X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2Flyxfind.cpp;h=73700cb4fcef54fc898044fb153f6ff8ff170ede;hb=90f7007a2e6c78ffd031e4636ff909ab1bc2ddec;hp=73b9df371885cb7a63c8e6f7d56f1bd62dc7befc;hpb=22d245312afb61f2267cf86714ab71f7926a1614;p=lyx.git diff --git a/src/lyxfind.cpp b/src/lyxfind.cpp index 73b9df3718..73700cb4fc 100644 --- a/src/lyxfind.cpp +++ b/src/lyxfind.cpp @@ -33,12 +33,14 @@ #include "TexRow.h" #include "Text.h" +#include "frontends/Application.h" #include "frontends/alert.h" #include "mathed/InsetMath.h" #include "mathed/InsetMathGrid.h" #include "mathed/InsetMathHull.h" #include "mathed/MathStream.h" +#include "mathed/MathSupport.h" #include "support/convert.h" #include "support/debug.h" @@ -92,7 +94,7 @@ private: int findForward(DocIterator & cur, MatchString const & match, - bool find_del = true) + bool find_del = true) { for (; cur; cur.forwardChar()) if (cur.inTexted()) { @@ -105,7 +107,7 @@ int findForward(DocIterator & cur, MatchString const & match, int findBackwards(DocIterator & cur, MatchString const & match, - bool find_del = true) + bool find_del = true) { while (cur) { cur.backwardChar(); @@ -147,7 +149,7 @@ bool searchAllowed(docstring const & str) bool findOne(BufferView * bv, docstring const & searchstr, - bool case_sens, bool whole, bool forward, bool find_del = true) + bool case_sens, bool whole, bool forward, bool find_del = true) { if (!searchAllowed(searchstr)) return false; @@ -156,8 +158,9 @@ bool findOne(BufferView * bv, docstring const & searchstr, MatchString const match(searchstr, case_sens, whole); - int match_len = forward ? findForward(cur, match, find_del) : - findBackwards(cur, match, find_del); + int match_len = forward + ? findForward(cur, match, find_del) + : findBackwards(cur, match, find_del); if (match_len > 0) bv->putSelectionAt(cur, match_len, !forward); @@ -191,11 +194,13 @@ int replaceAll(BufferView * bv, pos_type const pos = cur.pos(); Font const font = cur.paragraph().getFontSettings(buf.params(), pos); cur.recordUndo(); - int striked = ssize - cur.paragraph().eraseChars(pos, pos + match_len, - buf.params().trackChanges); + int striked = ssize - + cur.paragraph().eraseChars(pos, pos + match_len, + buf.params().trackChanges); cur.paragraph().insert(pos, replacestr, font, - Change(buf.params().trackChanges ? - Change::INSERTED : Change::UNCHANGED)); + Change(buf.params().trackChanges + ? Change::INSERTED + : Change::UNCHANGED)); for (int i = 0; i < rsize + striked; ++i) cur.forwardChar(); ++num; @@ -227,20 +232,20 @@ int replaceAll(BufferView * bv, // returns the number of replacements made (one, if any) and // whether anything at all was done. pair replaceOne(BufferView * bv, docstring searchstr, - docstring const & replacestr, bool case_sens, - bool whole, bool forward) + docstring const & replacestr, bool case_sens, + bool whole, bool forward, bool findnext) { Cursor & cur = bv->cursor(); if (!cur.selection()) { // no selection, non-empty search string: find it if (!searchstr.empty()) { findOne(bv, searchstr, case_sens, whole, forward); - return pair(true, 0); + return make_pair(true, 0); } // empty search string if (!cur.inTexted()) // bail in math - return pair(false, 0); + return make_pair(false, 0); // select current word and treat it as the search string cur.innerText()->selectWord(cur, WHOLE_WORD); searchstr = cur.selectionAsString(false); @@ -249,44 +254,43 @@ pair replaceOne(BufferView * bv, docstring searchstr, // if we still don't have a search string, report the error // and abort. if (!searchAllowed(searchstr)) - return pair(false, 0); + return make_pair(false, 0); bool have_selection = cur.selection(); docstring const selected = cur.selectionAsString(false); bool match = - case_sens ? searchstr == selected - : compare_no_case(searchstr, selected) == 0; + case_sens + ? searchstr == selected + : compare_no_case(searchstr, selected) == 0; // no selection or current selection is not search word: // just find the search word if (!have_selection || !match) { findOne(bv, searchstr, case_sens, whole, forward); - return pair(true, 0); + return make_pair(true, 0); } // we're now actually ready to replace. if the buffer is // read-only, we can't, though. if (bv->buffer().isReadonly()) - return pair(false, 0); + return make_pair(false, 0); - cap::replaceSelectionWithString(cur, replacestr, forward); + cap::replaceSelectionWithString(cur, replacestr); if (forward) { cur.pos() += replacestr.length(); LASSERT(cur.pos() <= cur.lastpos(), /* */); - } else { - cur.pos() -= replacestr.length(); - LASSERT(cur.pos() >= 0, /* */); } - findOne(bv, searchstr, case_sens, whole, forward, false); + if (findnext) + findOne(bv, searchstr, case_sens, whole, forward, false); - return pair(true, 1); + return