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::XETEX;
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::XETEX;
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 removeAccents();
1131 void setForDefaultLang(KeyInfo &defLang);
1132 int findclosing(int start, int end, char up, char down, int repeat);
1133 void handleParentheses(int lastpos, bool closingAllowed);
1135 int isOpeningPar(int pos);
1137 void output(ostringstream &os, int lastpos);
1138 // string show(int lastpos);
1141 int Intervall::isOpeningPar(int pos)
1143 if ((pos < 0) || (size_t(pos) >= par.size()))
1145 if (par[pos] != '{')
1147 if (size_t(pos) + 2 >= par.size())
1149 if (par[pos+2] != '}')
1151 if (par[pos+1] == '[' || par[pos+1] == ']')
1156 void Intervall::setForDefaultLang(KeyInfo &defLang)
1158 // Enable the use of first token again
1159 if (ignoreidx >= 0) {
1160 int value = defLang._tokenstart + defLang._tokensize;
1162 if (borders[0].low < value)
1163 borders[0].low = value;
1164 if (borders[0].upper < value)
1165 borders[0].upper = value;
1170 static void checkDepthIndex(int val)
1172 static int maxdepthidx = MAXOPENED-2;
1173 if (val > maxdepthidx) {
1175 LYXERR0("maxdepthidx now " << val);
1179 static void checkIgnoreIdx(int val)
1181 static int maxignoreidx = 2*MAXOPENED - 4;
1182 if (val > maxignoreidx) {
1184 LYXERR0("maxignoreidx now " << val);
1189 * Expand the region of ignored parts of the input latex string
1190 * The region is only relevant in output()
1192 void Intervall::addIntervall(int low, int upper)
1195 if (low == upper) return;
1196 for (idx = ignoreidx+1; idx > 0; --idx) {
1197 if (low > borders[idx-1].upper) {
1201 Border br(low, upper);
1202 if (idx > ignoreidx) {
1205 checkIgnoreIdx(ignoreidx);
1209 // Expand only if one of the new bound is inside the interwall
1210 // We know here that br.low > borders[idx-1].upper
1211 if (br.upper < borders[idx].low) {
1212 // We have to insert at this pos
1213 for (int i = ignoreidx+1; i > idx; --i) {
1214 borders[i] = borders[i-1];
1218 checkIgnoreIdx(ignoreidx);
1221 // Here we know, that we are overlapping
1222 if (br.low > borders[idx].low)
1223 br.low = borders[idx].low;
1224 // check what has to be concatenated
1226 for (int i = idx; i <= ignoreidx; i++) {
1227 if (br.upper >= borders[i].low) {
1229 if (br.upper < borders[i].upper)
1230 br.upper = borders[i].upper;
1236 // count should be >= 1 here
1239 for (int i = idx + count; i <= ignoreidx; i++) {
1240 borders[i-count+1] = borders[i];
1242 ignoreidx -= count - 1;
1248 typedef map<string, string> AccentsMap;
1249 static AccentsMap accents = map<string, string>();
1251 static void buildaccent(string name, string param, string values)
1254 for (size_t i = 0; i < param.size(); i++) {
1255 string key = name + "{" + param[i] + "}";
1256 // get the corresponding utf8-value
1257 if ((values[start] & 0xc0) != 0xc0) {
1258 // should not happen, utf8 encoding starts at least with 11xxxxxx
1262 for (int j = 1; ;j++) {
1263 if (start + j >= values.size())
1265 if ((values[start+j] & 0xc0) == 0xc0) {
1266 // This is the first byte of following utf8 char
1267 accents[key] = values.substr(start, j);
1275 static void buildAccentsMap()
1277 accents["imath"] = "ı";
1278 accents["ddot{\\imath}"] = "ï";
1279 accents["acute{\\imath}"] = "í";
1280 accents["lyxmathsym{ß}"] = "ß";
1281 buildaccent("ddot", "aeouyAEOUY", "äëöüÿÄËÖÜŸ");
1282 buildaccent("dot", "aeoyzAEOYZ", "ȧėȯẏżȦĖȮẎŻ");
1283 buildaccent("acute", "aeouyAEOUY", "äëöüÿÄËÖÜŸ");
1285 buildaccent("dacute", "oOuU", "őŐűŰ");
1286 buildaccent("H", "oOuU", "őŐűŰ"); // dacute in text
1288 buildaccent("mathring", "uU", "ůŮ");
1289 buildaccent("r", "uU", "ůŮ"); //mathring in text
1290 buildaccent("check", "cdnrszCDNRSZ", "čďřňšžČĎŘŇŠŽ");
1291 buildaccent("hat", "cCoOgGhHsS", "ĉĈôÔĝĜĥĤŝŜ");
1292 buildaccent("bar", "aAeE", "āĀēĒ");
1296 * Created accents in math or regexp environment
1297 * are macros, but we need the utf8 equivalent
1299 void Intervall::removeAccents()
1301 if (accents.empty())
1303 static regex const accre("\\\\((lyxmathsym|ddot|dot|acute|mathring|r|check|check|hat|bar)\\{[^\\{\\}]+\\}|imath)");
1305 for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
1307 string key = sub.str(1);
1308 if (accents.find(key) != accents.end()) {
1309 string val = accents[key];
1310 size_t pos = sub.position(0);
1311 for (size_t i = 0; i < val.size(); i++) {
1312 par[pos+i] = val[i];
1314 addIntervall(pos+val.size(), pos + sub.str(0).size());
1317 LYXERR0("Not added accent for \"" << key << "\"");
1322 void Intervall::handleOpenP(int i)
1325 depts[actualdeptindex] = i+1;
1326 closes[actualdeptindex] = -1;
1327 checkDepthIndex(actualdeptindex);
1330 void Intervall::handleCloseP(int i, bool closingAllowed)
1332 if (actualdeptindex <= 0) {
1333 if (! closingAllowed)
1334 LYXERR(Debug::FIND, "Bad closing parenthesis in latex"); /* should not happen, but the latex input may be wrong */
1335 // if we are at the very end
1336 addIntervall(i, i+1);
1339 closes[actualdeptindex] = i+1;
1344 void Intervall::resetOpenedP(int openPos)
1346 // Used as initializer for foreignlanguage entry
1347 actualdeptindex = 1;
1348 depts[1] = openPos+1;
1352 int Intervall::previousNotIgnored(int start)
1354 int idx = 0; /* int intervalls */
1355 for (idx = ignoreidx; idx >= 0; --idx) {
1356 if (start > borders[idx].upper)
1358 if (start >= borders[idx].low)
1359 start = borders[idx].low-1;
1364 int Intervall::nextNotIgnored(int start)
1366 int idx = 0; /* int intervalls */
1367 for (idx = 0; idx <= ignoreidx; idx++) {
1368 if (start < borders[idx].low)
1370 if (start < borders[idx].upper)
1371 start = borders[idx].upper;
1376 typedef map<string, KeyInfo> KeysMap;
1377 typedef vector< KeyInfo> Entries;
1378 static KeysMap keys = map<string, KeyInfo>();
1385 void buildKeys(bool);
1386 void buildEntries(bool);
1387 void makeKey(const string &, KeyInfo, bool isPatternString);
1388 void processRegion(int start, int region_end); /* remove {} parts */
1389 void removeHead(KeyInfo&, int count=0);
1392 LatexInfo(string par, bool isPatternString) : entidx(-1), interval(isPatternString) {
1394 interval.hasTitle = false;
1395 interval.titleValue = "";
1396 buildKeys(isPatternString);
1397 entries = vector<KeyInfo>();
1398 buildEntries(isPatternString);
1402 if (entries.empty()) {
1405 if (entries[0].keytype == KeyInfo::isTitle) {
1406 if (! entries[0].disabled) {
1407 interval.hasTitle = true;
1408 interval.titleValue = entries[0].head;
1411 interval.hasTitle = false;
1412 interval.titleValue = "";
1414 removeHead(entries[0]);
1415 if (entries.size() > 1)
1424 if (int(entries.size()) > entidx) {
1431 bool setNextKey(int idx) {
1432 if ((idx == entidx) && (entidx >= 0)) {
1439 int find(int start, KeyInfo::KeyType keytype) {
1443 while (tmpIdx < int(entries.size())) {
1444 if (entries[tmpIdx].keytype == keytype)
1450 int process(ostringstream &os, KeyInfo &actual);
1451 int dispatch(ostringstream &os, int previousStart, KeyInfo &actual);
1452 // string show(int lastpos) { return interval.show(lastpos);};
1453 int nextNotIgnored(int start) { return interval.nextNotIgnored(start);};
1454 KeyInfo &getKeyInfo(int keyinfo) {
1455 static KeyInfo invalidInfo = KeyInfo();
1456 if ((keyinfo < 0) || ( keyinfo >= int(entries.size())))
1459 return entries[keyinfo];
1461 void setForDefaultLang(KeyInfo &defLang) {interval.setForDefaultLang(defLang);};
1462 void addIntervall(int low, int up) { interval.addIntervall(low, up); };
1466 int Intervall::findclosing(int start, int end, char up = '{', char down = '}', int repeat = 1)
1471 for (int i = start; i < end; i += 1 + skip) {
1475 if (c == '\\') skip = 1;
1479 else if (c == down) {
1481 if ((repeat <= 0) || (par[i+1] != up))
1499 vector<MathEntry> entries;
1504 void insert(string wait, size_t start, size_t end) {
1505 MathEntry m = MathEntry();
1507 m.mathStart = start;
1509 m.mathSize = end - start;
1510 entries.push_back(m);
1512 bool empty() { return entries.empty(); };
1513 size_t getEndPos() {
1514 if (entries.empty() || (actualIdx >= entries.size())) {
1517 return entries[actualIdx].mathEnd;
1519 size_t getStartPos() {
1520 if (entries.empty() || (actualIdx >= entries.size())) {
1521 return 100000; /* definitely enough? */
1523 return entries[actualIdx].mathStart;
1525 size_t getFirstPos() {
1527 return getStartPos();
1530 if (entries.empty() || (actualIdx >= entries.size())) {
1533 return entries[actualIdx].mathSize;
1535 void incrEntry() { actualIdx++; };
1538 void LatexInfo::buildEntries(bool isPatternString)
1540 static regex const rmath("\\$|\\\\\\[|\\\\\\]|\\\\(begin|end)\\{((eqnarray|equation|flalign|gather|multline|align|alignat)\\*?)\\}");
