//
// C++ Implementation: RegexPreProcessor
//
// Description: performs operations or inspections on a string representing
// a valid regular expression
//
// Author: Lorenzo Bettini <http://www.lorenzobettini.it>, (C) 1999-2007
//
// Copyright: See COPYING file that comes with this distribution
//
//
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "regexpreprocessor.h"
#include <boost/regex.hpp>
#include <sstream>
#include <ctype.h>
#include <iostream>
using namespace std;
namespace srchilite {
// IMPORTANT: the following regular expressions assume that the
// regular expression they try to match is a valid regular expression
/// matches character sets in a regular expression
const boost::regex char_set_exp("\\[([^\\&#124;]*)\\]");
/// substitute a "(" with "(?:" if it's not followed by a ? and not preceeded by \\ char
const boost::regex from("(\\\\\\()&#124;(\\((?!\\?))");
/// substitute a "(" with "(?:" if it's not followed by a ? and not preceeded by \\ char
const string into = "(?1\\\\\\()(?2\\(\\?\\:)";
/// found actual marking parenthesis, i.e., not preceeded by \\ and not followed by ?
const boost::regex paren("(?<!\\\\)\\((?!\\?)"); /// regular expression matching a backreference, e.g., 1円 or inside a conditional (?(1)...) const boost::regex backreference("(\\\\([[:digit:]]))&#124;(\\(\\?\\(([[:digit:]]))"); /// index of the subexpression matching a slashed backreference #define SLASHED_BACKREFERENCE 1 /// index of the subexpression matching the number in a slashed backreference #define SLASHED_BACKREFERENCE_NUM 2 /// index of the subexpression matching a conditional backreference #define CONDITIONAL_BACKREFERENCE 3 /// index of the subexpression matching the number in a conditional backreference #define CONDITIONAL_BACKREFERENCE_NUM 4 /** regular expression matching a reference, e.g., @{digit} the @$ must not be preceeded by an escape char */ const boost::regex reference_exp("(?<!\\\\)@\\{([[:digit:]])\\}"); /// substitute a special char (i.e., .[{()\*+?&#124;^$)... const boost::regex special_char("(\\.&#124;\\[&#124;\\]&#124;\\{&#124;\\}&#124;\\(&#124;\\)&#124;\\\\&#124;\\*&#124;\\+&#124;\\?&#124;\\&#124;&#124;\\^&#124;\\$)"); /// ...with its escaped version const string special_char_escape = "(?1\\\\\1円)"; /// index of the subexpression matching the number backreference #define BACKREFERENCE_NUM 1 const string subexpressions_info::ERR_OUTER_UNMARKED = "unmarked subexpressions are allowed only inside marked subexpressions"; const string subexpressions_info::ERR_NESTED_SUBEXP = "subexpressions of subexpressions are not allowed"; const string subexpressions_info::ERR_UNBALANCED_PAREN = "unbalanced parenthesis"; const string subexpressions_info::ERR_OUTSIDE_SUBEXP = "parts outside marked subexpressions are not allowed"; RegexPreProcessor::RegexPreProcessor() { } RegexPreProcessor::~RegexPreProcessor() { } const string RegexPreProcessor::preprocess(const string &s) { // substitute a "(" with "(?:" if it's not followed by a ? and not preceeded by \\ char return boost::regex_replace(s, from, into, boost::match_default &#124; boost::format_all); } const string _make_nonsensitive(const string &s) { ostringstream result; for (string::const_iterator it = s.begin(); it != s.end(); ++it) if (isalpha(*it)) result << "["<< (char)toupper(*it) << (char)tolower(*it) << "]"; else result << *it; return result.str(); } const string RegexPreProcessor::make_nonsensitive(const string &s) { boost::sregex_iterator m1(s.begin(), s.end(), char_set_exp); boost::sregex_iterator m2; if (m1 == m2) return _make_nonsensitive(s); ostringstream buffer; string prefix; string suffix; for (boost::sregex_iterator it = m1; it != m2; ++it) { prefix = it->prefix();
 suffix = it->suffix();
 if (prefix.size()) {
 buffer << _make_nonsensitive(prefix); } buffer << (*it)[0]; } if (suffix.size()) { buffer << _make_nonsensitive(suffix); } return buffer.str(); } unsigned int RegexPreProcessor::num_of_subexpressions(const string &s) { boost::sregex_iterator m1(s.begin(), s.end(), paren); boost::sregex_iterator m2; int counter = 0; for (boost::sregex_iterator it = m1; it != m2; ++it) { ++counter; } return counter; } const subexpressions_strings *RegexPreProcessor::split_marked_subexpressions( const string &s) { boost::sregex_iterator m1(s.begin(), s.end(), paren); boost::sregex_iterator m2; // we don't need to parse it (we can use the regex) since we assume that // the regular expression represented by s is made up of only // marked subexpressions and no nested subexpressions and char outside subexpressions subexpressions_strings *split = new subexpressions_strings; for (boost::sregex_iterator it = m1; it != m2;) { string prefix = it->prefix();
 if (prefix.size())
 split->push_back("(" + prefix);
 string suffix = it->suffix();
 if (++it == m2)
 split->push_back("(" + suffix);
 }
 return split;
}
subexpressions_info RegexPreProcessor::num_of_marked_subexpressions(
 const string &s, bool allow_outer_char, bool allow_outer_nonmarked) {
 subexpressions_info sexps;
 // number of open parenthesis
 int open_paren_num = 0;
 // whether we're inside a marked subexpressions
 bool found_marked_subexp = false;
 // whether we're inside a char sequence [...]