make_pair(true, 1); } } // namespace anon docstring const find2string(docstring const & search, - bool casesensitive, bool matchword, bool forward) + bool casesensitive, bool matchword, bool forward) { odocstringstream ss; ss << search << '\n' @@ -298,8 +302,9 @@ docstring const find2string(docstring const & search, docstring const replace2string(docstring const & replace, - docstring const & search, bool casesensitive, bool matchword, - bool all, bool forward) + docstring const & search, + bool casesensitive, bool matchword, + bool all, bool forward, bool findnext) { odocstringstream ss; ss << replace << '\n' @@ -307,7 +312,8 @@ docstring const replace2string(docstring const & replace, << int(casesensitive) << ' ' << int(matchword) << ' ' << int(all) << ' ' - << int(forward); + << int(forward) << ' ' + << int(findnext); return ss.str(); } @@ -342,7 +348,7 @@ bool lyxreplace(BufferView * bv, // data is of the form // " // - // " + // " docstring search; docstring rplc; docstring howto = split(ev.argument(), rplc, '\n'); @@ -352,6 +358,7 @@ bool lyxreplace(BufferView * bv, bool matchword = parse_bool(howto); bool all = parse_bool(howto); bool forward = parse_bool(howto); + bool findnext = howto.empty() ? true : parse_bool(howto); int replace_count = 0; bool update = false; @@ -361,8 +368,8 @@ bool lyxreplace(BufferView * bv, replace_count = replaceAll(bv, search, rplc, casesensitive, matchword); update = replace_count > 0; } else { - pair rv = - replaceOne(bv, search, rplc, casesensitive, matchword, forward); + pair rv = + replaceOne(bv, search, rplc, casesensitive, matchword, forward, findnext); update = rv.first; replace_count = rv.second; } @@ -382,7 +389,7 @@ bool lyxreplace(BufferView * bv, buf.message(str); } } - } else { + } 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)) @@ -485,37 +492,65 @@ namespace { typedef vector > Escapes; /// A map of symbols and their escaped equivalent needed within a regex. +/// @note Beware of order Escapes const & get_regexp_escapes() { + typedef std::pair P; + static Escapes escape_map; if (escape_map.empty()) { - escape_map.push_back(pair("\\", "\\\\")); - escape_map.push_back(pair("^", "\\^")); - escape_map.push_back(pair("$", "\\$")); - escape_map.push_back(pair("{", "\\{")); - escape_map.push_back(pair("}", "\\}")); - escape_map.push_back(pair("[", "\\[")); - escape_map.push_back(pair("]", "\\]")); - escape_map.push_back(pair("(", "\\(")); - escape_map.push_back(pair(")", "\\)")); - escape_map.push_back(pair("+", "\\+")); - escape_map.push_back(pair("*", "\\*")); - escape_map.push_back(pair(".", "\\.")); + escape_map.push_back(P("$", "_x_$")); + escape_map.push_back(P("{", "_x_{")); + escape_map.push_back(P("}", "_x_}")); + escape_map.push_back(P("[", "_x_[")); + escape_map.push_back(P("]", "_x_]")); + escape_map.push_back(P("(", "_x_(")); + escape_map.push_back(P(")", "_x_)")); + escape_map.push_back(P("+", "_x_+")); + escape_map.push_back(P("*", "_x_*")); + escape_map.push_back(P(".", "_x_.")); + escape_map.push_back(P("\\", "(?:\\\\|\\\\backslash)")); + escape_map.push_back(P("~", "(?:\\\\textasciitilde|\\\\sim)")); + escape_map.push_back(P("^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\mathcircumflex)")); + escape_map.push_back(P("_x_", "\\")); } return escape_map; } /// A map of lyx escaped strings and their unescaped equivalent. -Escapes const & get_lyx_unescapes() { +Escapes const & get_lyx_unescapes() +{ + typedef std::pair P; + + static Escapes escape_map; + if (escape_map.empty()) { + escape_map.push_back(P("\\%", "%")); + escape_map.push_back(P("\\mathcircumflex ", "^")); + escape_map.push_back(P("\\mathcircumflex", "^")); + escape_map.push_back(P("\\backslash ", "\\")); + escape_map.push_back(P("\\backslash", "\\")); + escape_map.push_back(P("\\\\{", "_x_<")); + escape_map.push_back(P("\\\\}", "_x_>")); + escape_map.push_back(P("\\sim ", "~")); + escape_map.push_back(P("\\sim", "~")); + } + 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 P; + static Escapes escape_map; if (escape_map.empty()) { - escape_map.push_back(pair("{*}", "*")); - escape_map.push_back(pair("{[}", "[")); - escape_map.push_back(pair("\\$", "$")); - escape_map.push_back(pair("\\backslash{}", "\\")); - escape_map.push_back(pair("\\backslash", "\\")); - escape_map.push_back(pair("\\sim ", "~")); - escape_map.push_back(pair("\\^", "^")); + escape_map.push_back(P("\\\\", "(?:\\\\\\\\|\\\\backslash|\\\\textbackslash\\{\\})")); + escape_map.push_back(P("(first, pos)) < s.length()) { s.replace(pos, it->first.length(), it->second); -// LYXERR(Debug::FIND, "After escape: " << s); + LYXERR(Debug::FIND, "After escape: " << s); pos += it->second.length(); // LYXERR(Debug::FIND, "pos: " << pos); } @@ -541,67 +576,57 @@ string