1541 static regex const rkeys("\\$|\\\\\\[|\\\\\\]|\\\\((([a-zA-Z]+\\*?)(\\{([a-z]+\\*?)\\}|=[0-9]+[a-z]+)?))");
1542 static bool disableLanguageOverride = false;
1543 smatch sub, submath;
1544 bool evaluatingRegexp = false;
1546 bool evaluatingMath = false;
1547 bool evaluatingCode = false;
1549 bool evaluatingOptional = false;
1550 size_t optionalEnd = 0;
1553 bool math_end_waiting = false;
1554 size_t math_pos = 10000;
1557 interval.removeAccents();
1559 for (sregex_iterator itmath(interval.par.begin(), interval.par.end(), rmath), end; itmath != end; ++itmath) {
1561 if (math_end_waiting) {
1562 size_t pos = submath.position(size_t(0));
1563 if ((math_end == "$") &&
1564 (submath.str(0) == "$") &&
1565 (interval.par[pos-1] != '\\')) {
1566 mi.insert("$", math_pos, pos + 1);
1567 math_end_waiting = false;
1569 else if ((math_end == "\\]") &&
1570 (submath.str(0) == "\\]")) {
1571 mi.insert("\\]", math_pos, pos + 2);
1572 math_end_waiting = false;
1574 else if ((submath.str(1).compare("end") == 0) &&
1575 (submath.str(2).compare(math_end) == 0)) {
1576 mi.insert(math_end, math_pos, pos + submath.str(0).length());
1577 math_end_waiting = false;
1583 if (submath.str(1).compare("begin") == 0) {
1584 math_end_waiting = true;
1585 math_end = submath.str(2);
1586 math_pos = submath.position(size_t(0));
1588 else if (submath.str(0).compare("\\[") == 0) {
1589 math_end_waiting = true;
1591 math_pos = submath.position(size_t(0));
1593 else if (submath.str(0) == "$") {
1594 size_t pos = submath.position(size_t(0));
1595 if ((pos == 0) || (interval.par[pos-1] != '\\')) {
1596 math_end_waiting = true;
1603 // Ignore language if there is math somewhere in pattern-string
1604 if (isPatternString) {
1607 keys["foreignlanguage"].disabled = true;
1608 disableLanguageOverride = true;
1611 disableLanguageOverride = false;
1614 if (disableLanguageOverride) {
1615 keys["foreignlanguage"].disabled = true;
1618 math_pos = mi.getFirstPos();
1619 for (sregex_iterator it(interval.par.begin(), interval.par.end(), rkeys), end; it != end; ++it) {
1621 string key = sub.str(3);
1623 if (sub.str(0)[0] == '\\')
1624 key = sub.str(0)[1];
1628 size_t k_pos = sub.position(size_t(0));
1629 if ((k_pos > 0) && (interval.par[k_pos - 1] == '\\')) {
1630 // Escaped '$', ignoring
1636 if (evaluatingRegexp) {
1637 if (sub.str(1).compare("endregexp") == 0) {
1638 evaluatingRegexp = false;
1639 // found._tokenstart already set
1640 found._dataEnd = sub.position(size_t(0)) + 13;
1641 found._dataStart = found._dataEnd;
1642 found._tokensize = found._dataEnd - found._tokenstart;
1643 found.parenthesiscount = 0;
1644 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1651 if (evaluatingMath) {
1652 if (size_t(sub.position(size_t(0))) < mi.getEndPos())
1654 evaluatingMath = false;
1656 math_pos = mi.getStartPos();
1658 if (keys.find(key) == keys.end()) {
1659 found = KeyInfo(KeyInfo::isStandard, 0, true);
1660 if (isPatternString) {
1661 found.keytype = KeyInfo::isChar;
1662 found.disabled = false;
1669 if (key.compare("regexp") == 0) {
1670 evaluatingRegexp = true;
1671 found._tokenstart = sub.position(size_t(0));
1672 found._tokensize = 0;
1676 // Handle the other params of key
1677 if (found.keytype == KeyInfo::isIgnored)
1679 else if (found.keytype == KeyInfo::isMath) {
1680 if (size_t(sub.position(size_t(0))) == math_pos) {
1682 found._tokenstart = sub.position(size_t(0));
1683 found._tokensize = mi.getSize();
1684 found._dataEnd = found._tokenstart + found._tokensize;
1685 found._dataStart = found._dataEnd;
1686 found.parenthesiscount = 0;
1687 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1688 evaluatingMath = true;
1691 // begin|end of unknown env, discard
1692 // First handle tables
1693 // longtable|tabular
1694 bool discardComment;
1696 found.keytype = KeyInfo::doRemove;
1697 if ((sub.str(5).compare("longtable") == 0) ||
1698 (sub.str(5).compare("tabular") == 0)) {
1699 discardComment = true; /* '%' */
1702 discardComment = false;
1703 static regex const removeArgs("^(multicols|multipar|sectionbox|subsectionbox|tcolorbox)$");
1705 string token = sub.str(5);
1706 if (regex_match(token, sub2, removeArgs)) {
1707 found.keytype = KeyInfo::removeWithArg;
1710 // discard spaces before pos(0)
1711 int pos = sub.position(size_t(0));
1713 for (count = 0; pos - count > 0; count++) {
1714 char c = interval.par[pos-count-1];
1715 if (discardComment) {
1716 if ((c != ' ') && (c != '%'))
1722 found._tokenstart = pos - count;
1723 if (sub.str(1).compare(0, 5, "begin") == 0) {
1724 size_t pos1 = pos + sub.str(0).length();
1725 if (sub.str(5).compare("cjk") == 0) {
1726 pos1 = interval.findclosing(pos1+1, interval.par.length()) + 1;
1727 if ((interval.par[pos1] == '{') && (interval.par[pos1+1] == '}'))
1729 found.keytype = KeyInfo::isMain;
1730 found._dataStart = pos1;
1731 found._dataEnd = interval.par.length();
1732 found.disabled = keys["foreignlanguage"].disabled;
1733 found.used = keys["foreignlanguage"].used;
1734 found._tokensize = pos1 - found._tokenstart;
1735 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1738 // Swallow possible optional params
1739 while (interval.par[pos1] == '[') {
1740 pos1 = interval.findclosing(pos1+1, interval.par.length(), '[', ']')+1;
1742 // Swallow also the eventual parameter
1743 if (interval.par[pos1] == '{') {
1744 found._dataEnd = interval.findclosing(pos1+1, interval.par.length()) + 1;
1747 found._dataEnd = pos1;
1749 found._dataStart = found._dataEnd;
1750 found._tokensize = count + found._dataEnd - pos;
1751 found.parenthesiscount = 0;
1752 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1753 found.disabled = true;
1757 // Handle "\end{...}"
1758 found._dataStart = pos + sub.str(0).length();
1759 found._dataEnd = found._dataStart;
1760 found._tokensize = count + found._dataEnd - pos;
1761 found.parenthesiscount = 0;
1762 found.head = interval.par.substr(found._tokenstart, found._tokensize);
1763 found.disabled = true;
1767 else if (found.keytype != KeyInfo::isRegex) {
1768 found._tokenstart = sub.position(size_t(0));
1769 if (found.parenthesiscount == 0) {
1770 // Probably to be discarded
1771 size_t following_pos = sub.position(size_t(0)) + sub.str(3).length() + 1;
1772 char following = interval.par[following_pos];
1773 if (following == ' ')
1774 found.head = "\\" + sub.str(3) + " ";
1775 else if (following == '=') {
1776 // like \uldepth=1000pt
1777 found.head = sub.str(0);
1780 found.head = "\\" + key;
1781 found._tokensize = found.head.length();
1782 found._dataEnd = found._tokenstart + found._tokensize;
1783 found._dataStart = found._dataEnd;
1786 int params = found._tokenstart + key.length() + 1;
1787 if (evaluatingOptional) {
1788 if (size_t(found._tokenstart) > optionalEnd) {
1789 evaluatingOptional = false;
1792 found.disabled = true;
1795 int optend = params;
1796 while (interval.par[optend] == '[') {
1797 // discard optional parameters
1798 optend = interval.findclosing(optend+1, interval.par.length(), '[', ']') + 1;
1800 if (optend > params) {
1801 key += interval.par.substr(params, optend-params);
1802 evaluatingOptional = true;
1803 optionalEnd = optend;
1805 string token = sub.str(5);
1806 int closings = found.parenthesiscount;
1807 if (found.parenthesiscount == 1) {
1808 found.head = "\\" + key + "{";
1810 else if (found.parenthesiscount > 1) {
1812 found.head = sub.str(0) + "{";
1813 closings = found.parenthesiscount - 1;
1816 found.head = "\\" + key + "{";
1819 found._tokensize = found.head.length();
1820 found._dataStart = found._tokenstart + found.head.length();
1821 if (interval.par.substr(found._dataStart-1, 15).compare("\\endarguments{}") == 0) {
1822 found._dataStart += 15;
1824 size_t endpos = interval.findclosing(found._dataStart, interval.par.length(), '{', '}', closings);
1825 if (found.keytype == KeyInfo::isList) {
1826 // Check if it really is list env
1827 static regex const listre("^([a-z]+)$");
1829 if (!regex_match(token, sub2, listre)) {
1830 // Change the key of this entry. It is not in a list/item environment
1831 found.keytype = KeyInfo::endArguments;
1834 if (found.keytype == KeyInfo::noMain) {
1835 evaluatingCode = true;
1837 codeStart = found._dataStart;
1839 else if (evaluatingCode) {
1840 if (size_t(found._dataStart) > codeEnd)
1841 evaluatingCode = false;
1842 else if (found.keytype == KeyInfo::isMain) {
1843 // Disable this key, treate it as standard
1844 found.keytype = KeyInfo::isStandard;
1845 found.disabled = true;
1846 if ((codeEnd == interval.par.length()) &&
1847 (found._tokenstart == codeStart)) {
1848 // trickery, because the code inset starts
1849 // with \selectlanguage ...
