Re: Subversion Performance Benchmark

Greg Hudson <ghudson@MIT.EDU> writes:
> Still, this is pretty dumb when we're using ra_local. Not sure what the
> best fix is.

I think the solution is what Greg Stein has been advocating for a long
time: change from a push interface to a pull interface. Then the RA
layer would decide whether or not compression and/or diffing is called
for, and if so, what kind.

We used to have a document that described this in more detail (it was
listed as the infamous "Change #6" in that document). Now all we seem
to have is this issue

   http://subversion.tigris.org/issues/show_bug.cgi?id=510

which has an inadequate description. Greg S, do you still have the
detailed text somewhere? (Is this change applicable here, or am I
misremembering?)

-K

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Sep 24 19:59:22 2002