apply_escapes(string s, Escapes const & escape_map) return s; } -/** Return the position of the closing brace matching the open one at s[pos], - ** or s.size() if not found. - **/ -size_t find_matching_brace(string const & s, size_t pos) -{ - LASSERT(s[pos] == '{', /* */); - int open_braces = 1; - for (++pos; pos < s.size(); ++pos) { - if (s[pos] == '\\') - ++pos; - else if (s[pos] == '{') - ++open_braces; - else if (s[pos] == '}') { - --open_braces; - if (open_braces == 0) - return pos; - } - } - return s.size(); -} -/// Within \regexp{} apply get_regex_escapes(), while outside apply get_lyx_unescapes(). -string escape_for_regex(string 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) { size_t pos = 0; while (pos < s.size()) { - size_t new_pos = s.find("\\regexp{{{", pos); + 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 : " << t); + LYXERR(Debug::FIND, "t [lyx]: " << t); t = apply_escapes(t, get_regexp_escapes()); - LYXERR(Debug::FIND, "t : " << t); + 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; - size_t end_pos = s.find("}}}", new_pos + 10); // find_matching_brace(s, new_pos + 7); + // 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 = apply_escapes(s.substr(new_pos + 10, end_pos - (new_pos + 10)), get_lyx_unescapes()); - LYXERR(Debug::FIND, "t : " << t); + 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 [lyx]: " << t); + if (match_latex) { + t = apply_escapes(t, get_regexp_latex_escapes()); + LYXERR(Debug::FIND, "t in regexp [ltx]: " << t); + } if (end_pos == s.size()) { s.replace(new_pos, end_pos - new_pos, t); pos = s.size(); LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s); break; } - s.replace(new_pos, end_pos + 3 - new_pos, t); - LYXERR(Debug::FIND, "Regexp after \\regexp{} removal: " << s); + 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); } return s; } + /// Wrapper for lyx::regex_replace with simpler interface bool regex_replace(string const & s, string & t, string const & searchstr, - string const & replacestr) + string const & replacestr) { lyx::regex e(searchstr); ostringstream oss; @@ -613,6 +638,7 @@ bool regex_replace(string const & s, string & t, string const & searchstr, return rv; } + /** Checks if supplied string segment is well-formed from the standpoint of matching open-closed braces. ** ** Verify that closed braces exactly match open braces. This avoids that, for example, @@ -645,15 +671,16 @@ bool braces_match(string::const_iterator const & beg, } } if (open_pars != unmatched) { - LYXERR(Debug::FIND, "Found " << open_pars - << " instead of " << unmatched - << " unmatched open braces at the end of count"); - return false; + LYXERR(Debug::FIND, "Found " << open_pars + << " instead of " << unmatched + << " unmatched open braces at the end of count"); + return false; } LYXERR(Debug::FIND, "Braces match as expected"); return true; } + /** The class performing a match between a position in the document and the FindAdvOptions. **/ class MatchStringAdv { @@ -697,64 +724,155 @@ private: ** @todo Normalization should also expand macros, if the corresponding ** search option was checked. **/ - string normalize(docstring const & s) const; + string normalize(docstring const & s, bool hack_braces) const; // normalized string to search string par_as_string; // regular expression to use for searching lyx::regex regexp; // same as regexp, but prefixed with a ".*" lyx::regex regexp2; + // leading format material as string + string lead_as_string; + // 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; // number of (.*?) subexpressions added at end of search regexp for closing // environments, math mode, styles, etc... int close_wildcards; + // Are we searching with regular expressions ? + bool use_regexp; }; +static docstring buffer_to_latex(Buffer & buffer) +{ + OutputParams runparams(&buffer.params().encoding()); + TexRow texrow; + odocstringstream ods; + otexstream os(ods, texrow); + runparams.nice = true; + runparams.flavor = OutputParams::LATEX; + runparams.linelen = 80; //lyxrc.plaintext_linelen; + // No side effect of file copying and image conversion + runparams.dryrun = true; + pit_type const endpit = buffer.paragraphs().size(); + for (pit_type pit = 0; pit != endpit; ++pit) { + TeXOnePar(buffer, buffer.text(), pit, os, runparams); + LYXERR(Debug::FIND, "searchString up to here: " << ods.str()); + } + return ods.str(); +} + + +static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions const & opt) +{ + docstring str; + if (!opt.ignoreformat) { + str = buffer_to_latex(buffer); + } else { + OutputParams runparams(&buffer.params().encoding()); + runparams.nice = true; + runparams.flavor = OutputParams::LATEX; + runparams.linelen = 100000; //lyxrc.plaintext_linelen; + runparams.dryrun = true; + for (pos_type pit = pos_type(0); pit < (pos_type)buffer.paragraphs().size(); ++pit) { + Paragraph const & par = buffer.paragraphs().at(pit); + LYXERR(Debug::FIND, "Adding to search string: '" + << par.stringify(pos_type(0), par.size(), + AS_STR_INSETS, runparams) + << "'"); + str += par.stringify(pos_type(0), par.size(), + AS_STR_INSETS, runparams); + } + } + return str; +} + + +/// Return separation pos between the leading material and the rest +static size_t identifyLeading(string const & s) +{ + string t = s; + // @TODO Support \item[text] + while (regex_replace(t, t, "^\\\\(emph|textbf|subsubsection|subsection|section|subparagraph|paragraph|part)\\*?