<[EMAIL PROTECTED]> writes:
> On 6 Jun 2001, Jeff Trawick wrote:
>
> > <[EMAIL PROTECTED]> writes:
> >
> > > So yeah, if we added the logic to the sub-req-output-filter to do the
> > > conversion, this problem would be solved. The logic is already in
> > > core_output_filter, it's just a simple copy and paste.
> >
> > I'm running such logic now... But it seems there is another problem,
> > perhaps the one Ian suggested a few days back where there is missing
> > output.
> >
> > The instance of the content length filter which saved my included file
> > in its context is the one associated with the subrequest. That
> > instance of the content length filter never sees EOS so the contents
> > of my included file is lost forever.
> >
> > Why is the content length filter added for the subrequest? It
> > shouldn't be there at all, should it?
>
> There are two problems here. #1, the subrequest shouldn't be getting the
> C-L filter. #2, the C-L filter should have seen the EOS that was sent for
> the sub-request. The EOS shouldn't be removed until the
> sub_req_output_filter, and that should always be after the C-L
> filter.
Yep... see the patches that should show up any time in the next few
hours :)
The unnecessary/harmful filters on subrequests came about with
Graham's change to add the HTTP protocol filters in the insert filter
hook.
--
Jeff Trawick | [EMAIL PROTECTED] | PGP public key at web site:
http://www.geocities.com/SiliconValley/Park/9289/
Born in Roswell... married an alien...
