You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by "William A. Rowe, Jr." <wr...@rowe-clan.net> on 2001/02/27 16:57:19 UTC

Speaking of forgotten patches...

(thanks for reminding me Paul) ...

where are we with a Filtered ap_rput semantic v.s. a non-filtered, r->bb dependent
semantic?  We needed a patch either way to finish this, if I'm remembering right.

If we still can't agree, let me point out one thing.  If it's a filter, it needs
to be documented in the right places so folks don't go inserting a filter into
the wrong spot.  If it's r->bb, it needs to be documented so folks don't go creating
another brigade.

I'm at the point where the most obvious docs that will keep bugs reports to a 
minimum wins the day, in spite of my opposition to extra filter layers.

Bill



Re: Speaking of forgotten patches...

Posted by Greg Stein <gs...@lyra.org>.
On Tue, Feb 27, 2001 at 09:57:19AM -0600, William A. Rowe, Jr. wrote:
> (thanks for reminding me Paul) ...
> 
> where are we with a Filtered ap_rput semantic v.s. a non-filtered, r->bb dependent
> semantic?  We needed a patch either way to finish this, if I'm remembering right.
> 
> If we still can't agree, let me point out one thing.  If it's a filter, it needs
> to be documented in the right places so folks don't go inserting a filter into
> the wrong spot.  If it's r->bb, it needs to be documented so folks don't go creating
> another brigade.
> 
> I'm at the point where the most obvious docs that will keep bugs reports to a 
> minimum wins the day, in spite of my opposition to extra filter layers.

Right now, we use AP_FTYPE_CONTENT-1 for the type. That was always just a
short cut to get the stuff up and running. It should be replaced with its
own value, which I'd simply suggest as:

AP_FTYPE_INTERNAL  0  /* reserved for Apache-internal filters */

Nobody will get near it then :-)


Would anybody mind if I whacked the OLD_WRITE filter to use brigade
buffering? It will be totally invisible to all users/callers, so there won't
be an impact. It would at least demo my intent. Since there's no semantic
change, there shouldn't be a problem, and OLD_WRITE can always be yanked.

Cheers,
-g

-- 
Greg Stein, http://www.lyra.org/