// This causes a minor bug as undo will restore this selection,
// which the user did not create (#8986).
cur.innerText()->selectWord(cur, WHOLE_WORD);
- searchstr = cur.selectionAsString(false);
+ searchstr = cur.selectionAsString(false, true);
}
// if we still don't have a search string, report the error
return make_pair(false, 0);
bool have_selection = cur.selection();
- docstring const selected = cur.selectionAsString(false);
+ docstring const selected = cur.selectionAsString(false, true);
bool match =
case_sens
? searchstr == selected
bool matchword = parse_bool(howto);
bool forward = parse_bool(howto);
- return findOne(bv, search, casesensitive, matchword, forward, true, true);
+ return findOne(bv, search, casesensitive, matchword, forward, false, true);
}
-bool lyxreplace(BufferView * bv,
- FuncRequest const & ev, bool has_deleted)
+bool lyxreplace(BufferView * bv, FuncRequest const & ev)
{
if (!bv || ev.action() != LFUN_WORD_REPLACE)
return false;
bool update = false;
- if (!has_deleted) {
- int replace_count = 0;
- if (all) {
- replace_count = replaceAll(bv, search, rplc, casesensitive, matchword);
- update = replace_count > 0;
- } else {
- pair<bool, int> rv =
- replaceOne(bv, search, rplc, casesensitive, matchword, forward, findnext);
- update = rv.first;
- replace_count = rv.second;
- }
+ int replace_count = 0;
+ if (all) {
+ replace_count = replaceAll(bv, search, rplc, casesensitive, matchword);
+ update = replace_count > 0;
+ } else {
+ pair<bool, int> rv =
+ replaceOne(bv, search, rplc, casesensitive, matchword, forward, findnext);
+ update = rv.first;
+ replace_count = rv.second;
+ }
- Buffer const & buf = bv->buffer();
- if (!update) {
- // emit message signal.
- buf.message(_("String not found."));
+ Buffer const & buf = bv->buffer();
+ if (!update) {
+ // emit message signal.
+ buf.message(_("String not found."));
+ } else {
+ if (replace_count == 0) {
+ buf.message(_("String found."));
+ } else if (replace_count == 1) {
+ buf.message(_("String has been replaced."));
} else {
- if (replace_count == 0) {
- buf.message(_("String found."));
- } else if (replace_count == 1) {
- buf.message(_("String has been replaced."));
- } else {
- docstring const str =
- bformat(_("%1$d strings have been replaced."), replace_count);
- buf.message(str);
- }
+ docstring const str =
+ bformat(_("%1$d strings have been replaced."), replace_count);
+ buf.message(str);
}
- } else if (findnext) {
- // if we have deleted characters, we do not replace at all, but
- // rather search for the next occurence
- if (findOne(bv, search, casesensitive, matchword, forward, true, findnext))
- update = true;
- else
- bv->message(_("String not found."));
}
return update;
}
/// A map of symbols and their escaped equivalent needed within a regex.
/// @note Beware of order
+/*
Escapes const & get_regexp_escapes()
{
typedef std::pair<std::string, std::string> P;
}
return escape_map;
}
+*/
/// A map of lyx escaped strings and their unescaped equivalent.
+/*
Escapes const & get_lyx_unescapes()
{
typedef std::pair<std::string, std::string> P;
}
return escape_map;
}
+*/
/// A map of escapes turning a regexp matching text to one matching latex.
+/*
Escapes const & get_regexp_latex_escapes()
{
typedef std::pair<std::string, std::string> P;
}
return escape_map;
}
+*/
/** @todo Probably the maps need to be migrated to regexps, in order to distinguish if
** the found occurrence were escaped.
**/
+/*
string apply_escapes(string s, Escapes const & escape_map)
{
LYXERR(Debug::FIND, "Escaping: '" << s << "'");
LYXERR(Debug::FIND, "Escaped : '" << s << "'");
return s;
}
+*/
+
+
+string string2regex(string in)
+{
+ static std::regex specialChars { R"([-[\]{}()*+?.,\^$|#\s\$\\])" };
+ string temp = std::regex_replace(in, specialChars, R"(\$&)" );
+ string temp2("");
+ size_t lastpos = 0;
+ size_t fl_pos = 0;
+ int offset = 1;
+ while (fl_pos < temp.size()) {
+ fl_pos = temp.find("\\\\foreignlanguage", lastpos + offset);
+ if (fl_pos == string::npos)
+ break;
+ offset = 16;
+ temp2 += temp.substr(lastpos, fl_pos - lastpos);
+ temp2 += "\\n";
+ lastpos = fl_pos;
+ }
+ if (lastpos == 0)
+ return(temp);
+ if (lastpos < temp.size()) {
+ temp2 += temp.substr(lastpos, temp.size() - lastpos);
+ }
+ return temp2;
+}
+string correctRegex(string t)
+{
+ /* Convert \backslash => \
+ * and \{, \}, \[, \] => {, }, [, ]
+ */
+ string s("");
+ regex wordre("(\\\\)*(\\\\((backslash|mathcircumflex) ?|[\\[\\]\\{\\}]))");
+ size_t lastpos = 0;
+ smatch sub;
+ for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
