StringV.cc

Go to the documentation of this file.

/*---------------------------------------------------------------------\
|                          ____ _   __ __ ___                          |
|                         |__  / \ / / . \ . \                         |
|                           / / \ V /|  _/  _/                         |
|                          / /__ | | | | | |                           |
|                         /_____||_| |_| |_|                           |
|                                                                      |
\---------------------------------------------------------------------*/
#include <iostream>
#include <zypp-core/base/StringV.h>

namespace zypp
{
  namespace strv
  {
    namespace detail
    {
      unsigned _splitSimple( std::string_view line_r, const WordConsumer& fnc_r )
      {
        // callback stats
        unsigned fncCall = 0;
 
        // NOTE: line_r is not null-terminated!
        const char *const eol = line_r.data() + line_r.size();  // the '\0'
        const char * searchfrom = line_r.data();        // start of the next search (moves with each cycle!)
 
        // Empty sep: split at /[BLANK,TAB]+/ and report no-empty words
        auto isSep = []( char ch )->bool { return ch == ' '|| ch == '\t'; };
 
        auto skipSep = [eol,&isSep]( const char *& ptr )->bool {
          while ( ptr < eol && isSep( *ptr ) )
            ++ptr;
          return ptr < eol;     // whether we ended up at a new wordstart
        };
 
        auto skipWord = [eol,&isSep]( const char *& ptr )->void {
          while ( ptr < eol && ! isSep( *ptr ) )
            ++ptr;
        };
 
        // For the 'last' CB argument: we must remember a word
        // until we know whether another one is following
        std::string_view word;
 
        while ( skipSep( searchfrom ) ) {
          const char * wordstart = searchfrom;
          // Report a previous word found
          if ( ! word.empty() ) {
            if ( fnc_r ) {
              if ( ! fnc_r( word, fncCall, false/*more to come*/ ) ) {
                // Stop searching for further matches. Final report will
                // be the remaining line (right trimmed)
                const char * wordend = eol;
                while ( isSep( *(wordend-1) ) ) // noSep at wordstart stops loop
                  --wordend;
                word = std::string_view( wordstart, wordend-wordstart );
                ++fncCall;
                break;
              }
            }
            ++fncCall;
          }
          // remember new word
          ++searchfrom;
          skipWord( searchfrom );
          word = std::string_view( wordstart, searchfrom-wordstart );
        }
 
        // finally report the last word
        if ( ! word.empty() ) {
          if ( fnc_r ) {
            fnc_r( word, fncCall, true/*last*/ );
          }
          ++fncCall;
        }
 
        return fncCall;
      }
    } // namespace detail
 
    unsigned detail::_split( std::string_view line_r, std::string_view sep_r, Trim trim_r, WordConsumer && fnc_r )
    {
      if ( sep_r.empty() )
        return _splitSimple( line_r, std::move( fnc_r ) );
 
      // callback stats
      bool fncStop = false;
      unsigned fncCall = 0;
 
      using size_type = std::string_view::size_type;
      size_type wordstart = 0;  // start of the next string to be reported
      size_type searchfrom = 0; // start of the next search for a separator
 
      do {      // report lhs word of separator matches...
        searchfrom = line_r.find( sep_r, searchfrom );
        if ( fncStop || searchfrom == line_r.npos ) {
          break;
        }
 
        // Report lhs word of the match and advance...
        if ( fnc_r ) {
          if ( ! fnc_r( trim( line_r.substr(wordstart,searchfrom-wordstart), trim_r ), fncCall, false/*more to come*/ ) )
            fncStop= true;      // stop searching for further matches
        }
        ++fncCall;
 
        // Next wordstart is behind the separator match.
        searchfrom += sep_r.size();
        wordstart = searchfrom;
      } while( wordstart < line_r.size() );
 
      // finally report rhs word of the last separator match (or no match)
      if ( fnc_r ) {
        if ( wordstart < line_r.size() )
          fnc_r( trim( line_r.substr(wordstart,line_r.size()-wordstart), trim_r ), fncCall, true/*last*/ );
        else    // a final match at $ so a final empty word reported (no trim needed)
          fnc_r( std::string_view( line_r.data()+line_r.size(), 0 ), fncCall, true/*last*/ );
      }
      ++fncCall;
      return fncCall;
    }
 
    unsigned detail::_splitRx( std::string_view line_r, const regex & rx_r, const WordConsumer& fnc_r )
    {
      // callback stats
      bool fncStop = false;
      unsigned fncCall = 0;
 
      // report lhs word of separator matches...
      const char *const eol = line_r.data() + line_r.size();    // the '\0'
      bool trailingNL = line_r.size() && *(eol-1) == '\n';      // line ends with NL
      const char * wordstart = line_r.data();   // start of the next string to be reported
      const char * searchfrom = line_r.data();  // start of the next search for a separator (moves with each cycle!)
 
