X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2Flyxfind.cpp;h=c8fc40947e75643cb281589fc41b62b30508858c;hb=8f43f83ddfbe6adbb04a2ba86e62797c0a313324;hp=a4dd810e35d95ee48d9875ea28c219a803e4cea3;hpb=769df229dc2a95c0511f3e5510c772c02376c7f3;p=lyx.git diff --git a/src/lyxfind.cpp b/src/lyxfind.cpp index a4dd810e35..c8fc40947e 100644 --- a/src/lyxfind.cpp +++ b/src/lyxfind.cpp @@ -14,6 +14,7 @@ */ #include +#include #include "lyxfind.h" @@ -88,6 +89,8 @@ class IgnoreFormats { /// bool getShape() const { return ignoreShape_; } /// + bool getSize() const { return ignoreSize_; } + /// bool getUnderline() const { return ignoreUnderline_; } /// bool getMarkUp() const { return ignoreMarkUp_; } @@ -118,6 +121,8 @@ private: /// bool ignoreShape_ = false; /// + bool ignoreSize_ = true; + /// bool ignoreUnderline_ = false; /// bool ignoreMarkUp_ = false; @@ -166,6 +171,9 @@ void IgnoreFormats::setIgnoreFormat(string const & type, bool value, bool fromUs else if (type == "shape") { ignoreShape_ = value; } + else if (type == "size") { + ignoreSize_ = value; + } else if (type == "family") { ignoreFamily_ = value; } @@ -805,10 +813,48 @@ namespace { typedef vector > Escapes; +static string getRegexSpaceCount(int count) +{ + if (count > 0) { + if (count > 1) + return "\\s{" + std::to_string(count) + "}"; + else + return "\\s"; + } + return ""; +} + string string2regex(string in) { - static std::regex specialChars { R"([-[\]{}()*+?.,\^$|#\s\$\\])" }; - string temp = std::regex_replace(in, specialChars, R"(\$&)" ); + static std::regex specialChars { R"([-[\]{}()*+?.,\^$|#\$\\])" }; + string tempx = std::regex_replace(in, specialChars, R"(\$&)" ); + // Special handling for ' ' + string temp(""); + int blanks = 0; + for (unsigned i = 0; i < tempx.size(); i++) { + if (tempx[i] == ' ' || tempx[i] == '~' ) { + // normal blanks + blanks++; + } + else if ((tempx[i] == '\302' && tempx[i+1] == '\240') + || (tempx[i] == '\342' && tempx[i+1] == '\200')) { + // protected space + // thin space + blanks++; + i++; + } + else { + if (blanks > 0) { + temp += getRegexSpaceCount(blanks); + } + temp += tempx[i]; + blanks = 0; + } + } + if (blanks > 0) { + temp += getRegexSpaceCount(blanks); + } + string temp2(""); size_t lastpos = 0; size_t fl_pos = 0; @@ -838,7 +884,8 @@ string correctRegex(string t, bool withformat) * and \{, \}, \[, \] => {, }, [, ] */ string s(""); - regex wordre("(\\\\)*(\\\\(([A-Za-z]+|[\\{\\}%])( |\\{\\})?|[\\[\\]\\{\\}]))"); + static std::regex wordre("(\\\\)*(\\\\(( |[A-Za-z]+|[\\{\\}%])( |\\{\\})?|[\\[\\]\\{\\}]))"); + static std::regex protectedSpace { R"(~)" }; size_t lastpos = 0; smatch sub; bool backslashed = false; @@ -846,6 +893,7 @@ string correctRegex(string t, bool withformat) buildAccentsMap(); //LYXERR0("correctRegex input '" << t << "'"); + int skip = 0; for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) { sub = *it; string replace; @@ -858,10 +906,14 @@ string correctRegex(string t, bool withformat) { // transforms '\backslash \{' into '\{' string next = t.substr(sub.position(2) + sub.str(2).length(), 2); - if ((next == "\\{") || (next == "\\}")) { + if ((next == "\\{") || (next == "\\}") || (next == "\\ ")) { replace = ""; backslashed = true; } + else if (withformat && next[0] == '$') { + replace = accents["lyxdollar"]; + skip = 1; // Skip following '$' + } } } else if (sub.str(4) == "mathcircumflex") @@ -873,6 +925,8 @@ string correctRegex(string t, bool withformat) replace = accents["braceleft"]; else if (sub.str(3) == "}") replace = accents["braceright"]; + else if (sub.str(3) == " ") + replace = "\\ "; else { // else part should not exist LASSERT(0, /**/); @@ -883,6 +937,8 @@ string correctRegex(string t, bool withformat) replace = "\\{"; else if (sub.str(3) == "}") replace = "\\}"; + else if (sub.str(3) == " ") + replace = "\\ "; else { // else part should not exist LASSERT(0, /**/); @@ -895,6 +951,8 @@ string correctRegex(string t, bool withformat) replace = "}"; else if (sub.str(4) == "%") replace = "%"; + else if (sub.str(4) == " ") + replace = " "; else { AccentsIterator it_ac = accents.find(sub.str(4)); if (it_ac == accents.end()) { @@ -906,14 +964,15 @@ string correctRegex(string