On Sunday 22 June 2008 09:44, Florent Daigni?re wrote:
> * toad at freenetproject.org <toad at freenetproject.org> [2008-06-21
> 14:38:38]:
>
> > Author: toad
> > Date: 2008-06-21 14:38:38 +0000 (Sat, 21 Jun 2008)
> > New Revision: 20573
> >
> > Added:
> >
branches/db4o/freenet/src/freenet/support/io/BucketChainBucketFactory.java
> > Modified:
> > branches/db4o/freenet/src/freenet/client/async/InsertCompressor.java
> > branches/db4o/freenet/src/freenet/client/async/SplitFileInserter.java
> > branches/db4o/freenet/src/freenet/support/io/BucketChainBucket.java
> > branches/db4o/freenet/src/freenet/support/io/BucketTools.java
> > Log:
> > Write compressed data directly into a chain of buckets.
> > Saves copying, disk space, and time later on. (Currently this time is on
the database thread, because it's in the SplitFileFetcher constructor...).
> > BACKPORT.
> >
> > Modified: branches/db4o/freenet/src/freenet/support/io/BucketTools.java
> > ===================================================================
> > --- branches/db4o/freenet/src/freenet/support/io/BucketTools.java
2008-06-21 14:23:52 UTC (rev 20572)
> > +++ branches/db4o/freenet/src/freenet/support/io/BucketTools.java
2008-06-21 14:38:38 UTC (rev 20573)
> > @@ -334,13 +334,27 @@
> > * and the data written to them.
> > *
> > * Note that this method will allocate a buffer of size splitSize.
> > + * @param freeData
> > * @throws IOException If there is an error creating buckets, reading
from
> > * the provided bucket, or writing to created buckets.
> > */
> > - public static Bucket[] split(Bucket origData, int splitSize,
BucketFactory bf) throws IOException {
> > + public static Bucket[] split(Bucket origData, int splitSize,
BucketFactory bf, boolean freeData) throws IOException {
> > if(origData instanceof FileBucket) {
> > + if(freeData) {
> > + Logger.error(BucketTools.class, "Asked to free
> > data when splitting a
FileBucket ?!?!? Not freeing as this would clobber the split result...");
> > + }
> > return ((FileBucket)origData).split(splitSize);
> > }
> > + if(origData instanceof BucketChainBucket) {
> > + BucketChainBucket data = (BucketChainBucket)origData;
> > + if(data.bucketSize == splitSize) {
> > + if(freeData)
> > + data.clear();
> > + return data.getBuckets();
> > + } else {
> > + Logger.error(BucketTools.class, "Incompatible
> > split size splitting a
BucketChainBucket: his split size is "+data.bucketSize+" but mine
is "+splitSize+" - we will copy the data, but this suggests a bug", new
Exception("debug"));
> > + }
> > + }
> > long length = origData.size();
> > if(length > ((long)Integer.MAX_VALUE) * splitSize)
> > throw new IllegalArgumentException("Way too big!:
> > "+length+"
for "+splitSize);
> > @@ -374,6 +388,8 @@
> > else
> > is.close();
> > }
> > + if(freeData)
> > + origData.free();
> > return buckets;
> > }
>
> Is there a reason why we don't do the origData.free() in the finally
> block above ?
>
Yes, we don't want to free it if we don't return anything because we throw!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL:
<https://emu.freenetproject.org/pipermail/devl/attachments/20080624/73fdc1d9/attachment.pgp>