+ sub = *it;
+ string replace;
+ if ((sub.position(2) - sub.position(0)) % 2 == 1) {
+ continue;
+ }
+ else {
+ if (sub.str(4) == "backslash")
+ replace = "\\";
+ else if (sub.str(4) == "mathcircumflex")
+ replace = "^";
+ else if (sub.str(3) == "{")
+ replace = "\\braceleft";
+ else if (sub.str(3) == "}")
+ replace = "\\braceright";
+ else
+ replace = sub.str(3);
+ }
+ if (lastpos < (size_t) sub.position(2))
+ s += t.substr(lastpos, sub.position(2) - lastpos);
+ s += replace;
+ lastpos = sub.position(2) + sub.length(2);
+ }
+ if (lastpos == 0)
+ return t;
+ else if (lastpos < t.length())
+ s += t.substr(lastpos, t.length() - lastpos);
+ return 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)
+string escape_for_regex(string s)
{
- 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();
- string t;
- if (new_pos > pos) {
- // outside regexp
- LYXERR(Debug::FIND, "new_pos: " << new_pos);
- 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 after unescapes [lyx]: " << t);
- if (match_latex) {
- t = apply_escapes(t, get_regexp_latex_escapes());
- LYXERR(Debug::FIND, "t in regexp after latex_escapes [ltx]: " << t);
+ size_t lastpos = 0;
+ string result = "";
+ while (lastpos < s.size()) {
+ size_t regex_pos = s.find("\\regexp{", lastpos);
+ if (regex_pos == string::npos) {
+ regex_pos = s.size();
}
- if (end_pos == s.size()) {
- s.replace(new_pos, end_pos - new_pos, t);
- LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s);
- break;
+ if (regex_pos > lastpos) {
+ result += string2regex(s.substr(lastpos, regex_pos-lastpos));
+ lastpos = regex_pos;
+ if (lastpos == s.size())
+ 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);
+ size_t end_pos = s.find("\\endregexp{}}", regex_pos + 8);
+ result += correctRegex(s.substr(regex_pos + 8, end_pos -(regex_pos + 8)));
+ lastpos = end_pos + 13;
}
- return s;
+ return result;
}
int match2end;
int pos;
int leadsize;
+ int pos_len;
+ int searched_size;
vector <string> result = vector <string>();
- MatchResult(int len = 0): match_len(len),match_prefix(0),match2end(0), pos(0),leadsize(0) {};
+ MatchResult(int len = 0): match_len(len),match_prefix(0),match2end(0), pos(0),leadsize(0),pos_len(-1),searched_size(0) {};
};
static MatchResult::range interpretMatch(MatchResult &oldres, MatchResult &newres)
{
int range = oldres.match_len;
- if (range < 2) range = 2;
+ if (range > 0) range--;
if (newres.match2end < oldres.match2end - oldres.match_len)
return MatchResult::newIsTooFar;
if (newres.match_len < oldres.match_len)
class MatchStringAdv {
public:
- MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt);
+ MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & 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.
private:
/// Auxiliary find method (does not account for opt.matchword)
MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) 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 hack_braces) const;
+ string normalize(docstring const & s) const;
// normalized string to search
string par_as_string;
// regular expression to use for searching
public:
// Are we searching with regular expressions ?
bool use_regexp;
- int valid_matches;
- vector <string> matches = vector <string>(10);
+ static int valid_matches;
+ static vector <string> matches;
void FillResults(MatchResult &found_mr);
};
+int MatchStringAdv::valid_matches = 0;
+vector <string> MatchStringAdv::matches = vector <string>(10);
+
void MatchStringAdv::FillResults(MatchResult &found_mr)
{
if (found_mr.match_len > 0) {
accents["i"] = "ı";
accents["jmath"] = "ȷ";
accents["cdot"] = "·";
+ accents["textasciicircum"] = "^";
+ accents["mathcircumflex"] = "^";
+ accents["sim"] = "~";
accents["guillemotright"] = "»";
accents["guillemotleft"] = "«";
accents["hairspace"] = getutf8(0xf0000); // select from free unicode plane 15
buildAccentsMap();
static regex const accre("\\\\(([\\S]|grave|breve|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot|ogonek|"
"cedilla|subring|textsubring|subhat|textsubcircum|subtilde|textsubtilde|dgrave|textdoublegrave|rcap|textroundcap|slashed)\\{[^\\{\\}]+\\}"
- "|((i|imath|jmath|cdot|[a-z]+space)|((backslash )?([lL]y[xX]|[tT]e[xX]|[lL]a[tT]e[xX]e?|lyxarrow))|guillemot(left|right))(?![a-zA-Z]))");
+ "|((i|imath|jmath|cdot|[a-z]+space)|((backslash )?([lL]y[xX]|[tT]e[xX]|[lL]a[tT]e[xX]e?|lyxarrow))|guillemot(left|right)|textasciicircum|mathcircumflex|sim)(?![a-zA-Z]))");
smatch sub;
for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
sub = *itacc;
key += interval_.par.substr(params, optend-params);
