On Wednesday 29 April 2009 10:15:23 bo-le wrote: > Am Mittwoch, 29. April 2009 03:20:48 schrieb Matthew Toseland: > > On Tuesday 21 April 2009 15:16:55 saces at freenetproject.org wrote: > > > Author: saces > > > Date: 2009-04-21 14:16:55 +0000 (Tue, 21 Apr 2009) > > > New Revision: 27153 > > > > > > Modified: > > > trunk/freenet/src/freenet/client/async/BaseManifestPutter.java > > > Log: > > > indent: add comments > > > > > > Modified: trunk/freenet/src/freenet/client/async/BaseManifestPutter.java > > > =================================================================== > > > --- trunk/freenet/src/freenet/client/async/BaseManifestPutter.java > > > > 2009-04-21 12:45:14 UTC (rev 27152) > > > > > +++ trunk/freenet/src/freenet/client/async/BaseManifestPutter.java > > > > 2009-04-21 14:16:55 UTC (rev 27153) > > > > > @@ -1,3 +1,6 @@ > > > +/* This code is part of Freenet. It is distributed under the GNU General > > > + * Public License, version 2 (or at your option any later version). See > > > + * http://www.gnu.org/ for further details of the GPL. */ > > > package freenet.client.async; > > > > > > import java.io.IOException; > > > @@ -30,6 +33,20 @@ > > > import freenet.support.io.BucketTools; > > > import freenet.support.io.NativeThread; > > > > > > +/** > > > + * This class contains all the insert logic, but not any 'pack logic'. > > > + * The pack logic have to be implement in a subclass in makePutHandlers. > > > + * @see PlainManifestPutter and @see DefaultManifestPutter. > > > + * > > > + * Internal container redirect URIs: > > > + * The internal container URIs should be always redirects to CHKs, not > > > > just include the metadata into manifest only. > > > > > + * The (assumend) default behaviour is the reuse of containers between > > > > editions, > > > > > + * also ArchiveManger want to have a URI given, not Metadata. > > > + * This rule also makes site update code/logic much more easier. > > > > Okay, given that small files will always be inside a container, and given > > that we will implement DHKs, and internal URIs will then be DHKs i.e. they > > will not be dependant on a single block. If either of these conditions is > > false, then it makes sense to keep the top level inside the manifest. > > Using the archive hash (sha256) as identifier for archive manager is an > alternative, > this would allow matadata inclusion (container insert with > reportMatadataOnly=true)
I dunno. It seems wasteful to insert a separate CHK for a 2+2 block splitfile, and has poor redundancy, and if we have DHKs then we're inserting 7 blocks for that <64KB file! -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 835 bytes Desc: This is a digitally signed message part. URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20090604/e601f89e/attachment.pgp>