1851 if (entries.size() > 1) {
1852 entries[entries.size()-1]._dataEnd = codeEnd;
1857 if ((endpos == interval.par.length()) &&
1858 (found.keytype == KeyInfo::doRemove)) {
1859 // Missing closing => error in latex-input?
1860 // therefore do not delete remaining data
1861 found._dataStart -= 1;
1862 found._dataEnd = found._dataStart;
1865 found._dataEnd = endpos;
1867 if (isPatternString) {
1868 keys[key].used = true;
1871 entries.push_back(found);
1875 void LatexInfo::makeKey(const string &keysstring, KeyInfo keyI, bool isPatternString)
1877 stringstream s(keysstring);
1879 const char delim = '|';
1880 while (getline(s, key, delim)) {
1881 KeyInfo keyII(keyI);
1882 if (isPatternString) {
1885 else if ( !keys[key].used)
1886 keyII.disabled = true;
1891 void LatexInfo::buildKeys(bool isPatternString)
1894 static bool keysBuilt = false;
1895 if (keysBuilt && !isPatternString) return;
1897 // Known standard keys with 1 parameter.
1898 // Split is done, if not at start of region
1899 makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString);
1900 makeKey("textbf", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getSeries()), isPatternString);
1901 makeKey("textit|textsc|textsl", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getShape()), isPatternString);
1902 makeKey("uuline|uline|uwave", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getUnderline()), isPatternString);
1903 makeKey("emph|noun", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getMarkUp()), isPatternString);
1904 makeKey("sout|xout", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getStrikeOut()), isPatternString);
1906 makeKey("section|subsection|subsubsection|paragraph|subparagraph|minisec",
1907 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1908 makeKey("section*|subsection*|subsubsection*|paragraph*",
1909 KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1910 makeKey("part|part*|chapter|chapter*", KeyInfo(KeyInfo::isSectioning, 1, ignoreFormats.getSectioning()), isPatternString);
1911 makeKey("title|subtitle|author|subject|publishers|dedication|uppertitleback|lowertitleback|extratitle|lyxaddress|lyxrightaddress", KeyInfo(KeyInfo::isTitle, 1, ignoreFormats.getFrontMatter()), isPatternString);
1913 makeKey("regexp", KeyInfo(KeyInfo::isRegex, 1, false), isPatternString);
1915 // Split is done, if not at start of region
1916 makeKey("textcolor", KeyInfo(KeyInfo::isStandard, 2, ignoreFormats.getColor()), isPatternString);
1917 makeKey("latexenvironment", KeyInfo(KeyInfo::isStandard, 2, false), isPatternString);
1919 // Split is done always.
1920 makeKey("foreignlanguage", KeyInfo(KeyInfo::isMain, 2, ignoreFormats.getLanguage()), isPatternString);
1924 makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1925 makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1926 makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1927 makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1929 makeKey("quad|qquad|hfill|dotfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1930 makeKey("textvisiblespace|nobreakspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1931 makeKey("negthickspace|negmedspace|negthinspace", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1933 // makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1934 // Custom space/skip, remove the content (== length value)
1935 makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
1936 // Found in fr/UserGuide.lyx
1937 makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1939 makeKey("textquotedbl|quotesinglbase|lyxarrow", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1940 makeKey("textquotedblleft|textquotedblright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1941 // Known macros to remove (including their parameter)
1943 makeKey("input|inputencoding|label|ref|index|bibitem", KeyInfo(KeyInfo::doRemove, 1, false), isPatternString);
1944 makeKey("addtocounter|setlength", KeyInfo(KeyInfo::noContent, 2, true), isPatternString);
1945 // handle like standard keys with 1 parameter.
1946 makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString);
1948 // Macros to remove, but let the parameter survive
1950 makeKey("menuitem|textmd|textrm", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1952 // Remove language spec from content of these insets
1953 makeKey("code", KeyInfo(KeyInfo::noMain, 1, false), isPatternString);
1955 // Same effect as previous, parameter will survive (because there is no one anyway)
1957 makeKey("noindent|textcompwordmark|maketitle", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1958 // Remove table decorations
1959 makeKey("hline|tabularnewline|toprule|bottomrule|midrule", KeyInfo(KeyInfo::doRemove, 0, true), isPatternString);
1960 // Discard shape-header.
1961 // For footnote or shortcut too, because of lang settings
1962 // and wrong handling if used 'KeyInfo::noMain'
1963 makeKey("circlepar|diamondpar|heartpar|nutpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1964 makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1965 makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1966 makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1967 makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
1968 makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString);
1969 // like ('tiny{}' or '\tiny ' ... )
1970 makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString);
1972 // Survives, like known character
1973 makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
1974 makeKey("item|listitem", KeyInfo(KeyInfo::isList, 1, false), isPatternString);
1976 makeKey("begin|end", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1977 makeKey("[|]", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1978 makeKey("$", KeyInfo(KeyInfo::isMath, 1, false), isPatternString);
1980 makeKey("par|uldepth|ULdepth|protect|nobreakdash|medskip|relax", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1981 // Remove RTL/LTR marker
1982 makeKey("l|r|textlr|textfr|textar|beginl|endl", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1983 makeKey("lettrine", KeyInfo(KeyInfo::cleanToStart, 0, true), isPatternString);
1984 makeKey("lyxslide", KeyInfo(KeyInfo::isSectioning, 1, true), isPatternString);
1985 makeKey("endarguments", KeyInfo(KeyInfo::endArguments, 0, true), isPatternString);
1986 makeKey("twocolumn", KeyInfo(KeyInfo::removeWithArg, 2, true), isPatternString);
1987 makeKey("tnotetext|ead|fntext|cortext|address", KeyInfo(KeyInfo::removeWithArg, 0, true), isPatternString);
1988 makeKey("lyxend", KeyInfo(KeyInfo::isStandard, 0, true), isPatternString);
1989 if (isPatternString) {
1990 // Allow the first searched string to rebuild the keys too
1994 // no need to rebuild again
2000 * Keep the list of actual opened parentheses actual
2001 * (e.g. depth == 4 means there are 4 '{' not processed yet)
2003 void Intervall::handleParentheses(int lastpos, bool closingAllowed)
2006 for (int i = depts[actualdeptindex]; i < lastpos; i+= 1 + skip) {
2010 if (c == '\\') skip = 1;
2011 else if (c == '{') {
2014 else if (c == '}') {
2015 handleCloseP(i, closingAllowed);
2021 string Intervall::show(int lastpos)
2023 int idx = 0; /* int intervalls */
2026 for (idx = 0; idx <= ignoreidx; idx++) {
2027 while (i < lastpos) {
2029 if (i <= borders[idx].low) {
2030 if (borders[idx].low > lastpos)
2031 printsize = lastpos - i;
2033 printsize = borders[idx].low - i;
2034 s += par.substr(i, printsize);
2036 if (i >= borders[idx].low)
2037 i = borders[idx].upper;
2040 i = borders[idx].upper;
2046 s += par.substr(i, lastpos-i);
2052 void Intervall::output(ostringstream &os, int lastpos)
2054 // get number of chars to output
2055 int idx = 0; /* int intervalls */
2058 string startTitle = titleValue;
2059 for (idx = 0; idx <= ignoreidx; idx++) {
2061 if (i <= borders[idx].low) {
2063 if (borders[idx].low > lastpos)
2064 printsize = lastpos - i;
2066 printsize = borders[idx].low - i;
2067 if (printsize > 0) {
2068 os << startTitle << par.substr(i, printsize);
2070 printed += printsize;
2073 handleParentheses(i, false);
2074 if (i >= borders[idx].low)
2075 i = borders[idx].upper;
2078 i = borders[idx].upper;
2085 os << startTitle << par.substr(i, lastpos-i);
2086 printed += lastpos-i;
2088 handleParentheses(lastpos, false);
2089 for (int i = actualdeptindex; i > 0; --i) {
2092 if (hasTitle && (printed > 0))
2094 if (! isPatternString)
2096 handleParentheses(lastpos, true); /* extra closings '}' allowed here */
2099 void LatexInfo::processRegion(int start, int region_end)
2101 while (start < region_end) { /* Let {[} and {]} survive */
2102 int cnt = interval.isOpeningPar(start);
2104 // Closing is allowed past the region
2105 int closing = interval.findclosing(start+1, interval.par.length());
