*/
#include <config.h>
+#include <iterator>
#include "lyxfind.h"
//#define ResultsDebug
#define USE_QT_FOR_SEARCH
#if defined(USE_QT_FOR_SEARCH)
- #include <QtCore> // sets QT_VERSION
- #if (QT_VERSION >= 0x050000)
- #include <QRegularExpression>
- #define QTSEARCH 1
- #else
- #define QTSEARCH 0
- #endif
+ #include <QRegularExpression>
+ #define QTSEARCH 1
#else
#define QTSEARCH 0
#endif
///
bool getShape() const { return ignoreShape_; }
///
+ bool getSize() const { return ignoreSize_; }
+ ///
bool getUnderline() const { return ignoreUnderline_; }
///
bool getMarkUp() const { return ignoreMarkUp_; }
///
bool ignoreShape_ = false;
///
+ bool ignoreSize_ = true;
+ ///
bool ignoreUnderline_ = false;
///
bool ignoreMarkUp_ = false;
else if (type == "shape") {
ignoreShape_ = value;
}
+ else if (type == "size") {
+ ignoreSize_ = value;
+ }
else if (type == "family") {
ignoreFamily_ = value;
}
bool find_del, bool check_wrap, bool const auto_wrap,
bool instant, bool onlysel)
{
+ bool const had_selection = bv->cursor().selection();
+
// Clean up previous selections with empty searchstr on instant
if (searchstr.empty() && instant) {
- if (bv->cursor().selection()) {
+ if (had_selection) {
bv->setCursor(bv->cursor().selectionBegin());
bv->clearSelection();
}
if (!searchAllowed(searchstr))
return false;
- DocIterator const endcur = forward ? bv->cursor().selectionEnd() : bv->cursor().selectionBegin();
+ DocIterator const startcur = bv->cursor().selectionBegin();
+ DocIterator const endcur = bv->cursor().selectionEnd();
- if (onlysel && bv->cursor().selection()) {
+ if (onlysel && had_selection) {
docstring const matchstring = bv->cursor().selectionAsString(false);
docstring const lcmatchsting = support::lowercase(matchstring);
if (matchstring == searchstr || (!case_sens && lcmatchsting == lowercase(searchstr))) {
int match_len = forward
? findForward(cur, endcur, match, find_del, onlysel)
- : findBackwards(cur, endcur, match, find_del, onlysel);
+ : findBackwards(cur, startcur, match, find_del, onlysel);
if (match_len > 0)
bv->putSelectionAt(cur, match_len, !forward);
- else if (onlysel) {
+ else if (onlysel && had_selection) {
docstring q = _("The search string was not found within the selection.\n"
"Continue search outside?");
int search_answer = frontend::Alert::prompt(_("Search outside selection?"),
return false;
}
else if (check_wrap) {
- DocIterator cur_orig(bv->cursor());
bool wrap = auto_wrap;
if (!auto_wrap) {
docstring q;
find_del, false, false, false, false))
return true;
}
- bv->cursor().setCursor(cur_orig);
+ bv->setCursor(startcur);
+
+ // restore original selection
+ if (had_selection) {
+ bv->cursor().resetAnchor();
+ bv->setSelection(startcur, endcur);
+ }
return false;
}
if (had_selection) {
endcur.fixIfBroken();
bv->cursor().resetAnchor();
- bv->setCursorSelectionTo(endcur);
+ bv->setSelection(startcur, endcur);
}
return num;
typedef vector<pair<string, string> > 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') {
+ // Normal Space
+ blanks++;
+ i++;
+ }
+ else if (tempx[i] == '\342') {
+ if (tempx[i+1] == '\200') {
+ if ((tempx[i+2] == '\257')
+ || (tempx[i+2] == '\203')
+ || (tempx[i+2] == '\202')) {
+ // Non-breaking Thin (1/6 em)
+ // Quad(1 em), (Double quad counts as 2 blanks)
+ // Half Quad
+ blanks++;
+ i += 2;
+ }
+ else if (tempx[i+2] == '\213') {
+ // Ignoring parts of Medium and Thick
+ i += 2;
+ continue;
+ }
+ else if ((tempx[i+2] == '\204') || (tempx[i+2] == '\205')) {
+ // Thick
+ // Medium
+ blanks++;
+ i += 2;
+ }
+ }
+ else if (tempx[i+1] == '\201') {
+ if (tempx[i+2] == '\240') {
+ // Ignoring parts of half quad
+ i += 2;
+ continue;
+ }
+ }
+ else if ((tempx[i+1] == '\220') && (tempx[i+2] == '\243')) {
+ // Visible space
+ blanks++;
+ i += 2;
+ }
+ }
+ 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;
* 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;
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;
}
else {
if (sub.str(4) == "backslash") {
- replace = "\\";
+ replace = string("\\");
{
// 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")
replace = "^";
+ else if ((sub.str(4) == "negthinspace") || (sub.str(4) == "negmedspace") || (sub.str(4) == "negthickspace")) {
+ replace = accents[sub.str(4)+"{}"];
+ }
else if (backslashed) {
backslashed = false;
if (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, /**/);
replace = "\\{";
else if (sub.str(3) == "}")
replace = "\\}";
+ else if (sub.str(3) == " ")
+ replace = "\\ ";
else {
// else part should not exist
LASSERT(0, /**/);
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()) {
}
}
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 "\'"
** 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;
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.