t, bool withformat) } } if (lastpos < (size_t) sub.position(2)) - s += t.substr(lastpos, sub.position(2) - lastpos); + s += std::regex_replace(t.substr(lastpos, sub.position(2) - lastpos), protectedSpace, R"( )"); s += replace; - lastpos = sub.position(2) + sub.length(2); + lastpos = sub.position(2) + sub.length(2) + skip; + skip = 0; } if (lastpos == 0) - s = t; + s = std::regex_replace(t, protectedSpace, R"( )"); else if (lastpos < t.length()) - s += t.substr(lastpos, t.length() - lastpos); + s += std::regex_replace(t.substr(lastpos, t.length() - lastpos), protectedSpace, R"( )"); // Handle quotes in regex // substitute all '„', '“', '»', '«' with '"' // and all '‚', '‘', '›', '‹' with "\'" @@ -1008,13 +1067,19 @@ public: ** constructor as opt.search, under the opt.* options settings. ** ** @param at_begin - ** If set, then match is searched only against beginning of text starting at cur. - ** If unset, then match is searched anywhere in text starting at cur. + ** If set to MatchStringAdv::MatchFromStart, + ** then match is searched only against beginning of text starting at cur. + ** Otherwise the match is searched anywhere in text starting at cur. ** ** @return ** The length of the matching text, or zero if no match was found. **/ - MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const; + enum matchType { + MatchAnyPlace, + MatchFromStart + }; + string matchTypeAsString(matchType const x) const { return (x == MatchFromStart ? "MatchFromStart" : "MatchAnyPlace"); } + MatchResult operator()(DocIterator const & cur, int len, matchType at_begin) const; #if QTSEARCH bool regexIsValid; string regexError; @@ -1030,7 +1095,7 @@ public: private: /// Auxiliary find method (does not account for opt.matchword) - MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const; + MatchResult findAux(DocIterator const & cur, int len, matchType at_begin) const; void CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string = ""); /** Normalize a stringified or latexified LyX paragraph. @@ -1046,7 +1111,7 @@ private: ** @todo Normalization should also expand macros, if the corresponding ** search option was checked. **/ - string normalize(docstring const & s, bool ignore_fomat) const; + string convertLF2Space(docstring const & s, bool ignore_fomat) const; // normalized string to search string par_as_string; // regular expression to use for searching @@ -1091,7 +1156,7 @@ void MatchStringAdv::FillResults(MatchResult &found_mr) static void setFindParams(OutputParams &runparams) { runparams.flavor = Flavor::XeTeX; - runparams.use_polyglossia = true; + //runparams.use_polyglossia = true; runparams.linelen = 10000; //lyxrc.plaintext_linelen; // No side effect of file copying and image conversion runparams.dryrun = true; @@ -1106,11 +1171,11 @@ static docstring buffer_to_latex(Buffer & buffer) runparams.nice = true; setFindParams(runparams); if (ignoreFormats.getDeleted()) - runparams.for_search = OutputParams::SearchWithoutDeleted; + runparams.find_set_feature(OutputParams::SearchWithoutDeleted); else - runparams.for_search = OutputParams::SearchWithDeleted; + runparams.find_set_feature(OutputParams::SearchWithDeleted); if (ignoreFormats.getNonContent()) { - runparams.for_search |= OutputParams::SearchNonOutput; + runparams.find_add_feature(OutputParams::SearchNonOutput); } pit_type const endpit = buffer.paragraphs().size(); for (pit_type pit = 0; pit != endpit; ++pit) { @@ -1120,7 +1185,7 @@ static docstring buffer_to_latex(Buffer & buffer) return ods.str(); } -static string latexNamesToUtf8(docstring strIn) +static string latexNamesToUtf8(docstring strIn, bool withformat) { string addtmp = to_utf8(strIn); static regex const rmAcc("(\\\\)*(" @@ -1162,6 +1227,10 @@ static string latexNamesToUtf8(docstring strIn) add = addtmp; else if (addtmp.length() > lastpos) add += addtmp.substr(lastpos, addtmp.length() - lastpos); + if (!withformat) { + static std::regex repltilde { R"(~)" }; + add = std::regex_replace(add, repltilde, accents["lyxtilde"]); + } LYXERR(Debug::FINDVERBOSE, "Adding to search string: '" << add << "'"); return add; @@ -1180,20 +1249,20 @@ static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions co int option = AS_STR_INSETS |AS_STR_PLAINTEXT; if (ignoreFormats.getDeleted()) { option |= AS_STR_SKIPDELETE; - runparams.for_search = OutputParams::SearchWithoutDeleted; + runparams.find_set_feature(OutputParams::SearchWithoutDeleted); } else { - runparams.for_search = OutputParams::SearchWithDeleted; + runparams.find_set_feature(OutputParams::SearchWithDeleted); } if (ignoreFormats.getNonContent()) { - runparams.for_search |= OutputParams::SearchNonOutput; + runparams.find_add_feature(OutputParams::SearchNonOutput); } string t(""); for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) { Paragraph const & par = buffer.paragraphs().at(pit); string add = latexNamesToUtf8(par.asString(pos_type(0), par.size(), option, - &runparams)); + &runparams), !opt.ignoreformat); LYXERR(Debug::FINDVERBOSE, "Adding to search string: '" << add << "'"); t += add; @@ -1226,7 +1295,7 @@ static size_t identifyLeading(string const & s) || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\}", "")) ; LYXERR(Debug::FINDVERBOSE, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'"); - return s.find(t); + return s.size() - t.size(); } /* @@ -1366,8 +1435,8 @@ public: string par; int ignoreidx; static vector borders; - int depts[MAXOPENED]; - int closes[MAXOPENED]; + static vector depts; + static vector closes; int actualdeptindex; int previousNotIgnored(int) const; int nextNotIgnored(int) const; @@ -1379,6 +1448,7 @@ public: void removeAccents(); void setForDefaultLang(KeyInfo const & defLang) const; int findclosing(int start, int end, char up, char down, int repeat); + void removeInvalidClosings(void); void handleParentheses(int lastpos, bool closingAllowed); bool hasTitle; // Number of disabled language specs up @@ -1391,6 +1461,8 @@ public: }; vector Intervall::borders = vector(30); +vector Intervall::depts = vector(30); +vector Intervall::closes = vector(30); int Intervall::isOpeningPar(int pos) const { @@ -1424,6 +1496,8 @@ void Intervall::setForDefaultLang(KeyInfo const & defLang) const } } +#if 0 +// Not needed, because dpts and closes are now dynamically expanded static void checkDepthIndex(int val) { static int maxdepthidx = MAXOPENED-2; @@ -1437,6 +1511,7 @@ static void checkDepthIndex(int val) LYXERR(Debug::INFO, "maxdepthidx now " << val); } } +#endif #if 0 // Not needed, because borders are now dynamically expanded @@ -1888,7 +1963,6 @@ static void buildAccentsMap() accents["cdot"] = "·"; accents["textasciicircum"] = "^"; accents["mathcircumflex"] = "^"; - accents["sim"] = "~"; accents["guillemotright"] = "»"; accents["guillemotleft"] = "«"; accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15 @@ -1909,6 +1983,9 @@ static void buildAccentsMap() accents["lyxarrow"] = getutf8(0xf0020); accents["braceleft"] = getutf8(0xf0030); accents["braceright"] = getutf8(0xf0031); + accents["lyxtilde"] = getutf8(0xf0032); + accents["sim"] = getutf8(0xf0032); + accents["lyxdollar"] = getutf8(0xf0033); accents["backslash lyx"] = getutf8(0xf0010); // Used logos inserted with starting \backslash accents["backslash LyX"] = getutf8(0xf0010); accents["backslash tex"] = getutf8(0xf0011); @@ -1984,7 +2061,8 @@ void Intervall::removeAccents() if (accents.empty()) buildAccentsMap(); static regex const accre("\\\\(" - "([\\S]|[A-Za-z]+)\\{[^\\{\\}]+\\}" + "([\\S]|[A-Za-z]+)\\{[^\\\\\\{\\}]+\\}" + "|([\\S]|[A-Za-z]+)\\{\\\\[ij](math)?\\}" "|(" "(backslash ([lL]y[xX]|[tT]e[xX]|[lL]a[tT]e[xX]e?|lyxarrow))" "|[A-Za-z]+" @@ -2020,9 +2098,13 @@ void Intervall::removeAccents() void Intervall::handleOpenP(int i) { actualdeptindex++; + if ((size_t) actualdeptindex >= depts.size()) { + depts.resize(actualdeptindex + 30); + closes.resize(actualdeptindex + 30); + } depts[actualdeptindex] = i+1; closes[actualdeptindex] = -1; - checkDepthIndex(actualdeptindex); + // checkDepthIndex(actualdeptindex); } void Intervall::handleCloseP(int i, bool closingAllowed) @@ -2184,6 +2266,27 @@ int Intervall::findclosing(int start, int end, char up = '{', char down = '}', i return end; } +void Intervall::removeInvalidClosings(void) +{ + // this can happen, if there are deleted parts + int skip = 0; + int depth = 0; + for (unsigned i = 0; i < par.size(); i += 1 + skip) { + char c = par[i]; + skip = 0; + if (c == '\\') skip = 1; + else if (c == '{') + depth++; + else if (c == '}') { + if (depth == 0) { + addIntervall(i, i+1); + LYXERR(Debug::FINDVERBOSE, "removed invalid closing '}' at " << i); + } + else + --depth; + } + } +} class MathInfo { class MathEntry { public: @@ -2271,6 +2374,7 @@ void LatexInfo::buildEntries(bool isPatternString) static bool removeMathHull = false; interval_.removeAccents(); + interval_.removeInvalidClosings(); for (sregex_iterator itmath(interval_.par.begin(), interval_.par.end(), rmath), end; itmath != end; ++itmath) { submath = *itmath; @@ -2320,7 +2424,7 @@ void LatexInfo::buildEntries(bool isPatternString) } // Ignore language if there is math somewhere in pattern-string if (isPatternString) { - for (auto s: usedText) { + for (auto const & s: usedText) { // Remove entries created in previous search runs keys.erase(s); } @@ -2676,7 +2780,8 @@ void LatexInfo::buildKeys(bool isPatternString) if (keysBuilt && !isPatternString) return; // Keys to ignore in any case - makeKey("text|textcyrillic|lyxmathsym|ensuremath", KeyInfo(KeyInfo::headRemove, 1, true), true); + makeKey("text|lyxmathsym|ensuremath", KeyInfo(KeyInfo::headRemove, 1, true), true); + makeKey("nonumber|notag", KeyInfo(KeyInfo::headRemove, 0, true), true); // Known standard keys with 1 parameter. // Split is done, if not at start of region makeKey("textsf|textss|texttt", KeyInfo(KeyInfo::isStandard, 1, ignoreFormats.getFamily()), isPatternString); @@ -2729,9 +2834,14 @@ void LatexInfo::buildKeys(bool isPatternString) // handle like standard keys with 1 parameter. makeKey("url|href|vref|thanks", KeyInfo(KeyInfo::isStandard, 1, false), isPatternString); - // Ignore deleted text - makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString); - // but preserve added text + if (ignoreFormats.getDeleted()) { + // Ignore deleted text + makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 3, false), isPatternString); + } + else { + // but preserve added text + makeKey("lyxdeleted", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString); + } makeKey("lyxadded", KeyInfo(KeyInfo::doRemove, 2, false), isPatternString); // Macros to remove, but let the parameter survive @@ -2757,7 +2867,7 @@ void LatexInfo::buildKeys(bool isPatternString) makeKey("textgreek|textcyrillic", KeyInfo(KeyInfo::isStandard, 1, true), false); makeKey("parbox", KeyInfo(KeyInfo::doRemove, 1, true), isPatternString); // like ('tiny{}' or '\tiny ' ... ) - makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, false), isPatternString); + makeKey("footnotesize|tiny|scriptsize|small|large|Large|LARGE|huge|Huge", KeyInfo(KeyInfo::isSize, 0, ignoreFormats.getSize()), isPatternString); // Survives, like known character // makeKey("lyx|LyX|latex|LaTeX|latexe|LaTeXe|tex|TeX", KeyInfo(KeyInfo::isChar, 0, false), isPatternString); @@ -2814,6 +2924,8 @@ string Intervall::show(int lastpos) int idx = 0; /* int intervalls */ string s; int i = 0; + if ((unsigned) lastpos > par.size()) + lastpos = par.size(); for (idx = 0; idx <= ignoreidx; idx++) { while (i < lastpos) { int printsize; @@ -3360,10 +3472,12 @@ static string correctlanguagesetting(string par, bool isPatternString, bool with while ((parlen > 0) && (par[parlen-1] == '\n')) { parlen--; } +#if 0 if (isPatternString && (parlen > 0) && (par[parlen-1] == '~')) { // Happens to be there in case of description or labeling environment parlen--; } +#endif string result; if (withformat) { // Split the latex input into pieces which @@ -3545,7 +3659,7 @@ MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt) CreateRegexp(opt, "", "", ""); return; } - use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos; + use_regexp = ds.find(from_utf8("\\regexp{")) != std::string::npos; if (opt.replace_all && previous_single_replace) { previous_single_replace = false; num_replaced = 0; @@ -3555,7 +3669,7 @@ MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt) previous_single_replace = true; } // When using regexp, braces are hacked already by escape_for_regex() - par_as_string = normalize(ds, opt.ignoreformat); + par_as_string = convertLF2Space(ds, opt.ignoreformat); open_braces = 0; close_wildcards = 0; @@ -3634,7 +3748,7 @@ MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt) break; } if (lng < par_as_string.size()) - par_as_string = par_as_string.substr(0,lng); + par_as_string.resize(lng); } LYXERR(Debug::FINDVERBOSE, "par_as_string after correctRegex is '" << par_as_string << "'"); if ((lng > 0) && (par_as_string[0] == '^')) { @@ -3661,7 +3775,7 @@ MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt) } if (opt.matchword) { modifyRegexForMatchWord(par_as_string); - opt.matchword = false; + // opt.matchword = false; } regexp_str = "(" + lead_as_regexp + ")()" + par_as_string; regexp2_str = "(" + lead_as_regexp + ")(.*?)" + par_as_string; @@ -3672,18 +3786,15 @@ MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt) } } -MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const +MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, MatchStringAdv::matchType at_begin) const { MatchResult mres; mres.searched_size = len; - if (at_begin && - (opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) ) - return mres; docstring docstr = stringifyFromForSearch(opt, cur, len); string str; - str = normalize(docstr, opt.ignoreformat); + str = convertLF2Space(docstr, opt.ignoreformat); if (!opt.ignoreformat) { str = correctlanguagesetting(str, false, !opt.ignoreformat); // remove closing '}' and '\n' to allow for use of '$' in regex @@ -3693,22 +3804,25 @@ MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_be if (lng != str.size()) { str = str.substr(0, lng); } + // Replace occurences of '~' to ' ' + static std::regex specialChars { R"(~)" }; + str = std::regex_replace(str, specialChars, R"( )" ); } if (str.empty()) { mres.match_len = -1; return mres; } - LYXERR(Debug::FINDVERBOSE, "After normalization: Matching against:\n'" << str << "'"); + LYXERR(Debug::FINDVERBOSE|Debug::FIND, "After normalization: Matching against:\n'" << str << "'"); LASSERT(use_regexp, /**/); { // use_regexp always true - LYXERR(Debug::FINDVERBOSE, "Searching in regexp mode: at_begin=" << at_begin); + LYXERR(Debug::FINDVERBOSE, "Searching in regexp mode: at_begin=" << matchTypeAsString(at_begin)); #if QTSEARCH QString qstr = QString::fromStdString(str); QRegularExpression const *p_regexp; QRegularExpression::MatchType flags = QRegularExpression::NormalMatch; - if (at_begin) { + if (at_begin == MatchStringAdv::MatchFromStart) { p_regexp = ®exp; } else { p_regexp = ®exp2; @@ -3719,7 +3833,7 @@ MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_be #else regex const *p_regexp; regex_constants::match_flag_type flags; - if (at_begin) { + if (at_begin == MatchStringAdv::MatchFromStart) { flags = regex_constants::match_continuous; p_regexp = ®exp; } else { @@ -3758,7 +3872,7 @@ MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_be int matchend = match.capturedEnd(0); size_t strsize = qstr.size(); if (!opt.ignoreformat) { - while (mres.match_len > 0) { + while (mres.match_len > 1) { QChar c = qstr.at(matchend - 1); if ((c == '\n') || (c == '}') || (c == '{')) { mres.match_len--; @@ -3787,7 +3901,7 @@ MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_be int matchend = strend; size_t strsize = str.size(); if (!opt.ignoreformat) { - while (mres.match_len > 0) { + while (mres.match_len > 1) { char c = str.at(matchend - 1); if ((c == '\n') || (c == '}') || (c == '{')) { mres.match_len--; @@ -3833,23 +3947,20 @@ MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_be } -MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const +MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, MatchStringAdv::matchType at_begin) const { MatchResult mres = findAux(cur, len, at_begin); - int res = mres.match_len; LYXERR(Debug::FINDVERBOSE, - "res=" << res << ", at_begin=" << at_begin + "res=" << mres.match_len << ", at_begin=" << matchTypeAsString(at_begin) << ", matchAtStart=" << opt.matchAtStart << ", inTexted=" << cur.inTexted()); - if (opt.matchAtStart) { - if (cur.pos() != 0) - mres.match_len = 0; - else if (mres.match_prefix > 0) - mres.match_len = 0; - return mres; + if (mres.match_len > 0) { + if (opt.matchAtStart) { + if (cur.pos() > 0 || mres.match_prefix > 0) + mres.match_len = 0; + } } - else - return mres; + return mres; } #if 0 @@ -3877,63 +3988,73 @@ static bool simple_replace(string &t, string from, string to) } #endif -string MatchStringAdv::normalize(docstring const & s, bool ignore_format) const +string MatchStringAdv::convertLF2Space(docstring const &s, bool ignore_format) const { - string t; - t = lyx::to_utf8(s); - // Remove \n at begin - while (!t.empty() && t[0] == '\n') - t = t.substr(1); - // Remove \n* at end - while (!t.empty() && t[t.size() - 1] == '\n') { - t = t.substr(0, t.size() - 1); - } + // Using original docstring to handle '\n' + + if (s.size() == 0) return ""; + stringstream t; size_t pos; - // Handle all other '\n' - while ((pos = t.find("\n")) != string::npos) { - if (pos > 1 && t[pos-1] == '\\' && t[pos-2] == '\\' ) { - // Handle '\\\n' - if (isAlnumASCII(t[pos+1])) { - t.replace(pos-2, 3, " "); - } - else { - t.replace(pos-2, 3, ""); - } + size_t start = 0; + size_t end = s.size() - 1; + if (!ignore_format) { + while (s[start] == '\n' && start <= end) + start++; + while (end >= start && s[end] == '\n') + end--; + if (start >= end + 1) + return ""; + } + do { + bool dospace = true; + int skip = -1; + pos = s.find('\n', start); + if (pos >= end) { + t << lyx::to_utf8(s.substr(start, end + 1 - start)); + break; } - else { - if (!isAlnumASCII(t[pos+1]) || !isAlnumASCII(t[pos-1])) { - // '\n' adjacent to non-alpha-numerics, discard - t.replace(pos, 1, ""); - } - else { - // Replace all other \n with spaces - t.replace(pos, 1, " "); + if (!ignore_format) { + if ((pos > start + 1) && + s[pos-1] == '\\' && + s[pos-2] == '\\') { + skip = 2; + if ((pos > start + 2) && + (s[pos+1] == '~' || isSpace(s[pos+1]) || + s[pos-3] == '~' || isSpace(s[pos-3]))) { + // discard "\\\\\n", do not replace with space + dospace = false; + } } - if (!ignore_format) { - size_t count = 0; - while ((pos > count + 1) && (t[pos - 1 -count] == '%')) { - count++; + else if (pos > start) { + if (s[pos-1] == '%') { + skip = 1; + while ((pos > start+skip) && (s[pos-1-skip] == '%')) + skip++; + if ((pos > start+skip) && + (s[pos+1] == '~' || isSpace(s[pos+1]) || + s[pos-1-skip] == '~' || isSpace(s[pos-1-skip]))) { + // discard '%%%%%\n' + dospace = false; + } } - if (count > 0) { - t.replace(pos - count, count, ""); + else if (!isAlnumASCII(s[pos+1]) || !isAlnumASCII(s[pos-1])) { + dospace = false; + skip = 0; // remove the '\n' only } } } - } - // Remove stale empty \emph{}, \textbf{} and similar blocks from latexify - // Kornel: Added textsl, textsf, textit, texttt and noun - // + allow to seach for colored text too - LYXERR(Debug::FINDVERBOSE, "Removing stale empty macros from: " << t); - while (regex_replace(t, t, "\\\\(emph|noun|text(bf|sl|sf|it|tt)|(u|uu)line|(s|x)out|uwave)(\\{(\\{\\})?\\})+", "")) - LYXERR(Debug::FINDVERBOSE, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t); - while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", "")) - LYXERR(Debug::FINDVERBOSE, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t); - while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", "")); - - return t; + else { + dospace = true; + skip = 0; + } + t << lyx::to_utf8(s.substr(start, pos-skip-start)); + if (dospace) + t << ' '; + start = pos+1; + } while (start <= end); + return(t.str()); } - docstring stringifyFromCursor(DocIterator const & cur, int len) { LYXERR(Debug::FINDVERBOSE, "Stringifying with len=" << len << " from cursor at pos: " << cur); @@ -3950,20 +4071,20 @@ docstring stringifyFromCursor(DocIterator const & cur, int len) int option = AS_STR_INSETS | AS_STR_PLAINTEXT; if (ignoreFormats.getDeleted()) { option |= AS_STR_SKIPDELETE; - runparams.for_search = OutputParams::SearchWithoutDeleted; + runparams.find_set_feature(OutputParams::SearchWithoutDeleted); } else { - runparams.for_search = OutputParams::SearchWithDeleted; + runparams.find_set_feature(OutputParams::SearchWithDeleted); } if (ignoreFormats.getNonContent()) { - runparams.for_search |= OutputParams::SearchNonOutput; + runparams.find_add_feature(OutputParams::SearchNonOutput); } LYXERR(Debug::FINDVERBOSE, "Stringifying with cur: " << cur << ", from pos: " << cur.pos() << ", end: " << end); docstring res = from_utf8(latexNamesToUtf8(par.asString(cur.pos(), end, option, - &runparams))); - LYXERR(Debug::FIND, "Stringified text from pos(" << cur.pos() << ") len(" << len << "): " << res); + &runparams), false)); + LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Stringified text from pos(" << cur.pos() << ") len(" << len << "): " << res); return res; } else if (cur.inMathed()) { CursorSlice cs = cur.top(); @@ -3975,7 +4096,7 @@ docstring stringifyFromCursor(DocIterator const & cur, int len) MathData md2; for (MathData::const_iterator it = md.begin() + cs.pos(); it != it_end; ++it) md2.push_back(*it); - docstring res = from_utf8(latexNamesToUtf8(asString(md2))); + docstring res = from_utf8(latexNamesToUtf8(asString(md2), false)); LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Stringified math from pos(" << cur.pos() << ") len(" << len << "): " << res); return res; } @@ -4003,13 +4124,13 @@ docstring latexifyFromCursor(DocIterator const & cur, int len) runparams.nice = false; setFindParams(runparams); if (ignoreFormats.getDeleted()) { - runparams.for_search = OutputParams::SearchWithoutDeleted; + runparams.find_set_feature(OutputParams::SearchWithoutDeleted); } else { - runparams.for_search = OutputParams::SearchWithDeleted; + runparams.find_set_feature(OutputParams::SearchWithDeleted); } if (ignoreFormats.getNonContent()) { - runparams.for_search |= OutputParams::SearchNonOutput; + runparams.find_add_feature(OutputParams::SearchNonOutput); } if (cur.inTexted()) { @@ -4019,7 +4140,7 @@ docstring latexifyFromCursor(DocIterator const & cur, int len) endpos = cur.pos() + len; TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams, string(), cur.pos(), endpos, true); - LYXERR(Debug::FIND, "Latexified text from pos(" << cur.pos() << ") len(" << len << "): " << ods.str()); + LYXERR(Debug::FINDVERBOSE|Debug::FIND, "Latexified text from pos(" << cur.pos() << ") len(" << len << "): " << ods.str()); return(ods.str()); } else if (cur.inMathed()) { // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly @@ -4114,13 +4235,13 @@ MatchResult findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, Mat // either one sees "http://www.bla.bla" or nothing // so the search for "www" gives prefix_len = 7 (== sizeof("http://") // and although we search for only 3 chars, we find the whole hyperlink inset - bool at_begin = (expected.match_prefix == 0); + MatchStringAdv::matchType at_begin = (expected.match_prefix == 0) ? MatchStringAdv::MatchFromStart : MatchStringAdv::MatchAnyPlace; if (!match.opt.forward && match.opt.ignoreformat) { if (expected.pos > 0) return fail; } - LASSERT(at_begin, /**/); - if (expected.match_len > 0 && at_begin) { + LASSERT(at_begin == MatchStringAdv::MatchFromStart, /**/); + if (expected.match_len > 0 && at_begin == MatchStringAdv::MatchFromStart) { // Search for deepest match old_cur = cur; max_match = expected; @@ -4152,11 +4273,14 @@ MatchResult findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, Mat } else { // (expected.match_len <= 0) - mres = match(cur); /* match valid only if not searching whole words */ + mres = match(cur, -1, MatchStringAdv::MatchFromStart); /* match valid only if not searching whole words */ displayMres(mres, "Start with negative match", cur); max_match = mres; } - if (max_match.match_len <= 0) return fail; + // Only now we are really at_begin + if ((max_match.match_len <= 0) || + (match.opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed())) + return fail; LYXERR(Debug::FINDVERBOSE, "Ok"); // Compute the match length @@ -4254,7 +4378,7 @@ int findForwardAdv(DocIterator & cur, MatchStringAdv & match) while (!theApp()->longOperationCancelled() && cur) { //(void) findAdvForwardInnermost(cur); LYXERR(Debug::FINDVERBOSE, "findForwardAdv() cur: " << cur); - MatchResult mres = match(cur, -1, false); + MatchResult mres = match(cur, -1, MatchStringAdv::MatchAnyPlace); string msg = "Starting"; if (repeat) msg = "Repeated"; @@ -4292,7 +4416,7 @@ int findForwardAdv(DocIterator & cur, MatchStringAdv & match) continue; } cur.pos() = cur.pos() + increment; - MatchResult mres2 = match(cur, -1, false); + MatchResult mres2 = match(cur, -1, MatchStringAdv::MatchAnyPlace); displayMres(mres2, "findForwardAdv