On 08 Dec 2005 13:08:24 -0600, kfogel@collab.net <kfogel@collab.net> wrote:
> Greg Hudson <ghudson@MIT.EDU> writes:
> > I have three problems with this patch:
> >
> > * It doesn't reset the socket timeout after connecting, so it has
> > the unintended consequence of introducing a 30s timeout in ra_svn,
> > although that timeout gets unset for some operations once you go
> > into a pipelined edit.
> >
> > * We should not be introducing an unconfigurable 30-second timeout
> > on making network connections. The connect timeout is an OS
> > decision, not a we-hardcode-it-in-Subversion decision.
> > [ ... ]
> > Karl, please revert the change.
> On the basis of your comments, it sounds like acceptable solutions
> would be to:
> [...]
> 2) Revert r15909.
> I don't have time to do (1) and test it thoroughly, so I've done (2)
> instead, in r17697. [...]
Should we propose r17697 for backport to 1.3.0?
Cheers,
David
--
David James -- http://www.cs.toronto.edu/~james
Received on Thu Dec 8 22:22:11 2005