** @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
// 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;
+ int open_braces = 0;
// number of (.*?) subexpressions added at end of search regexp for closing
// environments, math mode, styles, etc...
- int close_wildcards;
+ int close_wildcards = 0;
public:
// Are we searching with regular expressions ?
bool use_regexp = false;
valid_matches = 0;
}
+static void setFindParams(OutputParams &runparams)
+{
+ runparams.flavor = Flavor::XeTeX;
+ //runparams.use_polyglossia = true;
+ runparams.linelen = 10000; //lyxrc.plaintext_linelen;
+ // No side effect of file copying and image conversion
+ runparams.dryrun = true;
+}
+
static docstring buffer_to_latex(Buffer & buffer)
{
//OutputParams runparams(&buffer.params().encoding());
odocstringstream ods;
otexstream os(ods);
runparams.nice = true;
- runparams.flavor = Flavor::XeTeX;
- runparams.linelen = 10000; //lyxrc.plaintext_linelen;
- // No side effect of file copying and image conversion
- runparams.dryrun = 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) {
- TeXOnePar(buffer, buffer.text(), pit, os, runparams);
+ TeXOnePar(buffer, buffer.text(), pit, os, runparams, string(), -1, -1, true);
LYXERR(Debug::FINDVERBOSE, "searchString up to here: " << ods.str());
}
return ods.str();
}
-static string latexNamesToUtf8(docstring strIn)
+static string latexNamesToUtf8(docstring strIn, bool withformat)
{
string addtmp = to_utf8(strIn);
static regex const rmAcc("(\\\\)*("
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;
// OutputParams runparams(&buffer.params().encoding());
OutputParams runparams(encodings.fromLyXName("utf8"));
runparams.nice = true;
- runparams.flavor = Flavor::XeTeX;
- runparams.linelen = 10000; //lyxrc.plaintext_linelen;
- runparams.dryrun = true;
+ setFindParams(runparams);
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;
|| 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();
}
/*
string par;
int ignoreidx;
static vector<Border> borders;
- int depts[MAXOPENED];
- int closes[MAXOPENED];
+ static vector<int> depts;
+ static vector<int> closes;
int actualdeptindex;
int previousNotIgnored(int) const;
int nextNotIgnored(int) const;
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
};
vector<Border> Intervall::borders = vector<Border>(30);
+vector<int> Intervall::depts = vector<int>(30);
+vector<int> Intervall::closes = vector<int>(30);
int Intervall::isOpeningPar(int pos) const
{
}
}
+#if 0
+// Not needed, because dpts and closes are now dynamically expanded
static void checkDepthIndex(int val)
{
static int maxdepthidx = MAXOPENED-2;
LYXERR(Debug::INFO, "maxdepthidx now " << val);
}
}
+#endif
#if 0
// Not needed, because borders are now dynamically expanded
accents["cdot"] = "·";
accents["textasciicircum"] = "^";
accents["mathcircumflex"] = "^";
- accents["sim"] = "~";
accents["guillemotright"] = "»";
accents["guillemotleft"] = "«";
accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
accents["thinspace"] = getutf8(0xf0002); // and used _only_ by findadv
- accents["negthinspace"] = getutf8(0xf0003); // to omit backslashed latex macros
+ accents["negthinspace{}"]= getutf8(0xf0003); // to omit backslashed latex macros
accents["medspace"] = getutf8(0xf0004); // See https://en.wikipedia.org/wiki/Private_Use_Areas
- accents["negmedspace"] = getutf8(0xf0005);
+ accents["negmedspace{}"] = getutf8(0xf0005);
accents["thickspace"] = getutf8(0xf0006);
- accents["negthickspace"] = getutf8(0xf0007);
+ accents["negthickspace{}"]= getutf8(0xf0007);
accents["lyx"] = getutf8(0xf0010); // Used logos
accents["LyX"] = getutf8(0xf0010);
accents["tex"] = getutf8(0xf0011);
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);
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]+"
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)
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:
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;
}
// 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);
}
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);
// 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
makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
makeKey("hphantom|vphantom|note|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), 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);
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;
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
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;
previous_single_replace = true;
}
// When using regexp, braces are hacked already by escape_for_regex()