2106 interval.addIntervall(start, start+1);
2107 interval.addIntervall(closing, closing+1);
2111 start = interval.nextNotIgnored(start+1);
2115 void LatexInfo::removeHead(KeyInfo &actual, int count)
2117 if (actual.parenthesiscount == 0) {
2118 // "{\tiny{} ...}" ==> "{{} ...}"
2119 interval.addIntervall(actual._tokenstart-count, actual._tokenstart + actual._tokensize);
2122 // Remove header hull, that is "\url{abcd}" ==> "abcd"
2123 interval.addIntervall(actual._tokenstart - count, actual._dataStart);
2124 interval.addIntervall(actual._dataEnd, actual._dataEnd+1);
2128 int LatexInfo::dispatch(ostringstream &os, int previousStart, KeyInfo &actual)
2131 switch (actual.keytype)
2133 case KeyInfo::isTitle: {
2135 nextKeyIdx = getNextKey();
2138 case KeyInfo::cleanToStart: {
2139 actual._dataEnd = actual._dataStart;
2140 nextKeyIdx = getNextKey();
2141 // Search for end of arguments
2142 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2144 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2145 entries[i].disabled = true;
2147 actual._dataEnd = entries[tmpIdx]._dataEnd;
2149 while (interval.par[actual._dataEnd] == ' ')
2151 interval.addIntervall(0, actual._dataEnd+1);
2152 interval.actualdeptindex = 0;
2153 interval.depts[0] = actual._dataEnd+1;
2154 interval.closes[0] = -1;
2157 case KeyInfo::noContent: { /* char like "\hspace{2cm}" */
2158 if (actual.disabled)
2159 interval.addIntervall(actual._tokenstart, actual._dataEnd);
2161 interval.addIntervall(actual._dataStart, actual._dataEnd);
2164 case KeyInfo::isChar: {
2165 nextKeyIdx = getNextKey();
2168 case KeyInfo::isSize: {
2169 if (actual.disabled || (interval.par[actual._dataStart] != '{') || (interval.par[actual._dataStart-1] == ' ')) {
2170 processRegion(actual._dataEnd, actual._dataEnd+1); /* remove possibly following {} */
2171 interval.addIntervall(actual._tokenstart, actual._dataEnd+1);
2172 nextKeyIdx = getNextKey();
2174 // Here _dataStart points to '{', so correct it
2175 actual._dataStart += 1;
2176 actual._tokensize += 1;
2177 actual.parenthesiscount = 1;
2178 if (interval.par[actual._dataStart] == '}') {
2179 // Determine the end if used like '{\tiny{}...}'
2180 actual._dataEnd = interval.findclosing(actual._dataStart+1, interval.par.length()) + 1;
2181 interval.addIntervall(actual._dataStart, actual._dataStart+1);
2184 // Determine the end if used like '\tiny{...}'
2185 actual._dataEnd = interval.findclosing(actual._dataStart, interval.par.length()) + 1;
2187 // Split on this key if not at start
2188 int start = interval.nextNotIgnored(previousStart);
2189 if (start < actual._tokenstart) {
2190 interval.output(os, actual._tokenstart);
2191 interval.addIntervall(start, actual._tokenstart);
2193 // discard entry if at end of actual
2194 nextKeyIdx = process(os, actual);
2198 case KeyInfo::endArguments:
2199 // Remove trailing '{}' too
2200 actual._dataStart += 1;
2201 actual._dataEnd += 1;
2202 interval.addIntervall(actual._tokenstart, actual._dataEnd+1);
2203 nextKeyIdx = getNextKey();
2205 case KeyInfo::noMain:
2207 case KeyInfo::isStandard: {
2208 if (actual.disabled) {
2210 processRegion(actual._dataStart, actual._dataStart+1);
2211 nextKeyIdx = getNextKey();
2213 // Split on this key if not at datastart of calling entry
2214 int start = interval.nextNotIgnored(previousStart);
2215 if (start < actual._tokenstart) {
2216 interval.output(os, actual._tokenstart);
2217 interval.addIntervall(start, actual._tokenstart);
2219 // discard entry if at end of actual
2220 nextKeyIdx = process(os, actual);
2224 case KeyInfo::removeWithArg: {
2225 nextKeyIdx = getNextKey();
2226 // Search for end of arguments
2227 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2229 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2230 entries[i].disabled = true;
2232 actual._dataEnd = entries[tmpIdx]._dataEnd;
2234 interval.addIntervall(actual._tokenstart, actual._dataEnd+1);
2237 case KeyInfo::doRemove: {
2238 // Remove the key with all parameters and following spaces
2240 for (pos = actual._dataEnd+1; pos < interval.par.length(); pos++) {
2241 if ((interval.par[pos] != ' ') && (interval.par[pos] != '%'))
2244 interval.addIntervall(actual._tokenstart, pos);
2245 nextKeyIdx = getNextKey();
2248 case KeyInfo::isList: {
2249 // Discard space before _tokenstart
2251 for (count = 0; count < actual._tokenstart; count++) {
2252 if (interval.par[actual._tokenstart-count-1] != ' ')
2255 nextKeyIdx = getNextKey();
2256 int tmpIdx = find(nextKeyIdx, KeyInfo::endArguments);
2258 // Special case: \item is not a list, but a command (like in Style Author_Biography in maa-monthly.layout)
2260 // How else can we catch this one?
2261 for (int i = nextKeyIdx; i <= tmpIdx; i++) {
2262 entries[i].disabled = true;
2264 actual._dataEnd = entries[tmpIdx]._dataEnd;
2266 else if (nextKeyIdx > 0) {
2267 // Ignore any lang entries inside data region
2268 for (int i = nextKeyIdx; i < int(entries.size()) && entries[i]._tokenstart < actual._dataEnd; i++) {
2269 if (entries[i].keytype == KeyInfo::isMain)
2270 entries[i].disabled = true;
2273 if (actual.disabled) {
2274 interval.addIntervall(actual._tokenstart-count, actual._dataEnd+1);
2277 interval.addIntervall(actual._tokenstart-count, actual._tokenstart);
2279 if (interval.par[actual._dataEnd+1] == '[') {
2280 int posdown = interval.findclosing(actual._dataEnd+2, interval.par.length(), '[', ']');
2281 if ((interval.par[actual._dataEnd+2] == '{') &&
2282 (interval.par[posdown-1] == '}')) {
2283 interval.addIntervall(actual._dataEnd+1,actual._dataEnd+3);
2284 interval.addIntervall(posdown-1, posdown+1);
2287 interval.addIntervall(actual._dataEnd+1, actual._dataEnd+2);
2288 interval.addIntervall(posdown, posdown+1);
2290 int blk = interval.nextNotIgnored(actual._dataEnd+1);
2291 if (blk > posdown) {
2292 // Discard at most 1 space after empty item
2294 for (count = 0; count < 1; count++) {
2295 if (interval.par[blk+count] != ' ')
2299 interval.addIntervall(blk, blk+count);
2304 case KeyInfo::isSectioning: {
2305 // Discard spaces before _tokenstart
2307 int val = actual._tokenstart;
2308 for (count = 0; count < actual._tokenstart;) {
2309 val = interval.previousNotIgnored(val-1);
2310 if (interval.par[val] != ' ')
2313 count = actual._tokenstart - val;
2316 if (actual.disabled) {
2317 removeHead(actual, count);
2318 nextKeyIdx = getNextKey();
2320 interval.addIntervall(actual._tokenstart-count, actual._tokenstart);
2321 nextKeyIdx = process(os, actual);
2325 case KeyInfo::isMath: {
2326 // Same as regex, use the content unchanged
2327 nextKeyIdx = getNextKey();
2330 case KeyInfo::isRegex: {
2331 // DO NOT SPLIT ON REGEX
2333 nextKeyIdx = getNextKey();
2336 case KeyInfo::isIgnored: {
2337 // Treat like a character for now
2338 nextKeyIdx = getNextKey();
2341 case KeyInfo::isMain: {
2342 if (interval.par.substr(actual._dataStart, 2) == "% ")
2343 interval.addIntervall(actual._dataStart, actual._dataStart+2);
2344 if (actual._tokenstart > 0) {
2345 int prev = interval.previousNotIgnored(actual._tokenstart - 1);
2346 if ((prev >= 0) && interval.par[prev] == '%')
2347 interval.addIntervall(prev, prev+1);
2349 if (actual.disabled) {
2351 if ((interval.par.substr(actual._dataStart, 3) == " \\[") ||
2352 (interval.par.substr(actual._dataStart, 8) == " \\begin{")) {
2353 // Discard also the space before math-equation
2354 interval.addIntervall(actual._dataStart, actual._dataStart+1);
2356 nextKeyIdx = getNextKey();
2357 // interval.resetOpenedP(actual._dataStart-1);
2360 if (actual._tokenstart < 26) {
2361 // for the first (and maybe dummy) language
2362 interval.setForDefaultLang(actual);
2364 interval.resetOpenedP(actual._dataStart-1);
2368 case KeyInfo::invalid:
2369 // This cannot happen, already handled
2372 // LYXERR0("Unhandled keytype");
2373 nextKeyIdx = getNextKey();
2380 int LatexInfo::process(ostringstream &os, KeyInfo &actual )
2382 int end = interval.nextNotIgnored(actual._dataEnd);
2383 int oldStart = actual._dataStart;
2384 int nextKeyIdx = getNextKey();
2386 if ((nextKeyIdx < 0) ||
2387 (entries[nextKeyIdx]._tokenstart >= actual._dataEnd) ||
2388 (entries[nextKeyIdx].keytype == KeyInfo::invalid)) {
2389 if (oldStart <= end) {
2390 processRegion(oldStart, end);
2395 KeyInfo &nextKey = getKeyInfo(nextKeyIdx);
2397 if ((nextKey.keytype == KeyInfo::isMain) && !nextKey.disabled) {
2398 (void) dispatch(os, actual._dataStart, nextKey);
2399 end = nextKey._tokenstart;
2402 processRegion(oldStart, nextKey._tokenstart);
2403 nextKeyIdx = dispatch(os, actual._dataStart, nextKey);
2405 oldStart = nextKey._dataEnd+1;
2407 // now nextKey is either invalid or is outside of actual._dataEnd
2408 // output the remaining and discard myself
2409 if (oldStart <= end) {
2410 processRegion(oldStart, end);
2412 if (interval.par[end] == '}') {
2414 // This is the normal case.