evaluatingOptional = true;
optionalEnd = optend;
+ if (found.keytype == KeyInfo::isSectioning) {
+ // Remove optional values (but still keep in header)
+ interval_.addIntervall(params, optend);
+ }
}
string token = sub.str(7);
int closings;
}
else {
found._dataStart = found._tokenstart + found._tokensize;
- }
+ }
closings = 0;
}
- if (interval_.par.substr(found._dataStart-1, 15).compare("\\endarguments{}") == 0) {
+ if (interval_.par.substr(found._dataStart, 15).compare("\\endarguments{}") == 0) {
found._dataStart += 15;
}
size_t endpos;
// Known charaters
// No split
makeKey("backslash|textbackslash|slash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
+ makeKey("braceleft|braceright", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
makeKey("textasciicircum|textasciitilde", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
makeKey("textasciiacute|texemdash", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
makeKey("dots|ldots", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
case KeyInfo::doRemove: {
// Remove the key with all parameters and following spaces
size_t pos;
- for (pos = actual._dataEnd+1; pos < interval_.par.length(); pos++) {
+ size_t start;
+ if (interval_.par[actual._dataEnd-1] == ' ')
+ start = actual._dataEnd;
+ else
+ start = actual._dataEnd+1;
+ for (pos = start; pos < interval_.par.length(); pos++) {
if ((interval_.par[pos] != ' ') && (interval_.par[pos] != '%'))
break;
}
return "";
}
}
+
}
else {
// LYXERR(Debug::INFO, "No regex formats");
static int num_replaced = 0;
static bool previous_single_replace = true;
-MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt)
+void MatchStringAdv::CreateRegexp(FindAndReplaceOptions const & opt, string regexp_str, string regexp2_str, string par_as_string)
+{
+#if QTSEARCH
+ // Handle \w properly
+ QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
+ if (! opt.casesensitive) {
+ popts |= QRegularExpression::CaseInsensitiveOption;
+ }
+ regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
+ regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
+ regexError = "";
+ if (regexp.isValid() && regexp2.isValid()) {
+ regexIsValid = true;
+ // Check '{', '}' pairs inside the regex
+ int balanced = 0;
+ int skip = 1;
+ for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
+ char c = par_as_string[i];
+ if (c == '\\') {
+ skip = 2;
+ continue;
+ }
+ if (c == '{')
+ balanced++;
+ else if (c == '}') {
+ balanced--;
+ if (balanced < 0)
+ break;
+ }
+ skip = 1;
+ }
+ if (balanced != 0) {
+ regexIsValid = false;
+ regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
+ }
+ }
+ else {
+ regexIsValid = false;
+ if (!regexp.isValid())
+ regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
+ else
+ regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
+ }
+#else
+ if (opt.casesensitive) {
+ regexp = regex(regexp_str);
+ regexp2 = regex(regexp2_str);
+ }
+ else {
+ regexp = regex(regexp_str, std::regex_constants::icase);
+ regexp2 = regex(regexp2_str, std::regex_constants::icase);
+ }
+#endif
+}
+
+static void modifyRegexForMatchWord(string &t)
+{
+ string s("");
+ regex wordre("(\\\\)*((\\.|\\\\b))");
+ size_t lastpos = 0;
+ smatch sub;
+ for (sregex_iterator it(t.begin(), t.end(), wordre), end; it != end; ++it) {
+ sub = *it;
+ if ((sub.position(2) - sub.position(0)) % 2 == 1) {
+ continue;
+ }
+ else if (sub.str(2) == "\\\\b")
+ return;
+ if (lastpos < (size_t) sub.position(2))
+ s += t.substr(lastpos, sub.position(2) - lastpos);
+ s += "\\S";
+ lastpos = sub.position(2) + sub.length(2);
+ }
+ if (lastpos == 0) {
+ s = "\\b" + t + "\\b";
+ t = s;
+ return;
+ }
+ else if (lastpos < t.length())
+ s += t.substr(lastpos, t.length() - lastpos);
+ t = "\\b" + s + "\\b";
+}
+
+MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions & opt)
: p_buf(&buf), p_first_buf(&buf), opt(opt)
{
Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true);
previous_single_replace = true;
}
// When using regexp, braces are hacked already by escape_for_regex()
- par_as_string = normalize(ds, !use_regexp);
+ par_as_string = normalize(ds);
open_braces = 0;
close_wildcards = 0;
size_t lead_size = 0;
// correct the language settings
par_as_string = correctlanguagesetting(par_as_string, true, !opt.ignoreformat);
- if (opt.ignoreformat) {
- if (!use_regexp) {
- // if par_as_string_nolead were emty,
- // the following call to findAux will always *find* the string
- // in the checked data, and thus always using the slow
- // examining of the current text part.
