On Thu, 13 Mar 2003 06:34:37 -0800 Justin Erenkrantz <[EMAIL PROTECTED]> wrote:
[...] > > It would be cool if we could extend mod_proxy, but if this is impossible > > (because of some technical issues I'm not aware of) we might do our own > > thing. Writting a small, customized proxy in C/APR, Perl, Python, whatever > > schouldn't be all that hard. > > If you use httpd-2.0 as a framework, I think you could get away with > something like the following: install an input filter that intercepts all > inbound data At first I though about extending mod_proxy, but filters seem to be a bit better choice. I don't like the idea of swallowing all the data. We could at least allow to specify some URL regexp like http://www.site.com/(.*), and store only bits matching such patern. > and writes the input body to a file in a special format that is > essentially like flood's URL XML syntax for a urllist. I was thinking about storing all the data in memory and dumping them on demand, but this is hard (if possible at all) to achieve, so we might just stick to file writes. Fine editing of such urllist would be the job of flood's GUI. If there are no special reasons against, I would like such module to be configurable only through custom response handler. I get the feeling that a bunch of server directives would trigger too much server restarts. But that's just my opinion. > If you'd like to pursue this, let me know and I can try to give more > specifics. -- justin I've never written anything for 2.0 (only two small, custom modules for 1.3), so a tip or two from httpd-2.0 guru is very welcome :) regards, Jacek Prucia
