual._dataStart, interval.par.length()) + 1;
2109 // Split on this key if not at start
2110 int start = interval.nextNotIgnored(previousStart);
2111 if (start < actual._tokenstart) {
2112 interval.output(os, actual._tokenstart);
2113 interval.addIntervall(start, actual._tokenstart);
2115 // discard entry if at end of actual
2116 nextKeyIdx = process(os, actual);
2120 case KeyInfo::endArguments:
2121 // Remove trailing '{}' too
2122 actual._dataStart += 1;
2123 actual._dataEnd += 1;
2124 interval.addIntervall(actual._tokenstart, actual._dataEnd+1);
2125 nextKeyIdx = getNextKey();
2127 case KeyInfo::noMain:
2129 case KeyInfo::isStandard: {
2130 if (actual.disabled) {
2132 processRegion(actual._dataStart, actual._dataStart+1);
2133 nextKeyIdx = getNextKey();
2135 // Split on this key if not at datastart of calling entry
2136 int start = interval.nextNotIgnored(previousStart);
2137 if (start < actual._tokenstart) {
2138 interval.output(os, actual._tokenstart);
2139 interval.addIntervall(start, actual._tokenstart);
2141 // discard entry if at end of actual
2142 nextKeyIdx = process(os, actual);
2146 case KeyInfo::removeWithArg: {
2147 nextKeyIdx = getNextKey();
2148 // Search for end of arguments
2149 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2151 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2152 entries[i].disabled = true;
2154 actual._dataEnd = entries[tmpIdx]._dataEnd;
2156 interval.addIntervall(actual._tokenstart, actual._dataEnd+1);
2159 case KeyInfo::doRemove: {
2160 // Remove the key with all parameters and following spaces
2162 for (pos = actual._dataEnd+1; pos < interval.par.length(); pos++) {
2163 if ((interval.par[pos] != ' ') && (interval.par[pos] != '%'))
2166 interval.addIntervall(actual._tokenstart, pos);
2167 nextKeyIdx = getNextKey();
2170 case KeyInfo::isList: {
2171 // Discard space before _tokenstart
2173 for (count = 0; count < actual._tokenstart; count++) {
2174 if (interval.par[actual._tokenstart-count-1] != ' ')
2177 nextKeyIdx = getNextKey();
2178 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2180 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2182 // How else can we catch this one?
2183 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2184 entries[i].disabled = true;
2186 actual._dataEnd = entries[tmpIdx]._dataEnd;
2188 else if (nextKeyIdx > 0) {
2189 // Ignore any lang entries inside data region
2190 for (int i = nextKeyIdx; i < int(entries.size()) && entries[i]._tokenstart < actual._dataEnd; i++) {
2191 if (entries[i].keytype == KeyInfo::isMain)
2192 entries[i].disabled = true;
2195 if (actual.disabled) {
2196 interval.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2199 interval.addIntervall(actual._tokenstart-count, actual._tokenstart);
2201 if (interval.par[actual._dataEnd+1] == '[') {
2202 int posdown = interval.findclosing(actual._dataEnd+2, interval.par.length(), '[', ']');
2203 if ((interval.par[actual._dataEnd+2] == '{') &&
2204 (interval.par[posdown-1] == '}')) {
2205 interval.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
2206 interval.addIntervall(posdown-1, posdown+1);
2209 interval.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
2210 interval.addIntervall(posdown, posdown+1);
2212 int blk = interval.nextNotIgnored(actual._dataEnd+1);
2213 if (blk > posdown) {
2214 // Discard at most 1 space after empty item
2216 for (count = 0; count < 1; count++) {
2217 if (interval.par[blk+count] != ' ')
2221 interval.addIntervall(blk, blk+count);
2226 case KeyInfo::isSectioning: {
2227 // Discard spaces before _tokenstart
2229 int val = actual._tokenstart;
2230 for (count = 0; count < actual._tokenstart;) {
2231 val = interval.previousNotIgnored(val-1);
2232 if (interval.par[val] != ' ')
2235 count = actual._tokenstart - val;
2238 if (actual.disabled) {
2239 removeHead(actual, count);
2240 nextKeyIdx = getNextKey();
2242 interval.addIntervall(actual._tokenstart-count, actual._tokenstart);
2243 nextKeyIdx = process(os, actual);
2247 case KeyInfo::isMath: {
2248 // Same as regex, use the content unchanged
2249 nextKeyIdx = getNextKey();
2252 case KeyInfo::isRegex: {
2253 // DO NOT SPLIT ON REGEX
2255 nextKeyIdx = getNextKey();
2258 case KeyInfo::isIgnored: {
2259 // Treat like a character for now
2260 nextKeyIdx = getNextKey();
2263 case KeyInfo::isMain: {
2264 if (interval.par.substr(actual._dataStart, 2) == "% ")
2265 interval.addIntervall(actual._dataStart, actual._dataStart+2);
2266 if (actual._tokenstart > 0) {
2267 int prev = interval.previousNotIgnored(actual._tokenstart - 1);
2268 if ((prev >= 0) && interval.par[prev] == '%')
2269 interval.addIntervall(prev, prev+1);
2271 if (actual.disabled) {