- par_as_string_nolead = par_as_string;
+ opt.matchstart = false;
+ if (!use_regexp) {
+ identifyClosing(par_as_string); // Removes math closings ($, ], ...) at end of string
+ if (opt.ignoreformat) {
+ lead_size = 0;
}
- } else {
+ else {
+ lead_size = identifyLeading(par_as_string);
+ }
+ lead_as_string = par_as_string.substr(0, lead_size);
+ string lead_as_regex_string = string2regex(lead_as_string);
+ par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
+ string par_as_regex_string_nolead = string2regex(par_as_string_nolead);
+ /* Handle whole words too in this case
+ */
+ if (opt.matchword) {
+ par_as_regex_string_nolead = "\\b" + par_as_regex_string_nolead + "\\b";
+ opt.matchword = false;
+ }
+ string regexp_str = "(" + lead_as_regex_string + ")()" + par_as_regex_string_nolead;
+ string regexp2_str = "(" + lead_as_regex_string + ")(.*?)" + par_as_regex_string_nolead;
+ CreateRegexp(opt, regexp_str, regexp2_str);
+ use_regexp = true;
+ LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
+ LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
+ return;
+ }
+
+ if (!opt.ignoreformat) {
lead_size = identifyLeading(par_as_string);
LYXERR(Debug::FIND, "Lead_size: " << lead_size);
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);
- LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'");
- } else {
+ // Here we are using regexp
+ LASSERT(use_regexp, /**/);
+ {
string lead_as_regexp;
if (lead_size > 0) {
- // @todo No need to search for \regexp{} insets in leading material
- static std::regex specialChars { R"([-[\]{}()*+?.,\^$|#\s\\])" };
- lead_as_regexp = std::regex_replace(par_as_string.substr(0, lead_size), specialChars, R"(\$&)" );
- // lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat);
+ lead_as_regexp = string2regex(par_as_string.substr(0, lead_size));
+ regex_replace(par_as_string_nolead, par_as_string_nolead, "}$", "");
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 << "'");
}
LYXERR(Debug::FIND, "par_as_string before escape_for_regex() is '" << par_as_string << "'");
- par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
+ par_as_string = escape_for_regex(par_as_string);
// 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")
- // 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$3$4")
- // Insert .* before trailing '\}' ('}' has been escaped by escape_for_regex)
- || regex_replace(par_as_string, par_as_string, "(.*[^\\\\])(\\\\\\})\\'", "$1(.*?)$2")
- ) {
- ++close_wildcards;
- }
+ LYXERR(Debug::FIND, "par_as_string after correctRegex is '" << par_as_string << "'");
+ ++close_wildcards;
+ size_t lng = par_as_string.size();
if (!opt.ignoreformat) {
// Remove extra '\}' at end if not part of \{\.\}
- size_t lng = par_as_string.size();
while(lng > 2) {
if (par_as_string.substr(lng-2, 2).compare("\\}") == 0) {
if (lng >= 6) {
}
if (lng < par_as_string.size())
par_as_string = par_as_string.substr(0,lng);
- /*
- // save '\.'
- regex_replace(par_as_string, par_as_string, "\\\\\\.", "_xxbdotxx_");
- // handle '.' -> '[^]', replace later as '[^\}\{\\]'
- regex_replace(par_as_string, par_as_string, "\\.", "[^]");
- // replace '[^...]' with '[^...\}\{\\]'
- regex_replace(par_as_string, par_as_string, "\\[\\^([^\\\\\\]]*)\\]", "_xxbrlxx_$1\\}\\{\\\\_xxbrrxx_");
- regex_replace(par_as_string, par_as_string, "_xxbrlxx_", "[^");
- regex_replace(par_as_string, par_as_string, "_xxbrrxx_", "]");
- // restore '\.'
- regex_replace(par_as_string, par_as_string, "_xxbdotxx_", "\\.");
- */
+ }
+ if ((lng > 0) && (par_as_string[0] == '^')) {
+ par_as_string = par_as_string.substr(1);
+ --lng;
+ opt.matchstart = true;
}
LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
LYXERR(Debug::FIND, "Open braces: " << open_braces);
- LYXERR(Debug::FIND, "Close .*? : " << close_wildcards);
LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
// If entered regexp must match at begin of searched string buffer
string dest = "\\" + std::to_string(i+2);
while (regex_replace(par_as_string, par_as_string, orig, dest));
}
+ if (opt.matchword) {
+ modifyRegexForMatchWord(par_as_string);
+ opt.matchword = false;
+ }
regexp_str = "(" + lead_as_regexp + ")()" + par_as_string;
regexp2_str = "(" + lead_as_regexp + ")(.*?)" + par_as_string;
}
LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
-#if QTSEARCH
- // Handle \w properly
- QRegularExpression::PatternOptions popts = QRegularExpression::UseUnicodePropertiesOption | QRegularExpression::MultilineOption;
- if (! opt.casesensitive) {
- popts |= QRegularExpression::CaseInsensitiveOption;
- }
- regexp = QRegularExpression(QString::fromStdString(regexp_str), popts);