\\{", "") + || regex_replace(t, t, "^\\$", "") + || regex_replace(t, t, "^\\\\\\[ ", "") + || regex_replace(t, t, "^\\\\item ", "") + || regex_replace(t, t, "^\\\\begin\\{[a-zA-Z_]*\\*?\\} ", "")) + LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: '" << t << "'"); + return s.find(t); +} + + +// Remove trailing closure of math, macros and environments, so to catch parts of them. +static int identifyClosing(string & t) +{ + int open_braces = 0; + do { + LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'"); + if (regex_replace(t, t, "(.*[^\\\\])\\$\\'", "$1")) + continue; + if (regex_replace(t, t, "(.*[^\\\\]) \\\\\\]\\'", "$1")) + continue; + if (regex_replace(t, t, "(.*[^\\\\]) \\\\end\\{[a-zA-Z_]*\\*?\\}\\'", "$1")) + continue; + if (regex_replace(t, t, "(.*[^\\\\])\\}\\'", "$1")) { + ++open_braces; + continue; + } + break; + } while (true); + return open_braces; +} + + MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt) : p_buf(&buf), p_first_buf(&buf), opt(opt) { - par_as_string = normalize(opt.search); + Buffer & find_buf = *theBufferList().getBuffer(FileName(to_utf8(opt.find_buf_name)), true); + docstring const & ds = stringifySearchBuffer(find_buf, opt); + use_regexp = lyx::to_utf8(ds).find("\\regexp{") != std::string::npos; + // When using regexp, braces are hacked already by escape_for_regex() + par_as_string = normalize(ds, !use_regexp); open_braces = 0; close_wildcards = 0; - if (! opt.regexp) { - // Remove trailing closure of math, macros and environments, so to catch parts of them. - do { - LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'"); - if (regex_replace(par_as_string, par_as_string, "(.*)[[:blank:]]\\'", "$1")) - continue; - if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\]) ?\\$\\'", "$1")) - continue; - // @todo need to account for open square braces as well ? - if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\]) ?\\\\\\]\\'", "$1")) - continue; - if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\]) ?\\\\end\\{[a-zA-Z_]*\\}\\'", "$1")) - continue; - if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\]) ?\\}\\'", "$1")) { - ++open_braces; - continue; - } - break; - } while (true); + size_t lead_size = 0; + if (!opt.ignoreformat) { + lead_size = identifyLeading(par_as_string); + lead_as_string = par_as_string.substr(0, lead_size); + par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size); + } + + if (!use_regexp) { + open_braces = identifyClosing(par_as_string); + identifyClosing(par_as_string_nolead); LYXERR(Debug::FIND, "Open braces: " << open_braces); LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'"); } else { - par_as_string = escape_for_regex(par_as_string); + string lead_as_regexp; + if (lead_size > 0) { + // @todo No need to search for \regexp{} insets in leading material + lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat); + par_as_string = par_as_string_nolead; + LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'"); + LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'"); + } + par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat); // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them. LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'"); if ( // Insert .* before trailing '\$' ('$' 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") - // 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 now is '" << par_as_string << "'"); @@ -762,9 +880,14 @@ MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & 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 - regexp = lyx::regex(string("\\`") + par_as_string); + string regexp_str = string("\\`") + lead_as_regexp + par_as_string; + LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'"); + regexp = lyx::regex(regexp_str); + // If entered regexp may match wherever in searched string buffer - regexp2 = lyx::regex(string("\\`.*") + par_as_string); + string regexp2_str = string("\\`.*") + lead_as_regexp + ".