2273 if ((interval.par.substr(actual._dataStart, 3) == " \\[") ||
2274 (interval.par.substr(actual._dataStart, 8) == " \\begin{")) {
2275 // Discard also the space before math-equation
2276 interval.addIntervall(actual._dataStart, actual._dataStart+1);
2278 nextKeyIdx = getNextKey();
2279 // interval.resetOpenedP(actual._dataStart-1);
2282 if (actual._tokenstart < 26) {
2283 // for the first (and maybe dummy) language
2284 interval.setForDefaultLang(actual);
2286 interval.resetOpenedP(actual._dataStart-1);
2290 case KeyInfo::invalid:
2291 // This cannot happen, already handled
2294 // LYXERR0("Unhandled keytype");
2295 nextKeyIdx = getNextKey();
2302 int LatexInfo::process(ostringstream &os, KeyInfo &actual )
2304 int end = interval.nextNotIgnored(actual._dataEnd);
2305 int oldStart = actual._dataStart;
2306 int nextKeyIdx = getNextKey();
2308 if ((nextKeyIdx < 0) ||
2309 (entries[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
2310 (entries[nextKeyIdx].keytype == KeyInfo::invalid)) {
2311 if (oldStart <= end) {
2312 processRegion(oldStart, end);
2317 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
2319 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
2320 (void) dispatch(os, actual._dataStart, nextKey);
2321 end = nextKey._tokenstart;
2324 processRegion(oldStart, nextKey._tokenstart);
2325 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
2327 oldStart = nextKey._dataEnd+1;
2329 // now nextKey is either invalid or is outside of actual._dataEnd
2330 // output the remaining and discard myself
2331 if (oldStart <= end) {
2332 processRegion(oldStart, end);
2334 if (interval.par[end] == '}') {
2336 // This is the normal case.
2337 // But if using the firstlanguage, the closing may be missing
2339 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
2341 if (actual._dataEnd < end)
2342 output_end = interval.nextNotIgnored(actual._dataEnd);
2344 output_end = interval.nextNotIgnored(end);
2345 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
2346 interval.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
2348 // Remove possible empty data
2349 int dstart = interval.nextNotIgnored(actual._dataStart);
2350 while (interval.isOpeningPar(dstart) == 1) {
2351 interval.addIntervall(dstart, dstart+1);
2352 int dend = interval.findclosing(dstart+1, output_end);
2353 interval.addIntervall(dend, dend+1);
2354 dstart = interval.nextNotIgnored(dstart+1);
2356 if (dstart < output_end)
2357 interval.output(os, output_end);
2358 interval.addIntervall(actual._tokenstart, end);
2362 string splitOnKnownMacros(string par, bool isPatternString)
2365 LatexInfo li(par, isPatternString);
2366 // LYXERR0("Berfore split: " << par);
2367 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
2369 DummyKey._tokensize = 0;
2370 DummyKey._dataStart = 0;
2371 DummyKey._dataEnd = par.length();
2372 DummyKey.disabled = true;
2373 int firstkeyIdx = li.getFirstKey();
2375 if (firstkeyIdx >= 0) {
2376 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
2377 DummyKey._tokenstart = firstKey._tokenstart;
2379 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
2380 // Use dummy firstKey
2381 firstKey = DummyKey;
2382 (void) li.setNextKey(firstkeyIdx);
2385 if (par.substr(firstKey._dataStart, 2) == "% ")
2386 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
2388 nextkeyIdx = li.process(os, firstKey);
2389 while (nextkeyIdx >= 0) {
2390 // Check for a possible gap between the last
2391 // entry and this one
2392 int datastart = li.nextNotIgnored(firstKey._dataStart);
2393 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
2394 if ((nextKey._tokenstart > datastart)) {
2396 firstKey._dataStart = datastart;
2397 firstKey._dataEnd = par.length();
2398 (void) li.setNextKey(nextkeyIdx);
2399 // Fake the last opened parenthesis
2400 li.setForDefaultLang(firstKey);
2401 nextkeyIdx = li.process(os, firstKey);
2404 if (nextKey.keytype != KeyInfo::isMain) {
2405 firstKey._dataStart = datastart;
2406 firstKey._dataEnd = nextKey._dataEnd+1;
2407 (void) li.setNextKey(nextkeyIdx);
2408 li.setForDefaultLang(firstKey);
2409 nextkeyIdx = li.process(os, firstKey);
2412 nextkeyIdx = li.process(os, nextKey);
2416 // Handle the remaining
2417 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
2418 firstKey._dataEnd = par.length();
2420 if ((firstKey._dataStart < firstKey._dataEnd) &&
2421 (par[firstKey._dataStart] != '}')) {
2422 li.setForDefaultLang(firstKey);
2423 (void) li.process(os, firstKey);
2427 // return string definitelly impossible to match
2428 s = "\\foreignlanguage{ignore}{ }";
2432 s = par; /* no known macros found */
2433 // LYXERR0("After split: " << s);
2438 * Try to unify the language specs in the latexified text.
