Greg Hudson wrote:
>By sacrificing random access, we could get the following benefits:
>
> * We could get rid of delta windowing, which is a source of
>complexity. (Obviously not before 2.0, but I'm not sure if we can
>realize any of these benefits before 2.0.) Eliminating windowing also
>means we could get much better results for some changes, like "insert
>128K of data at the beginning of the file".
>
>
I'm not sure that's the case. Xdelta and diff both use windows because
they'd otherwise have to index all the source and target data at once,
and streaminess is still a requirement. The restriction might possibly
make it easier to implement sliding windows, though.
-- Brane
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Feb 11 00:12:36 2005