2415 // But if using the firstlanguage, the closing may be missing
2417 // get minimum of 'end' and 'actual._dataEnd' in case that the nextKey.keytype was 'KeyInfo::isMain'
2419 if (actual._dataEnd < end)
2420 output_end = interval.nextNotIgnored(actual._dataEnd);
2422 output_end = interval.nextNotIgnored(end);
2423 if ((actual.keytype == KeyInfo::isMain) && actual.disabled) {
2424 interval.addIntervall(actual._tokenstart, actual._tokenstart+actual._tokensize);
2426 // Remove possible empty data
2427 int dstart = interval.nextNotIgnored(actual._dataStart);
2428 while (interval.isOpeningPar(dstart) == 1) {
2429 interval.addIntervall(dstart, dstart+1);
2430 int dend = interval.findclosing(dstart+1, output_end);
2431 interval.addIntervall(dend, dend+1);
2432 dstart = interval.nextNotIgnored(dstart+1);
2434 if (dstart < output_end)
2435 interval.output(os, output_end);
2436 interval.addIntervall(actual._tokenstart, end);
2440 string splitOnKnownMacros(string par, bool isPatternString)
2443 LatexInfo li(par, isPatternString);
2444 // LYXERR0("Berfore split: " << par);
2445 KeyInfo DummyKey = KeyInfo(KeyInfo::KeyType::isMain, 2, true);
2447 DummyKey._tokensize = 0;
2448 DummyKey._dataStart = 0;
2449 DummyKey._dataEnd = par.length();
2450 DummyKey.disabled = true;
2451 int firstkeyIdx = li.getFirstKey();
2453 if (firstkeyIdx >= 0) {
2454 KeyInfo firstKey = li.getKeyInfo(firstkeyIdx);
2455 DummyKey._tokenstart = firstKey._tokenstart;
2457 if ((firstKey.keytype != KeyInfo::isMain) || firstKey.disabled) {
2458 // Use dummy firstKey
2459 firstKey = DummyKey;
2460 (void) li.setNextKey(firstkeyIdx);
2463 if (par.substr(firstKey._dataStart, 2) == "% ")
2464 li.addIntervall(firstKey._dataStart, firstKey._dataStart+2);
2466 nextkeyIdx = li.process(os, firstKey);
2467 while (nextkeyIdx >= 0) {
2468 // Check for a possible gap between the last
2469 // entry and this one
2470 int datastart = li.nextNotIgnored(firstKey._dataStart);
2471 KeyInfo &nextKey = li.getKeyInfo(nextkeyIdx);
2472 if ((nextKey._tokenstart > datastart)) {
2474 firstKey._dataStart = datastart;
2475 firstKey._dataEnd = par.length();
2476 (void) li.setNextKey(nextkeyIdx);
2477 // Fake the last opened parenthesis
2478 li.setForDefaultLang(firstKey);
2479 nextkeyIdx = li.process(os, firstKey);
2482 if (nextKey.keytype != KeyInfo::isMain) {
2483 firstKey._dataStart = datastart;
2484 firstKey._dataEnd = nextKey._dataEnd+1;
2485 (void) li.setNextKey(nextkeyIdx);
2486 li.setForDefaultLang(firstKey);
2487 nextkeyIdx = li.process(os, firstKey);
2490 nextkeyIdx = li.process(os, nextKey);
2494 // Handle the remaining
2495 firstKey._dataStart = li.nextNotIgnored(firstKey._dataStart);
2496 firstKey._dataEnd = par.length();
2498 if ((firstKey._dataStart < firstKey._dataEnd) &&
2499 (par[firstKey._dataStart] != '}')) {
2500 li.setForDefaultLang(firstKey);
2501 (void) li.process(os, firstKey);
2505 // return string definitelly impossible to match
2506 s = "\\foreignlanguage{ignore}{ }";
2510 s = par; /* no known macros found */
2511 // LYXERR0("After split: " << s);
2516 * Try to unify the language specs in the latexified text.
2517 * Resulting modified string is set to "", if
2518 * the searched tex does not contain all the features in the search pattern
2520 static string correctlanguagesetting(string par, bool isPatternString, bool withformat)
2522 static Features regex_f;
2523 static int missed = 0;
2524 static bool regex_with_format = false;
2526 int parlen = par.length();
2528 while ((parlen > 0) && (par[parlen-1] == '\n')) {
2531 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) {
2532 // Happens to be there in case of description or labeling environment
2537 // Split the latex input into pieces which
2538 // can be digested by our search engine
2539 LYXERR(Debug::FIND, "input: \"" << par << "\"");
2540 result = splitOnKnownMacros(par.substr(0,parlen), isPatternString);
2541 LYXERR(Debug::FIND, "After split: \"" << result << "\"");
2544 result = par.substr(0, parlen);
2545 if (isPatternString) {
2548 regex_f = identifyFeatures(result);
2549 string features = "";
2550 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2551 string a = it->first;
2552 regex_with_format = true;
2553 features += " " + a;
2554 // LYXERR0("Identified regex format:" << a);
2556 LYXERR(Debug::FIND, "Identified Features" << features);
2559 } else if (regex_with_format) {
2560 Features info = identifyFeatures(result);
2561 for (auto it = regex_f.cbegin(); it != regex_f.cend(); ++it) {
2562 string a = it->first;
2563 bool b = it->second;
2564 if (b && ! info[a]) {
2566 LYXERR(Debug::FIND, "Missed(" << missed << " " << a <<", srclen = " << parlen );
2572 // LYXERR0("No regex formats");
2578 // Remove trailing closure of math, macros and environments, so to catch parts of them.
2579 static int identifyClosing(string & t)
2581 int open_braces = 0;
2583 LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
2584 if (regex_replace(t, t, "(.*[^\\\\])\\$" REGEX_EOS, "$1"))
2586 if (regex_replace(t, t, "(.*[^\\\\]) \\\\\\]" REGEX_EOS, "$1"))
2588 if (regex_replace(t, t, "(.*[^\\\\]) \\\\end\\{[a-zA-Z_]*\\*?\\}" REGEX_EOS, "$1"))
2590 if (regex_replace(t, t, "(.*[^\\\\])\\}" REGEX_EOS, "$1")) {
2600 MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt)
2601 : p_buf(&buf), p_first_buf(&buf), opt(opt)
2603 Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
2604 docstring const & ds = stringifySearchBuffer(find_buf, opt);
2605 use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos;
2606 // When using regexp, braces are hacked already by escape_for_regex()
2607 par_as_string = normalize(ds, !use_regexp);
2609 close_wildcards = 0;
2611 size_t lead_size = 0;
2612 // correct the language settings
2613 par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat);
2614 if (opt.ignoreformat) {
2616 // if par_as_string_nolead were emty,
2617 // the following call to findAux will always *find* the string
2618 // in the checked data, and thus always using the slow
2619 // examining of the current text part.