2439 * Resulting modified string is set to "", if
2440 * the searched tex does not contain all the features in the search pattern
2442 static string correctlanguagesetting(string par, bool isPatternString, bool withformat)
2444 static Features regex_f;
2445 static int missed = 0;
2446 static bool regex_with_format = false;
2448 int parlen = par.length();
2450 while ((parlen > 0) && (par[parlen-1] == '\n')) {
2453 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
2454 // Happens to be there in case of description or labeling environment
2459 // Split the latex input into pieces which
2460 // can be digested by our search engine
2461 LYXERR(Debug::FIND, "input: \"" << par << "\"");
2462 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
2463 LYXERR(Debug::FIND, "After split: \"" << result << "\"");
2466 result = par.substr(0, parlen);
2467 if (isPatternString) {
2470 regex_f = identifyFeatures(result);
2471 string features = "";
2472 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2473 string a = it->first;
2474 regex_with_format = true;
2475 features += " " + a;
2476 // LYXERR0("Identified regex format:" << a);
2478 LYXERR(Debug::FIND, "Identified Features" << features);
2481 } else if (regex_with_format) {
2482 Features info = identifyFeatures(result);
2483 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2484 string a = it->first;
2485 bool b = it->second;
2486 if (b && ! info[a]) {
2488 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
2494 // LYXERR0("No regex formats");
2500 // Remove trailing closure of math, macros and environments, so to catch parts of them.
2501 static int identifyClosing(string & t)
2503 int open_braces = 0;
2505 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
2506 if (regex_replace(t, t, "(.*[^\\\\])\\$" REGEX_EOS, "$1"))
2508 if (regex_replace(t, t, "(.*[^\\\\]) \\\\\\]" REGEX_EOS, "$1"))
2510 if (regex_replace(t, t, "(.*[^\\\\]) \\\\end\\{[a-zA-Z_]*\\*?\\}" REGEX_EOS, "$1"))
2512 if (regex_replace(t, t, "(.*[^\\\\])\\}" REGEX_EOS, "$1")) {
2522 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt)
2523 : p_buf(&buf), p_first_buf(&buf), opt(opt)
2525 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
2526 docstring const & ds = stringifySearchBuffer(find_buf, opt);
2527 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
2528 // When using regexp, braces are hacked already by escape_for_regex()
2529 par_as_string = normalize(ds, !use_regexp);
2531 close_wildcards = 0;
2533 size_t lead_size = 0;
2534 // correct the language settings
2535 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat);
2536 if (opt.ignoreformat) {
2538 // if par_as_string_nolead were emty,
2539 // the following call to findAux will always *find* the string
2540 // in the checked data, and thus always using the slow
2541 // examining of the current text part.
2542 par_as_string_nolead = par_as_string;
2545 lead_size = identifyLeading(par_as_string);
2546 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
2547 lead_as_string = par_as_string.substr(0, lead_size);
2548 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
2552 open_braces = identifyClosing(par_as_string);
2553 identifyClosing(par_as_string_nolead);
2554 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2555 LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'");
2557 string lead_as_regexp;
2558 if (lead_size > 0) {
2559 // @todo No need to search for \regexp{} insets in leading material
2560 lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat);
2561 par_as_string = par_as_string_nolead;
2562 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
2563 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2565 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
2566 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
2567 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2569 // Insert .* before trailing '\$' ('$' has been escaped by escape_for_regex)
2570 regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\$)\\'", "$1(.*?)$2")
2571 // Insert .* before trailing '\\\]' ('\]' has been escaped by escape_for_regex)
2572 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])( \\\\\\\\\\\\\\])\\'", "$1(.*?)$2")
2573 // Insert .* before trailing '\\end\{...}' ('\end{...}' has been escaped by escape_for_regex)
2574 || regex_replace(par_as_string, par_as_string,
2575 "(.*[^\\\\])( \\\\\\\\end\\\\\\{[a-zA-Z_]*)(\\\\\\*)?(\\\\\\})\\'", "$1(.*?)$2$3$4")
2576 // Insert .* before trailing '\}' ('}' has been escaped by escape_for_regex)
2577 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\})\\'", "$1(.*?)$2")
2581 if (!opt.ignoreformat) {
2582 // Remove extra '\}' at end if not part of \{\.\}
2583 size_t lng = par_as_string.size();
2585 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
2587 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
2596 if (lng < par_as_string.size())
2597 par_as_string = par_as_string.substr(0,lng);
2600 regex_replace(par_as_string, par_as_string, "\\\\\\.", "_xxbdotxx_");
2601 // handle '.' -> '[^]', replace later as '[^\}\{\\]'