- regexp2 = QRegularExpression(QString::fromStdString(regexp2_str), popts);
- regexError = "";
- if (regexp.isValid() && regexp2.isValid()) {
- regexIsValid = true;
- // Check '{', '}' pairs inside the regex
- int balanced = 0;
- int skip = 1;
- for (unsigned i = 0; i < par_as_string.size(); i+= skip) {
- char c = par_as_string[i];
- if (c == '\\') {
- skip = 2;
- continue;
- }
- if (c == '{')
- balanced++;
- else if (c == '}') {
- balanced--;
- if (balanced < 0)
- break;
- }
- skip = 1;
- }
- if (balanced != 0) {
- regexIsValid = false;
- regexError = "Unbalanced curly brackets in regexp \"" + regexp_str + "\"";
- }
- }
- else {
- regexIsValid = false;
- if (!regexp.isValid())
- regexError += "Invalid regexp \"" + regexp_str + "\", error = " + regexp.errorString().toStdString();
- if (!regexp2.isValid())
- regexError += "Invalid regexp2 \"" + regexp2_str + "\", error = " + regexp2.errorString().toStdString();
- }
-#else
- if (opt.casesensitive) {
- regexp = regex(regexp_str);
- regexp2 = regex(regexp2_str);
- }
- else {
- regexp = regex(regexp_str, std::regex_constants::icase);
- regexp2 = regex(regexp2_str, std::regex_constants::icase);
- }
-#endif
+ CreateRegexp(opt, regexp_str, regexp2_str, par_as_string);
}
}
{
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;
if (use_regexp || opt.casesensitive)
- str = normalize(docstr, true);
+ str = normalize(docstr);
else
- str = normalize(lowercase(docstr), true);
+ str = normalize(lowercase(docstr));
if (!opt.ignoreformat) {
str = correctlanguagesetting(str, false, !opt.ignoreformat);
}
int matchend = match.capturedEnd(0);
while (mres.match_len > 0) {
QChar c = qstr.at(matchend - 1);
- if ((c == '\n') || (c == '}') || (c == '{')) {
+ if (c == '\n') {
mres.match_len--;
matchend--;
}
int res = mres.match_len;
LYXERR(Debug::FIND,
"res=" << res << ", at_begin=" << at_begin
- << ", matchword=" << opt.matchword
+ << ", matchstart=" << opt.matchstart
<< ", inTexted=" << cur.inTexted());
- if (res == 0 || !at_begin || !opt.matchword || !cur.inTexted())
+ if (opt.matchstart) {
+ if (cur.pos() != 0)
+ mres.match_len = 0;
+ else if (mres.match_prefix > 0)
+ mres.match_len = 0;
return mres;
- if ((len > 0) && (res < len)) {
- mres.match_len = 0;
- return mres;
- }
- Paragraph const & par = cur.paragraph();
- bool ws_left = (cur.pos() > 0)
- ? par.isWordSeparator(cur.pos() - 1)
- : true;
- bool ws_right;
- if (len < 0)
- ws_right = true;
- else {
- ws_right = (cur.pos() + len < par.size())
- ? par.isWordSeparator(cur.pos() + len)
- : true;
}
- LYXERR(Debug::FIND,
- "cur.pos()=" << cur.pos() << ", res=" << res
- << ", separ: " << ws_left << ", " << ws_right
- << ", len: " << len
- << endl);
- if (ws_left && ws_right) {
- // Check for word separators inside the found 'word'
- for (int i = 0; i < len; i++) {
- if (par.isWordSeparator(cur.pos() + i)) {
- mres.match_len = 0;
- return mres;
- }
- }
- return mres;
- }
- mres.match_len = 0;
- return mres;
+ else
+ return mres;
}
+#if 0
static bool simple_replace(string &t, string from, string to)
{
regex repl("(\\\\)*(" + from + ")");
t = s;
return true;
}
+#endif
-string MatchStringAdv::normalize(docstring const & s, bool hack_braces) const
+string MatchStringAdv::normalize(docstring const & s) const
{
string t;
t = lyx::to_utf8(s);
while (regex_replace(t, t, "\\\\((sub)?(((sub)?section)|paragraph)|part)\\*?(\\{(\\{\\})?\\})+", ""))
LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t);
while (regex_replace(t, t, "\\\\(foreignlanguage|textcolor|item)\\{[a-z]+\\}(\\{(\\{\\})?\\})+", ""));
- // 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 {
- simple_replace(t, "\\\\\\{", "_x_<");
- simple_replace(t, "\\\\\\}", "_x_>");
- LYXERR(Debug::FIND, "After {} replacement: '" << t << "'");
- }
- }
return t;
}
#if defined(ResultsDebug)
// Debugging output
-static void displayMResult(MatchResult &mres, int increment)
+static void displayMResult(MatchResult &mres, string from, DocIterator & cur)
{
- LYXERR0( "pos: " << mres.pos << " increment " << increment);
- LYXERR0( "leadsize: " << mres.leadsize);
- LYXERR0( "match_len: " << mres.match_len);
- LYXERR0( "match_prefix: " << mres.match_prefix);
- LYXERR0( "match2end: " << mres.match2end);
- for (size_t i = 0; i < mres.result.size(); i++)
- LYXERR0( "Match " << i << " = \"" << mres.result[i] << "\"");
-}
- #define displayMres(s,i) displayMResult(s,i);
-#else
- #define displayMres(s,i)
-#endif
-
-static bool findAdvForwardInnermost(DocIterator & cur)
-{
- size_t d;
- DocIterator old_cur(cur.buffer());
- int forwardCount = 0;
- do {
- d = cur.depth();
- old_cur = cur;
- cur.forwardPos();
- if (!cur) {
- break;
- }
- if (cur.depth() > d) {