2620 par_as_string_nolead = par_as_string;
2623 lead_size = identifyLeading(par_as_string);
2624 LYXERR(Debug::FIND, "Lead_size: " << lead_size);
2625 lead_as_string = par_as_string.substr(0, lead_size);
2626 par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
2630 open_braces = identifyClosing(par_as_string);
2631 identifyClosing(par_as_string_nolead);
2632 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2633 LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'");
2635 string lead_as_regexp;
2636 if (lead_size > 0) {
2637 // @todo No need to search for \regexp{} insets in leading material
2638 lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat);
2639 par_as_string = par_as_string_nolead;
2640 LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
2641 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2643 par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
2644 // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
2645 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2647 // Insert .* before trailing '\$' ('$' has been escaped by escape_for_regex)
2648 regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\$)\\'", "$1(.*?)$2")
2649 // Insert .* before trailing '\\\]' ('\]' has been escaped by escape_for_regex)
2650 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])( \\\\\\\\\\\\\\])\\'", "$1(.*?)$2")
2651 // Insert .* before trailing '\\end\{...}' ('\end{...}' has been escaped by escape_for_regex)
2652 || regex_replace(par_as_string, par_as_string,
2653 "(.*[^\\\\])( \\\\\\\\end\\\\\\{[a-zA-Z_]*)(\\\\\\*)?(\\\\\\})\\'", "$1(.*?)$2$3$4")
2654 // Insert .* before trailing '\}' ('}' has been escaped by escape_for_regex)
2655 || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\})\\'", "$1(.*?)$2")
2659 if (!opt.ignoreformat) {
2660 // Remove extra '\}' at end if not part of \{\.\}
2661 size_t lng = par_as_string.size();
2663 if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
2665 if (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
2674 if (lng < par_as_string.size())
2675 par_as_string = par_as_string.substr(0,lng);
2678 regex_replace(par_as_string, par_as_string, "\\\\\\.", "_xxbdotxx_");
2679 // handle '.' -> '[^]', replace later as '[^\}\{\\]'
2680 regex_replace(par_as_string, par_as_string, "\\.", "[^]");
2681 // replace '[^...]' with '[^...\}\{\\]'
2682 regex_replace(par_as_string, par_as_string, "\\[\\^([^\\\\\\]]*)\\]", "_xxbrlxx_$1\\}\\{\\\\_xxbrrxx_");
2683 regex_replace(par_as_string, par_as_string, "_xxbrlxx_", "[^");
2684 regex_replace(par_as_string, par_as_string, "_xxbrrxx_", "]");
2686 regex_replace(par_as_string, par_as_string, "_xxbdotxx_", "\\.");
2689 LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
2690 LYXERR(Debug::FIND, "Open braces: " << open_braces);
2691 LYXERR(Debug::FIND, "Close .*? : " << close_wildcards);
2692 LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
2694 // If entered regexp must match at begin of searched string buffer
2695 // Kornel: Added parentheses to use $1 for size of the leading string
2699 // TODO: Adapt '\[12345678]' in par_as_string to acount for the first '()
2700 // Unfortunately is '\1', '\2', etc not working for strings with extra format
2701 // so the convert has no effect in that case
2702 for (int i = 8; i > 0; --i) {
2703 string orig = "\\\\" + std::to_string(i);
2704 string dest = "\\" + std::to_string(i+1);
2705 while (regex_replace(par_as_string, par_as_string, orig, dest));
2707 regexp_str = "(" + lead_as_regexp + ")" + par_as_string;
2708 regexp2_str = "(" + lead_as_regexp + ").*?" + par_as_string;
2710 LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
2711 regexp = lyx::regex(regexp_str);
2713 LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
2714 regexp2 = lyx::regex(regexp2_str);
2719 // Count number of characters in string
2723 // \\[a-zA-Z]+ ==> 1
2724 static int computeSize(string s, int len)
2730 for (int i = 0; i < len; i += skip, count++) {
2733 if (isalpha(s[i+1])) {
2734 for (int j = 2; i+j < len; j++) {
2735 if (! isalpha(s[i+j])) {
2738 else if ((s[i+j] == '{') && s[i+j+1] == '}')
2740 else if ((s[i+j] == '{') && (i + j + 1 >= len))
2748 else if (s[i] == '{') {
2754 else if (s[i] == '-') {
2755 if (s[i+1] == '-') {
2771 MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
2776 (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
2779 docstring docstr = stringifyFromForSearch(opt, cur, len);
2780 string str = normalize(docstr, true);
2781 if (!opt.ignoreformat) {
2782 str = correctlanguagesetting(str, false, !opt.ignoreformat);
2785 mres.match_len = -1;
2788 LYXERR(Debug::FIND, "Matching against '" << lyx::to_utf8(docstr) << "'");
2789 LYXERR(Debug::FIND, "After normalization: '" << str << "'");
2792 LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin);
2793 regex const *p_regexp;
2794 regex_constants::match_flag_type flags;
2796 flags = regex_constants::match_continuous;
2799 flags = regex_constants::match_default;
2800 p_regexp = ®exp2;
2802 sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
2803 if (re_it == sregex_iterator())
2805 match_results<string::const_iterator> const & m = *re_it;
2807 if (0) { // Kornel Benko: DO NOT CHECKK
2808 // Check braces on the segment that matched the entire regexp expression,
2809 // plus the last subexpression, if a (.*?) was inserted in the constructor.
2810 if (!braces_match(m[0].first, m[0].second, open_braces))
2814 // Check braces on segments that matched all (.*?) subexpressions,
2815 // except the last "padding" one inserted by lyx.
2816 for (size_t i = 1; i < m.size() - 1; ++i)
2817 if (!braces_match(m[i].first, m[i].second, open_braces))
2820 // Exclude from the returned match length any length
2821 // due to close wildcards added at end of regexp
2822 // and also the length of the leading (e.g. '\emph{')
2824 // Whole found string, including the leading: m[0].second - m[0].first
2825 // Size of the leading string: m[1].second - m[1].first
2826 int leadingsize = 0;
2828 leadingsize = m[1].second - m[1].first;
2830 for (size_t i = 0; i < m.size(); i++) {
2831 LYXERR(Debug::FIND, "Match " << i << " is " << m[i].second - m[i].first << " long");
2833 if (close_wildcards == 0)
2834 result = m[0].second - m[0].first;
2837 result = m[m.size() - close_wildcards].first - m[0].first;
2839 size_t pos = m.position(size_t(0));
2840 // Ignore last closing characters
2841 while (result > 0) {
2842 if (str[pos+result-1] == '}')
2847 if (result > leadingsize)
2848 result -= leadingsize;
2851 mres.match_len = computeSize(str.substr(pos+leadingsize,result), result);
2852 mres.match2end = str.size() - pos - leadingsize;
2853 mres.pos = pos+leadingsize;
2857 // else !use_regexp: but all code paths above return
2858 LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='"
2859 << par_as_string << "', str='" << str << "'");
2860 LYXERR(Debug::FIND, "Searching in normal mode: lead_as_string='"
2861 << lead_as_string << "', par_as_string_nolead='"
2862 << par_as_string_nolead << "'");
2865 LYXERR(Debug::FIND, "size=" << par_as_string.size()
2866 << ", substr='" << str.substr(0, par_as_string.size()) << "'");
2867 if (str.substr(0, par_as_string.size()) == par_as_string) {
2868 mres.match_len = par_as_string.size();
2869 mres.match2end = str.size();
2874 size_t pos = str.find(par_as_string_nolead);
2875 if (pos != string::npos) {
2876 mres.match_len = par_as_string.size();
2877 mres.match2end = str.size() - pos;
2886 MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
2888 MatchResult mres = findAux(cur, len, at_begin);
2889 int res = mres.match_len;
2891 "res=" << res << ", at_begin=" << at_begin
2892 << ", matchword=" << opt.matchword
2893 << ", inTexted=" << cur.inTexted());
2894 if (res == 0 || !at_begin || !opt.matchword || !cur.inTexted())
2896 if ((len > 0) && (res < len)) {
2900 Paragraph const & par = cur.paragraph();
2901 bool ws_left = (cur.pos() > 0)
2902 ? par.isWordSeparator(cur.pos() - 1)
2904 bool ws_right = (cur.pos() + len < par.size())
2905 ? par.isWordSeparator(cur.pos() + len)
2908 "cur.pos()=" << cur.pos() << ", res=" << res
2909 << ", separ: " << ws_left << ", " << ws_right
2912 if (ws_left && ws_right) {
2913 // Check for word separators inside the found 'word'
2914 for (int i = 0; i < len; i++) {
2915 if (par.isWordSeparator(cur.pos() + i)) {
2927 string MatchStringAdv::normalize(docstring const & s, bool hack_braces) const
2930 if (! opt.casesensitive)
2931 t = lyx::to_utf8(lowercase(s));
2933 t = lyx::to_utf8(s);
2934 // Remove \n at begin
2935 while (!t.empty() && t[0] == '\n')
2938 while (!t.empty() && t[t.size() - 1] == '\n')
2939 t = t.substr(0, t.size() - 1);
2941 // Handle all other '\n'
2942 while ((pos = t.find("\n")) != string::npos) {
2943 if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) {
2945 if (isAlnumASCII(t[pos+1])) {
2946 t.replace(pos-2, 3, " ");
2949 t.replace(pos-2, 3, "");
2952 else if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) {
2953 // '\n' adjacent to non-alpha-numerics, discard
2954 t.replace(pos, 1, "");
2957 // Replace all other \n with spaces
2958 t.replace(pos, 1, " ");
2961 // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify
2962 // Kornel: Added textsl, textsf, textit, texttt and noun
2963 // + allow to seach for colored text too
2964 LYXERR(Debug::FIND, "Removing stale empty \\emph{}, \\textbf{}, \\*section{} macros from: " << t);
2965 while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", ""))
2966 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
2967 while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
2968 LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
2970 while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
2971 // FIXME - check what preceeds the brace
2973 if (opt.ignoreformat)
2974 while (regex_replace(t, t, "\\{", "_x_<")
2975 || regex_replace(t, t, "\\}", "_x_>"))
2976 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
2978 while (regex_replace(t, t, "\\\\\\{", "_x_<")
2979 || regex_replace(t, t, "\\\\\\}", "_x_>"))
2980 LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
2987 docstring stringifyFromCursor(DocIterator const & cur, int len)
2989 LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur);
2990 if (cur.inTexted()) {
2991 Paragraph const & par = cur.paragraph();
2992 // TODO what about searching beyond/across paragraph breaks ?