- par_as_string = normalize(ds, opt.ignoreformat);
- open_braces = 0;
- close_wildcards = 0;
+ par_as_string = convertLF2Space(ds, opt.ignoreformat);
size_t lead_size = 0;
// correct the language settings
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] == '^')) {
}
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;
}
}
-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
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;
#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 {
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--;
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--;
}
-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
}
#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);
// OutputParams runparams(&cur.buffer()->params().encoding());
OutputParams runparams(encodings.fromLyXName("utf8"));
runparams.nice = true;
- runparams.flavor = Flavor::XeTeX;
- runparams.linelen = 10000; //lyxrc.plaintext_linelen;
- // No side effect of file copying and image conversion
- runparams.dryrun = true;
+ setFindParams(runparams);
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();
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;
}
//OutputParams runparams(&buf.params().encoding());
OutputParams runparams(encodings.fromLyXName("utf8"));
runparams.nice = false;
- runparams.flavor = Flavor::XeTeX;
- runparams.linelen = 8000; //lyxrc.plaintext_linelen;
- // No side effect of file copying and image conversion
- runparams.dryrun = 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);
}
if (cur.inTexted()) {
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 from pos(" << cur.pos() << ") len(" << len << "): " << ods.str());
+ string(), cur.pos(), endpos, true);
+ 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
// 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;
}
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
{
if (!cur)
return 0;
- bool repeat = false;
+ int repeat = 0;
DocIterator orig_cur; // to be used if repeat not successful
MatchResult orig_mres;
+ do {
+ orig_cur = cur;
+ cur.forwardPos();
+ } while (cur.depth() > orig_cur.depth());
+ cur = orig_cur;
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)
+ if (repeat > 0)
msg = "Repeated";
displayMres(mres, msg + " findForwardAdv", cur)
int match_len = mres.match_len;
match_len = 0;
}
if (match_len <= 0) {
- // This should exit nested insets, if any, or otherwise undefine the currsor.
- cur.pos() = cur.lastpos();
+ if (repeat > 0) {
+ repeat--;
+ }
+ else {
+ // This should exit nested insets, if any, or otherwise undefine the currsor.
+ cur.pos() = cur.lastpos();
+ }
LYXERR(Debug::FINDVERBOSE, "Advancing pos: cur=" << cur);
cur.forwardPos();
}
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)) {
+ switch (interpretMatch(mres, mres2)) {
case MatchResult::newIsTooFar:
// behind the expected match
firstInvalid = increment;
increment /= 2;
break;
case MatchResult::newIsBetter:
- // not reached yet, but cur.pos()+increment is bettert
+ // not reached yet, but cur.pos()+increment is better
mres = mres2;
firstInvalid -= increment;
if (increment > firstInvalid*3/4)
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;
if (mres.match_len > 0) {
if (mres.match_prefix + mres.pos - mres.leadsize > 0) {
// The match seems to indicate some deeper level
- repeat = true;
+ repeat = 2;
orig_cur = cur;
orig_mres = mres;
cur.forwardPos();
continue;
}
}
- else if (repeat) {
+ else if (repeat > 0) {
// should never be reached.
cur = orig_cur;
mres = orig_mres;
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)
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)
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) {
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
// OutputParams runparams(&repl_buffer.params().encoding());
OutputParams runparams(encodings.fromLyXName("utf8"));
runparams.nice = false;
- runparams.flavor = Flavor::XeTeX;
- runparams.linelen = 8000; //lyxrc.plaintext_linelen;
- runparams.dryrun = true;
- TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams);
+ setFindParams(runparams);
+ TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams, string(), -1, -1, true);
//repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false);
docstring repl_latex = ods.str();
LYXERR(Debug::FINDVERBOSE, "Latexified replace_buffer: '" << repl_latex << "'");
return 1;
}
+static bool isWordChar(char_type c)
+{
+ return isLetterChar(c) || isNumberChar(c);
+}
/// Perform a FindAdv operation.
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) {
projects / lyx.git / blobdiff