- forwardCount++;
- continue;
+ LYXERR0( "from:\t\t\t" << from);
+ string status;
+ if (mres.pos_len > 0) {
+ // Set in finalize
+ status = "FINALSEARCH";
+ }
+ else {
+ if (mres.match_len > 0) {
+ if ((mres.match_prefix == 0) && (mres.pos == mres.leadsize))
+ status = "Good Match";
+ else
+ status = "Matched in";
}
- if (cur.depth() == d)
- break;
- } while(1);
- cur = old_cur;
- if (forwardCount > 0) {
- LYXERR(Debug::FIND, "Forwarded " << forwardCount << " step(s) (searching for innermost match)");
- return true;;
+ else
+ status = "MissedSearch";
}
- else
- return false;
+
+ LYXERR0( status << "(" << cur.pos() << " ... " << mres.searched_size + cur.pos() << ") cur.lastpos(" << cur.lastpos() << ")");
+ if ((mres.leadsize > 0) || (mres.match_len > 0) || (mres.match2end > 0))
+ LYXERR0( "leadsize(" << mres.leadsize << ") match_len(" << mres.match_len << ") match2end(" << mres.match2end << ")");
+ if ((mres.pos > 0) || (mres.match_prefix > 0))
+ LYXERR0( "pos(" << mres.pos << ") match_prefix(" << mres.match_prefix << ")");
+ for (size_t i = 0; i < mres.result.size(); i++)
+ LYXERR0( "Match " << i << " = \"" << mres.result[i] << "\"");
}
+ #define displayMres(s, txt, cur) displayMResult(s, txt, cur);
+#else
+ #define displayMres(s, txt, cur)
+#endif
/** 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
+ ** Return the cur.pos() difference between start and end of found match
**/
MatchResult &findAdvFinalize(DocIterator & cur, MatchStringAdv const & match, MatchResult const & expected = MatchResult(-1))
{
// 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);
- if (findAdvForwardInnermost(cur)) {
- mres = match(cur, -1, at_begin);
- displayMres(mres, 0);
- if (expected.match_len > 0) {
+ LASSERT(at_begin, /**/);
+ if (expected.match_len > 0 && at_begin) {
+ // Search for deepest match
+ old_cur = cur;
+ max_match = expected;
+ do {
+ size_t d = cur.depth();
+ cur.forwardPos();
+ if (!cur)
+ break;
+ if (cur.depth() < d)
+ break;
+ if (cur.depth() == d)
+ break;
+ size_t lastd = d;
+ while (cur && cur.depth() > lastd) {
+ lastd = cur.depth();
+ mres = match(cur, -1, at_begin);
+ displayMres(mres, "Checking innermost", cur);
+ if (mres.match_len > 0)
+ break;
+ // maybe deeper?
+ cur.forwardPos();
+ }
if (mres.match_len < expected.match_len)
- return fail;
- }
- else {
- if (mres.match_len <= 0)
- return fail;
- }
- max_match = mres.match_len;
+ break;
+ max_match = mres;
+ old_cur = cur;;
+ } while(1);
+ cur = old_cur;
}
- else if (expected.match_len < 0) {
+ else {
+ // (expected.match_len <= 0)
mres = match(cur); /* match valid only if not searching whole words */
- displayMres(mres, 0);
+ displayMres(mres, "Start with negative match", cur);
max_match = mres;
}
- else {
- max_match = expected;
- }
if (max_match.match_len <= 0) return fail;
LYXERR(Debug::FIND, "Ok");
// Compute the match length
- int len = 1;
+ int len = 1;
if (cur.pos() + len > cur.lastpos())
return fail;
// regexp should use \w+, \S+, or \b(some string)\b
// to search for whole words
if (match.opt.matchword && !match.use_regexp) {
- LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
- while (cur.pos() + len <= cur.lastpos() && match(cur, len).match_len <= 0) {
- ++len;
- LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
- }
- // Length of matched text (different from len param)
- static MatchResult old_match = match(cur, len, at_begin);
- if (old_match.match_len < 0)
- old_match = fail;
- MatchResult new_match;
- // Greedy behaviour while matching regexps
- while ((new_match = match(cur, len + 1, at_begin)).match_len > old_match.match_len) {
- ++len;
- old_match = new_match;
- LYXERR(Debug::FIND, "verifying match with len = " << len);
- }
+ LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
+ while (cur.pos() + len <= cur.lastpos() && match(cur, len).match_len <= 0) {
+ ++len;
+ LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
+ }
+ // Length of matched text (different from len param)
+ static MatchResult old_match = match(cur, len, at_begin);
+ if (old_match.match_len < 0)
+ old_match = fail;
+ MatchResult new_match;
+ // Greedy behaviour while matching regexps
+ while ((new_match = match(cur, len + 1, at_begin)).match_len > old_match.match_len) {
+ ++len;
+ old_match = new_match;
+ LYXERR(Debug::FIND, "verifying match with len = " << len);
+ }
+ displayMres(old_match, "SEARCH RESULT", cur)
return old_match;
- }
- else {
+ }
+ else {
int minl = 1;
int maxl = cur.lastpos() - cur.pos();
// Greedy behaviour while matching regexps
while (maxl > minl) {
- MatchResult mres2;
- mres2 = match(cur, len, at_begin);
- displayMres(mres2, len);
- int actual_match = mres2.match_len;