2993 // TODO Try adding a AS_STR_INSERTS as last arg
2994 pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ?
2995 int(par.size()) : cur.pos() + len;
2996 OutputParams runparams(&cur.buffer()->params().encoding());
2997 runparams.nice = true;
2998 runparams.flavor = OutputParams::XETEX;
2999 runparams.linelen = 10000; //lyxrc.plaintext_linelen;
3000 // No side effect of file copying and image conversion
3001 runparams.dryrun = true;
3002 runparams.for_search = true;
3003 LYXERR(Debug::FIND, "Stringifying with cur: "
3004 << cur << ", from pos: " << cur.pos() << ", end: " << end);
3005 return par.asString(cur.pos(), end,
3006 AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
3008 } else if (cur.inMathed()) {
3009 CursorSlice cs = cur.top();
3010 MathData md = cs.cell();
3011 MathData::const_iterator it_end =
3012 (( len == -1 || cs.pos() + len > int(md.size()))
3014 : md.begin() + cs.pos() + len );
3016 for (MathData::const_iterator it = md.begin() + cs.pos();
3019 docstring s = asString(md2);
3020 LYXERR(Debug::FIND, "Stringified math: '" << s << "'");
3023 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3028 /** Computes the LaTeX export of buf starting from cur and ending len positions
3029 * after cur, if len is positive, or at the paragraph or innermost inset end
3032 docstring latexifyFromCursor(DocIterator const & cur, int len)
3034 LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur);
3035 LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow="
3036 << cur.lastrow() << ", cur.lastcol=" << cur.lastcol());
3037 Buffer const & buf = *cur.buffer();
3039 odocstringstream ods;
3041 OutputParams runparams(&buf.params().encoding());
3042 runparams.nice = false;
3043 runparams.flavor = OutputParams::XETEX;
3044 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3045 // No side effect of file copying and image conversion
3046 runparams.dryrun = true;
3047 runparams.for_search = true;
3049 if (cur.inTexted()) {
3050 // @TODO what about searching beyond/across paragraph breaks ?
3051 pos_type endpos = cur.paragraph().size();
3052 if (len != -1 && endpos > cur.pos() + len)
3053 endpos = cur.pos() + len;
3054 TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams,
3055 string(), cur.pos(), endpos);
3056 string s = lyx::to_utf8(ods.str());
3057 LYXERR(Debug::FIND, "Latexified +modified text: '" << s << "'");
3058 return(lyx::from_utf8(s));
3059 } else if (cur.inMathed()) {
3060 // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly
3061 for (int s = cur.depth() - 1; s >= 0; --s) {
3062 CursorSlice const & cs = cur[s];
3063 if (cs.asInsetMath() && cs.asInsetMath()->asHullInset()) {
3065 cs.asInsetMath()->asHullInset()->header_write(ws);
3070 CursorSlice const & cs = cur.top();
3071 MathData md = cs.cell();
3072 MathData::const_iterator it_end =
3073 ((len == -1 || cs.pos() + len > int(md.size()))
3075 : md.begin() + cs.pos() + len);
3077 for (MathData::const_iterator it = md.begin() + cs.pos();
3081 ods << asString(md2);
3082 // Retrieve the math environment type, and add '$' or '$]'
3083 // or others (\end{equation}) accordingly
3084 for (int s = cur.depth() - 1; s >= 0; --s) {
3085 CursorSlice const & cs2 = cur[s];
3086 InsetMath * inset = cs2.asInsetMath();
3087 if (inset && inset->asHullInset()) {
3089 inset->asHullInset()->footer_write(ws);
3093 LYXERR(Debug::FIND, "Latexified math: '" << lyx::to_utf8(ods.str()) << "'");
3095 LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur);
3101 /** Finalize an advanced find operation, advancing the cursor to the innermost
3102 ** position that matches, plus computing the length of the matching text to
3105 int findAdvFinalize(DocIterator & cur, MatchStringAdv const & match)
3107 // Search the foremost position that matches (avoids find of entire math
3108 // inset when match at start of it)
3110 DocIterator old_cur(cur.buffer());
3112 LYXERR(Debug::FIND, "Forwarding one step (searching for innermost match)");
3116 } while (cur && cur.depth() > d && match(cur).match_len > 0);
3118 int max_match = match(cur).match_len; /* match valid only if not searching whole words */
3119 if (max_match <= 0) return 0;
3120 LYXERR(Debug::FIND, "Ok");
3122 // Compute the match length
3124 if (cur.pos() + len > cur.lastpos())
3126 if (match.opt.matchword) {
3127 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3128 while (cur.pos() + len <= cur.lastpos() && match(cur, len).match_len <= 0) {
3130 LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
3132 // Length of matched text (different from len param)
3133 int old_match = match(cur, len).match_len;
3137 // Greedy behaviour while matching regexps
3138 while ((new_match = match(cur, len + 1).match_len) > old_match) {
3140 old_match = new_match;
3141 LYXERR(Debug::FIND, "verifying match with len = " << len);
3148 int maxl = cur.lastpos() - cur.pos();
3149 // Greedy behaviour while matching regexps
3150 while (maxl > minl) {
3151 int actual_match = match(cur, len).match_len;
3152 if (actual_match >= max_match) {
3153 // actual_match > max_match _can_ happen,
3154 // if the search area splits
3155 // some following word so that the regex
3156 // (e.g. 'r.*r\b' matches 'r' from the middle of the
3158 // This means, the len value is too big
3160 len = (int)((maxl + minl)/2);
3163 // (actual_match < max_match)
3165 len = (int)((maxl + minl)/2);
3169 // Search for real start of matched characters
3174 } while (cur.depth() > old_cur.depth()); /* Skip inner insets */
3175 if (cur.depth() < old_cur.depth()) {
3177 LYXERR0("cur.depth() < old_cur.depth(), this should never happen");
3180 if (cur.pos() != old_cur.pos()) {
3181 // OK, forwarded 1 pos in actual inset
3182 actual_match = match(cur, len-1).match_len;
3183 if (actual_match == max_match) {
3184 // Ha, got it! The shorter selection has the same match length
3189 // OK, the shorter selection matches less chars, revert to previous value
3195 LYXERR0("cur.pos() == old_cur.pos(), this should never happen");
3196 actual_match = match(cur, len).match_len;
3197 if (actual_match == max_match)
3207 int findForwardAdv(DocIterator & cur, MatchStringAdv & match)
3211 while (!theApp()->longOperationCancelled() && cur) {
3212 LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
3213 MatchResult mres = match(cur, -1, false);
3214 int match_len = mres.match_len;
3215 LYXERR(Debug::FIND, "match_len: " << match_len);
3216 if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
3217 LYXERR0("BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
3220 if (match_len > 0) {
3221 // Try to find the begin of searched string
3222 int increment = mres.pos/2;
3223 while (mres.pos > 5 && (increment > 5)) {
3224 DocIterator old_cur = cur;
3225 for (int i = 0; i < increment && cur; cur.forwardPos(), i++) {
3232 MatchResult mres2 = match(cur, -1, false);
3233 if ((mres2.match2end < mres.match2end) ||
3234 (mres2.match_len < mres.match_len)) {
3241 if (increment > mres.pos/2)
3242 increment = mres.pos/2;
3246 int match_len_zero_count = 0;
3247 for (int i = 0; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) {
3249 int remaining_len = match(cur, -1, false).match_len;
3250 if (remaining_len <= 0) {
3251 // Apparently the searched string is not in the remaining part
3258 LYXERR(Debug::FIND, "Advancing cur: " << cur);
3259 int match_len3 = match(cur, 1).match_len;
3262 int match_len2 = match(cur).match_len;
3263 LYXERR(Debug::FIND, "match_len2: " << match_len2);
3264 if (match_len2 > 0) {
3265 // Sometimes in finalize we understand it wasn't a match
3266 // and we need to continue the outest loop
3267 int len = findAdvFinalize(cur, match);
3272 if (match_len2 >= 0) {
3273 if (match_len2 == 0)
3274 match_len_zero_count++;
3276 match_len_zero_count = 0;
3279 if (++match_len_zero_count > 3) {
3280 LYXERR(Debug::FIND, "match_len2_zero_count: " << match_len_zero_count << ", match_len was " << match_len);
3281 match_len_zero_count = 0;
3289 if (match_len >= 0 && cur.pit() < cur.lastpit()) {
3290 LYXERR(Debug::FIND, "Advancing par: cur=" << cur);
3293 // This should exit nested insets, if any, or otherwise undefine the currsor.
3294 cur.pos() = cur.lastpos();
3295 LYXERR(Debug::FIND, "Advancing pos: cur=" << cur);
3303 /// Find the most backward consecutive match within same paragraph while searching backwards.