2602 regex_replace(par_as_string, par_as_string, "\\.", "[^]");
2603 // replace '[^...]' with '[^...\}\{\\]'
2604 regex_replace(par_as_string, par_as_string, "\\[\\^([^\\\\\\]]*)\\]", "_xxbrlxx_$1\\}\\{\\\\_xxbrrxx_");
2605 regex_replace(par_as_string, par_as_string, "_xxbrlxx_", "[^");
2606 regex_replace(par_as_string, par_as_string, "_xxbrrxx_", "]");
2608 regex_replace(par_as_string, par_as_string, "_xxbdotxx_", "\\.");
2611 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2612 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2613 LYXERR(Debug::FIND, "Close .*? : " << close_wildcards);
2614 LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
2616 // If entered regexp must match at begin of searched string buffer
2617 // Kornel: Added parentheses to use $1 for size of the leading string
2621 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
2622 // Unfortunately is '\1', '\2', etc not working for strings with extra format
2623 // so the convert has no effect in that case
2624 for (int i = 8; i > 0; --i) {
2625 string orig = "\\\\" + std::to_string(i);
2626 string dest = "\\" + std::to_string(i+1);
2627 while (regex_replace(par_as_string, par_as_string, orig, dest));
2629 regexp_str = "(" + lead_as_regexp + ")" + par_as_string;
2630 regexp2_str = "(" + lead_as_regexp + ").*?" + par_as_string;
2632 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
2633 regexp = lyx::regex(regexp_str);
2635 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
2636 regexp2 = lyx::regex(regexp2_str);
2641 // Count number of characters in string
2645 // \\[a-zA-Z]+ ==> 1
2646 static int computeSize(string s, int len)
2652 for (int i = 0; i < len; i += skip, count++) {
2655 if (isalpha(s[i+1])) {
2656 for (int j = 2; i+j < len; j++) {
2657 if (! isalpha(s[i+j])) {
2660 else if ((s[i+j] == '{') && s[i+j+1] == '}')
2662 else if ((s[i+j] == '{') && (i + j + 1 >= len))
2670 else if (s[i] == '{') {
2676 else if (s[i] == '-') {
2677 if (s[i+1] == '-') {
2693 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
2698 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
2701 docstring docstr = stringifyFromForSearch(opt, cur, len);
2702 string str = normalize(docstr, true);
2703 if (!opt.ignoreformat) {
2704 str = correctlanguagesetting(str, false, !opt.ignoreformat);
2707 mres.match_len = -1;
2710 LYXERR(Debug::FIND, "Matching against '" << lyx::to_utf8(docstr) << "'");
2711 LYXERR(Debug::FIND, "After normalization: '" << str << "'");
2714 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
2715 regex const *p_regexp;
2716 regex_constants::match_flag_type flags;
2718 flags = regex_constants::match_continuous;
2721 flags = regex_constants::match_default;
2722 p_regexp = ®exp2;
2724 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
2725 if (re_it == sregex_iterator())
2727 match_results<string::const_iterator> const & m = *re_it;
2729 if (0) { // Kornel Benko: DO NOT CHECKK
2730 // Check braces on the segment that matched the entire regexp expression,
2731 // plus the last subexpression, if a (.*?) was inserted in the constructor.
2732 if (!braces_match(m[0].first, m[0].second, open_braces))
2736 // Check braces on segments that matched all (.*?) subexpressions,
2737 // except the last "padding" one inserted by lyx.
2738 for (size_t i = 1; i < m.size() - 1; ++i)
2739 if (!braces_match(m[i].first, m[i].second, open_braces))
2742 // Exclude from the returned match length any length
2743 // due to close wildcards added at end of regexp
2744 // and also the length of the leading (e.g. '\emph{')
2746 // Whole found string, including the leading: m[0].second - m[0].first
2747 // Size of the leading string: m[1].second - m[1].first
2748 int leadingsize = 0;
2750 leadingsize = m[1].second - m[1].first;
2752 for (size_t i = 0; i < m.size(); i++) {
2753 LYXERR(Debug::FIND, "Match " << i << " is " << m[i].second - m[i].first << " long");
2755 if (close_wildcards == 0)
2756 result = m[0].second - m[0].first;
2759 result = m[m.size() - close_wildcards].first - m[0].first;
2761 size_t pos = m.position(size_t(0));
2762 // Ignore last closing characters
2763 while (result > 0) {
2764 if (str[pos+result-1] == '}')
2769 if (result > leadingsize)
2770 result -= leadingsize;
2773 mres.match_len = computeSize(str.substr(pos+leadingsize,result), result);
2774 mres.match2end = str.size() - pos - leadingsize;
2775 mres.pos = pos+leadingsize;
2779 // else !use_regexp: but all code paths above return
2780 LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='"
2781 << par_as_string << "', str='" << str << "'");
2782 LYXERR(Debug::FIND, "Searching in normal mode: lead_as_string='"
2783 << lead_as_string << "', par_as_string_nolead='"
2784 << par_as_string_nolead << "'");
2787 LYXERR(Debug::FIND, "size=" << par_as_string.size()
2788 << ", substr='" << str.substr(0, par_as_string.size()) << "'");
2789 if (str.substr(0, par_as_string.size()) == par_as_string) {
2790 mres.match_len = par_as_string.size();
2791 mres.match2end = str.size();