- if (actual_match >= max_match.match_len) {
- // actual_match > max_match _can_ happen,
+ MatchResult mres2;
+ mres2 = match(cur, len, at_begin);
+ displayMres(mres2, "Finalize loop", cur);
+ int actual_match_len = mres2.match_len;
+ if (actual_match_len >= max_match.match_len) {
+ // actual_match_len > max_match _can_ happen,
// if the search area splits
// some following word so that the regex
// (e.g. 'r.*r\b' matches 'r' from the middle of the
// splitted word)
// This means, the len value is too big
+ actual_match_len = max_match.match_len;
+ max_match = mres2;
+ max_match.match_len = actual_match_len;
maxl = len;
if (maxl - minl < 4)
len = (int)((maxl + minl)/2);
len = (int)(minl + (maxl - minl + 3)/4);
}
else {
- // (actual_match < max_match)
+ // (actual_match_len < max_match.match_len)
minl = len + 1;
len = (int)((maxl + minl)/2);
}
}
+ len = minl;
old_cur = cur;
// Search for real start of matched characters
while (len > 1) {
// Ha, got it! The shorter selection has the same match length
len--;
old_cur = cur;
+ max_match = actual_match;
}
else {
// OK, the shorter selection matches less chars, revert to previous value
actual_match = match(cur, len, at_begin);
if (actual_match.match_len == max_match.match_len) {
old_cur = cur;
- max_match = actual_match;
- }
+ max_match = actual_match;
+ }
}
}
- if (len == 0)
- return fail;
- else
- return max_match;
- }
+ if (len == 0)
+ return fail;
+ else {
+ max_match.pos_len = len;
+ displayMres(max_match, "SEARCH RESULT", cur)
+ return max_match;
+ }
+ }
}
/// Finds forward
{
if (!cur)
return 0;
+ bool repeat = false;
while (!theApp()->longOperationCancelled() && cur) {
- (void) findAdvForwardInnermost(cur);
+ //(void) findAdvForwardInnermost(cur);
LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
MatchResult mres = match(cur, -1, false);
- displayMres(mres,-1)
+ string msg = "Starting";
+ if (repeat)
+ msg = "Repeated";
+ displayMres(mres, msg + " findForwardAdv", cur)
int match_len = mres.match_len;
if ((mres.pos > 100000) || (mres.match2end > 100000) || (match_len > 100000)) {
LYXERR(Debug::INFO, "BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
match_len = 0;
}
- if (match_len > 0) {
+ if (match_len <= 0) {
+ // 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();
+ }
+ else { // match_len > 0
// Try to find the begin of searched string
int increment;
int firstInvalid = 100000;
- if (mres.match_prefix + mres.pos - mres.leadsize > 1)
- increment = (mres.match_prefix + mres.pos - mres.leadsize + 1)*3/4;
- else
- increment = 10;
+ {
+ int incrmatch = (mres.match_prefix + mres.pos - mres.leadsize + 1)*3/4;
+ int incrcur = (cur.lastpos() - cur.pos() + 1 )*3/4;
+ if (incrcur < incrmatch)
+ increment = incrcur;
+ else
+ increment = incrmatch;
+ if (increment < 1)
+ increment = 1;
+ }
LYXERR(Debug::FIND, "Set increment to " << increment);
while (increment > 0) {
DocIterator old_cur = cur;
- for (int i = 0; i < increment && cur; cur.forwardPos(), i++) {
+ size_t skipping = cur.depth();
+ for (int i = 0; i < increment && cur; i++) {
+ cur.forwardPos();
+ while (cur && cur.depth() > skipping) {
+ cur.pos() = cur.lastpos();
+ cur.forwardPos();
+ }
}
if (! cur || (cur.pit() > old_cur.pit())) {
// Are we outside of the paragraph?
}
else {
MatchResult mres2 = match(cur, -1, false);
- displayMres(mres2,increment)
+ displayMres(mres2, "findForwardAdv loop", cur)
switch (interpretMatch(mres, mres2)) {
case MatchResult::newIsTooFar:
// behind the expected match
increment /= 2;
break;
case MatchResult::newIsBetter:
- // not reached ye, but cur.pos()+increment is bettert
+ // not reached yet, but cur.pos()+increment is bettert
mres = mres2;
firstInvalid -= increment;
if (increment > firstInvalid*3/4)
default:
// Todo@
// Handle not like MatchResult::newIsTooFar
- // LYXERR0( "Something is wrong: Increment = " << increment << " match_prefix = " << mres.match_prefix);
+ LYXERR0( "Something is wrong: Increment = " << increment << " match_prefix = " << mres.match_prefix);
firstInvalid--;
increment = increment*3/4;
cur = old_cur;
}
}
}
+ if (mres.match_len > 0 && mres.match_prefix + mres.pos - mres.leadsize > 0) {
+ repeat = true;
+ cur.forwardPos();
+ continue;
+ }
// LYXERR0("Leaving first loop");
- {
- LYXERR(Debug::FIND, "Finalizing 1");
- MatchResult found_match = findAdvFinalize(cur, match, mres);
- if (found_match.match_len > 0) {
- match.FillResults(found_match);
- return found_match.match_len;
- }
- else {
- // try next possible match
- cur.forwardPos();
- continue;
- }
+ LYXERR(Debug::FIND, "Finalizing 1");
+ MatchResult found_match = findAdvFinalize(cur, match, mres);