 bool char_sequence_state = false;
 // whether we're inside a char class name, e.g., [:alpha:]
 bool char_class_name_state = false;
 // len of string
 int len = s.size();
 // char we're examining
 char c;
 for (int i = 0; i < len; ++i) { c = s[i]; if (!char_sequence_state && c == '[' && (i == 0 &#124;&#124; s[i-1] != '\\') ) { char_sequence_state = true; } else if (char_sequence_state) { // here i cannot be 0 if (char_class_name_state) { if (c == ']' && s[i-1] == ':') { // exit the char class name state char_class_name_state = false; } } else if (!char_class_name_state && c == '[' && s[i-1] != '\\' && (i+1) < len && s[i+1] == ':') { char_class_name_state = true; } else if (c == ']' && s[i-1] != '\\') { // exit the char sequence state char_sequence_state = false; } // else simply go on skipping chars } else if (c == '\\' && (i+1) < len && (s[i+1] == '('&#124;&#124; s[i+1] == ')')) { // skip the escaped paren ++i; } else if (c == '(') { // we found a subexp ++open_paren_num; if ((i+1) < len && s[i+1] == '?') { if (!found_marked_subexp && !allow_outer_nonmarked) { // outer subexpressions must be marked sexps.errors = subexpressions_info::ERR_OUTER_UNMARKED; return sexps; } } else { // it's a marked subexp if (found_marked_subexp) { // we don't allow nested subexpressions sexps.errors = subexpressions_info::ERR_NESTED_SUBEXP; return sexps; } // we found a marked subexp found_marked_subexp = true; ++(sexps.marked); } } else if (c == ')') { if (!open_paren_num) { // unbalanced parenthesis sexps.errors = subexpressions_info::ERR_UNBALANCED_PAREN; return sexps; } --open_paren_num; // end of marked subexp if (!open_paren_num && found_marked_subexp) found_marked_subexp = false; } else { // we don't allow non marked parts if (!found_marked_subexp && !allow_outer_char) { sexps.errors = subexpressions_info::ERR_OUTSIDE_SUBEXP; return sexps; } } } // check that all paren are closed if (open_paren_num) sexps.errors = subexpressions_info::ERR_UNBALANCED_PAREN; return sexps; } bool RegexPreProcessor::contains_backreferences(const std::string &s) { return boost::regex_search(s, backreference); } backreference_info RegexPreProcessor::num_of_backreferences(const string &s) { boost::sregex_iterator m1(s.begin(), s.end(), backreference); boost::sregex_iterator m2; backreference_info info(std::make_pair(0, 0)); for (boost::sregex_iterator it = m1; it != m2; ++it) { ++(info.first); // for converting a string into the number stringstream buffer; int backreference_num; if ((*it)[SLASHED_BACKREFERENCE].matched) { buffer << (*it)[SLASHED_BACKREFERENCE_NUM]; } else { buffer << (*it)[CONDITIONAL_BACKREFERENCE_NUM]; } // retrieve the number of the backreference buffer>> backreference_num;
 if (backreference_num> info.second) {
 // we update the number of the highest backreference
 info.second = backreference_num;
 }
 }
 return info;
}
backreference_info RegexPreProcessor::num_of_references(const string &s) {
 boost::sregex_iterator m1(s.begin(), s.end(), reference_exp);
 boost::sregex_iterator m2;
 backreference_info info(std::make_pair(0, 0));
 for (boost::sregex_iterator it = m1; it != m2; ++it) {
 ++(info.first);
 // for converting a string into the number
 stringstream buffer;
 int backreference_num;
 // the digit of the reference
 buffer << (*it)[BACKREFERENCE_NUM]; // retrieve the number of the backreference buffer>> backreference_num;
 if (backreference_num> info.second) {
 // we update the number of the highest backreference
 info.second = backreference_num;
 }
 }
 return info;
}
const std::string RegexPreProcessor::replace_backreferences(
 const std::string &original, const backreference_replacements &replace) {