2796 size_t pos = str.find(par_as_string_nolead);
2797 if (pos != string::npos) {
2798 mres.match_len = par_as_string.size();
2799 mres.match2end = str.size() - pos;
2808 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
2810 MatchResult mres = findAux(cur, len, at_begin);
2811 int res = mres.match_len;
2813 "res=" << res << ", at_begin=" << at_begin
2814 << ", matchword=" << opt.matchword
2815 << ", inTexted=" << cur.inTexted());
2816 if (res == 0 || !at_begin || !opt.matchword || !cur.inTexted())
2818 if ((len > 0) && (res < len)) {
2822 Paragraph const & par = cur.paragraph();
2823 bool ws_left = (cur.pos() > 0)
2824 ? par.isWordSeparator(cur.pos() - 1)
2826 bool ws_right = (cur.pos() + len < par.size())
2827 ? par.isWordSeparator(cur.pos() + len)
2830 "cur.pos()=" << cur.pos() << ", res=" << res
2831 << ", separ: " << ws_left << ", " << ws_right
2834 if (ws_left && ws_right) {
2835 // Check for word separators inside the found 'word'
2836 for (int i = 0; i < len; i++) {
2837 if (par.isWordSeparator(cur.pos() + i)) {
2849 string MatchStringAdv::normalize(docstring const & s, bool hack_braces) const
2852 if (! opt.casesensitive)
2853 t = lyx::to_utf8(lowercase(s));
2855 t = lyx::to_utf8(s);
2856 // Remove \n at begin
2857 while (!t.empty() && t[0] == '\n')
2860 while (!t.empty() && t[t.size() - 1] == '\n')
2861 t = t.substr(0, t.size() - 1);
2863 // Handle all other '\n'
2864 while ((pos = t.find("\n")) != string::npos) {
2865 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
2867 if (isAlnumASCII(t[pos+1])) {
2868 t.replace(pos-2, 3, " ");
2871 t.replace(pos-2, 3, "");
2874 else if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
2875 // '\n' adjacent to non-alpha-numerics, discard
2876 t.replace(pos, 1, "");
2879 // Replace all other \n with spaces
2880 t.replace(pos, 1, " ");
2883 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
2884 // Kornel: Added textsl, textsf, textit, texttt and noun
2885 // + allow to seach for colored text too
2886 LYXERR(Debug::FIND, "Removing stale empty \\emph{}, \\textbf{}, \\*section{} macros from: " << t);
2887 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
2888 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
2889 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
2890 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
2892 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
2893 // FIXME - check what preceeds the brace
2895 if (opt.ignoreformat)
2896 while (regex_replace(t, t, "\\{", "_x_<")
2897 || regex_replace(t, t, "\\}", "_x_>"))
2898 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
2900 while (regex_replace(t, t, "\\\\\\{", "_x_<")
2901 || regex_replace(t, t, "\\\\\\}", "_x_>"))
2902 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
2909 docstring stringifyFromCursor(DocIterator const & cur, int len)
2911 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
2912 if (cur.inTexted()) {
2913 Paragraph const & par = cur.paragraph();
2914 // TODO what about searching beyond/across paragraph breaks ?
2915 // TODO Try adding a AS_STR_INSERTS as last arg
2916 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
2917 int(par.size()) : cur.pos() + len;
2918 OutputParams runparams(&cur.buffer()->params().encoding());
2919 runparams.nice = true;
2920 runparams.flavor = OutputParams::LATEX;
2921 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
2922 // No side effect of file copying and image conversion
2923 runparams.dryrun = true;
2924 runparams.for_search = true;
2925 LYXERR(Debug::FIND, "Stringifying with cur: "
2926 << cur << ", from pos: " << cur.pos() << ", end: " << end);
2927 return par.asString(cur.pos(), end,
2928 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
2930 } else if (cur.inMathed()) {
2931 CursorSlice cs = cur.top();
2932 MathData md = cs.cell();
2933 MathData::const_iterator it_end =
2934 (( len == -1 || cs.pos() + len > int(md.size()))
2936 : md.begin() + cs.pos() + len );
2938 for (MathData::const_iterator it = md.begin() + cs.pos();
2941 docstring s = asString(md2);
2942 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
2945 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
2950 /** Computes the LaTeX export of buf starting from cur and ending len positions
2951 * after cur, if len is positive, or at the paragraph or innermost inset end
2954 docstring latexifyFromCursor(DocIterator const & cur, int len)
2956 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
2957 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
2958 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
2959 Buffer const & buf = *cur.buffer();
2961 odocstringstream ods;
2963 OutputParams runparams(&buf.params().encoding());
2964 runparams.nice = false;
2965 runparams.flavor = OutputParams::LATEX;
2966 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
2967 // No side effect of file copying and image conversion
2968 runparams.dryrun = true;
2969 runparams.for_search = true;
2971 if (cur.inTexted()) {
2972 // @TODO what about searching beyond/across paragraph breaks ?
