X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2Flyxfind.cpp;h=b67f915d7c815d75f6fed65f7861d9f68f0e7437;hb=55a3dd7b346d29a52ba305a4558e9e380ef50f47;hp=ff2d41d952d51e2690df956232fd58825252dbdc;hpb=4304901faf55f352de8e56b18eb6e180166f50d4;p=lyx.git diff --git a/src/lyxfind.cpp b/src/lyxfind.cpp index ff2d41d952..b67f915d7c 100644 --- a/src/lyxfind.cpp +++ b/src/lyxfind.cpp @@ -18,36 +18,38 @@ #include "Buffer.h" #include "buffer_funcs.h" +#include "BufferList.h" #include "BufferParams.h" #include "BufferView.h" #include "Changes.h" #include "Cursor.h" #include "CutAndPaste.h" #include "FuncRequest.h" -#include "OutputParams.h" +#include "LyX.h" #include "output_latex.h" +#include "OutputParams.h" #include "Paragraph.h" #include "ParIterator.h" #include "TexRow.h" #include "Text.h" -#include "FuncRequest.h" -#include "LyXFunc.h" + +#include "frontends/Application.h" +#include "frontends/alert.h" #include "mathed/InsetMath.h" #include "mathed/InsetMathGrid.h" #include "mathed/InsetMathHull.h" #include "mathed/MathStream.h" - -#include "frontends/alert.h" +#include "mathed/MathSupport.h" #include "support/convert.h" #include "support/debug.h" #include "support/docstream.h" #include "support/gettext.h" -#include "support/lstrings.h" #include "support/lassert.h" +#include "support/lstrings.h" -#include +#include "support/regex.h" #include using namespace std; @@ -67,51 +69,55 @@ bool parse_bool(docstring & howto) } -class MatchString : public binary_function +class MatchString : public binary_function { public: MatchString(docstring const & str, bool cs, bool mw) - : str(str), cs(cs), mw(mw) + : str(str), case_sens(cs), whole_words(mw) {} // returns true if the specified string is at the specified position // del specifies whether deleted strings in ct mode will be considered - bool operator()(Paragraph const & par, pos_type pos, bool del = true) const + int operator()(Paragraph const & par, pos_type pos, bool del = true) const { - return par.find(str, cs, mw, pos, del); + return par.find(str, case_sens, whole_words, pos, del); } private: // search string docstring str; // case sensitive - bool cs; + bool case_sens; // match whole words only - bool mw; + bool whole_words; }; -bool findForward(DocIterator & cur, MatchString const & match, +int findForward(DocIterator & cur, MatchString const & match, bool find_del = true) { for (; cur; cur.forwardChar()) - if (cur.inTexted() && - match(cur.paragraph(), cur.pos(), find_del)) - return true; - return false; + if (cur.inTexted()) { + int len = match(cur.paragraph(), cur.pos(), find_del); + if (len > 0) + return len; + } + return 0; } -bool findBackwards(DocIterator & cur, MatchString const & match, +int findBackwards(DocIterator & cur, MatchString const & match, bool find_del = true) { while (cur) { cur.backwardChar(); - if (cur.inTexted() && - match(cur.paragraph(), cur.pos(), find_del)) - return true; + if (cur.inTexted()) { + int len = match(cur.paragraph(), cur.pos(), find_del); + if (len > 0) + return len; + } } - return false; + return 0; } @@ -132,7 +138,7 @@ bool findChange(DocIterator & cur, bool next) } -bool searchAllowed(BufferView * /*bv*/, docstring const & str) +bool searchAllowed(docstring const & str) { if (str.empty()) { frontend::Alert::error(_("Search error"), _("Search string is empty")); @@ -142,36 +148,38 @@ bool searchAllowed(BufferView * /*bv*/, docstring const & str) } -bool find(BufferView * bv, docstring const & searchstr, - bool cs, bool mw, bool fw, bool find_del = true) +bool findOne(BufferView * bv, docstring const & searchstr, + bool case_sens, bool whole, bool forward, bool find_del = true) { - if (!searchAllowed(bv, searchstr)) + if (!searchAllowed(searchstr)) return false; DocIterator cur = bv->cursor(); - MatchString const match(searchstr, cs, mw); + MatchString const match(searchstr, case_sens, whole); - bool found = fw ? findForward(cur, match, find_del) : + int match_len = forward ? findForward(cur, match, find_del) : findBackwards(cur, match, find_del); - if (found) - bv->putSelectionAt(cur, searchstr.length(), !fw); + if (match_len > 0) + bv->putSelectionAt(cur, match_len, !forward); - return found; + return match_len > 0; } int replaceAll(BufferView * bv, docstring const & searchstr, docstring const & replacestr, - bool cs, bool mw) + bool case_sens, bool whole) { Buffer & buf = bv->buffer(); - if (!searchAllowed(bv, searchstr) || buf.isReadonly()) + if (!searchAllowed(searchstr) || buf.isReadonly()) return 0; - MatchString const match(searchstr, cs, mw); + DocIterator cur_orig(bv->cursor()); + + MatchString const match(searchstr, case_sens, whole); int num = 0; int const rsize = replacestr.size(); @@ -179,12 +187,13 @@ int replaceAll(BufferView * bv, Cursor cur(*bv); cur.setCursor(doc_iterator_begin(&buf)); - while (findForward(cur, match, false)) { + int match_len = findForward(cur, match, false); + while (match_len > 0) { // Backup current cursor position and font. pos_type const pos = cur.pos(); Font const font = cur.paragraph().getFontSettings(buf.params(), pos); cur.recordUndo(); - int striked = ssize - cur.paragraph().eraseChars(pos, pos + ssize, + int striked = ssize - cur.paragraph().eraseChars(pos, pos + match_len, buf.params().trackChanges); cur.paragraph().insert(pos, replacestr, font, Change(buf.params().trackChanges ? @@ -192,57 +201,88 @@ int replaceAll(BufferView * bv, for (int i = 0; i < rsize + striked; ++i) cur.forwardChar(); ++num; + match_len = findForward(cur, match, false); } - buf.updateLabels(); bv->putSelectionAt(doc_iterator_begin(&buf), 0, false); - if (num) - buf.markDirty(); + + cur_orig.fixIfBroken(); + bv->setCursor(cur_orig); + return num; } -bool stringSelected(BufferView * bv, docstring & searchstr, - bool cs, bool mw, bool fw) +// the idea here is that we are going to replace the string that +// is selected IF it is the search string. +// if there is a selection, but it is not the search string, then +// we basically ignore it. (FIXME We ought to replace only within +// the selection.) +// if there is no selection, then: +// (i) if some search string has been provided, then we find it. +// (think of how the dialog works when you hit "replace" the +// first time.) +// (ii) if no search string has been provided, then we treat the +// word the cursor is in as the search string. (why? i have no +// idea.) but this only works in text? +// +// returns the number of replacements made (one, if any) and +// whether anything at all was done. +pair replaceOne(BufferView * bv, docstring searchstr, + docstring const & replacestr, bool case_sens, + bool whole, bool forward, bool findnext) { - // if nothing selected and searched string is empty, this - // means that we want to search current word at cursor position. - if (!bv->cursor().selection() && searchstr.empty()) { - bv->cursor().innerText()->selectWord(bv->cursor(), WHOLE_WORD); - searchstr = bv->cursor().selectionAsString(false); - return true; + Cursor & cur = bv->cursor(); + if (!cur.selection()) { + // no selection, non-empty search string: find it + if (!searchstr.empty()) { + findOne(bv, searchstr, case_sens, whole, forward); + return pair(true, 0); + } + // empty search string + if (!cur.inTexted()) + // bail in math + return pair(false, 0); + // select current word and treat it as the search string + cur.innerText()->selectWord(cur, WHOLE_WORD); + searchstr = cur.selectionAsString(false); } - - // if nothing selected or selection does not equal search - // string search and select next occurance and return - docstring const & str1 = searchstr; - docstring const str2 = bv->cursor().selectionAsString(false); - if ((cs && str1 != str2) || compare_no_case(str1, str2) != 0) { - find(bv, searchstr, cs, mw, fw); - return false; + + // if we still don't have a search string, report the error + // and abort. + if (!searchAllowed(searchstr)) + return pair(false, 0); + + bool have_selection = cur.selection(); + docstring const selected = cur.selectionAsString(false); + bool match = + case_sens ? searchstr == selected + : compare_no_case(searchstr, selected) == 0; + + // no selection or current selection is not search word: + // just find the search word + if (!have_selection || !match) { + findOne(bv, searchstr, case_sens, whole, forward); + return pair(true, 0); } - return true; -} + // we're now actually ready to replace. if the buffer is + // read-only, we can't, though. + if (bv->buffer().isReadonly()) + return pair(false, 0); + cap::replaceSelectionWithString(cur, replacestr, forward); + if (forward) { + cur.pos() += replacestr.length(); + LASSERT(cur.pos() <= cur.lastpos(), /* */); + } else { + cur.pos() -= replacestr.length(); + LASSERT(cur.pos() >= 0, /* */); + } + if (findnext) + findOne(bv, searchstr, case_sens, whole, forward, false); -int replace(BufferView * bv, docstring & searchstr, - docstring const & replacestr, bool cs, bool mw, bool fw) -{ - if (!stringSelected(bv, searchstr, cs, mw, fw)) - return 0; - - if (!searchAllowed(bv, searchstr) || bv->buffer().isReadonly()) - return 0; - - Cursor & cur = bv->cursor(); - cap::replaceSelectionWithString(cur, replacestr, fw); - bv->buffer().markDirty(); - find(bv, searchstr, cs, mw, fw, false); - bv->buffer().updateMacros(); - bv->processUpdateFlags(Update::Force | Update::FitCursor); - - return 1; + return pair(true, 1); } } // namespace anon @@ -262,7 +302,7 @@ docstring const find2string(docstring const & search, docstring const replace2string(docstring const & replace, docstring const & search, bool casesensitive, bool matchword, - bool all, bool forward) + bool all, bool forward, bool findnext) { odocstringstream ss; ss << replace << '\n' @@ -270,14 +310,15 @@ docstring const replace2string(docstring const & replace, << int(casesensitive) << ' ' << int(matchword) << ' ' << int(all) << ' ' - << int(forward); + << int(forward) << ' ' + << int(findnext); return ss.str(); } -bool find(BufferView * bv, FuncRequest const & ev) +bool lyxfind(BufferView * bv, FuncRequest const & ev) { - if (!bv || ev.action != LFUN_WORD_FIND) + if (!bv || ev.action() != LFUN_WORD_FIND) return false; //lyxerr << "find called, cmd: " << ev << endl; @@ -292,19 +333,20 @@ bool find(BufferView * bv, FuncRequest const & ev) bool matchword = parse_bool(howto); bool forward = parse_bool(howto); - return find(bv, search, casesensitive, matchword, forward); + return findOne(bv, search, casesensitive, matchword, forward); } -void replace(BufferView * bv, FuncRequest const & ev, bool has_deleted) +bool lyxreplace(BufferView * bv, + FuncRequest const & ev, bool has_deleted) { - if (!bv || ev.action != LFUN_WORD_REPLACE) - return; + if (!bv || ev.action() != LFUN_WORD_REPLACE) + return false; // data is of the form // " // - // " + // " docstring search; docstring rplc; docstring howto = split(ev.argument(), rplc, '\n'); @@ -314,35 +356,46 @@ void replace(BufferView * bv, FuncRequest const & ev, bool has_deleted) bool matchword = parse_bool(howto); bool all = parse_bool(howto); bool forward = parse_bool(howto); + bool findnext = howto.empty() ? true : parse_bool(howto); + + int replace_count = 0; + bool update = false; if (!has_deleted) { - int const replace_count = all - ? replaceAll(bv, search, rplc, casesensitive, matchword) - : replace(bv, search, rplc, casesensitive, matchword, forward); + if (all) { + replace_count = replaceAll(bv, search, rplc, casesensitive, matchword); + update = replace_count > 0; + } else { + pair rv = + replaceOne(bv, search, rplc, casesensitive, matchword, forward, findnext); + update = rv.first; + replace_count = rv.second; + } - Buffer & buf = bv->buffer(); - if (replace_count == 0) { + Buffer const & buf = bv->buffer(); + if (!update) { // emit message signal. buf.message(_("String not found!")); } else { - if (replace_count == 1) { - // emit message signal. + if (replace_count == 0) { + buf.message(_("String found.")); + } else if (replace_count == 1) { buf.message(_("String has been replaced.")); } else { - docstring str = convert(replace_count); - str += _(" strings have been replaced."); - // emit message signal. + docstring const str = + bformat(_("%1$d strings have been replaced."), replace_count); buf.message(str); } } - } else { + } else if (findnext) { // if we have deleted characters, we do not replace at all, but // rather search for the next occurence - if (find(bv, search, casesensitive, matchword, forward)) - bv->showCursor(); + if (findOne(bv, search, casesensitive, matchword, forward)) + update = true; else bv->message(_("String not found!")); } + return update; } @@ -363,7 +416,7 @@ bool findChange(BufferView * bv, bool next) if (bv->cursor().selection()) { // set the cursor at the beginning or at the end of the selection // before searching. Otherwise, the current change will be found. - if (next != (bv->cursor().top() > bv->cursor().anchor())) + if (next != (bv->cursor().top() > bv->cursor().normalAnchor())) bv->cursor().setCursorToAnchor(); } @@ -373,11 +426,18 @@ bool findChange(BufferView * bv, bool next) // clear the selection and search the other way around (see the end // of this function). This will avoid changes to be selected half. bool search_both_sides = false; - if (cur.pos() > 1) { - Change change_next_pos - = cur.paragraph().lookupChange(cur.pos()); + DocIterator tmpcur = cur; + // Leave math first + while (tmpcur.inMathed()) + tmpcur.pop_back(); + Change change_next_pos + = tmpcur.paragraph().lookupChange(tmpcur.pos()); + if (change_next_pos.changed() && cur.inMathed()) { + cur = tmpcur; + search_both_sides = true; + } else if (tmpcur.pos() > 0 && tmpcur.inTexted()) { Change change_prev_pos - = cur.paragraph().lookupChange(cur.pos() - 1); + = tmpcur.paragraph().lookupChange(tmpcur.pos() - 1); if (change_next_pos.isSimilarTo(change_prev_pos)) search_both_sides = true; } @@ -398,14 +458,14 @@ bool findChange(BufferView * bv, bool next) if (next) { for (; !tip.at_end(); tip.forwardPos()) { Change change = tip.paragraph().lookupChange(tip.pos()); - if (change != orig_change) + if (!change.isSimilarTo(orig_change)) break; } } else { for (; !tip.at_begin();) { tip.backwardPos(); Change change = tip.paragraph().lookupChange(tip.pos()); - if (change != orig_change) { + if (!change.isSimilarTo(orig_change)) { // take a step forward to correctly set the selection tip.forwardPos(); break; @@ -430,12 +490,11 @@ namespace { typedef vector > Escapes; /// A map of symbols and their escaped equivalent needed within a regex. +/// @note Beware of order Escapes const & get_regexp_escapes() { static Escapes escape_map; if (escape_map.empty()) { - escape_map.push_back(pair("\\", "\\\\")); - escape_map.push_back(pair("^", "\\^")); escape_map.push_back(pair("$", "\\$")); escape_map.push_back(pair("{", "\\{")); escape_map.push_back(pair("}", "\\}")); @@ -446,6 +505,9 @@ Escapes const & get_regexp_escapes() escape_map.push_back(pair("+", "\\+")); escape_map.push_back(pair("*", "\\*")); escape_map.push_back(pair(".", "\\.")); + escape_map.push_back(pair("\\", "(?:\\\\|\\\\backslash)")); + escape_map.push_back(pair("~", "(?:\\\\textasciitilde|\\\\sim)")); + escape_map.push_back(pair("^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\mathcircumflex)")); } return escape_map; } @@ -454,13 +516,30 @@ Escapes const & get_regexp_escapes() Escapes const & get_lyx_unescapes() { static Escapes escape_map; if (escape_map.empty()) { - escape_map.push_back(pair("{*}", "*")); - escape_map.push_back(pair("{[}", "[")); - escape_map.push_back(pair("\\$", "$")); - escape_map.push_back(pair("\\backslash{}", "\\")); + escape_map.push_back(pair("\\%", "%")); + escape_map.push_back(pair("\\mathcircumflex ", "^")); + escape_map.push_back(pair("\\mathcircumflex", "^")); + escape_map.push_back(pair("\\backslash ", "\\")); escape_map.push_back(pair("\\backslash", "\\")); + escape_map.push_back(pair("\\\\{", "_x_<")); + escape_map.push_back(pair("\\\\}", "_x_>")); escape_map.push_back(pair("\\sim ", "~")); - escape_map.push_back(pair("\\^", "^")); + escape_map.push_back(pair("\\sim", "~")); + } + return escape_map; +} + +/// A map of escapes turning a regexp matching text to one matching latex. +Escapes const & get_regexp_latex_escapes() { + static Escapes escape_map; + if (escape_map.empty()) { + escape_map.push_back(pair("\\\\", "(?:\\\\\\\\|\\\\backslash|\\\\textbackslash\\{\\})")); + escape_map.push_back(pair("(("(("\\[", "\\{\\[\\}")); + escape_map.push_back(pair("\\]", "\\{\\]\\}")); + escape_map.push_back(pair("\\^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\mathcircumflex)")); + escape_map.push_back(pair("%", "\\\\\\%")); } return escape_map; } @@ -477,7 +556,7 @@ string apply_escapes(string s, Escapes const & escape_map) unsigned int pos = 0; while (pos < s.length() && (pos = s.find(it->first, pos)) < s.length()) { s.replace(pos, it->first.length(), it->second); -// LYXERR(Debug::FIND, "After escape: " << s); + LYXERR(Debug::FIND, "After escape: " << s); pos += it->second.length(); // LYXERR(Debug::FIND, "pos: " << pos); } @@ -486,71 +565,61 @@ string apply_escapes(string s, Escapes const & escape_map) return s; } -/** Return the position of the closing brace matching the open one at s[pos], - ** or s.size() if not found. - **/ -size_t find_matching_brace(string const & s, size_t pos) -{ - LASSERT(s[pos] == '{', /* */); - int open_braces = 1; - for (++pos; pos < s.size(); ++pos) { - if (s[pos] == '\\') - ++pos; - else if (s[pos] == '{') - ++open_braces; - else if (s[pos] == '}') { - --open_braces; - if (open_braces == 0) - return pos; - } - } - return s.size(); -} -/// Within \regexp{} apply get_regex_escapes(), while outside apply get_lyx_unescapes(). -string escape_for_regex(string s) +/// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string), +/// while outside apply get_lyx_unescapes()+get_regexp_escapes(). +/// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well. +string escape_for_regex(string s, bool match_latex) { size_t pos = 0; while (pos < s.size()) { - size_t new_pos = s.find("\\regexp{", pos); - if (new_pos == string::npos) - new_pos = s.size(); - LYXERR(Debug::FIND, "new_pos: " << new_pos); - string t = apply_escapes(s.substr(pos, new_pos - pos), get_lyx_unescapes()); - LYXERR(Debug::FIND, "t : " << t); - t = apply_escapes(t, get_regexp_escapes()); - s.replace(pos, new_pos - pos, t); - new_pos = pos + t.size(); - LYXERR(Debug::FIND, "Regexp after escaping: " << s); - LYXERR(Debug::FIND, "new_pos: " << new_pos); - if (new_pos == s.size()) - break; - size_t end_pos = find_matching_brace(s, new_pos + 7); - LYXERR(Debug::FIND, "end_pos: " << end_pos); - t = apply_escapes(s.substr(new_pos + 8, end_pos - (new_pos + 8)), get_lyx_unescapes()); - LYXERR(Debug::FIND, "t : " << t); - if (end_pos == s.size()) { - s.replace(new_pos, end_pos - new_pos, t); - pos = s.size(); - LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s); - break; - } - s.replace(new_pos, end_pos + 1 - new_pos, t); + size_t new_pos = s.find("\\regexp{", pos); + if (new_pos == string::npos) + new_pos = s.size(); + LYXERR(Debug::FIND, "new_pos: " << new_pos); + string t = apply_escapes(s.substr(pos, new_pos - pos), get_lyx_unescapes()); + LYXERR(Debug::FIND, "t [lyx]: " << t); + t = apply_escapes(t, get_regexp_escapes()); + LYXERR(Debug::FIND, "t [rxp]: " << t); + s.replace(pos, new_pos - pos, t); + new_pos = pos + t.size(); + LYXERR(Debug::FIND, "Regexp after escaping: " << s); + LYXERR(Debug::FIND, "new_pos: " << new_pos); + if (new_pos == s.size()) + break; + // Might fail if \\endregexp{} is preceeded by unexpected stuff (weird escapes) + size_t end_pos = s.find("\\endregexp{}}", new_pos + 8); + LYXERR(Debug::FIND, "end_pos: " << end_pos); + t = s.substr(new_pos + 8, end_pos - (new_pos + 8)); + LYXERR(Debug::FIND, "t in regexp : " << t); + t = apply_escapes(t, get_lyx_unescapes()); + LYXERR(Debug::FIND, "t in regexp [lyx]: " << t); + if (match_latex) { + t = apply_escapes(t, get_regexp_latex_escapes()); + LYXERR(Debug::FIND, "t in regexp [ltx]: " << t); + } + if (end_pos == s.size()) { + s.replace(new_pos, end_pos - new_pos, t); + pos = s.size(); LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s); - pos = new_pos + t.size(); - LYXERR(Debug::FIND, "pos: " << pos); + break; + } + s.replace(new_pos, end_pos + 13 - new_pos, t); + LYXERR(Debug::FIND, "Regexp after \\regexp{...