gpg: Signature made Fri May 9 22:03:24 2008 UTC using DSA key ID E43DA450
gpg: Good signature from "Matthew John Toseland <toad@amphibian.dyndns.org>"
On Sunday 04 May 2008 14:11, you wrote:
> Author: j16sdiz
> Date: 2008-05-04 13:11:55 +0000 (Sun, 04 May 2008)
> New Revision: 19733
>
> Added:
>
branches/saltedhashstore/freenet/src/freenet/store/SaltedHashFreenetStore.java
> Log:
> SaltedHashFreenetStore basic function (without locking/resizing)
>
>
> Added:
branches/saltedhashstore/freenet/src/freenet/store/SaltedHashFreenetStore.java
> ===================================================================
> ---
branches/saltedhashstore/freenet/src/freenet/store/SaltedHashFreenetStore.java
(rev 0)
> +++
branches/saltedhashstore/freenet/src/freenet/store/SaltedHashFreenetStore.java
2008-05-04 13:11:55 UTC (rev 19733)
> +
> + public StorableBlock getStorableBlock(byte[] routingKey, byte[] fullKey)
throws KeyVerifyException {
> + if (!decrypt(routingKey))
> + return null;
> +
> + StorableBlock block = callback.construct(data, header, routingKey,
fullKey);
> + byte[] blockRoutingKey = block.getRoutingKey();
> +
> + if (!Arrays.equals(blockRoutingKey, routingKey)) {
> + // either the data is corrupted or we have found a SHA-1 collision
> + // can't recover, as decrypt() depends on a correct route key
> + return null;
> + }
> +
> + return block;
> + }
If we can't be sure that the Entry isn't re-encrypted, we should copy the
data/headers here. If we can, we should document it.
> +
> + /**
> + * Check if a block is free
> + *
> + * @param offset
> + * @throws IOException
> + */
> + private boolean isFree(long offset) throws IOException {
> + int split = (int) (offset % FILE_SPLIT);
> + long rawOffset = (offset / FILE_SPLIT) * entryTotalLength;
> +
> + ByteBuffer bf = ByteBuffer.allocate((int) ENTRY_HEADER_LENGTH);
> +
> + do {
> + int status = storeFC[split].write(bf, rawOffset + bf.position());
> + if (status == -1)
> + throw new EOFException();
> + } while (bf.hasRemaining());
> +
> + return (bf.getLong(0x30) & ENTRY_FLAG_OCCUPIED) == 0;
> + }
Why read the whole entry just to check one byte (or one long)?
