size_t new_pos = s.find("\\regexp{", pos);
if (new_pos == string::npos)
new_pos = s.size();
- LYXERR(Debug::FIND, "new_pos: " << new_pos);
- string t = apply_escapes(s.substr(pos, new_pos - pos), get_lyx_unescapes());
- LYXERR(Debug::FIND, "t [lyx]: " << t);
- t = apply_escapes(t, get_regexp_escapes());
- LYXERR(Debug::FIND, "t [rxp]: " << t);
- s.replace(pos, new_pos - pos, t);
- new_pos = pos + t.size();
- LYXERR(Debug::FIND, "Regexp after escaping: " << s);
- LYXERR(Debug::FIND, "new_pos: " << new_pos);
- if (new_pos == s.size())
- break;
+ string t;
+ if (new_pos > pos) {
+ 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);
}
+class MatchResult {
+public:
+ int match_len;
+ int match2end;
+ int pos;
+ MatchResult(): match_len(0),match2end(0), pos(0) {};
+};
+
/** The class performing a match between a position in the document and the FindAdvOptions.
**/
class MatchStringAdv {
** @return
** The length of the matching text, or zero if no match was found.
**/
- int operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
+ MatchResult operator()(DocIterator const & cur, int len = -1, bool at_begin = true) const;
public:
/// buffer
private:
/// Auxiliary find method (does not account for opt.matchword)
- int findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
+ MatchResult findAux(DocIterator const & cur, int len = -1, bool at_begin = true) const;
/** Normalize a stringified or latexified LyX paragraph.
**
otexstream os(ods);
runparams.nice = true;
runparams.flavor = OutputParams::LATEX;
- runparams.linelen = 100000; //lyxrc.plaintext_linelen;
+ runparams.linelen = 10000; //lyxrc.plaintext_linelen;
// No side effect of file copying and image conversion
runparams.dryrun = true;
runparams.for_search = true;
OutputParams runparams(&buffer.params().encoding());
runparams.nice = true;
runparams.flavor = OutputParams::LATEX;
- runparams.linelen = 100000; //lyxrc.plaintext_linelen;
+ runparams.linelen = 10000; //lyxrc.plaintext_linelen;
runparams.dryrun = true;
runparams.for_search = true;
for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) {
int findclosing(int start, int end, char up, char down, int repeat);
void handleParentheses(int lastpos, bool closingAllowed);
bool hasTitle;
+ int isOpeningPar(int pos);
string titleValue;
void output(ostringstream &os, int lastpos);
// string show(int lastpos);
};
+int Intervall::isOpeningPar(int pos)
+{
+ if ((pos < 0) || (size_t(pos) >= par.size()))
+ return 0;
+ if (par[pos] != '{')
+ return 0;
+ if (size_t(pos) + 2 >= par.size())
+ return 1;
+ if (par[pos+2] != '}')
+ return 1;
+ if (par[pos+1] == '[' || par[pos+1] == ']')
+ return 3;
+ return 1;
+}
+
void Intervall::setForDefaultLang(KeyInfo &defLang)
{
// Enable the use of first token again
found._dataStart = found._dataEnd;
found._tokensize = found._dataEnd - found._tokenstart;
found.parenthesiscount = 0;
+ found.head = interval.par.substr(found._tokenstart, found._tokensize);
}
+ else
+ continue;
}
else {
if (evaluatingMath) {
found._dataEnd = found._tokenstart + found._tokensize;
found._dataStart = found._dataEnd;
found.parenthesiscount = 0;
+ found.head = interval.par.substr(found._tokenstart, found._tokensize);
evaluatingMath = true;
}
else {
found.head = interval.par.substr(found._tokenstart, found._tokensize);
}
else {
+ // Swallow possible optional params
while (interval.par[pos1] == '[') {
pos1 = interval.findclosing(pos1+1, interval.par.length(), '[', ']')+1;
}
+ // Swallow also the eventual parameter
if (interval.par[pos1] == '{') {
found._dataEnd = interval.findclosing(pos1+1, interval.par.length()) + 1;
}
found._dataStart = found._dataEnd;
found._tokensize = count + found._dataEnd - pos;
found.parenthesiscount = 0;
+ found.head = interval.par.substr(found._tokenstart, found._tokensize);
found.disabled = true;
}
}