\\endregexp{}} removal: " << s); + pos = new_pos + t.size(); + LYXERR(Debug::FIND, "pos: " << pos); } return s; } -/// Wrapper for boost::regex_replace with simpler interface +/// Wrapper for lyx::regex_replace with simpler interface bool regex_replace(string const & s, string & t, string const & searchstr, string const & replacestr) { - boost::regex e(searchstr); + lyx::regex e(searchstr); ostringstream oss; ostream_iterator it(oss); - boost::regex_replace(it, s.begin(), s.end(), e, replacestr); + lyx::regex_replace(it, s.begin(), s.end(), e, replacestr); // tolerate t and s be references to the same variable bool rv = (s != oss.str()); t = oss.str(); @@ -566,7 +635,8 @@ bool regex_replace(string const & s, string & t, string const & searchstr, ** Number of open braces that must remain open at the end for the verification to succeed. **/ bool braces_match(string::const_iterator const & beg, - string::const_iterator const & end, int unmatched = 0) + string::const_iterator const & end, + int unmatched = 0) { int open_pars = 0; string::const_iterator it = beg; @@ -588,8 +658,10 @@ bool braces_match(string::const_iterator const & beg, } } if (open_pars != unmatched) { - LYXERR(Debug::FIND, "Found " << open_pars << " instead of " << unmatched << " unmatched open braces at the end of count"); - return false; + LYXERR(Debug::FIND, "Found " << open_pars + << " instead of " << unmatched + << " unmatched open braces at the end of count"); + return false; } LYXERR(Debug::FIND, "Braces match as expected"); return true; @@ -599,7 +671,7 @@ bool braces_match(string::const_iterator const & beg, **/ class MatchStringAdv { public: - MatchStringAdv(lyx::Buffer const & buf, FindAndReplaceOptions const & opt); + MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt); /** Tests if text starting at the supplied position matches with the one provided to the MatchStringAdv ** constructor as opt.search, under the opt.* options settings. @@ -615,11 +687,16 @@ public: public: /// buffer - lyx::Buffer const & buf; + lyx::Buffer * p_buf; + /// first buffer on which search was started + lyx::Buffer * const p_first_buf; /// options FindAndReplaceOptions const & opt; private: + /// Auxiliary find method (does not account for opt.matchword) + int findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const; + /** Normalize a stringified or latexified LyX paragraph. ** ** Normalize means: @@ -633,61 +710,149 @@ private: ** @todo Normalization should also expand macros, if the corresponding ** search option was checked. **/ - string normalize(docstring const & s) const; + string normalize(docstring const & s, bool hack_braces) const; // normalized string to search string par_as_string; // regular expression to use for searching - boost::regex regexp; + lyx::regex regexp; // same as regexp, but prefixed with a ".*" - boost::regex regexp2; + lyx::regex regexp2; + // leading format material as string + string lead_as_string; + // par_as_string after removal of lead_as_string + string par_as_string_nolead; // unmatched open braces in the search string/regexp int open_braces; // number of (.*?) subexpressions added at end of search regexp for closing // environments, math mode, styles, etc... int close_wildcards; + // Are we searching with regular expressions ? + bool use_regexp; }; -MatchStringAdv::MatchStringAdv(lyx::Buffer const & buf, FindAndReplaceOptions const & opt) - : buf(buf), opt(opt) +static docstring buffer_to_latex(Buffer & buffer) +{ + OutputParams runparams(&buffer.params().encoding()); + TexRow texrow; + odocstringstream ods; + otexstream os(ods, texrow); + runparams.nice = true; + runparams.flavor = OutputParams::LATEX; + runparams.linelen = 80; //lyxrc.plaintext_linelen; + // No side effect of file copying and image conversion + runparams.dryrun = true; + pit_type const endpit = buffer.paragraphs().size(); + for (pit_type pit = 0; pit != endpit; ++pit) { + TeXOnePar(buffer, buffer.text(), pit, os, runparams); + LYXERR(Debug::FIND, "searchString up to here: " << ods.str()); + } + return ods.str(); +} + + +static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt) { + docstring str; + if (!opt.ignoreformat) { + str = buffer_to_latex(buffer); + } else { + OutputParams runparams(&buffer.params().encoding()); + runparams.nice = true; + runparams.flavor = OutputParams::LATEX; + runparams.linelen = 100000; //lyxrc.plaintext_linelen; + runparams.dryrun = true; + for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) { + Paragraph const & par = buffer.paragraphs().at(pit); + LYXERR(Debug::FIND, "Adding to search string: '" + << par.stringify(pos_type(0), par.size(), + AS_STR_INSETS, runparams) + << "'"); + str += par.stringify(pos_type(0), par.size(), + AS_STR_INSETS, runparams); + } + } + return str; +} + + +/// Return separation pos between the leading material and the rest +static size_t identifyLeading(string const & s) { + string t = s; + // @TODO Support \item[text] + while (regex_replace(t, t, "^\\\\(emph|textbf|subsubsection|subsection|section|subparagraph|paragraph|part)\\*?\\{", "") + || regex_replace(t, t, "^\\$", "") + || regex_replace(t, t, "^\\\\\\[ ", "") + || regex_replace(t, t, "^\\\\item ", "") + || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\} ", "")) + LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'"); + return s.find(t); +} + + +// Remove trailing closure of math, macros and environments, so to catch parts of them. +static int identifyClosing(string & t) { + int open_braces = 0; + do { + LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'"); + if (regex_replace(t, t, "(.*[^\\\\])\\$\\'", "$1")) + continue; + if (regex_replace(t, t, "(.*[^\\\\]) \\\\\\]\\'", "$1")) + continue; + if (regex_replace(t, t, "(.*[^\\\\]) \\\\end\\{[a-zA-Z_]*\\*?\\}\\'", "$1")) + continue; + if (regex_replace(t, t, "(.*[^\\\\])\\}\\'", "$1")) { + ++open_braces; + continue; + } + break; + } while (true); + return open_braces; +} + + +MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt) + : p_buf(&buf), p_first_buf(&buf), opt(opt) { - par_as_string = normalize(opt.search); + Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true); + docstring const & ds = stringifySearchBuffer(find_buf, opt); + use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos; + // When using regexp, braces are hacked already by escape_for_regex() + par_as_string = normalize(ds, !use_regexp); open_braces = 0; close_wildcards = 0; - if (! opt.regexp) { - // Remove trailing closure of math, macros and environments, so to catch parts of them. - do { - LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'"); - if (regex_replace(par_as_string, par_as_string, "(.*)[[:blank:]]\\'", "$1")) - continue; - if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\]) ?\\$\\'", "$1")) - continue; - // @todo need to account for open square braces as well ? - if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\]) ?\\\\\\]\\'", "$1")) - continue; - if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\]) ?\\\\end\\{[a-zA-Z_]*\\}\\'", "$1")) - continue; - if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\]) ?