hashirama/goldendict-ng
hashirama/goldendict-ng
0
0
Fork 0
mirror of https://github.com/xiaoyifang/goldendict-ng.git synced 2024-11-27 19:24:08 +00:00
Code Issues Actions 1 Projects Releases Wiki Activity

Zim: Handle clusters in random order case

Browse source
This commit is contained in:
Abs62 2020-10-06 17:52:30 +03:00
parent b70c3e8c88
commit 8e13789e8f
1 changed files with 37 additions and 11 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

48
zim.cc
View file

@ -191,6 +191,7 @@ private:
ZIM_header zimHeader;
Cache cache[ CACHE_SIZE ];
int stamp;
QVector< QPair< quint64, quint32 > > clusterOffsets;
void clearCache();
};
@ -270,6 +271,27 @@ bool ZimFile::open()
if( read( reinterpret_cast< char * >( &zimHeader ), sizeof( zimHeader ) ) != sizeof( zimHeader ) )
return false;
// Clusters in zim file may be placed in random order.
// We create sorted offsets list to calculate clusters size.
clusterOffsets.resize( zimHeader.clusterCount );
QVector< quint64 > offs;
offs.resize( zimHeader.clusterCount );
seek( zimHeader.clusterPtrPos );
qint64 size = zimHeader.clusterCount * sizeof( quint64 );
if( read( reinterpret_cast< char * >( offs.data() ), size) != size )
{
vector< string > names;
getFilenames( names );
throw exCantReadFile( names[ 0 ] );
}
for( quint32 i = 0; i < zimHeader.clusterCount; i++ )
clusterOffsets[ i ] = QPair< quint64, quint32 >( offs.at( i ), i );
qSort( clusterOffsets );
return true;
}
@ -312,24 +334,25 @@ string ZimFile::getClusterData( quint32 cluster_nom )
// Cache miss, read data from file
// Read cluster pointers
quint64 clusters[ 2 ];
seek( zimHeader.clusterPtrPos + cluster_nom * 8 );
if( read( reinterpret_cast< char * >( clusters ), sizeof(clusters) ) != sizeof(clusters) )
return string();
// Calculate cluster size
quint64 clusterSize;
if( cluster_nom < zimHeader.clusterCount - 1 )
clusterSize = clusters[ 1 ] - clusters[ 0 ];
quint32 nom;
for( nom = 0; nom < zimHeader.clusterCount; nom++ )
if( clusterOffsets.at( nom ).second == cluster_nom )
break;
if( nom >= zimHeader.clusterCount ) // Invalid cluster nom
return string();
if( nom < zimHeader.clusterCount - 1 )
clusterSize = clusterOffsets.at( nom + 1 ).first - clusterOffsets.at( nom ).first;
else
clusterSize = size() - clusters[ 0 ];
clusterSize = size() - clusterOffsets.at( nom ).first;
// Read cluster data
seek( clusters[ 0 ] );
seek( clusterOffsets.at( nom ).first );
char compressionType;
if( !getChar( &compressionType ) )
@ -356,6 +379,9 @@ string ZimFile::getClusterData( quint32 cluster_nom )
else
return string();
if( decompressedData.empty() )
return string();
// Check BLOBs number in the cluster
// We cache multi-element clusters only