+ if (found_match.match_len > 0) {
+ LASSERT(found_match.pos_len > 0, /**/);
+ match.FillResults(found_match);
+ return found_match.pos_len;
}
- // The following code is newer reached
- // but parts of it may be needed in future
- int match_len_zero_count = 0;
- MatchResult mres3;
- for (int i = 0; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) {
- if (i++ > 3) {
- mres3 = match(cur, -1, false);
- displayMres(mres3, 1)
- int remaining_len = mres3.match_len;
- if (remaining_len <= 0) {
- // Apparently the searched string is not in the remaining part
- break;
- }
- else {
- i = 0;
- }
- }
- LYXERR(Debug::FIND, "Advancing cur: " << cur);
- mres3 = match(cur, 1);
- displayMres(mres3, 1)
- int match_len3 = mres3.match_len;
- if (match_len3 < 0)
- continue;
- mres3 = match(cur);
- displayMres(mres3, 1)
- int match_len2 = mres3.match_len;
- LYXERR(Debug::FIND, "match_len2: " << match_len2);
- if (match_len2 > 0) {
- // Sometimes in finalize we understand it wasn't a match
- // and we need to continue the outest loop
- LYXERR(Debug::FIND, "Finalizing 2");
- MatchResult mres4 = findAdvFinalize(cur, match, mres.match_len);
- if (mres4.match_len > 0) {
- match.FillResults(mres4);
- return mres4.match_len;
- }
- }
- if (match_len2 > 0)
- match_len_zero_count = 0;
- else if (match_len2 == 0)
- match_len_zero_count++;
- if (match_len2 < 0) {
- if (++match_len_zero_count > 3) {
- LYXERR(Debug::FIND, "match_len2_zero_count: " << match_len_zero_count << ", match_len was " << match_len);
- }
- break;
- }
+ else {
+ // try next possible match
+ cur.forwardPos();
+ repeat = false;
+ continue;
}
- if (!cur)
- return 0;
- }
- if (match_len >= 0 && 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();
}
}
return 0;
if (found_match) {
MatchResult found_mr = findMostBackwards(cur, match);
match.FillResults(found_mr);
- return found_mr.match_len;
+ LASSERT(found_mr.pos_len > 0, /**/);
+ return found_mr.pos_len;
}
// Stop if begin of document reached
right = pit->size();
pit->changeCase(buffer.params(), pos_type(1), right, others_case);
}
-
} // namespace
+#if 1
+static bool replaceMatches(string &t, int maxmatchnum, vector <string> const & replacements)
+{
+ // Should replace the string "$" + std::to_string(matchnum) with replacement
+ // if the char '$' is not prefixed with odd number of char '\\'
+ static regex const rematch("(\\\\)*(\\$\\$([0-9]))");
+ string s;
+ size_t lastpos = 0;
+ smatch sub;
+ for (sregex_iterator it(t.begin(), t.end(), rematch), end; it != end; ++it) {
+ sub = *it;
+ if ((sub.position(2) - sub.position(0)) % 2 == 1)
+ continue;
+ int num = stoi(sub.str(3), nullptr, 10);
+ if (num >= maxmatchnum)
+ continue;
+ if (lastpos < (size_t) sub.position(2))
+ s += t.substr(lastpos, sub.position(2) - lastpos);
+ s += replacements[num];
+ lastpos = sub.position(2) + sub.length(2);
+ }
+ if (lastpos == 0)
+ return false;
+ else if (lastpos < t.length())
+ s += t.substr(lastpos, t.length() - lastpos);
+ t = s;
+ return true;
+}
+#endif
+
///
static int findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv)
{
ostringstream oss;
repl_buffer_orig.write(oss);
string lyx = oss.str();
+ if (matchAdv.valid_matches > 0) {
+ replaceMatches(lyx, matchAdv.valid_matches, matchAdv.matches);
+ }
Buffer repl_buffer("", false);
repl_buffer.setUnnamed(true);
LASSERT(repl_buffer.readString(lyx), return 0);
/// Perform a FindAdv operation.
-bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt)
+bool findAdv(BufferView * bv, FindAndReplaceOptions & opt)
{
DocIterator cur;
- int match_len = 0;
+ int pos_len = 0;
// e.g., when invoking word-findadv from mini-buffer wither with
// wrong options syntax or before ever opening advanced F&R pane
num_replaced += findAdvReplace(bv, opt, matchAdv);
cur = bv->cursor();
if (opt.forward)
- match_len = findForwardAdv(cur, matchAdv);
+ pos_len = findForwardAdv(cur, matchAdv);
else
- match_len = findBackwardsAdv(cur, matchAdv);
+ pos_len = findBackwardsAdv(cur, matchAdv);
} catch (exception & ex) {
bv->message(from_utf8(ex.what()));
return false;
}
- if (match_len == 0) {
+ if (pos_len == 0) {
if (num_replaced > 0) {
switch (num_replaced)
{
else
bv->message(_("Match found."));
- LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << match_len);
- bv->putSelectionAt(cur, match_len, !opt.forward);
+ if (cur.pos() + pos_len > cur.lastpos()) {
+ // Prevent crash in bv->putSelectionAt()
+ // Should never happen, maybe LASSERT() here?
+ pos_len = cur.lastpos() - cur.pos();
+ }
+ LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << pos_len);
+ bv->putSelectionAt(cur, pos_len, !opt.forward);
return true;
}
projects / lyx.git / blobdiff