Re: [Monotone-devel] Re: Extensions to automation, again

[Nathaniel Smith]

On Wed, Oct 18, 2006 at 11:00:14PM +0100, Bruce Stephens wrote:
> Nathaniel Smith <address@hidden> writes:
> 
> [...]
> 
> > Oh, uh -- good point.  We even had a function to do this
> > (xdelta.cc:measure_delta_target_size) that got stripped out a few
> > months ago because there were no callers.
> >
> > Objections withdrawn.
> 
> Cool, though that still leaves full-text versions, I guess (or can one
> determine the uncompressed size in some way)?

Not really -- I guess we actually do store it in the gzip header right
now, for whatever _that's_ worth (that we put gzip headers on the db
blobs at all is kind of dumb).  You might also have diffs that are
large like fulltexts -- sometimes there's just nothing in common.

I guess it's true, with this approach, the common case (where you're
looking at a head) doesn't gain anything at all, since there the files
are stored as fulltexts anyway, so if you want to dump the file sizes
for the full tree you do have to churn through all the data in the
tree too.

If we switched to forward deltas that would be different, of course...
so many options :-).

-- Nathaniel

-- 
"But suppose I am not willing to claim that.  For in fact pianos
are heavy, and very few persons can carry a piano all by themselves."