On Fri, May 2, 2008 at 5:48 PM, Hyrum K. Wright
<hyrum_wright_at_mail.utexas.edu> wrote:
>
> Hyrum K. Wright wrote:
> > * r30963: Not technically a blocker, but a pretty big performance
> > booster. Would be nice to have.
> >
>
> Mark and I will look at this.
I tested this pretty thoroughly on his branch and am in favor of it.
However, it does not merge cleanly to 1.5.x and requires a backport
branch.
I'd agree it is probably not a true blocker, but it is more than a
performance booster. Whatever he did that fixed the performance also
fixed all of the main bugs that we had been planning to postpone. For
example, log -g on the root of the repository not only runs quickly
(if given a smallish range) but it also shows merge results which the
current code does not do.
Also, while the performance boost is great from the client
perspective, my bigger worry is people running log -g on their entire
repository and killing the server. I do not think with the current
code it would be a problem. All this is just to say I think it is
really worth getting into RC5.
I'd also like to make a pitch for the SASL bug that glasser fixed. I
was banging my head on this problem for a solid day so I was able to
see the before and after that he in fact fixed the problem. epg also
tested this, so hopefully he will see this and add his +1 to STATUS so
we can backport it. SASL digest mode does not work without his patch,
and that is probably the recommended mode (at least until using
Kerberos becomes easy).
--
Thanks
Mark Phippard
http://markphip.blogspot.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-05-03 01:46:05 CEST