On Wed, October 25, 2006 4:14 pm, Davi Arnaut wrote:
> && (rv == APR_SUCCESS && rv2 == APR_SUCCESS) {
>
> Otherwise we may keep walking through the brigade unnecessarily.This is intentional - we don't want a failure on store_body() to cause a premature abort on the network write. The inverse of not wanting a failure on network write to affect store_body() also needs to be fleshed out, and needs to be configurable, but that's a job for another patch. Regards, Graham --