\\}\\'", "$1")) { - ++open_braces; - continue; - } - break; - } while (true); + size_t lead_size = 0; + if (!opt.ignoreformat) { + lead_size = identifyLeading(par_as_string); + lead_as_string = par_as_string.substr(0, lead_size); + par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size); + } + + if (!use_regexp) { + open_braces = identifyClosing(par_as_string); + identifyClosing(par_as_string_nolead); LYXERR(Debug::FIND, "Open braces: " << open_braces); - LASSERT(braces_match(par_as_string.begin(), par_as_string.end(), open_braces), /* */); LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'"); } else { - par_as_string = escape_for_regex(par_as_string); + string lead_as_regexp; + if (lead_size > 0) { + // @todo No need to search for \regexp{} insets in leading material + lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat); + par_as_string = par_as_string_nolead; + LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'"); + LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'"); + } + par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat); // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them. LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'"); if ( // Insert .* before trailing '\$' ('$' has been escaped by escape_for_regex) regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\$)\\'", "$1(.*?)$2") // Insert .* before trailing '\\\]' ('\]' has been escaped by escape_for_regex) - || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\\\\\\\\\])\\'", "$1(.*?)$2") + || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])( \\\\\\\\\\\\\\])\\'", "$1(.*?)$2") // Insert .* before trailing '\\end\{...}' ('\end{...}' has been escaped by escape_for_regex) - || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\\\end\\\\\\{[a-zA-Z_]*\\\\\\})\\'", "$1(.*?)$2") + || regex_replace(par_as_string, par_as_string, + "(.*[^\\\\])( \\\\\\\\end\\\\\\{[a-zA-Z_]*)(\\\\\\*)?(\\\\\\})\\'", "$1(.*?)$2$3$4") // Insert .* before trailing '\}' ('}' has been escaped by escape_for_regex) || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\})\\'", "$1(.*?)$2") ) { @@ -696,47 +861,58 @@ MatchStringAdv::MatchStringAdv(lyx::Buffer const & buf, FindAndReplaceOptions co LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'"); LYXERR(Debug::FIND, "Open braces: " << open_braces); LYXERR(Debug::FIND, "Close .*? : " << close_wildcards); - LASSERT(braces_match(par_as_string.begin(), par_as_string.end(), open_braces), /* */); LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string); // If entered regexp must match at begin of searched string buffer - regexp = boost::regex(string("\\`") + par_as_string); + string regexp_str = string("\\`") + lead_as_regexp + par_as_string; + LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'"); + regexp = lyx::regex(regexp_str); + // If entered regexp may match wherever in searched string buffer - regexp2 = boost::regex(string("\\`.*") + par_as_string); + string regexp2_str = string("\\`.*") + lead_as_regexp + ".*" + par_as_string; + LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'"); + regexp2 = lyx::regex(regexp2_str); } } -int MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const +int MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const { docstring docstr = stringifyFromForSearch(opt, cur, len); LYXERR(Debug::FIND, "Matching against '" << lyx::to_utf8(docstr) << "'"); - string str = normalize(docstr); + string str = normalize(docstr, true); LYXERR(Debug::FIND, "After normalization: '" << str << "'"); - if (! opt.regexp) { + if (! use_regexp) { + LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='" << par_as_string << "', str='" << str << "'"); + LYXERR(Debug::FIND, "Searching in normal mode: lead_as_string='" << lead_as_string << "', par_as_string_nolead='" << par_as_string_nolead << "'"); if (at_begin) { + LYXERR(Debug::FIND, "size=" << par_as_string.size() << ", substr='" << str.substr(0, par_as_string.size()) << "'"); if (str.substr(0, par_as_string.size()) == par_as_string) return par_as_string.size(); } else { - size_t pos = str.find(par_as_string); + size_t pos = str.find(par_as_string_nolead); if (pos != string::npos) return par_as_string.size(); } } else { - // Try all possible regexp matches, until one that verifies the braces match test is found - boost::regex const *p_regexp = at_begin ? ®exp : ®exp2; - boost::sregex_iterator re_it(str.begin(), str.end(), *p_regexp); - boost::sregex_iterator re_it_end; + LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin); + // Try all possible regexp matches, + //until one that verifies the braces match test is found + regex const *p_regexp = at_begin ? ®exp : ®exp2; + sregex_iterator re_it(str.begin(), str.end(), *p_regexp); + sregex_iterator re_it_end; for (; re_it != re_it_end; ++re_it) { - boost::match_results const & m = *re_it; + match_results const & m = *re_it; // Check braces on the segment that matched the entire regexp expression, // plus the last subexpression, if a (.*?) was inserted in the constructor. - if (! braces_match(m[0].first, m[0].second, open_braces)) + if (!braces_match(m[0].first, m[0].second, open_braces)) return 0; - // Check braces on segments that matched all (.*?) subexpressions. - for (size_t i = 1; i < m.size(); ++i) - if (! braces_match(m[i].first, m[i].second)) + // Check braces on segments that matched all (.*?) subexpressions, + // except the last "padding" one inserted by lyx. + for (size_t i = 1; i < m.size() - 1; ++i) + if (!braces_match(m[i].first, m[i].second)) return false; - // Exclude from the returned match length any length due to close wildcards added at end of regexp + // Exclude from the returned match length any length + // due to close wildcards added at end of regexp if (close_wildcards == 0) return m[0].second - m[0].first; else @@ -747,7 +923,29 @@ int MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) } -string MatchStringAdv::normalize(docstring const & s) const +int MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const +{ + int res = findAux(cur, len, at_begin); + LYXERR(Debug::FIND, + "res=" << res << ", at_begin=" << at_begin << ", matchword=" << opt.matchword << ", inTexted=" << cur.inTexted()); + if (res == 0 || !at_begin || !opt.matchword || !cur.inTexted()) + return res; + Paragraph const & par = cur.paragraph(); + bool ws_left = cur.pos() > 0 ? + par.isWordSeparator(cur.pos() - 1) : true; + bool ws_right = cur.pos() + res < par.size() ? + par.isWordSeparator(cur.pos() + res) : true; + LYXERR(Debug::FIND, + "cur.pos()=" << cur.pos() << ", res=" << res + << ", separ: " << ws_left << ", " << ws_right + << endl); + if (ws_left && ws_right) + return res; + return 0; +} + + +string MatchStringAdv::normalize(docstring const & s, bool hack_braces) const { string t; if (! opt.casesensitive) @@ -765,9 +963,22 @@ string MatchStringAdv::normalize(docstring const & s) const while ((pos = t.find("\n")) != string::npos) t.replace(pos, 1, " "); // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify - LYXERR(Debug::FIND, "Removing stale empty \\emph{}, \\textbf{} macros from: " << t); - while (regex_replace(t, t, "\\\\(emph|textbf)(\\{\\})+", "")) + LYXERR(Debug::FIND, "Removing stale empty \\emph{}, \\textbf{}, \\*section{} macros from: " << t); + while (regex_replace(t, t, "\\\\(emph|textbf|subsubsection|subsection|section|subparagraph|paragraph|part)(\\{\\})+", "")) LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t); + + // FIXME - check what preceeds the brace + if (hack_braces) { + if (opt.ignoreformat) + while (regex_replace(t, t, "\\{", "_x_<") + || regex_replace(t, t, "\\}", "_x_>")) + LYXERR(Debug::FIND, "After {} replacement: '" << t << "'"); + else + while (regex_replace(t, t, "\\\\\\{", "_x_<") + || regex_replace(t, t, "\\\\\\}", "_x_>")) + LYXERR(Debug::FIND, "After {} replacement: '" << t << "'"); + } + return t; } @@ -779,7 +990,8 @@ docstring stringifyFromCursor(DocIterator const & cur, int len) Paragraph const & par = cur.paragraph(); // TODO what about searching beyond/across paragraph breaks ? // TODO Try adding a AS_STR_INSERTS as last arg - pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ? int(par.size()) : cur.pos() + len; + pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ? + int(par.size()) : cur.pos() + len; OutputParams runparams(&cur.buffer()->params().encoding()); odocstringstream os; runparams.nice = true; @@ -787,36 +999,41 @@ docstring stringifyFromCursor(DocIterator const & cur, int len) runparams.linelen = 100000; //lyxrc.plaintext_linelen; // No side effect of file copying and image conversion runparams.dryrun = true; - LYXERR(Debug::FIND, "Stringifying with cur: " << cur << ", from pos: " << cur.pos() << ", end: " << end); + LYXERR(Debug::FIND, "Stringifying with cur: " + << cur << ", from pos: " << cur.pos() << ", end: " << end); return par.stringify(cur.pos(), end, AS_STR_INSETS, runparams); } else if (cur.inMathed()) { - odocstringstream os; + docstring s; CursorSlice cs = cur.top(); MathData md = cs.cell(); - MathData::const_iterator it_end = ( ( len == -1 || cs.pos() + len > int(md.size()) ) ? md.end() : md.begin() + cs.pos() + len ); + MathData::const_iterator it_end = + ( ( len == -1 || cs.pos() + len > int(md.size()) ) + ? md.end() : md.begin() + cs.pos() + len ); for (MathData::const_iterator it = md.begin() + cs.pos(); it != it_end; ++it) - os << *it; - return os.str(); + s = s + asString(*it); + LYXERR(Debug::FIND, "Stringified math: '" << s << "'"); + return s; } LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur); return docstring(); } + /** Computes the LaTeX export of buf starting from cur and ending len positions * after cur, if len is positive, or at the paragraph or innermost inset end * if len is -1. */ - docstring latexifyFromCursor(DocIterator const & cur, int len) { LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur); LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow=" << cur.lastrow() << ", cur.lastcol=" << cur.lastcol()); Buffer const & buf = *cur.buffer(); - LASSERT(buf.isLatex(), /* */); + LASSERT(buf.params().isLatex(), /* */); TexRow texrow; odocstringstream ods; + otexstream os(ods, texrow); OutputParams runparams(&buf.params().encoding()); runparams.nice = false; runparams.flavor = OutputParams::LATEX; @@ -825,17 +1042,12 @@ docstring latexifyFromCursor(DocIterator const & cur, int len) runparams.dryrun = true; if (cur.inTexted()) { - // @TODO what about searching beyond/across paragraph breaks ? - ParagraphList::const_iterator pit = cur.innerText()->paragraphs().begin(); - for (int i = 0; i < cur.pit(); ++i) - ++pit; -// ParagraphList::const_iterator pit_end = pit; -// ++pit_end; -// lyx::latexParagraphs(buf, cur.innerText()->paragraphs(), ods, texrow, runparams, string(), pit, pit_end); - pos_type const endpos = (len == -1 || cur.pos() + len > int(pit->size())) - ? pit->size() : cur.pos() + len; - TeXOnePar(buf, *cur.innerText(), pit, ods, texrow, runparams, string(), - cur.pos(), endpos); + // @TODO what about searching beyond/across paragraph breaks ? + pos_type endpos = cur.paragraph().size(); + if (len != -1 && endpos > cur.pos() + len) + endpos = cur.pos() + len; + TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams, + string(), cur.pos(), endpos); LYXERR(Debug::FIND, "Latexified text: '" << lyx::to_utf8(ods.str()) << "'"); } else if (cur.inMathed()) { // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly @@ -853,14 +1065,7 @@ docstring latexifyFromCursor(DocIterator const & cur, int len) MathData::const_iterator it_end = ( ( len == -1 || cs.pos() + len > int(md.size()) ) ? md.end() : md.begin() + cs.pos() + len ); for (MathData::const_iterator it = md.begin() + cs.pos(); it != it_end; ++it) - ods << *it; - - // MathData md = cur.cell(); - // MathData::const_iterator it_end = ( ( len == -1 || cur.pos() + len > int(md.size()) ) ? md.end() : md.begin() + cur.pos() + len ); - // for (MathData::const_iterator it = md.begin() + cur.pos(); it != it_end; ++it) { - // MathAtom const & ma = *it; - // ma.nucleus()->latex(buf, ods, runparams); - // } + ods << asString(*it); // Retrieve the math environment type, and add '$' or '$]' // or others (\end{equation}) accordingly @@ -880,6 +1085,7 @@ docstring latexifyFromCursor(DocIterator const & cur, int len) return ods.str(); } + /** Finalize an advanced find operation, advancing the cursor to the innermost ** position that matches, plus computing the length of the matching text to ** be selected @@ -902,6 +1108,8 @@ int findAdvFinalize(DocIterator & cur, MatchStringAdv const & match) // Compute the match length int len = 1; + if (cur.pos() + len > cur.lastpos()) + return 0; LYXERR(Debug::FIND, "verifying unmatch with len = " << len); while (cur.pos() + len <= cur.lastpos() && match(cur, len) == 0) { ++len; @@ -921,127 +1129,124 @@ int findAdvFinalize(DocIterator & cur, MatchStringAdv const & match) /// Finds forward -int findForwardAdv(DocIterator & cur, MatchStringAdv const & match) +int findForwardAdv(DocIterator & cur, MatchStringAdv & match) { if (!cur) return 0; - int wrap_answer; - do { - while (cur && !match(cur, -1, false)) { - if (cur.pit() < cur.lastpit()) - cur.forwardPar(); - else { - cur.forwardPos(); + while (cur) { + LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur); + int match_len = match(cur, -1, false); + LYXERR(Debug::FIND, "match_len: " << match_len); + if (match_len) { + for (; cur; cur.forwardPos()) { + LYXERR(Debug::FIND, "Advancing cur: " << cur); + int match_len = match(cur); + LYXERR(Debug::FIND, "match_len: " << match_len); + if (match_len) { + // Sometimes in finalize we understand it wasn't a match + // and we need to continue the outest loop + int len = findAdvFinalize(cur, match); + if (len > 0) + return len; + } } + if (!cur) + return 0; } - for (; cur; cur.forwardPos()) { - if (match(cur)) - return findAdvFinalize(cur, match); + if (cur.pit() < cur.lastpit()) { + LYXERR(Debug::FIND, "Advancing par: cur=" << cur); + cur.forwardPar(); + } else { + // This should exit nested insets, if any, or otherwise undefine the currsor. + cur.pos() = cur.lastpos(); + LYXERR(Debug::FIND, "Advancing pos: cur=" << cur); + cur.forwardPos(); } - wrap_answer = frontend::Alert::prompt( - _("Wrap search ?"), - _("End of document reached while searching forward\n" - "\n" - "Continue searching from beginning ?"), - 0, 1, _("&Yes"), _("&No")); - cur.clear(); - cur.push_back(CursorSlice(match.buf.inset())); - } while (wrap_answer == 0); + } return 0; } + /// Find the most backward consecutive match within same paragraph while searching backwards. -void findMostBackwards(DocIterator & cur, MatchStringAdv const & match, int & len) { +int findMostBackwards(DocIterator & cur, MatchStringAdv const & match) +{ DocIterator cur_begin = doc_iterator_begin(cur.buffer()); - len = findAdvFinalize(cur, match); - if (cur != cur_begin) { - Inset & inset = cur.inset(); - int old_len; - DocIterator old_cur; - DocIterator dit2; - do { - old_cur = cur; - old_len = len; - cur.backwardPos(); - LYXERR(Debug::FIND, "findMostBackwards(): old_cur=" << old_cur << ", old_len=" << len << ", cur=" << cur); - dit2 = cur; - } while (cur != cur_begin && &cur.inset() == &inset && match(cur) - && (len = findAdvFinalize(dit2, match)) > old_len); - cur = old_cur; - len = old_len; + DocIterator tmp_cur = cur; + int len = findAdvFinalize(tmp_cur, match); + Inset & inset = cur.inset(); + for (; cur != cur_begin; cur.backwardPos()) { + LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur); + DocIterator new_cur = cur; + new_cur.backwardPos(); + if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur)) + break; + int new_len = findAdvFinalize(new_cur, match); + if (new_len == len) + break; + len = new_len; } - LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur); + LYXERR(Debug::FIND, "findMostBackwards(): exiting with cur=" << cur); + return len; } + /// Finds backwards -int findBackwardsAdv(DocIterator & cur, MatchStringAdv const & match) { +int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match) { if (! cur) return 0; - DocIterator cur_orig(cur); + // Backup of original position DocIterator cur_begin = doc_iterator_begin(cur.buffer()); -/* if (match(cur_orig)) */ -/* findAdvFinalize(cur_orig, match); */ - int wrap_answer = 0; + if (cur == cur_begin) + return 0; + cur.backwardPos(); + DocIterator cur_orig(cur); bool found_match; + bool pit_changed = false; + found_match = false; do { - bool pit_changed = false; - found_match = false; - // Search in current par occurs from start to end, but in next loop match is discarded if pos > original pos cur.pos() = 0; found_match = match(cur, -1, false); - LYXERR(Debug::FIND, "findBackAdv0: found_match=" << found_match << ", cur: " << cur); - while (cur != cur_begin) { - if (found_match) - break; - if (cur.pit() > 0) - --cur.pit(); - else - cur.backwardPos(); - pit_changed = true; - // Search in previous pars occurs from start to end - cur.pos() = 0; - found_match = match(cur, -1, false); - LYXERR(Debug::FIND, "findBackAdv1: found_match=" << found_match << ", cur: " << cur); - } - if (pit_changed) - cur.pos() = cur.lastpos(); - else - cur.pos() = cur_orig.pos(); - LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur); + if (found_match) { - while (true) { - found_match=match(cur); - LYXERR(Debug::FIND, "findBackAdv3: found_match=" << found_match << ", cur: " << cur); - if (found_match) { - int len; - findMostBackwards(cur, match, len); - if (&cur.inset() != &cur_orig.inset() || !(cur.pit()==cur_orig.pit()) || cur.pos() < cur_orig.pos()) - return len; - } + if (pit_changed) + cur.pos() = cur.lastpos(); + else + cur.pos() = cur_orig.pos(); + LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur); + DocIterator cur_prev_iter; + do { + found_match = match(cur); + LYXERR(Debug::FIND, "findBackAdv3: found_match=" + << found_match << ", cur: " << cur); + if (found_match) + return findMostBackwards(cur, match); + + // Stop if begin of document reached if (cur == cur_begin) break; + cur_prev_iter = cur; cur.backwardPos(); - }; + } while (true); } - wrap_answer = frontend::Alert::prompt( - _("Wrap search ?"), - _("Beginning of document reached while searching backwards\n" - "\n" - "Continue searching from end ?"), - 0, 1, _("&Yes"), _("&No")); - cur = doc_iterator_end(&match.buf); - cur.backwardPos(); - } while (wrap_answer == 0); - cur = cur_orig; + if (cur == cur_begin) + break; + if (cur.pit() > 0) + --cur.pit(); + else + cur.backwardPos(); + pit_changed = true; + } while (true); return 0; } + } // anonym namespace docstring stringifyFromForSearch(FindAndReplaceOptions const & opt, DocIterator const & cur, int len) { + LASSERT(cur.pos() >= 0 && cur.pos() <= cur.lastpos(), /* */); if (!opt.ignoreformat) return latexifyFromCursor(cur, len); else @@ -1049,38 +1254,162 @@ docstring stringifyFromForSearch(FindAndReplaceOptions const & opt, } -lyx::FindAndReplaceOptions::FindAndReplaceOptions(docstring const & search, bool casesensitive, +FindAndReplaceOptions::FindAndReplaceOptions(docstring const & find_buf_name, bool casesensitive, bool matchword, bool forward, bool expandmacros, bool ignoreformat, - bool regexp, docstring const & replace) - : search(search), casesensitive(casesensitive), matchword(matchword), + docstring const & repl_buf_name, bool keep_case, + SearchScope scope) + : find_buf_name(find_buf_name), casesensitive(casesensitive), matchword(matchword), forward(forward), expandmacros(expandmacros), ignoreformat(ignoreformat), - regexp(regexp), replace(replace) + repl_buf_name(repl_buf_name), keep_case(keep_case), scope(scope) +{ +} + + +namespace { + + +/** Check if 'len' letters following cursor are all non-lowercase */ +static bool allNonLowercase(DocIterator const & cur, int len) { + pos_type end_pos = cur.pos() + len; + for (pos_type pos = cur.pos(); pos != end_pos; ++pos) + if (isLowerCase(cur.paragraph().getChar(pos))) + return false; + return true; +} + + +/** Check if first letter is upper case and second one is lower case */ +static bool firstUppercase(DocIterator const & cur) { + char_type ch1, ch2; + if (cur.pos() >= cur.lastpos() - 1) { + LYXERR(Debug::FIND, "No upper-case at cur: " << cur); + return false; + } + ch1 = cur.paragraph().getChar(cur.pos()); + ch2 = cur.paragraph().getChar(cur.pos()+1); + bool result = isUpperCase(ch1) && isLowerCase(ch2); + LYXERR(Debug::FIND, "firstUppercase(): " + << "ch1=" << ch1 << "(" << char(ch1) << "), ch2=" + << ch2 << "(" << char(ch2) << ")" + << ", result=" << result << ", cur=" << cur); + return result; +} + + +/** Make first letter of supplied buffer upper-case, and the rest