      // Whether to match the ^ at beginning of the line or after an \n:
      auto matchAtBOL = [&]() {
        return searchfrom == line_r.data() || *(searchfrom-1) == '\n' ? regex::none : regex::not_bol;
      };
      do {      // report lhs word of separator matches...
        smatch match;
        if ( fncStop || ! rx_r.matches( searchfrom, match, matchAtBOL() ) ) {
          break;
        }
        if ( trailingNL && searchfrom+match.begin(0) == eol )
          break;        // don't report matches behind a trailing NL
 
        if ( match.end(0) == 0 && searchfrom == wordstart && searchfrom != line_r.data() ) {
          // An empty(!) separator found at wordstart is ignored unless we're at the very beginning.
          // If searchfrom == wordstart we just skipped over a not-empty separator. If wordstart is
          // not part of a 2nd not-empty separator, we want the next separator to its right.
          // Example: Rx:"b*" Str:"0b2" - at pos 2 Rx matches empty; the previous cycle found 'b' and reported the '0'.
          ++searchfrom;
        } else {
          // Report lhs word of the match and advance...
          if ( fnc_r ) {
            if ( ! fnc_r( std::string_view( wordstart, searchfrom+match.begin(0) - wordstart ), fncCall, false/*more to come*/ ) )
              fncStop= true;    // stop searching for further matches
          }
          ++fncCall;
          // Next wordstart is behind the separator match.
          // Next searchfrom also, but advances at least by 1.
          wordstart = searchfrom+match.end(0);
          searchfrom += match.end(0) ? match.end(0) : 1;
        }
      } while ( searchfrom <= eol );    // incl. '== eol' as there might be an (empty) match at $
 
      // finally report rhs word of the last separator match (or no match)
      if ( fnc_r ) {
        if ( wordstart < eol )
          fnc_r( std::string_view( wordstart, eol-wordstart ), fncCall, true/*last*/ );
        else    // a final match at $ so a final empty word reported
          fnc_r( std::string_view( eol, 0 ), fncCall, true/*last*/ );
      }
      ++fncCall;
      return fncCall;
    }
 
 
    std::string doBSEscape( std::string_view word_r, std::string_view special_r )
    {
      // 1. Calculate the required capacity to avoid reallocations
      size_t extra = 0;
      for ( char c : word_r ) {
        if ( c == '\\' || special_r.find(c) != std::string_view::npos ) {
          extra++;
        }
      }
 
      if (extra == 0) return std::string(word_r);
 
      // 2. Construct the string with the exact size needed
      std::string result;
      result.reserve( word_r.size() + extra );
 
      for ( char c : word_r ) {
        if ( c == '\\' || special_r.find(c) != std::string_view::npos ) {
          result.push_back( '\\' );
        }
        result.push_back( c );
      }
 
      return result;
    }
 
    std::string unBSEscape( std::string_view word_r )
    {
      std::string result;
      result.reserve( word_r.size() ); // Optimization: avoid multiple reallocations
 
      for ( size_t i = 0; i < word_r.size(); ++i ) {
        if ( word_r[i] == '\\' ) {
          // Move to the escaped character
          if ( ++i < word_r.size() ) {
            result.push_back( word_r[i] );
          }
        } else {
          result.push_back( word_r[i] );
        }
      }
      result.shrink_to_fit();
      return result;
    }
 
    std::vector<std::string> unBSEscape( const std::vector<std::string_view> & words_r )
    {
      std::vector<std::string> result;
      result.reserve( words_r.size() );
 
      for ( const auto& word : words_r ) {
        result.push_back( unBSEscape( word ) );
      }
      return result;
    }
 
    std::vector<std::string_view> splitBSEscaped( std::string_view line_r, std::string_view chars_r, unsigned maxSplits_r )
    {
      std::vector<std::string_view> fragments;
      unsigned splitsDone = 0;
      size_t start = 0;
 
      for ( size_t i = 0; i < line_r.size(); ++i ) {
        // Check for maxSplits limit
        if ( maxSplits_r > 0 && splitsDone >= maxSplits_r ) {
          break;
        }
 
        if ( line_r[i] == '\\' ) {
          if ( ++i < line_r.size() ) {
            ; // Skip the next character regardless of what it is
          }
          continue;
        }
 
        if ( chars_r.find( line_r[i] ) != std::string_view::npos ) {
          // Found an unescaped separator
          fragments.emplace_back( line_r.substr( start, i - start ) );
          start = i + 1;
          splitsDone++;
        }
      }
 
      fragments.emplace_back( line_r.substr( start ) );
 
      return fragments;
    }
 
  } // namespace strv
} // namespace zypp