found._dataEnd = found._dataStart;
found._tokensize = count + found._dataEnd - pos;
found.parenthesiscount = 0;
+ found.head = interval.par.substr(found._tokenstart, found._tokensize);
found.disabled = true;
}
}
// Skip
// makeKey("enskip|smallskip|medskip|bigskip|vfill", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
// Custom space/skip, remove the content (== length value)
- makeKey("vspace|hspace|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
+ makeKey("vspace|vspace*|hspace|hspace*|mspace", KeyInfo(KeyInfo::noContent, 1, false), isPatternString);
// Found in fr/UserGuide.lyx
makeKey("og|fg", KeyInfo(KeyInfo::isChar, 0, false), isPatternString);
// quotes
makeKey("trianglerightpar|hexagonpar|starpar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
makeKey("triangleuppar|triangledownpar|droppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
makeKey("triangleleftpar|shapepar|dropuppar", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
- makeKey("hphantom|footnote|shortcut|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
+ makeKey("hphantom|vphantom|footnote|shortcut|include|includegraphics", KeyInfo(KeyInfo::isStandard, 1, true), isPatternString);
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);
void LatexInfo::processRegion(int start, int region_end)
{
while (start < region_end) { /* Let {[} and {]} survive */
- if ((interval.par[start] == '{') &&
- (interval.par[start+1] != ']') &&
- (interval.par[start+1] != '[')) {
+ int cnt = interval.isOpeningPar(start);
+ if (cnt == 1) {
// Closing is allowed past the region
int closing = interval.findclosing(start+1, interval.par.length());
interval.addIntervall(start, start+1);
interval.addIntervall(closing, closing+1);
}
+ else if (cnt == 3)
+ start += 2;
start = interval.nextNotIgnored(start+1);
}
}
}
// Remove possible empty data
int dstart = interval.nextNotIgnored(actual._dataStart);
- while ((dstart < output_end) && (interval.par[dstart] == '{')) {
+ while (interval.isOpeningPar(dstart) == 1) {
interval.addIntervall(dstart, dstart+1);
int dend = interval.findclosing(dstart+1, output_end);
interval.addIntervall(dend, dend+1);
++close_wildcards;
}
if (!opt.ignoreformat) {
- // Remove extra '\}' at end
- while ( regex_replace(par_as_string, par_as_string, "(.*)\\\\}$", "$1")) {
- open_braces++;
+ // 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 (par_as_string.substr(lng-6,3).compare("\\{\\") == 0)
+ break;
+ }
+ lng -= 2;
+ open_braces++;
+ }
+ else
+ break;
}
+ 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_");
return count;
}
-int MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
+MatchResult MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) const
{
+ MatchResult mres;
+
if (at_begin &&
(opt.restr == FindAndReplaceOptions::R_ONLY_MATHS && !cur.inMathed()) )
- return 0;
+ return mres;
docstring docstr = stringifyFromForSearch(opt, cur, len);
string str = normalize(docstr, true);
if (!opt.ignoreformat) {
str = correctlanguagesetting(str, false, !opt.ignoreformat);
}
- if (str.empty()) return(-1);
+ if (str.empty()) {
+ mres.match_len = -1;
+ return mres;
+ }
LYXERR(Debug::FIND, "Matching against '" << lyx::to_utf8(docstr) << "'");
LYXERR(Debug::FIND, "After normalization: '" << str << "'");
}
sregex_iterator re_it(str.begin(), str.end(), *p_regexp, flags);
if (re_it == sregex_iterator())
- return 0;
+ return mres;
match_results<string::const_iterator> const & m = *re_it;
if (0) { // Kornel Benko: DO NOT CHECKK
// Check braces on the segment that matched the entire regexp expression,
// plus the last subexpression, if a (.*?) was inserted in the constructor.
if (!braces_match(m[0].first, m[0].second, open_braces))
- return 0;
+ return mres;
}
// Check braces on segments that matched all (.*?) subexpressions,
// except the last "padding" one inserted by lyx.
