httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brian Pane <bp...@pacbell.net>
Subject Re: cvs commit: httpd-2.0/server/mpm/worker worker.c
Date Tue, 13 Nov 2001 06:16:23 GMT
Ryan Bloom wrote:

>On Monday 12 November 2001 09:48 pm, Justin Erenkrantz wrote:
>
[...]

>>modules/proxy/proxy_connect.c does raw socket writes (see line
>>308).  I think the idea here is that mod_proxy wants to bypass
>>everyone.  Not the greatest of ideas (quite bogus actually) -
>>perhaps we can setup a minimal filter output stack.
>>
>>Thoughts?  I haven't had time to digest or even look at your
>>patch yet, so I'm not really sure what you did.  -- justin
>>
>
>The proxy is bogus currently.  However, the part you are talking about
>is the client side of the proxy, not the server.  When Apache is acting as
>a client, it can use network primitives directly.  The important part of this
>patch, is that it is possible to accept requests from transport layers that 
>aren't sockets.
>

Yep, the raw socket writes aren't an issue because they're on the
client side of the proxy.  For a hypothetical future event-based
MPM, I suppose we'd just need an interface to let mod_proxy add
its client-side socket to the MPM's set of descriptors to poll

--Brian



Mime
View raw message