Merging upstream version 0.10.

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

1 files changed, 29 insertions, 104 deletions

diff --git a/extract.cc b/extract.cc
index e25f5b6..f85cf67 100644
--- a/extract.cc
+++ b/extract.cc
@@ -37,7 +37,6 @@
 #include <lzlib.h>
 
 #include "arg_parser.h"
-#include "lzip.h"
 #include "lzip_index.h"
 #include "tarlz.h"
 
@@ -268,19 +267,19 @@ void format_member_name( const Extended & extended, const Tar_header header,
     for( int i = 0; i < 2; ++i )
       {
       const int len = snprintf( rbuf() + offset, rbuf.size() - offset,
-                  " %9llu %4d-%02u-%02u %02u:%02u %s%s%s\n",
-                  extended.size, 1900 + tm->tm_year, 1 + tm->tm_mon,
-                  tm->tm_mday, tm->tm_hour, tm->tm_min, extended.path.c_str(),
-                  link_string, !islink ? "" : extended.linkpath.c_str() );
+                " %9llu %4d-%02u-%02u %02u:%02u %s%s%s\n",
+                extended.file_size(), 1900 + tm->tm_year, 1 + tm->tm_mon,
+                tm->tm_mday, tm->tm_hour, tm->tm_min, extended.path().c_str(),
+                link_string, !islink ? "" : extended.linkpath().c_str() );
       if( (int)rbuf.size() > len + offset ) break;
       else rbuf.resize( len + offset + 1 );
       }
     }
   else
     {
-    if( rbuf.size() < extended.path.size() + 2 )
-      rbuf.resize( extended.path.size() + 2 );
-    snprintf( rbuf(), rbuf.size(), "%s\n", extended.path.c_str() );
+    if( rbuf.size() < extended.path().size() + 2 )
+      rbuf.resize( extended.path().size() + 2 );
+    snprintf( rbuf(), rbuf.size(), "%s\n", extended.path().c_str() );
     }
   }
 
@@ -303,8 +302,8 @@ int list_member( const int infd, const Extended & extended,
 
   const unsigned bufsize = 32 * header_size;
   uint8_t buf[bufsize];
-  unsigned long long rest = extended.size;
-  const int rem = extended.size % header_size;
+  unsigned long long rest = extended.file_size();
+  const int rem = rest % header_size;
   const int padding = rem ? header_size - rem : 0;
   while( rest > 0 )
     {
@@ -331,7 +330,7 @@ bool contains_dotdot( const char * const filename )
 int extract_member( const int infd, const Extended & extended,
                     const Tar_header header, const bool keep_damaged )
   {
-  const char * const filename = extended.path.c_str();
+  const char * const filename = extended.path().c_str();
   if( contains_dotdot( filename ) )
     {
     show_file_error( filename, "Contains a '..' component, skipping." );
@@ -357,7 +356,7 @@ int extract_member( const int infd, const Extended & extended,
     case tf_link:
     case tf_symlink:
       {
-      const char * const linkname = extended.linkpath.c_str();
+      const char * const linkname = extended.linkpath().c_str();
 /*      if( contains_dotdot( linkname ) )
         {
         show_file_error( filename,
@@ -421,8 +420,8 @@ int extract_member( const int infd, const Extended & extended,
 
   const unsigned bufsize = 32 * header_size;
   uint8_t buf[bufsize];
-  unsigned long long rest = extended.size;
-  const int rem = extended.size % header_size;
+  unsigned long long rest = extended.file_size();
+  const int rem = rest % header_size;
   const int padding = rem ? header_size - rem : 0;
   while( rest > 0 )
     {
@@ -501,42 +500,6 @@ bool compare_tslash( const char * const name1, const char * const name2 )
 
 namespace {
 
-unsigned long long parse_decimal( const char * const ptr,
-                                  const char ** const tailp,
-                                  const unsigned long long size )
-  {
-  unsigned long long result = 0;
-  unsigned long long i = 0;
-  while( i < size && std::isspace( ptr[i] ) ) ++i;
-  if( !std::isdigit( (unsigned char)ptr[i] ) )
-    { if( tailp ) *tailp = ptr; return 0; }
-  for( ; i < size && std::isdigit( (unsigned char)ptr[i] ); ++i )
-    {
-    const unsigned long long prev = result;
-    result *= 10; result += ptr[i] - '0';
-    if( result < prev || result > LLONG_MAX )		// overflow
-      { if( tailp ) *tailp = ptr; return 0; }
-    }
-  if( tailp ) *tailp = ptr + i;
-  return result;
-  }
-
-
-uint32_t parse_record_crc( const char * const ptr )
-  {
-  uint32_t crc = 0;
-  for( int i = 0; i < 8; ++i )
-    {
-    crc <<= 4;
-    if( ptr[i] >= '0' && ptr[i] <= '9' ) crc += ptr[i] - '0';
-    else if( ptr[i] >= 'A' && ptr[i] <= 'F' ) crc += ptr[i] + 10 - 'A';
-    else if( ptr[i] >= 'a' && ptr[i] <= 'f' ) crc += ptr[i] + 10 - 'a';
-    else { crc = 0; break; }		// invalid digit in crc string
-    }
-  return crc;
-  }
-
-
 bool parse_records( const int infd, Extended & extended,
                     const Tar_header header, const bool permissive )
   {
@@ -602,48 +565,6 @@ unsigned long long parse_octal( const uint8_t * const ptr, const int size )
   }
 