loop", cur) switch (interpretMatch(mres, mres2)) { case MatchResult::newIsTooFar: @@ -4315,7 +4439,7 @@ int findForwardAdv(DocIterator & cur, MatchStringAdv & match) default: // Todo@ // Handle not like MatchResult::newIsTooFar - LYXERR0( "Probably too far: Increment = " << increment << " match_prefix = " << mres.match_prefix); + LYXERR(Debug::FINDVERBOSE, "Probably too far: Increment = " << increment << " match_prefix = " << mres.match_prefix); firstInvalid--; increment = increment*3/4; cur = old_cur; @@ -4368,7 +4492,8 @@ MatchResult findMostBackwards(DocIterator & cur, MatchStringAdv const & match, M LYXERR(Debug::FINDVERBOSE, "findMostBackwards(): cur=" << cur); DocIterator new_cur = cur; new_cur.backwardPos(); - if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len) + if (new_cur == cur || &new_cur.inset() != &inset + || match(new_cur, -1, MatchStringAdv::MatchFromStart).match_len <= 0) break; MatchResult new_mr = findAdvFinalize(new_cur, match, expected); if (new_mr.match_len == mr.match_len) @@ -4394,7 +4519,7 @@ int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match) bool pit_changed = false; do { cur.pos() = 0; - MatchResult found_match = match(cur, -1, false); + MatchResult found_match = match(cur, -1, MatchStringAdv::MatchAnyPlace); if (found_match.match_len > 0) { if (pit_changed) @@ -4404,7 +4529,7 @@ int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match) LYXERR(Debug::FINDVERBOSE, "findBackAdv2: cur: " << cur); DocIterator cur_prev_iter; do { - found_match = match(cur); + found_match = match(cur, -1, MatchStringAdv::MatchFromStart); LYXERR(Debug::FINDVERBOSE, "findBackAdv3: found_match=" << (found_match.match_len > 0) << ", cur: " << cur); if (found_match.match_len > 0) { @@ -4566,7 +4691,7 @@ static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, Ma return 0; LASSERT(sel_len > 0, return 0); - if (!matchAdv(sel_beg, sel_len).match_len) + if (matchAdv(sel_beg, sel_len, MatchStringAdv::MatchFromStart).match_len <= 0) return 0; // Build a copy of the replace buffer, adapted to the KeepCase option @@ -4634,6 +4759,10 @@ static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, Ma return 1; } +static bool isWordChar(char_type c) +{ + return isLetterChar(c) || isNumberChar(c); +} /// Perform a FindAdv operation. bool findAdv(BufferView * bv, FindAndReplaceOptions & opt) @@ -4659,8 +4788,52 @@ bool findAdv(BufferView * bv, FindAndReplaceOptions & opt) bv->putSelectionAt(bv->cursor().selectionBegin(), length, !opt.forward); num_replaced += findAdvReplace(bv, opt, matchAdv); cur = bv->cursor(); - if (opt.forward) + if (opt.forward) { + if (opt.matchword && cur.pos() > 0) { // Skip word-characters if we are in the mid of a word + if (cur.inTexted()) { + Paragraph const & par = cur.paragraph(); + int len_limit, new_pos; + if (cur.lastpos() < par.size()) + len_limit = cur.lastpos(); + else + len_limit = par.size(); + for (new_pos = cur.pos() - 1; new_pos < len_limit; new_pos++) { + if (!isWordChar(par.getChar(new_pos))) + break; + } + if (new_pos > cur.pos()) + cur.pos() = new_pos; + } + else if (cur.inMathed()) { + // Check if 'cur.pos()-1' and 'cur.pos()' both point to a letter, + // I am not sure, we should consider the selection + bool sel = bv->cursor().selection(); + if (!sel && cur.pos() < cur.lastpos()) { + CursorSlice const & cs = cur.top(); + MathData md = cs.cell(); + int len = -1; + MathData::const_iterator it_end = md.end(); + MathData md2; + // Start the check with one character before actual cursor position + for (MathData::const_iterator it = md.begin() + cs.pos() - 1; + it != it_end; ++it) + md2.push_back(*it); + docstring inp = asString(md2); + for (len = 0; (unsigned) len < inp.size() && len + cur.pos() <= cur.lastpos(); len++) { + if (!isWordChar(inp[len])) + break; + } + // len == 0 means previous char was a word separator + // len == 1 search starts with a word separator + // len == 2 ... we have to skip len -1 chars + if (len > 1) + cur.pos() = cur.pos() + len - 1; + } + } + opt.matchword = false; + } pos_len = findForwardAdv(cur, matchAdv); + } else pos_len = findBackwardsAdv(cur, matchAdv); } catch (exception & ex) {