[freenet-support] Freenet 0.7 build 1181
Freenet 0.7 build 1181 is now available. It will be mandatory on Sunday. Please upgrade. The main change in 1181 is a partial fix (workaround) to the download corruption bug. All sites should now load without decompression errors and without silent data corruption. However, we implement this by ignoring the last data block of each segment, so there may be a small cost to performance and/or data reachability. The origin of the problem is that we have apparently had 3 different padding algorithms for the last block of the splitfile, all of them based on different incorrect byte[] to int[] conversions (yes, I know this is incredibly dumb). The new freenet-ext.jar (#26) finally implements the correct conversion algorithm. Unfortunately, old data will be padded incorrectly, in one of three different ways, and there's no easy way to identify which. Meaning if we download a splitfile and by chance happen to include the last data block in the decode, we can get data corruption. So the solution is to ignore the last data block when FEC decoding. 1182 will provide a more permanent solution, involving a new metadata format and enforced post-decode post-decompress checksums. However 1181 should fix the problem. PLEASE UPGRADE! If your node has problems upgrading (it might well!), please run update.sh / update.cmd manually. If you find any bugs in 1181, please let us know, preferably via the bug tracker: https://bugs.freenetproject.org/ -- next part -- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 827 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/support/attachments/20081121/9f35e5a7/attachment.pgp>
[freenet-support] Freenet 0.7 build 1181
Freenet 0.7 build 1181 is now available. It will be mandatory on Sunday. Please upgrade. The main change in 1181 is a partial fix (workaround) to the download corruption bug. All sites should now load without decompression errors and without silent data corruption. However, we implement this by ignoring the last data block of each segment, so there may be a small cost to performance and/or data reachability. The origin of the problem is that we have apparently had 3 different padding algorithms for the last block of the splitfile, all of them based on different incorrect byte[] to int[] conversions (yes, I know this is incredibly dumb). The new freenet-ext.jar (#26) finally implements the correct conversion algorithm. Unfortunately, old data will be padded incorrectly, in one of three different ways, and there's no easy way to identify which. Meaning if we download a splitfile and by chance happen to include the last data block in the decode, we can get data corruption. So the solution is to ignore the last data block when FEC decoding. 1182 will provide a more permanent solution, involving a new metadata format and enforced post-decode post-decompress checksums. However 1181 should fix the problem. PLEASE UPGRADE! If your node has problems upgrading (it might well!), please run update.sh / update.cmd manually. If you find any bugs in 1181, please let us know, preferably via the bug tracker: https://bugs.freenetproject.org/ pgp06gLX7t651.pgp Description: PGP signature ___ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]