On 2/23/07, William A. Rowe, Jr. <[EMAIL PROTECTED]> wrote:
Jonathan Gilbert wrote:
>
> The updated version of the patch only enforces chunked writes specifically
> for console output, and it does it using the existing buffering
> functionality, so no new code is required, unlike the original patch.
But unlike your original patch, you begin buffering output that the
author didn't intend to buffer?
You assume stdhandles are consoles. Other objects can be similarly
pipe-oriented streams that will blow up the same way, we might as well
fix all these cases now if we are investing the time in this patch.
Yet stdhandles aren't necessarily consoles and your patch tromps those
cases of redirected standard handles with this assumption.
It appears your proposed patch is moving in the wrong direction from
where you started, IMHO. Can we return to the earliest comments and
react to the case from the moment where we first flood NT's pipe buffer,
and we must back down to 32k writes?
Jonathan,
Any progress on the matter? I'd love to be able to close the issue in
Subversion by pointing to the right APR version!
bye,
Erik.
