On Wed, 16 Dec 2009, Goswin von Brederlow wrote:
How much does that help? Do those bzip2 files converge again so you only need to store some kb delta or is basically everything after the first difference changed?
I seem to remember to compare them with vbindiff, and I believe a few bytes matched in common positions, but most were different.
I suspect -- since the input to the two compressor versions is the same and the first stage RLE [0] has presumably not changed between 0.9.5d and 1.0 -- that the input to the BWT is identical. Based on the changelog, something strictly after the first stage RLE could have been changed. As the compressed blocks are bit-aligned in the output, there is a good chance, in my opinion, that the two files will never resync after the first difference, even if some compressed blocks are otherwise identical.
This is just speculation, though. Cheers, lacos [0] http://en.wikipedia.org/wiki/Bzip2#Compression_stack -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