lower-case. + ** + ** \fixme What to do with possible further paragraphs in replace buffer ? + **/ +static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case) { + ParagraphList::iterator pit = buffer.paragraphs().begin(); + pos_type right = pos_type(1); + pit->changeCase(buffer.params(), pos_type(0), right, first_case); + right = pit->size() + 1; + pit->changeCase(buffer.params(), right, right, others_case); +} +} // anon namespace + +/// +static void findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv) { + Cursor & cur = bv->cursor(); + if (opt.repl_buf_name == docstring()) + return; + + DocIterator sel_beg = cur.selectionBegin(); + DocIterator sel_end = cur.selectionEnd(); + if (&sel_beg.inset() != &sel_end.inset() + || sel_beg.pit() != sel_end.pit()) + return; + int sel_len = sel_end.pos() - sel_beg.pos(); + LYXERR(Debug::FIND, "sel_beg: " << sel_beg << ", sel_end: " << sel_end + << ", sel_len: " << sel_len << endl); + if (sel_len == 0) + return; + LASSERT(sel_len > 0, /**/); + + if (!matchAdv(sel_beg, sel_len)) + return; + + // Build a copy of the replace buffer, adapted to the KeepCase option + Buffer & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true); + ostringstream oss; + repl_buffer_orig.write(oss); + string lyx = oss.str(); + Buffer repl_buffer("", false); + repl_buffer.setUnnamed(true); + LASSERT(repl_buffer.readString(lyx), /**/); + if (opt.keep_case && sel_len >= 2) { + if (cur.inTexted()) { + if (firstUppercase(cur)) + changeFirstCase(repl_buffer, text_uppercase, text_lowercase); + else if (allNonLowercase(cur, sel_len)) + changeFirstCase(repl_buffer, text_uppercase, text_uppercase); + } + } + cap::cutSelection(cur, false, false); + if (cur.inTexted()) { + repl_buffer.changeLanguage( + repl_buffer.language(), + cur.getFont().language()); + LYXERR(Debug::FIND, "Replacing by pasteParagraphList()ing repl_buffer"); + LYXERR(Debug::FIND, "Before pasteParagraphList() cur=" << cur << endl); + cap::pasteParagraphList(cur, repl_buffer.paragraphs(), + repl_buffer.params().documentClassPtr(), + bv->buffer().errorList("Paste")); + LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl); + sel_len = repl_buffer.paragraphs().begin()->size(); + } else if (cur.inMathed()) { + TexRow texrow; + odocstringstream ods; + otexstream os(ods, texrow); + OutputParams runparams(&repl_buffer.params().encoding()); + runparams.nice = false; + runparams.flavor = OutputParams::LATEX; + runparams.linelen = 8000; //lyxrc.plaintext_linelen; + runparams.dryrun = true; + TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams); + //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false); + docstring repl_latex = ods.str(); + LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'"); + string s; + regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1"); + regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1"); + repl_latex = from_utf8(s); + LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth()); + MathData ar(cur.buffer()); + asArray(repl_latex, ar, Parse::NORMAL); + cur.insert(ar); + sel_len = ar.size(); + LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len); + } + if (cur.pos() >= sel_len) + cur.pos() -= sel_len; + else + cur.pos() = 0; + LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len); + bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward); + bv->processUpdateFlags(Update::Force); + bv->buffer().updatePreviews(); } + /// Perform a FindAdv operation. bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt) { - DocIterator cur = bv->cursor(); + DocIterator cur; int match_len = 0; - if (opt.search.empty()) { - bv->message(_("Search text is empty!")); - return false; - } -// if (! bv->buffer()) { -// bv->message(_("No open document !")); -// return false; -// } - try { - MatchStringAdv const matchAdv(bv->buffer(), opt); + MatchStringAdv matchAdv(bv->buffer(), opt); + findAdvReplace(bv, opt, matchAdv); + cur = bv->cursor(); if (opt.forward) match_len = findForwardAdv(cur, matchAdv); else match_len = findBackwardsAdv(cur, matchAdv); } catch (...) { - // This may only be raised by boost::regex() + // This may only be raised by lyx::regex() bv->message(_("Invalid regular expression!")); return false; } @@ -1090,46 +1419,34 @@ bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt) return false; } - LYXERR(Debug::FIND, "Putting selection at " << cur << " with len: " << match_len); - bv->putSelectionAt(cur, match_len, ! opt.forward); bv->message(_("Match found!")); - if (opt.replace != docstring(from_utf8(LYX_FR_NULL_STRING))) { - dispatch(FuncRequest(LFUN_SELF_INSERT, opt.replace)); - } - - return true; -} - -void findAdv(BufferView * bv, FuncRequest const & ev) -{ - if (!bv || ev.action != LFUN_WORD_FINDADV) - return; + LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << match_len); + bv->putSelectionAt(cur, match_len, !opt.forward); - FindAndReplaceOptions opt; - istringstream iss(to_utf8(ev.argument())); - iss >> opt; - findAdv(bv, opt); + return true; } -ostringstream & operator<<(ostringstream & os, lyx::FindAndReplaceOptions const & opt) +ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt) { - os << to_utf8(opt.search) << "\nEOSS\n" + os << to_utf8(opt.find_buf_name) << "\nEOSS\n" << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' ' << opt.expandmacros << ' ' << opt.ignoreformat << ' ' - << opt.regexp << ' ' - << to_utf8(opt.replace) << "\nEOSS\n"; + << to_utf8(opt.repl_buf_name) << "\nEOSS\n" + << opt.keep_case << ' ' + << int(opt.scope); LYXERR(Debug::FIND, "built: " << os.str()); return os; } -istringstream & operator>>(istringstream & is, lyx::FindAndReplaceOptions & opt) + +istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt) { LYXERR(Debug::FIND, "parsing"); string s; @@ -1143,9 +1460,9 @@ istringstream & operator>>(istringstream & is, lyx::FindAndReplaceOptions & opt) break; getline(is, line); } - LYXERR(Debug::FIND, "searching for: '" << s << "'"); - opt.search = from_utf8(s); - is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.regexp; + LYXERR(Debug::FIND, "file_buf_name: '" << s << "'"); + opt.find_buf_name = from_utf8(s); + is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat; is.get(); // Waste space before replace string s = ""; getline(is, line); @@ -1157,10 +1474,14 @@ istringstream & operator>>(istringstream & is, lyx::FindAndReplaceOptions & opt) break; getline(is, line); } + LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'"); + opt.repl_buf_name = from_utf8(s); + is >> opt.keep_case; + int i; + is >> i; + opt.scope = FindAndReplaceOptions::SearchScope(i); LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' ' - << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.regexp); - LYXERR(Debug::FIND, "replacing with: '" << s << "'"); - opt.replace = from_utf8(s); + << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case); return is; }