On 5/26/05, Greg Hudson <ghudson@mit.edu> wrote:
> On Thu, 2005-05-26 at 22:56 +0100, Colin JN Breame wrote:
> > A timeout would be useful here; if the client does not receive any response
> > from the server after x seconds (when expecting a response), report an error
> > and fail.
>
> We've had some pretty nasty problems in the past on the ra_dav side
> because our default timeout used to be too short, at I think 120
> seconds. Sometimes a big operation really does take that long. So I'm
> not a big fan of timeouts. Even if we could find the magic number which
> always means it's time to give up, it still doesn't tell the user
> whether it's a server problem or a network problem; all it does is help
> rule out a client-side problem.
>
> Which is not to say that it's not a good idea to provide users of some
> indication that things are progressing. We do that in some cases and
> not others. In many of the cases where we don't, we'd have to extend
> our protocols to provide progress indications; it could get to be a lot
> of work.
Why not add it as a parameter that can be configured for different
repositories. With no default you get the current behavior, but for
someone like me who has a work machine that only accesses svn on our
internal network, we can set a timeout that would reflect the lower
network latency.
I don't know how much effort it would be to add, but it would allow
both of your positions and give the clients more control.
Josh
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri May 27 02:31:24 2005