Brandon <blanu at uts.cc.utexas.edu> writes:

> So are you saying that the CHKs for two identical files with
> differing metadata will be different? Because the point is that the
> CHKs for two identical files with differing metadata should be the
> same unless they're control documents, in which case they should be
> different, which is the fundamental problem here.

I'm pretty sure that right now the CHK hash is generated based on the
entire payload.  So long as we're doing that, it doesn't make any
sense whatsoever to generate the encryption key based on just the data
part.

But you have a point that it would be nice to have *both* the hash and
the encryption key generated based only on the data part, and that the
only thing this messes up is control documents, since they put their
data in the metadata part.  That's a lame idea anyway, so let's just
not do that.  Instead of identifying control documents by their 0 data
length, let's just have a metadata header that says "this is a control
document", and put the data in the data part where it belongs.  That
way, duplicate CHK control documents will be avoided just like
duplicates of everything else.

-S

_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to