for (size_t i = 1; i < m.size() - 1; ++i)
if (!braces_match(m[i].first, m[i].second, open_braces))
- return 0;
+ return mres;
// Exclude from the returned match length any length
// due to close wildcards added at end of regexp
result -= leadingsize;
else
result = 0;
- return computeSize(str.substr(pos+leadingsize,result), result);
+ mres.match_len = computeSize(str.substr(pos+leadingsize,result), result);
+ mres.match2end = str.size() - pos - leadingsize;
+ mres.pos = pos+leadingsize;
+ return mres;
}
// else !use_regexp: but all code paths above return
if (at_begin) {
LYXERR(Debug::FIND, "size=" << par_as_string.size()
<< ", substr='" << str.substr(0, par_as_string.size()) << "'");
- if (str.substr(0, par_as_string.size()) == par_as_string)
- return par_as_string.size();
+ if (str.substr(0, par_as_string.size()) == par_as_string) {
+ mres.match_len = par_as_string.size();
+ mres.match2end = str.size();
+ mres.pos = 0;
+ return mres;
+ }
} else {
size_t pos = str.find(par_as_string_nolead);
- if (pos != string::npos)
- return par_as_string.size();
+ if (pos != string::npos) {
+ mres.match_len = par_as_string.size();
+ mres.match2end = str.size() - pos;
+ mres.pos = pos;
+ return mres;
+ }
}
- return 0;
+ return mres;
}
-int MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
+MatchResult MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const
{
- int res = findAux(cur, len, at_begin);
+ MatchResult mres = findAux(cur, len, at_begin);
+ int res = mres.match_len;
LYXERR(Debug::FIND,
"res=" << res << ", at_begin=" << at_begin
<< ", matchword=" << opt.matchword
<< ", inTexted=" << cur.inTexted());
if (res == 0 || !at_begin || !opt.matchword || !cur.inTexted())
- return res;
- if ((len > 0) && (res < len))
- return 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)
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))
- return 0;
+ if (par.isWordSeparator(cur.pos() + i)) {
+ mres.match_len = 0;
+ return mres;
+ }
}
- return res;
+ return mres;
}
- return 0;
+ mres.match_len = 0;
+ return mres;
}
OutputParams runparams(&cur.buffer()->params().encoding());
runparams.nice = true;
runparams.flavor = OutputParams::LATEX;
- runparams.linelen = 100000; //lyxrc.plaintext_linelen;
+ runparams.linelen = 10000; //lyxrc.plaintext_linelen;
// No side effect of file copying and image conversion
runparams.dryrun = true;
LYXERR(Debug::FIND, "Stringifying with cur: "
<< cur << ", from pos: " << cur.pos() << ", end: " << end);
- return par.asString(cur.pos(), end,
+ docstring result = par.asString(cur.pos(), end,
AS_STR_INSETS | AS_STR_SKIPDELETE | AS_STR_PLAINTEXT,
&runparams);
+ return result;
} else if (cur.inMathed()) {
CursorSlice cs = cur.top();
MathData md = cs.cell();
d = cur.depth();
old_cur = cur;
cur.forwardPos();
- } while (cur && cur.depth() > d && match(cur) > 0);
+ } while (cur && cur.depth() > d && match(cur).match_len > 0);
cur = old_cur;
- int max_match = match(cur); /* match valid only if not searching whole words */
+ int max_match = match(cur).match_len; /* match valid only if not searching whole words */
if (max_match <= 0) return 0;
LYXERR(Debug::FIND, "Ok");
return 0;
if (match.opt.matchword) {
LYXERR(Debug::FIND, "verifying unmatch with len = " << len);
- while (cur.pos() + len <= cur.lastpos() && match(cur, len) <= 0) {
+ 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)
- int old_match = match(cur, len);
+ int old_match = match(cur, len).match_len;
if (old_match < 0)
old_match = 0;
int new_match;
// Greedy behaviour while matching regexps
- while ((new_match = match(cur, len + 1)) > old_match) {
