public void AddProperty(Property p)
{
if (p != null)
{
properties.Add(p);
}
if (p.Directive == (byte)Directives.Attributes)
{
attributes = OperaHistory.ParseRow(p.Content).Properties;
}
}
public bool HasNextIndexable ()
{
if (history_enumerator == null)
return false;
while (history_enumerator.MoveNext ()) {
OperaHistory.Row row = (OperaHistory.Row) history_enumerator.Current;
if (Allowed (row) && ! IsUpToDate (row))
return true;
}
history_enumerator = null;
history = null; // Help the GC here
return false;
}
public static void Main(string[] args)
{
if (args.Length != 1)
{
return;
}
OperaHistory op = new OperaHistory(args [0]);
IEnumerator iter = op.GetEnumerator();
while (iter.MoveNext())
{
Row row = (Row)iter.Current;
Uri uri = row.Address;
Console.WriteLine(uri);
}
}
public OperaIndexableGenerator(OperaIndexer indexer, string cache_dir)
{
this.cache_dir = cache_dir;
this.indexer = indexer;
if(history != null && history.GetLastRead() >= Directory.GetLastWriteTime(cache_dir)){
history_enumerator = history.GetEnumerator();
return;
}
try {
history = new OperaHistory (Path.Combine (cache_dir, "dcache4.url"));
history_enumerator = history.GetEnumerator ();
} catch (Exception e) {
Logger.Log.Error (e, "Failed to list cache objects in {0}",
Path.Combine (cache_dir, "dcache4.url"));
}
}
public OperaIndexableGenerator(OperaIndexer indexer, string cache_dir)
{
this.cache_dir = cache_dir;
this.indexer = indexer;
if (history != null && history.GetLastRead() >= Directory.GetLastWriteTime(cache_dir))
{
history_enumerator = history.GetEnumerator();
return;
}
try {
history = new OperaHistory(Path.Combine(cache_dir, "dcache4.url"));
history_enumerator = history.GetEnumerator();
} catch (Exception e) {
Logger.Log.Error(e, "Failed to list cache objects in {0}",
Path.Combine(cache_dir, "dcache4.url"));
}
}
public bool HasNextIndexable()
{
if (history_enumerator == null)
{
return(false);
}
while (history_enumerator.MoveNext())
{
OperaHistory.Row row = (OperaHistory.Row)history_enumerator.Current;
if (Allowed(row) && !IsUpToDate(row))
{
return(true);
}
}
history_enumerator = null;
history = null; // Help the GC here
return(false);
}
private Indexable OperaRowToIndexable (OperaHistory.Row row)
{
// It's unsafe to index secure content since it may contain sensitive data
if (row.Address.Scheme == Uri.UriSchemeHttps)
return null;
Indexable indexable = new Indexable (row.Address);
indexable.HitType = "WebHistory";
indexable.MimeType = row.MimeType;
indexable.Timestamp = row.LastVisited;
indexable.AddProperty(Beagle.Property.New ("fixme:host",row.Address.Host));
indexable.AddProperty (Beagle.Property.NewUnsearched ("fixme:size", row.Length));
// hint for the filter about the charset
indexable.AddProperty (Property.NewUnsearched (StringFu.UnindexedNamespace + "charset", row.Encoding.ToString ()));
if(row.Compression == "gzip")
indexable.SetBinaryStream (new GZipInputStream (File.OpenRead (Path.Combine (cache_dir, row.LocalFileName))));
else
indexable.ContentUri = new Uri (Path.Combine (cache_dir, row.LocalFileName));
indexer.AttributeStore.AttachLastWriteTime (Path.Combine (cache_dir, row.LocalFileName), DateTime.UtcNow);
return indexable;
}
public bool Allowed (OperaHistory.Row row)
{
foreach (string mime in indexed_mimetypes) {
if (row.MimeType == mime)
return true;
}
return false;
}
public bool IsUpToDate (OperaHistory.Row row)
{
return indexer.AttributeStore.IsUpToDate (Path.Combine (cache_dir, row.LocalFileName));
}
public static void Main (string[] args)
{
if (args.Length != 1)
return;
OperaHistory op = new OperaHistory (args [0]);
IEnumerator iter = op.GetEnumerator ();
while (iter.MoveNext ()) {
Row row = (Row) iter.Current;
Uri uri = row.Address;
Console.WriteLine (uri);
}
}