*" + par_as_string; + LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'"); + regexp2 = lyx::regex(regexp2_str); } } @@ -773,26 +896,22 @@ int MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) con { docstring docstr = stringifyFromForSearch(opt, cur, len); LYXERR(Debug::FIND, "Matching against '" << lyx::to_utf8(docstr) << "'"); - string str = normalize(docstr); + string str = normalize(docstr, true); LYXERR(Debug::FIND, "After normalization: '" << str << "'"); - if (! opt.regexp) { + if (! use_regexp) { LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='" << par_as_string << "', str='" << str << "'"); + LYXERR(Debug::FIND, "Searching in normal mode: lead_as_string='" << lead_as_string << "', par_as_string_nolead='" << par_as_string_nolead << "'"); 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(); } else { - string t = par_as_string; - while (regex_replace(t, t, "\\\\(emph|textbf|subsubsection|subsection|section|subparagraph|paragraph|part)\\{", "") - || regex_replace(t, t, "^\\$", "") - || regex_replace(t, t, "^\\\\\\[ ", "")) - LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: " << t); - size_t pos = str.find(t); + size_t pos = str.find(par_as_string_nolead); if (pos != string::npos) return par_as_string.size(); } } else { - LYXERR(Debug::FIND, "Searching in regexp mode"); + LYXERR(Debug::FIND, "Searching in regexp mode: at_begin=" << at_begin); // Try all possible regexp matches, //until one that verifies the braces match test is found regex const *p_regexp = at_begin ? ®exp : ®exp2; @@ -804,8 +923,9 @@ int MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) con // plus the last subexpression, if a (.*?) was inserted in the constructor. if (!braces_match(m[0].first, m[0].second, open_braces)) return 0; - // Check braces on segments that matched all (.*?) subexpressions. - for (size_t i = 1; i < m.size(); ++i) + // 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)) return false; // Exclude from the returned match length any length @@ -823,13 +943,19 @@ int MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) con int MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) const { int res = findAux(cur, len, at_begin); + 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; Paragraph const & par = cur.paragraph(); - bool ws_left = cur.pos() > 0 ? - par.isWordSeparator(cur.pos() - 1) : true; - bool ws_right = cur.pos() + res < par.size() ? - par.isWordSeparator(cur.pos() + res) : true; + bool ws_left = (cur.pos() > 0) + ? par.isWordSeparator(cur.pos() - 1) + : true; + bool ws_right = (cur.pos() + res < par.size()) + ? par.isWordSeparator(cur.pos() + res) + : true; LYXERR(Debug::FIND, "cur.pos()=" << cur.pos() << ", res=" << res << ", separ: " << ws_left << ", " << ws_right @@ -840,7 +966,7 @@ int MatchStringAdv::operator()(DocIterator const & cur, int len, bool at_begin) } -string MatchStringAdv::normalize(docstring const & s) const +string MatchStringAdv::normalize(docstring const & s, bool hack_braces) const { string t; if (! opt.casesensitive) @@ -848,10 +974,10 @@ string MatchStringAdv::normalize(docstring const & s) const else t = lyx::to_utf8(s); // Remove \n at begin - while (t.size() > 0 && t[0] == '\n') + while (!t.empty() && t[0] == '\n') t = t.substr(1); // Remove \n at end - while (t.size() > 0 && t[t.size() - 1] == '\n') + while (!t.empty() && t[t.size() - 1] == '\n') t = t.substr(0, t.size() - 1); size_t pos; // Replace all other \n with spaces @@ -861,6 +987,19 @@ string MatchStringAdv::normalize(docstring const & s) const LYXERR(Debug::FIND, "Removing stale empty \\emph{}, \\textbf{}, \\*section{} macros from: " << t); while (regex_replace(t, t, "\\\\(emph|textbf|subsubsection|subsection|section|subparagraph|paragraph|part)(\\{\\})+", "")) LYXERR(Debug::FIND, " further removing stale empty \\emph{}, \\textbf{} macros from: " << t); + + // 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 + while (regex_replace(t, t, "\\\\\\{", "_x_<") + || regex_replace(t, t, "\\\\\\}", "_x_>")) + LYXERR(Debug::FIND, "After {} replacement: '" << t << "'"); + } + return t; } @@ -869,31 +1008,34 @@ docstring stringifyFromCursor(DocIterator const & cur, int len) { LYXERR(Debug::FIND, "Stringifying with len=" << len << " from cursor at pos: " << cur); if (cur.inTexted()) { - Paragraph const & par = cur.paragraph(); - // TODO what about searching beyond/across paragraph breaks ? - // TODO Try adding a AS_STR_INSERTS as last arg - pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ? - int(par.size()) : cur.pos() + len; - OutputParams runparams(&cur.buffer()->params().encoding()); - odocstringstream os; - runparams.nice = true; - runparams.flavor = OutputParams::LATEX; - runparams.linelen = 100000; //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.stringify(cur.pos(), end, AS_STR_INSETS, runparams); + Paragraph const & par = cur.paragraph(); + // TODO