 boost::sregex_iterator m1(original.begin(), original.end(), backreference);
 boost::sregex_iterator m2;
 if (m1 == m2) {
 // no backreference in the current string
 return original;
 }
 ostringstream result;
 for (boost::sregex_iterator it = m1; it != m2;) {
 string prefix = it->prefix();
 if (prefix.size())
 result << prefix; // for converting a string into the number stringstream buffer; int backreference_num; // substitute the backreference number with the corresponding replace // string if ((*it)[SLASHED_BACKREFERENCE].matched) { buffer << (*it)[SLASHED_BACKREFERENCE_NUM]; // retrieve the number of the backreference buffer>> backreference_num;
 result << replace[backreference_num-1]; } else { buffer << (*it)[CONDITIONAL_BACKREFERENCE_NUM]; // retrieve the number of the backreference buffer>> backreference_num;
 result << "(?("<< replace[backreference_num-1]; } string suffix = it->suffix();
 if (++it == m2)
 result << suffix; } return result.str(); } const std::string RegexPreProcessor::replace_backreferences( const std::string &original, const regex_match_results &results) { // build the replacement vector backreference_replacements replace(9); // fill it with matched results for (unsigned i = 1; i < results.size(); ++i) { // remember that 0 is the whole expressions replace[i-1] = results[i]; } return replace_backreferences(original, replace); } const string RegexPreProcessor::replace_references( const std::string &original, const backreference_replacements &replace) { boost::sregex_iterator m1(original.begin(), original.end(), reference_exp); boost::sregex_iterator m2; if (m1 == m2) { // no backreference in the current string return original; } ostringstream result; for (boost::sregex_iterator it = m1; it != m2;) { string prefix = it->prefix();
 if (prefix.size())
 result << prefix; // for converting a string into the number stringstream buffer; int backreference_num; buffer << (*it)[BACKREFERENCE_NUM]; // retrieve the number of the backreference buffer>> backreference_num;
 // substitute the backreference number with the corresponding replace
 // string, after escaping non-alphanumerical characters
 // this is necessary since this string will be used to create a regular
 // expression
 result << boost::regex_replace(replace[backreference_num-1], special_char, special_char_escape, boost::match_default &#124; boost::format_all); string suffix = it->suffix();
 if (++it == m2)
 result << suffix;
 }
 return result.str();
}
const string RegexPreProcessor::replace_references(
 const std::string &original, const regex_match_results &results) {
 // build the replacement vector
 backreference_replacements replace(9);
 // fill it with matched results
 for (unsigned i = 1; i < results.size(); ++i) {
 // remember that 0 is the whole expressions
 replace[i-1] = results[i];
 }
 return replace_references(original, replace);
}
}
</div><div class="naked_ctrl">
<form action="/index.cgi/contrast" method="get" name="gate">
<p><a href="http://altstyle.alfasado.net">AltStyle</a> によって変換されたページ <a href="https://cgit.git.savannah.gnu.org/cgit/src-highlite.git/plain/lib/srchilite/regexpreprocessor.cpp">(-&gt;オリジナル)</a>
/ <label>アドレス: <input type="text" name="naked_post_url" value="https://cgit.git.savannah.gnu.org/cgit/src-highlite.git/plain/lib/srchilite/regexpreprocessor.cpp" size="22" /></label> <label>モード: <select name="naked_post_mode">
<option value="default">デフォルト</option>
<option value="speech">音声ブラウザ</option>
<option value="ruby">ルビ付き</option>
<option value="contrast" selected="selected">配色反転</option>
<option value="larger-text">文字拡大</option>
<option value="mobile">モバイル</option>
</select>
<input type="submit" value="表示" />
</p>
</form>
</div>