2973 pos_type endpos = cur.paragraph().size();
2974 if (len != -1 && endpos > cur.pos() + len)
2975 endpos = cur.pos() + len;
2976 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
2977 string(), cur.pos(), endpos);
2978 string s = lyx::to_utf8(ods.str());
2979 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
2980 return(lyx::from_utf8(s));
2981 } else if (cur.inMathed()) {
2982 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
2983 for (int s = cur.depth() - 1; s >= 0; --s) {
2984 CursorSlice const & cs = cur[s];
2985 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
2987 cs.asInsetMath()->asHullInset()->header_write(ws);
2992 CursorSlice const & cs = cur.top();
2993 MathData md = cs.cell();
2994 MathData::const_iterator it_end =
2995 ((len == -1 || cs.pos() + len > int(md.size()))
2997 : md.begin() + cs.pos() + len);
2999 for (MathData::const_iterator it = md.begin() + cs.pos();
3003 ods << asString(md2);
3004 // Retrieve the math environment type, and add '$' or '$]'
3005 // or others (\end{equation}) accordingly
3006 for (int s = cur.depth() - 1; s >= 0; --s) {
3007 CursorSlice const & cs2 = cur[s];
3008 InsetMath * inset = cs2.asInsetMath();
3009 if (inset && inset->asHullInset()) {
3011 inset->asHullInset()->footer_write(ws);
3015 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3017 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3023 /** Finalize an advanced find operation, advancing the cursor to the innermost
3024 ** position that matches, plus computing the length of the matching text to
3027 int findAdvFinalize(DocIterator & cur, MatchStringAdv const & match)
3029 // Search the foremost position that matches (avoids find of entire math
3030 // inset when match at start of it)
3032 DocIterator old_cur(cur.buffer());
3034 LYXERR(Debug::FIND, "Forwarding one step (searching for innermost match)");
3038 } while (cur && cur.depth() > d && match(cur).match_len > 0);
3040 int max_match = match(cur).match_len; /* match valid only if not searching whole words */
3041 if (max_match <= 0) return 0;
3042 LYXERR(Debug::FIND, "Ok");
3044 // Compute the match length
3046 if (cur.pos() + len > cur.lastpos())
3048 if (match.opt.matchword) {
3049 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3050 while (cur.pos() + len <= cur.lastpos() && match(cur, len).match_len <= 0) {
3052 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3054 // Length of matched text (different from len param)
3055 int old_match = match(cur, len).match_len;
3059 // Greedy behaviour while matching regexps
3060 while ((new_match = match(cur, len + 1).match_len) > old_match) {
3062 old_match = new_match;
3063 LYXERR(Debug::FIND, "verifying match with len = " << len);
3070 int maxl = cur.lastpos() - cur.pos();
3071 // Greedy behaviour while matching regexps
3072 while (maxl > minl) {
3073 int actual_match = match(cur, len).match_len;
3074 if (actual_match >= max_match) {
3075 // actual_match > max_match _can_ happen,
3076 // if the search area splits
3077 // some following word so that the regex
3078 // (e.g. 'r.*r\b' matches 'r' from the middle of the
3080 // This means, the len value is too big
3082 len = (int)((maxl + minl)/2);
3085 // (actual_match < max_match)
3087 len = (int)((maxl + minl)/2);
3091 // Search for real start of matched characters
3096 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
3097 if (cur.depth() < old_cur.depth()) {
3099 LYXERR0("cur.depth() < old_cur.depth(), this should never happen");
3102 if (cur.pos() != old_cur.pos()) {
3103 // OK, forwarded 1 pos in actual inset
3104 actual_match = match(cur, len-1).match_len;
3105 if (actual_match == max_match) {
3106 // Ha, got it! The shorter selection has the same match length
3111 // OK, the shorter selection matches less chars, revert to previous value
3117 LYXERR0("cur.pos() == old_cur.pos(), this should never happen");
3118 actual_match = match(cur, len).match_len;
3119 if (actual_match == max_match)
3129 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
3133 while (!theApp()->longOperationCancelled() && cur) {
3134 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
3135 MatchResult mres = match(cur, -1, false);
3136 int match_len = mres.match_len;
3137 LYXERR(Debug::FIND, "match_len: " << match_len);
3138 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
3139 LYXERR0("BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
3142 if (match_len > 0) {
3143 // Try to find the begin of searched string
3144 int increment = mres.pos/2;
3145 while (mres.pos > 5 && (increment > 5)) {
3146 DocIterator old_cur = cur;
3147 for (int i = 0; i < increment && cur; cur.forwardPos(), i++) {
3154 MatchResult mres2 = match(cur, -1, false);
3155 if ((mres2.match2end < mres.match2end) ||
3156 (mres2.match_len < mres.match_len)) {
3163 if (increment > mres.pos/2)
3164 increment = mres.pos/2;