what about searching beyond/across paragraph breaks ? + // TODO Try adding a AS_STR_INSERTS as last arg + pos_type end = ( len == -1 || cur.pos() + len > int(par.size()) ) ? + int(par.size()) : cur.pos() + len; + OutputParams runparams(&cur.buffer()->params().encoding()); + odocstringstream os; + runparams.nice = true; + runparams.flavor = OutputParams::LATEX; + runparams.linelen = 100000; //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.stringify(cur.pos(), end, AS_STR_INSETS, runparams); } else if (cur.inMathed()) { - odocstringstream os; - CursorSlice cs = cur.top(); - MathData md = cs.cell(); - MathData::const_iterator it_end = - ( ( len == -1 || cs.pos() + len > int(md.size()) ) - ? md.end() : md.begin() + cs.pos() + len ); - for (MathData::const_iterator it = md.begin() + cs.pos(); it != it_end; ++it) - os << *it; - return os.str(); + docstring s; + CursorSlice cs = cur.top(); + MathData md = cs.cell(); + MathData::const_iterator it_end = + (( len == -1 || cs.pos() + len > int(md.size())) + ? md.end() + : md.begin() + cs.pos() + len ); + for (MathData::const_iterator it = md.begin() + cs.pos(); + it != it_end; ++it) + s = s + asString(*it); + LYXERR(Debug::FIND, "Stringified math: '" << s << "'"); + return s; } LYXERR(Debug::FIND, "Don't know how to stringify from here: " << cur); return docstring(); @@ -908,12 +1050,13 @@ docstring latexifyFromCursor(DocIterator const & cur, int len) { LYXERR(Debug::FIND, "Latexifying with len=" << len << " from cursor at pos: " << cur); LYXERR(Debug::FIND, " with cur.lastpost=" << cur.lastpos() << ", cur.lastrow=" - << cur.lastrow() << ", cur.lastcol=" << cur.lastcol()); + << cur.lastrow() << ", cur.lastcol=" << cur.lastcol()); Buffer const & buf = *cur.buffer(); - LASSERT(buf.isLatex(), /* */); + LASSERT(buf.params().isLatex(), /* */); TexRow texrow; odocstringstream ods; + otexstream os(ods, texrow); OutputParams runparams(&buf.params().encoding()); runparams.nice = false; runparams.flavor = OutputParams::LATEX; @@ -926,26 +1069,29 @@ docstring latexifyFromCursor(DocIterator const & cur, int len) pos_type endpos = cur.paragraph().size(); if (len != -1 && endpos > cur.pos() + len) endpos = cur.pos() + len; - TeXOnePar(buf, *cur.innerText(), cur.pit(), ods, texrow, runparams, - string(), cur.pos(), endpos); + TeXOnePar(buf, *cur.innerText(), cur.pit(), os, runparams, + string(), cur.pos(), endpos); LYXERR(Debug::FIND, "Latexified text: '" << lyx::to_utf8(ods.str()) << "'"); } else if (cur.inMathed()) { // Retrieve the math environment type, and add '$' or '$[' or others (\begin{equation}) accordingly for (int s = cur.depth() - 1; s >= 0; --s) { - CursorSlice const & cs = cur[s]; - if (cs.asInsetMath() && cs.asInsetMath() && cs.asInsetMath()->asHullInset()) { - WriteStream ws(ods); - cs.asInsetMath()->asHullInset()->header_write(ws); - break; - } + CursorSlice const & cs = cur[s]; + if (cs.asInsetMath() && cs.asInsetMath() && cs.asInsetMath()->asHullInset()) { + WriteStream ws(ods); + cs.asInsetMath()->asHullInset()->header_write(ws); + break; + } } CursorSlice const & cs = cur.top(); MathData md = cs.cell(); - MathData::const_iterator it_end = ( ( len == -1 || cs.pos() + len > int(md.size()) ) - ? md.end() : md.begin() + cs.pos() + len ); - for (MathData::const_iterator it = md.begin() + cs.pos(); it != it_end; ++it) - ods << *it; + MathData::const_iterator it_end = + ((len == -1 || cs.pos() + len > int(md.size())) + ? md.end() + : md.begin() + cs.pos() + len); + for (MathData::const_iterator it = md.begin() + cs.pos(); + it != it_end; ++it) + ods << asString(*it); // Retrieve the math environment type, and add '$' or '$]' // or others (\end{equation}) accordingly @@ -1013,17 +1159,36 @@ int findForwardAdv(DocIterator & cur, MatchStringAdv & match) { if (!cur) return 0; - while (cur && !match(cur, -1, false)) { - if (cur.pit() < cur.lastpit()) + while (!theApp()->longOperationCancelled() && cur) { + LYXERR(Debug::FIND, "findForwardAdv() cur: " << cur); + int match_len = match(cur, -1, false); + LYXERR(Debug::FIND, "match_len: " << match_len); + if (match_len) { + for (; !theApp()->longOperationCancelled() && cur; cur.forwardPos()) { + LYXERR(Debug::FIND, "Advancing cur: " << cur); + int match_len = match(cur); + LYXERR(Debug::FIND, "match_len: " << match_len); + if (match_len) { + // Sometimes in finalize we understand it wasn't a match + // and we need to continue the outest loop + int len = findAdvFinalize(cur, match); + if (len > 0) + return len; + } + } + if (!cur) + return 0; + } + if (cur.pit() < cur.lastpit()) { + LYXERR(Debug::FIND, "Advancing par: cur=" << cur); cur.forwardPar(); - else { + } 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(); } } - for (; cur; cur.forwardPos()) { - if (match(cur)) - return findAdvFinalize(cur, match); - } return 0; } @@ -1052,7 +1217,8 @@ int findMostBackwards(DocIterator & cur, MatchStringAdv const & match) /// Finds backwards -int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match) { +int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match) +{ if (! cur) return 0; // Backup of original position @@ -1096,7 +1262,7 @@ int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match) { else cur.backwardPos(); pit_changed = true; - } while (true); + } while (!theApp()->longOperationCancelled()); return 0; } @@ -1105,8 +1271,9 @@ int findBackwardsAdv(DocIterator & cur, MatchStringAdv & match) { docstring stringifyFromForSearch(FindAndReplaceOptions const & opt, - DocIterator const & cur, int len) + DocIterator const & cur, int len) { + LASSERT(cur.pos() >= 0 && cur.pos() <= cur.lastpos(), /* */); if (!opt.ignoreformat) return latexifyFromCursor(cur, len); else @@ -1114,13 +1281,14 @@ docstring stringifyFromForSearch(FindAndReplaceOptions const & opt, } -FindAndReplaceOptions::FindAndReplaceOptions(docstring const & search, bool casesensitive, +FindAndReplaceOptions::FindAndReplaceOptions( + docstring const & find_buf_name, bool casesensitive, bool matchword, bool forward, bool expandmacros, bool ignoreformat, - bool regexp, docstring const & replace, bool keep_case, + docstring const & repl_buf_name, bool keep_case, SearchScope scope) - : search(search), casesensitive(casesensitive), matchword(matchword), - forward(forward), expandmacros(expandmacros), ignoreformat(ignoreformat), - regexp(regexp), replace(replace), keep_case(keep_case), scope(scope) + : find_buf_name(find_buf_name), casesensitive(casesensitive), matchword(matchword), + forward(forward), expandmacros(expandmacros), ignoreformat(ignoreformat), + repl_buf_name(repl_buf_name), keep_case(keep_case), scope(scope) { } @@ -1129,7 +1297,8 @@ namespace { /** Check if 'len' letters following cursor are all non-lowercase */ -static bool allNonLowercase(DocIterator const & cur, int len) { +static bool allNonLowercase(DocIterator const & cur, int len) +{ pos_type end_pos = cur.pos() + len; for (pos_type pos = cur.pos(); pos != end_pos; ++pos) if (isLowerCase(cur.paragraph().getChar(pos))) @@ -1139,7 +1308,8 @@ static bool allNonLowercase(DocIterator const & cur, int len) { /** Check if first letter is upper case and second one is lower case */ -static bool firstUppercase(DocIterator const & cur) { +static bool firstUppercase(DocIterator const & cur) +{ char_type ch1, ch2; if (cur.pos() >= cur.lastpos() - 1) { LYXERR(Debug::FIND, "No upper-case at cur: " << cur); @@ -1160,21 +1330,24 @@ static bool firstUppercase(DocIterator const & cur) { ** ** \fixme What to do with possible further paragraphs in replace buffer ? **/ -static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case) { +static void changeFirstCase(Buffer & buffer, TextCase first_case, TextCase others_case) +{ ParagraphList::iterator pit = buffer.paragraphs().begin(); pos_type right = pos_type(1); pit->changeCase(buffer.params(), pos_type(0), right, first_case); right = pit->size() + 1; pit->changeCase(buffer.params(), right, right, others_case); } + } // anon namespace /// static void findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, MatchStringAdv & matchAdv) { Cursor & cur = bv->cursor(); - if (opt.replace == docstring(from_utf8(LYX_FR_NULL_STRING))) + if (opt.repl_buf_name == docstring()) return; + DocIterator sel_beg = cur.selectionBegin(); DocIterator sel_end = cur.selectionEnd(); if (&sel_beg.inset() != &sel_end.inset() @@ -1190,9 +1363,11 @@ static void findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, M if (!matchAdv(sel_beg, sel_len)) return; - string lyx = to_utf8(opt.replace); - // FIXME: Seems so stupid to me to rebuild a buffer here, - // when we already have one (replace_work_area_.buffer()) + // Build a copy of the replace buffer, adapted to the KeepCase option + Buffer & repl_buffer_orig = *theBufferList().getBuffer(FileName(to_utf8(opt.repl_buf_name)), true); + ostringstream oss; + repl_buffer_orig.write(oss); + string lyx = oss.str(); Buffer repl_buffer("", false); repl_buffer.setUnnamed(true); LASSERT(repl_buffer.readString(lyx), /**/); @@ -1205,7 +1380,7 @@ static void findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, M } } cap::cutSelection(cur, false, false); - if (!cur.inMathed()) { + if (cur.inTexted()) { repl_buffer.changeLanguage( repl_buffer.language(), cur.getFont().language()); @@ -1215,15 +1390,17 @@ static void findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, M repl_buffer.params().documentClassPtr(), bv->buffer().errorList("Paste")); LYXERR(Debug::FIND, "After pasteParagraphList() cur=" << cur << endl); - } else { + sel_len = repl_buffer.paragraphs().begin()->size(); + } else if (cur.inMathed()) { + TexRow texrow; odocstringstream ods; + otexstream os(ods, texrow); OutputParams runparams(&repl_buffer.params().encoding()); runparams.nice = false; runparams.flavor = OutputParams::LATEX; runparams.linelen = 8000; //lyxrc.plaintext_linelen; runparams.dryrun = true; - TexRow texrow; - TeXOnePar(repl_buffer, repl_buffer.text(), 0, ods, texrow, runparams); + TeXOnePar(repl_buffer, repl_buffer.text(), 0, os, runparams); //repl_buffer.getSourceCode(ods, 0, repl_buffer.paragraphs().size(), false); docstring repl_latex = ods.str(); LYXERR(Debug::FIND, "Latexified replace_buffer: '" << repl_latex << "'"); @@ -1231,13 +1408,21 @@ static void findAdvReplace(BufferView * bv, FindAndReplaceOptions const & opt, M regex_replace(to_utf8(repl_latex), s, "\\$(.*)\\$", "$1"); regex_replace(s, s, "\\\\\\[(.*)\\\\\\]", "$1"); repl_latex = from_utf8(s); - LYXERR(Debug::FIND, "Replacing by niceInsert()ing latex: '" << repl_latex << "'"); - cur.niceInsert(repl_latex); + LYXERR(Debug::FIND, "Replacing by insert()ing latex: '" << repl_latex << "' cur=" << cur << " with depth=" << cur.depth()); + MathData ar(cur.buffer()); + asArray(repl_latex, ar, Parse::NORMAL); + cur.insert(ar); + sel_len = ar.size(); + LYXERR(Debug::FIND, "After insert() cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len); } - cur.pos() -= repl_buffer.paragraphs().begin()->size(); - LYXERR(Debug::FIND, "Putting selection at cur=" << cur << " with len: " << repl_buffer.paragraphs().begin()->size()); - bv->putSelectionAt(DocIterator(cur), repl_buffer.paragraphs().begin()->size(), !opt.forward); + if (cur.pos() >= sel_len) + cur.pos() -= sel_len; + else + cur.pos() = 0; + LYXERR(Debug::FIND, "After pos adj cur=" << cur << " with depth: " << cur.depth() << " and len: " << sel_len); + bv->putSelectionAt(DocIterator(cur), sel_len, !opt.forward); bv->processUpdateFlags(Update::Force); + bv->buffer().updatePreviews(); } @@ -1247,19 +1432,14 @@ bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt) DocIterator cur; int match_len = 0; - if (opt.search.empty()) { - bv->message(_("Search text is empty!")); - return false; - } - try { MatchStringAdv matchAdv(bv->buffer(), opt); findAdvReplace(bv, opt, matchAdv); cur = bv->cursor(); if (opt.forward) - match_len = findForwardAdv(cur, matchAdv); + match_len = findForwardAdv(cur, matchAdv); else - match_len = findBackwardsAdv(cur, matchAdv); + match_len = findBackwardsAdv(cur, matchAdv); } catch (...) { // This may only be raised by lyx::regex() bv->message(_("Invalid regular expression!")); @@ -1282,14 +1462,13 @@ bool findAdv(BufferView * bv, FindAndReplaceOptions const & opt) ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt) { - os << to_utf8(opt.search) << "\nEOSS\n" + os << to_utf8(opt.find_buf_name) << "\nEOSS\n" << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' ' << opt.expandmacros << ' ' << opt.ignoreformat << ' ' - << opt.regexp << ' ' - << to_utf8(opt.replace) << "\nEOSS\n" + << to_utf8(opt.repl_buf_name) << "\nEOSS\n" << opt.keep_case << ' ' << int(opt.scope); @@ -1298,6 +1477,7 @@ ostringstream & operator<<(ostringstream & os, FindAndReplaceOptions const & opt return os; } + istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt) { LYXERR(Debug::FIND, "parsing"); @@ -1306,34 +1486,34 @@ istringstream & operator>>(istringstream & is, FindAndReplaceOptions & opt) getline(is, line); while (line != "EOSS") { if (! s.empty()) - s = s + "\n"; + s = s + "\n"; s = s + line; if (is.eof()) // Tolerate malformed request - break; + break; getline(is, line); } - LYXERR(Debug::FIND, "searching for: '" << s << "'"); - opt.search = from_utf8(s); - is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat >> opt.regexp; + LYXERR(Debug::FIND, "file_buf_name: '" << s << "'"); + opt.find_buf_name = from_utf8(s); + is >> opt.casesensitive >> opt.matchword >> opt.forward >> opt.expandmacros >> opt.ignoreformat; is.get(); // Waste space before replace string s = ""; getline(is, line); while (line != "EOSS") { if (! s.empty()) - s = s + "\n"; + s = s + "\n"; s = s + line; if (is.eof()) // Tolerate malformed request - break; + break; getline(is, line); } + LYXERR(Debug::FIND, "repl_buf_name: '" << s << "'"); + opt.repl_buf_name = from_utf8(s); is >> opt.keep_case; int i; is >> i; opt.scope = FindAndReplaceOptions::SearchScope(i); LYXERR(Debug::FIND, "parsed: " << opt.casesensitive << ' ' << opt.matchword << ' ' << opt.forward << ' ' - << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.regexp << ' ' << opt.keep_case); - LYXERR(Debug::FIND, "replacing with: '" << s << "'"); - opt.replace = from_utf8(s); + << opt.expandmacros << ' ' << opt.ignoreformat << ' ' << opt.keep_case); return is; }