X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2Flyxfind.cpp;h=609f38c7815fce43eed133435c94a27cde1ee14b;hb=d12bf6f9792a73ef6b8714747c67a70b7b23a116;hp=8c609a7980119fc709cd7577c9993b5a9ed7c996;hpb=8e256fbf0fb1fdffafe5bb35e18a0746d3ce8d1f;p=features.git diff --git a/src/lyxfind.cpp b/src/lyxfind.cpp index 8c609a7980..609f38c781 100644 --- a/src/lyxfind.cpp +++ b/src/lyxfind.cpp @@ -892,6 +892,7 @@ 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; @@ -908,6 +909,10 @@ string correctRegex(string t, bool withformat) replace = ""; backslashed = true; } + else if (withformat && next[0] == '$') { + replace = accents["lyxdollar"]; + skip = 1; // Skip following '$' + } } } else if (sub.str(4) == "mathcircumflex") @@ -960,7 +965,8 @@ string correctRegex(string t, bool withformat) if (lastpos < (size_t) sub.position(2)) 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 = std::regex_replace(t, protectedSpace, R"( )"); @@ -1422,8 +1428,8 @@ 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; @@ -1435,6 +1441,7 @@ 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 @@ -1447,6 +1454,8 @@ public: }; vector Intervall::borders = vector(30); +vector Intervall::depts = vector(30); +vector Intervall::closes = vector(30); int Intervall::isOpeningPar(int pos) const { @@ -1480,6 +1489,8 @@ 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; @@ -1493,6 +1504,7 @@ static void checkDepthIndex(int val) LYXERR(Debug::INFO, "maxdepthidx now " << val); } } +#endif #if 0 // Not needed, because borders are now dynamically expanded @@ -1966,6 +1978,7 @@ static void buildAccentsMap() 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); @@ -2041,7 +2054,8 @@ 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]+" @@ -2077,9 +2091,13 @@ 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) @@ -2241,6 +2259,27 @@ 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: @@ -2328,6 +2367,7 @@ 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; @@ -2733,7 +2773,8 @@ 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); @@ -3611,7 +3652,7 @@ 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; @@ -3700,7 +3741,7 @@ 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] == '^')) { @@ -3743,9 +3784,6 @@ MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_be 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; @@ -3946,8 +3984,79 @@ static bool simple_replace(string &t, string from, string to) } #endif -string MatchStringAdv::normalize(docstring const & s, bool ignore_format) const +#if 1 +static string convertLF2Space(docstring const &s, bool ignore_format) { + // Using original docstring to handle '\n' + + if (s.size() == 0) return ""; + stringstream t; + size_t pos; + 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; + } + 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; + } + } + 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; + } + } + else if (!isAlnumASCII(s[pos+1]) || !isAlnumASCII(s[pos-1])) { + dospace = false; + skip = 0; // remove the '\n' only + } + } + } + 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()); +} + +#else +static string convertLF2Space(docstring const & s, bool ignore_format) +{ + // Using utf8-converted string to handle '\n' + string t; t = lyx::to_utf8(s); // Remove \n at begin @@ -3990,19 +4099,29 @@ string MatchStringAdv::normalize(docstring const & s, bool ignore_format) const } } } - // 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 \\section{}, \\part{}, \\paragraph{} macros from: " << t); - while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", "")); + return(t); - return t; } +#endif +string MatchStringAdv::normalize(docstring const & s, bool ignore_format) const +{ + string t = convertLF2Space(s, ignore_format); + + // The following replaces are not appropriate in non-format-search mode + if (!ignore_format) { + // 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 \\section{}, \\part{}, \\paragraph{} macros from: " << t); + while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", "")); + } + return t; +} docstring stringifyFromCursor(DocIterator const & cur, int len) { @@ -4226,7 +4345,10 @@ MatchResult findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, Mat 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