3168 int match_len_zero_count = 0;
3169 for (int i = 0; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) {
3171 int remaining_len = match(cur, -1, false).match_len;
3172 if (remaining_len <= 0) {
3173 // Apparently the searched string is not in the remaining part
3180 LYXERR(Debug::FIND, "Advancing cur: " << cur);
3181 int match_len3 = match(cur, 1).match_len;
3184 int match_len2 = match(cur).match_len;
3185 LYXERR(Debug::FIND, "match_len2: " << match_len2);
3186 if (match_len2 > 0) {
3187 // Sometimes in finalize we understand it wasn't a match
3188 // and we need to continue the outest loop
3189 int len = findAdvFinalize(cur, match);
3194 if (match_len2 >= 0) {
3195 if (match_len2 == 0)
3196 match_len_zero_count++;
3198 match_len_zero_count = 0;
3201 if (++match_len_zero_count > 3) {
3202 LYXERR(Debug::FIND, "match_len2_zero_count: " << match_len_zero_count << ", match_len was " << match_len);
3203 match_len_zero_count = 0;
3211 if (match_len >= 0 && cur.pit() < cur.lastpit()) {
3212 LYXERR(Debug::FIND, "Advancing par: cur=" << cur);
3215 // This should exit nested insets, if any, or otherwise undefine the currsor.
3216 cur.pos() = cur.lastpos();
3217 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
3225 /// Find the most backward consecutive match within same paragraph while searching backwards.
3226 int findMostBackwards(DocIterator & cur, MatchStringAdv const & match)
3228 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3229 DocIterator tmp_cur = cur;
3230 int len = findAdvFinalize(tmp_cur, match);
3231 Inset & inset = cur.inset();
3232 for (; cur != cur_begin; cur.backwardPos()) {
3233 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
3234 DocIterator new_cur = cur;
3235 new_cur.backwardPos();
3236 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
3238 int new_len = findAdvFinalize(new_cur, match);
3243 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
3249 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
3253 // Backup of original position
3254 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3255 if (cur == cur_begin)
3258 DocIterator cur_orig(cur);
3259 bool pit_changed = false;
3262 bool found_match = (match(cur, -1, false).match_len > 0);
3266 cur.pos() = cur.lastpos();
3268 cur.pos() = cur_orig.pos();
3269 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
3270 DocIterator cur_prev_iter;
3272 found_match = (match(cur).match_len > 0);
3273 LYXERR(Debug::FIND, "findBackAdv3: found_match="
3274 << found_match << ", cur: " << cur);
3276 return findMostBackwards(cur, match);
3278 // Stop if begin of document reached
3279 if (cur == cur_begin)
3281 cur_prev_iter = cur;
3285 if (cur == cur_begin)
3292 } while (!theApp()->longOperationCancelled());
3300 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
3301 DocIterator const & cur, int len)
3303 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
3305 if (!opt.ignoreformat)
3306 return latexifyFromCursor(cur, len);
3308 return stringifyFromCursor(cur, len);
3312 FindAndReplaceOptions::FindAndReplaceOptions(
3313 docstring const & find_buf_name, bool casesensitive,
3314 bool matchword, bool forward, bool expandmacros, bool ignoreformat,
3315 docstring const & repl_buf_name, bool keep_case,
3316 SearchScope scope, SearchRestriction restr)
3317 : find_buf_name(find_buf_name), casesensitive(casesensitive), matchword(matchword),
3318 forward(forward), expandmacros(expandmacros), ignoreformat(ignoreformat),
3319 repl_buf_name(repl_buf_name), keep_case(keep_case), scope(scope), restr(restr)
3327 /** Check if 'len' letters following cursor are all non-lowercase */
3328 static bool allNonLowercase(Cursor const & cur, int len)
3330 pos_type beg_pos = cur.selectionBegin().pos();
3331 pos_type end_pos = cur.selectionBegin().pos() + len;
3332 if (len > cur.lastpos() + 1 - beg_pos) {
3333 LYXERR(Debug::FIND, "This should not happen, more debug needed");
3334 len = cur.lastpos() + 1 - beg_pos;
3335 end_pos = beg_pos + len;
3337 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
3338 if (isLowerCase(cur.paragraph().getChar(pos)))
3344 /** Check if first letter is upper case and second one is lower case */
3345 static bool firstUppercase(Cursor const & cur)
3348 pos_type pos = cur.selectionBegin().pos();
3349 if (pos >= cur.lastpos() - 1) {
3350 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
3353 ch1 = cur.paragraph().getChar(pos);
3354 ch2 = cur.paragraph().getChar(pos + 1);
3355 bool result = isUpperCase(ch1) && isLowerCase(ch2);
3356 LYXERR(Debug::FIND, "firstUppercase(): "
3357 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
3358 << ch2 << "(" << char(ch2) << ")"
3359 << ", result=" << result << ", cur=" << cur);
3364 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
3366 ** \fixme What to do with possible further paragraphs in replace buffer ?