+ while ((new_match = match(cur, len + 1).match_len) > old_match) {
++len;
old_match = new_match;
LYXERR(Debug::FIND, "verifying match with len = " << len);
int maxl = cur.lastpos() - cur.pos();
// Greedy behaviour while matching regexps
while (maxl > minl) {
- int actual_match = match(cur, len);
+ int actual_match = match(cur, len).match_len;
if (actual_match >= max_match) {
// actual_match > max_match _can_ happen,
// if the search area splits
}
if (cur.pos() != old_cur.pos()) {
// OK, forwarded 1 pos in actual inset
- actual_match = match(cur, len-1);
+ actual_match = match(cur, len-1).match_len;
if (actual_match == max_match) {
// Ha, got it! The shorter selection has the same match length
len--;
}
else {
LYXERR0("cur.pos() == old_cur.pos(), this should never happen");
- actual_match = match(cur, len);
+ actual_match = match(cur, len).match_len;
if (actual_match == max_match)
old_cur = cur;
}
return 0;
while (!theApp()->longOperationCancelled() && cur) {
LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur);
- int match_len = match(cur, -1, false);
+ MatchResult mres = match(cur, -1, false);
+ int match_len = mres.match_len;
LYXERR(Debug::FIND, "match_len: " << match_len);
+ if ((mres.pos > 10000) || (mres.match2end > 10000) || (match_len > 10000)) {
+ LYXERR0("BIG LENGTHS: " << mres.pos << ", " << match_len << ", " << mres.match2end);
+ match_len = 0;
+ }
if (match_len > 0) {
+ // Try to find the begin of searched string
+ int increment = mres.pos/2;
+ while (mres.pos > 5 && (increment > 5)) {
+ DocIterator old_cur = cur;
+ for (int i = 0; i < increment && cur; cur.forwardPos(), i++) {
+ /*
+ while (cur && cur.depth() != old_cur.depth())
+ cur.forwardPos();
+ */
+ }
+ if (! cur) {
+ cur = old_cur;
+ increment /= 2;
+ }
+ else {
+ MatchResult mres2 = match(cur, -1, false);
+ if ((mres2.match2end < mres.match2end) ||
+ (mres2.match_len < mres.match_len)) {
+ cur = old_cur;
+ increment /= 2;
+ }
+ else {
+ mres = mres2;
+ increment -= 2;
+ if (increment > mres.pos/2)
+ increment = mres.pos/2;
+ }
+ }
+ }
int match_len_zero_count = 0;
- for (; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) {
+ for (int i = 0; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) {
+ if (i++ > 10) {
+ int remaining_len = match(cur, -1, false).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);
- int match_len3 = match(cur, 1);
+ int match_len3 = match(cur, 1).match_len;
if (match_len3 < 0)
continue;
- int match_len2 = match(cur);
+ int match_len2 = match(cur).match_len;
LYXERR(Debug::FIND, "match_len2: " << match_len2);
if (match_len2 > 0) {
// Sometimes in finalize we understand it wasn't a match
LYXERR(Debug::FIND, "findMostBackwards(): cur=" << cur);
DocIterator new_cur = cur;
new_cur.backwardPos();
- if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur))
+ if (new_cur == cur || &new_cur.inset() != &inset || !match(new_cur).match_len)
break;
int new_len = findAdvFinalize(new_cur, match);
if (new_len == len)
bool pit_changed = false;
do {
cur.pos() = 0;
- bool found_match = match(cur, -1, false);
+ bool found_match = (match(cur, -1, false).match_len > 0);
if (found_match) {
if (pit_changed)
LYXERR(Debug::FIND, "findBackAdv2: cur: " << cur);
DocIterator cur_prev_iter;
do {
- found_match = match(cur);
+ found_match = (match(cur).match_len > 0);
LYXERR(Debug::FIND, "findBackAdv3: found_match="
<< found_match << ", cur: " << cur);
if (found_match)
return;
LASSERT(sel_len > 0, return);
- if (!matchAdv(sel_beg, sel_len))
+ if (!matchAdv(sel_beg, sel_len).match_len)
return;
// Build a copy of the replace buffer, adapted to the KeepCase option
projects / lyx.git / blobdiff