 
-bool Extended::parse( const char * const buf, const unsigned long long edsize,
-                      const bool permissive )
-  {
-  for( unsigned long long pos = 0; pos < edsize; )	// parse records
-    {
-    const char * tail;
-    const unsigned long long rsize =
-      parse_decimal( buf + pos, &tail, edsize - pos );
-    if( rsize == 0 || rsize > edsize - pos || tail[0] != ' ' ||
-        buf[pos+rsize-1] != '\n' ) return false;
-    ++tail;	// point to keyword
-    // rest = length of (keyword + '=' + value) without the final newline
-    const unsigned long long rest = ( buf + ( pos + rsize - 1 ) ) - tail;
-    if( rest > 5 && std::memcmp( tail, "path=", 5 ) == 0 )
-      { if( path.size() && !permissive ) return false;
-        path.assign( tail + 5, rest - 5 ); }
-    else if( rest > 9 && std::memcmp( tail, "linkpath=", 9 ) == 0 )
-      { if( linkpath.size() && !permissive ) return false;
-        linkpath.assign( tail + 9, rest - 9 ); }
-    else if( rest > 5 && std::memcmp( tail, "size=", 5 ) == 0 )
-      {
-      if( size != 0 && !permissive ) return false;
-      size = parse_decimal( tail + 5, &tail, rest - 5 );
-      // parse error or size fits in ustar header
-      if( size < 1ULL << 33 || tail != buf + ( pos + rsize - 1 ) ) return false;
-      }
-    else if( rest > 10 && std::memcmp( tail, "GNU.crc32=", 10 ) == 0 )
-      {
-      if( crc_present && !permissive ) return false;
-      if( rsize != 22 ) return false;
-      const uint32_t stored_crc = parse_record_crc( tail + 10 );
-      const uint32_t computed_crc =
-        crc32c.windowed_crc( (const uint8_t *)buf, pos + rsize - 9, edsize );
-      crc_present = true;
-      if( stored_crc != computed_crc ) return false;
-      }
-    pos += rsize;
-    }
-  return true;
-  }
-
-
 int decode( const std::string & archive_name, const Arg_parser & parser,
             const int filenames, const int num_workers, const int debug_level,
             const bool keep_damaged, const bool listing, const bool missing_crc,
@@ -722,23 +643,27 @@ int decode( const std::string & archive_name, const Arg_parser & parser,
       if( !parse_records( infd, extended, header, permissive ) )
         { show_error( "Error in extended records. Skipping to next header." );
           extended.reset(); gretval = 2; }
-      else if( !extended.crc_present && missing_crc )
+      else if( !extended.crc_present() && missing_crc )
         { show_error( "Missing CRC in extended records.", 0, true ); return 2; }
       prev_extended = true;
       continue;
       }
     prev_extended = false;
 
-    if( extended.linkpath.empty() )	// copy linkpath from ustar header
+    if( extended.linkpath().empty() )	// copy linkpath from ustar header
       {
-      for( int i = 0; i < linkname_l && header[linkname_o+i]; ++i )
-        extended.linkpath += header[linkname_o+i];
-      while( extended.linkpath.size() > 1 &&		// trailing '/'
-             extended.linkpath[extended.linkpath.size()-1] == '/' )
-        extended.linkpath.resize( extended.linkpath.size() - 1 );
+      int len = 0;
+      while( len < linkname_l && header[linkname_o+len] ) ++len;
+      while( len > 1 && header[linkname_o+len-1] == '/' ) --len; // trailing '/'
+      if( len > 0 )
+        {
+        const uint8_t c = header[linkname_o+len]; header[linkname_o+len] = 0;
+        extended.linkpath( (const char *)header + linkname_o );
+        header[linkname_o+len] = c;
+        }
       }
 
-    if( extended.path.empty() )		// copy path from ustar header
+    if( extended.path().empty() )		// copy path from ustar header
       {
       char stored_name[prefix_l+1+name_l+1];
       int len = 0;
@@ -749,9 +674,9 @@ int decode( const std::string & archive_name, const Arg_parser & parser,
         { stored_name[len] = header[name_o+i]; ++len; }
       while( len > 0 && stored_name[len-1] == '/' ) --len;	// trailing '/'
       stored_name[len] = 0;
-      extended.path = remove_leading_slash( stored_name );
+      extended.path( remove_leading_slash( stored_name ) );
       }
-    const char * const filename = extended.path.c_str();
+    const char * const filename = extended.path().c_str();
 
     bool skip = filenames > 0;
     if( skip )
@@ -765,9 +690,9 @@ int decode( const std::string & archive_name, const Arg_parser & parser,
             { skip = false; name_pending[i] = false; break; }
           }
 
-    if( extended.size == 0 &&
+    if( extended.file_size() == 0 &&
         ( typeflag == tf_regular || typeflag == tf_hiperf ) )
-      extended.size = parse_octal( header + size_o, size_l );
+      extended.file_size( parse_octal( header + size_o, size_l ) );
 
     if( listing || skip )
       retval = list_member( infd, extended, header, skip );