3368 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
3370 ParagraphList::iterator pit = buffer.paragraphs().begin();
3371 LASSERT(pit->size() >= 1, /**/);
3372 pos_type right = pos_type(1);
3373 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
3374 right = pit->size();
3375 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
3381 static void findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
3383 Cursor & cur = bv->cursor();
3384 if (opt.repl_buf_name == docstring()
3385 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
3386 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3389 DocIterator sel_beg = cur.selectionBegin();
3390 DocIterator sel_end = cur.selectionEnd();
3391 if (&sel_beg.inset() != &sel_end.inset()
3392 || sel_beg.pit() != sel_end.pit()
3393 || sel_beg.idx() != sel_end.idx())
3395 int sel_len = sel_end.pos() - sel_beg.pos();
3396 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
3397 << ", sel_len: " << sel_len << endl);
3400 LASSERT(sel_len > 0, return);
3402 if (!matchAdv(sel_beg, sel_len).match_len)
3405 // Build a copy of the replace buffer, adapted to the KeepCase option
3406 Buffer & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
3408 repl_buffer_orig.write(oss);
3409 string lyx = oss.str();
3410 Buffer repl_buffer("", false);
3411 repl_buffer.setUnnamed(true);
3412 LASSERT(repl_buffer.readString(lyx), return);
3413 if (opt.keep_case && sel_len >= 2) {
3414 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
3415 if (cur.inTexted()) {
3416 if (firstUppercase(cur))
3417 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
3418 else if (allNonLowercase(cur, sel_len))
3419 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
3422 cap::cutSelection(cur, false);
3423 if (cur.inTexted()) {
3424 repl_buffer.changeLanguage(
3425 repl_buffer.language(),
3426 cur.getFont().language());
3427 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
3428 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
3429 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
3430 repl_buffer.params().documentClassPtr(),
3431 bv->buffer().errorList("Paste"));
3432 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
3433 sel_len = repl_buffer.paragraphs().begin()->size();
3434 } else if (cur.inMathed()) {
3435 odocstringstream ods;
3437 OutputParams runparams(&repl_buffer.params().encoding());
3438 runparams.nice = false;
3439 runparams.flavor = OutputParams::LATEX;
3440 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3441 runparams.dryrun = true;
3442 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
3443 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
3444 docstring repl_latex = ods.str();
3445 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
3447 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
3448 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
3449 repl_latex = from_utf8(s);
3450 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
3451 MathData ar(cur.buffer());
3452 asArray(repl_latex, ar, Parse::NORMAL);
3454 sel_len = ar.size();
3455 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3457 if (cur.pos() >= sel_len)
3458 cur.pos() -= sel_len;
3461 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3462 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
3463 bv->processUpdateFlags(Update::Force);
3467 /// Perform a FindAdv operation.
3468 bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt)
3473 // e.g., when invoking word-findadv from mini-buffer wither with
3474 // wrong options syntax or before ever opening advanced F&R pane
3475 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3479 MatchStringAdv matchAdv(bv->buffer(), opt);
3480 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
3482 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
3483 findAdvReplace(bv, opt, matchAdv);
3486 match_len = findForwardAdv(cur, matchAdv);
3488 match_len = findBackwardsAdv(cur, matchAdv);
3490 // This may only be raised by lyx::regex()
3491 bv->message(_("Invalid regular expression!"));
3495 if (match_len == 0) {
3496 bv->message(_("Match not found!"));
3500 bv->message(_("Match found!"));
3502 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << match_len);
3503 bv->putSelectionAt(cur, match_len, !opt.forward);
3509 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
3511 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
3512 << opt.casesensitive << ' '
3513 << opt.matchword << ' '
3514 << opt.forward << ' '
3515 << opt.expandmacros << ' '
3516 << opt.ignoreformat << ' '
3517 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
3518 << opt.keep_case << ' '
3519 << int(opt.scope) << ' '
3522 LYXERR(Debug::FIND, "built: " << os.str());
3528 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
3530 LYXERR(Debug::FIND, "parsing");
3534 while (line != "EOSS") {
3538 if (is.eof()) // Tolerate malformed request
3542 LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
3543 opt.find_buf_name = from_utf8(s);
3544 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat;
3545 is.get(); // Waste space before replace string
3548 while (line != "EOSS") {
3552 if (is.eof()) // Tolerate malformed request
3556 LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
3557 opt.repl_buf_name = from_utf8(s);
3558 is >> opt.keep_case;
3561 opt.scope = FindAndReplaceOptions::SearchScope(i);
3563 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
3565 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
3566 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
3567 << opt.scope << ' ' << opt.restr);