3304 int findMostBackwards(DocIterator & cur, MatchStringAdv const & match)
3306 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3307 DocIterator tmp_cur = cur;
3308 int len = findAdvFinalize(tmp_cur, match);
3309 Inset & inset = cur.inset();
3310 for (; cur != cur_begin; cur.backwardPos()) {
3311 LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
3312 DocIterator new_cur = cur;
3313 new_cur.backwardPos();
3314 if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
3316 int new_len = findAdvFinalize(new_cur, match);
3321 LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur);
3327 int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match)
3331 // Backup of original position
3332 DocIterator cur_begin = doc_iterator_begin(cur.buffer());
3333 if (cur == cur_begin)
3336 DocIterator cur_orig(cur);
3337 bool pit_changed = false;
3340 bool found_match = (match(cur, -1, false).match_len > 0);
3344 cur.pos() = cur.lastpos();
3346 cur.pos() = cur_orig.pos();
3347 LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
3348 DocIterator cur_prev_iter;
3350 found_match = (match(cur).match_len > 0);
3351 LYXERR(Debug::FIND, "findBackAdv3: found_match="
3352 << found_match << ", cur: " << cur);
3354 return findMostBackwards(cur, match);
3356 // Stop if begin of document reached
3357 if (cur == cur_begin)
3359 cur_prev_iter = cur;
3363 if (cur == cur_begin)
3370 } while (!theApp()->longOperationCancelled());
3378 docstring stringifyFromForSearch(FindAndReplaceOptions const & opt,
3379 DocIterator const & cur, int len)
3381 if (cur.pos() < 0 || cur.pos() > cur.lastpos())
3383 if (!opt.ignoreformat)
3384 return latexifyFromCursor(cur, len);
3386 return stringifyFromCursor(cur, len);
3390 FindAndReplaceOptions::FindAndReplaceOptions(
3391 docstring const & find_buf_name, bool casesensitive,
3392 bool matchword, bool forward, bool expandmacros, bool ignoreformat,
3393 docstring const & repl_buf_name, bool keep_case,
3394 SearchScope scope, SearchRestriction restr)
3395 : find_buf_name(find_buf_name), casesensitive(casesensitive), matchword(matchword),
3396 forward(forward), expandmacros(expandmacros), ignoreformat(ignoreformat),
3397 repl_buf_name(repl_buf_name), keep_case(keep_case), scope(scope), restr(restr)
3405 /** Check if 'len' letters following cursor are all non-lowercase */
3406 static bool allNonLowercase(Cursor const & cur, int len)
3408 pos_type beg_pos = cur.selectionBegin().pos();
3409 pos_type end_pos = cur.selectionBegin().pos() + len;
3410 if (len > cur.lastpos() + 1 - beg_pos) {
3411 LYXERR(Debug::FIND, "This should not happen, more debug needed");
3412 len = cur.lastpos() + 1 - beg_pos;
3413 end_pos = beg_pos + len;
3415 for (pos_type pos = beg_pos; pos != end_pos; ++pos)
3416 if (isLowerCase(cur.paragraph().getChar(pos)))
3422 /** Check if first letter is upper case and second one is lower case */
3423 static bool firstUppercase(Cursor const & cur)
3426 pos_type pos = cur.selectionBegin().pos();
3427 if (pos >= cur.lastpos() - 1) {
3428 LYXERR(Debug::FIND, "No upper-case at cur: " << cur);
3431 ch1 = cur.paragraph().getChar(pos);
3432 ch2 = cur.paragraph().getChar(pos + 1);
3433 bool result = isUpperCase(ch1) && isLowerCase(ch2);
3434 LYXERR(Debug::FIND, "firstUppercase(): "
3435 << "ch1=" << ch1 << "(" << char(ch1) << "), ch2="
3436 << ch2 << "(" << char(ch2) << ")"
3437 << ", result=" << result << ", cur=" << cur);
3442 /** Make first letter of supplied buffer upper-case, and the rest lower-case.
3444 ** \fixme What to do with possible further paragraphs in replace buffer ?
3446 static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case)
3448 ParagraphList::iterator pit = buffer.paragraphs().begin();
3449 LASSERT(pit->size() >= 1, /**/);
3450 pos_type right = pos_type(1);
3451 pit->changeCase(buffer.params(), pos_type(0), right, first_case);
3452 right = pit->size();
3453 pit->changeCase(buffer.params(), pos_type(1), right, others_case);
3459 static void findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
3461 Cursor & cur = bv->cursor();
3462 if (opt.repl_buf_name == docstring()
3463 || theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true) == 0
3464 || theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3467 DocIterator sel_beg = cur.selectionBegin();
3468 DocIterator sel_end = cur.selectionEnd();
3469 if (&sel_beg.inset() != &sel_end.inset()
3470 || sel_beg.pit() != sel_end.pit()
3471 || sel_beg.idx() != sel_end.idx())
3473 int sel_len = sel_end.pos() - sel_beg.pos();
3474 LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end
3475 << ", sel_len: " << sel_len << endl);
3478 LASSERT(sel_len > 0, return);
3480 if (!matchAdv(sel_beg, sel_len).match_len)
3483 // Build a copy of the replace buffer, adapted to the KeepCase option
3484 Buffer & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true);
3486 repl_buffer_orig.write(oss);
3487 string lyx = oss.str();
3488 Buffer repl_buffer("", false);
3489 repl_buffer.setUnnamed(true);
3490 LASSERT(repl_buffer.readString(lyx), return);
3491 if (opt.keep_case && sel_len >= 2) {
3492 LYXERR(Debug::FIND, "keep_case true: cur.pos()=" << cur.pos() << ", sel_len=" << sel_len);
3493 if (cur.inTexted()) {
3494 if (firstUppercase(cur))
3495 changeFirstCase(repl_buffer, text_uppercase, text_lowercase);
3496 else if (allNonLowercase(cur, sel_len))
3497 changeFirstCase(repl_buffer, text_uppercase, text_uppercase);
3500 cap::cutSelection(cur, false);
3501 if (cur.inTexted()) {
3502 repl_buffer.changeLanguage(
3503 repl_buffer.language(),
3504 cur.getFont().language());
3505 LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer");
3506 LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl);
3507 cap::pasteParagraphList(cur, repl_buffer.paragraphs(),
3508 repl_buffer.params().documentClassPtr(),
3509 bv->buffer().errorList("Paste"));
3510 LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl);
3511 sel_len = repl_buffer.paragraphs().begin()->size();
3512 } else if (cur.inMathed()) {
3513 odocstringstream ods;
3515 OutputParams runparams(&repl_buffer.params().encoding());
3516 runparams.nice = false;
3517 runparams.flavor = OutputParams::XETEX;
3518 runparams.linelen = 8000; //lyxrc.plaintext_linelen;
3519 runparams.dryrun = true;
3520 TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
3521 //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
3522 docstring repl_latex = ods.str();
3523 LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'");
3525 (void)regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1");
3526 (void)regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1");
3527 repl_latex = from_utf8(s);
3528 LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth());
3529 MathData ar(cur.buffer());
3530 asArray(repl_latex, ar, Parse::NORMAL);
3532 sel_len = ar.size();
3533 LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3535 if (cur.pos() >= sel_len)
3536 cur.pos() -= sel_len;
3539 LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len);
3540 bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward);
3541 bv->processUpdateFlags(Update::Force);
3545 /// Perform a FindAdv operation.
3546 bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt)
3551 // e.g., when invoking word-findadv from mini-buffer wither with
3552 // wrong options syntax or before ever opening advanced F&R pane
3553 if (theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true) == 0)
3557 MatchStringAdv matchAdv(bv->buffer(), opt);
3558 int length = bv->cursor().selectionEnd().pos() - bv->cursor().selectionBegin().pos();
3560 bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward);
3561 findAdvReplace(bv, opt, matchAdv);
3564 match_len = findForwardAdv(cur, matchAdv);
3566 match_len = findBackwardsAdv(cur, matchAdv);
3568 // This may only be raised by lyx::regex()
3569 bv->message(_("Invalid regular expression!"));
3573 if (match_len == 0) {
3574 bv->message(_("Match not found!"));
3578 bv->message(_("Match found!"));
3580 LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << match_len);
3581 bv->putSelectionAt(cur, match_len, !opt.forward);
3587 ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt)
3589 os << to_utf8(opt.find_buf_name) << "\nEOSS\n"
3590 << opt.casesensitive << ' '
3591 << opt.matchword << ' '
3592 << opt.forward << ' '
3593 << opt.expandmacros << ' '
3594 << opt.ignoreformat << ' '
3595 << to_utf8(opt.repl_buf_name) << "\nEOSS\n"
3596 << opt.keep_case << ' '
3597 << int(opt.scope) << ' '
3600 LYXERR(Debug::FIND, "built: " << os.str());
3606 istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt)
3608 LYXERR(Debug::FIND, "parsing");
3612 while (line != "EOSS") {
3616 if (is.eof()) // Tolerate malformed request
3620 LYXERR(Debug::FIND, "file_buf_name: '" << s << "'");
3621 opt.find_buf_name = from_utf8(s);
3622 is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat;
3623 is.get(); // Waste space before replace string
3626 while (line != "EOSS") {
3630 if (is.eof()) // Tolerate malformed request
3634 LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'");
3635 opt.repl_buf_name = from_utf8(s);
3636 is >> opt.keep_case;
3639 opt.scope = FindAndReplaceOptions::SearchScope(i);
3641 opt.restr = FindAndReplaceOptions::SearchRestriction(i);
3643 LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' '
3644 << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case << ' '
3645 << opt.scope << ' ' << opt.restr);