Return-Path: Delivered-To: new-httpd-archive@hyperreal.org Received: (qmail 10361 invoked by uid 6000); 4 Feb 1998 08:00:15 -0000 Received: (qmail 10354 invoked from network); 4 Feb 1998 08:00:14 -0000 Received: from staff2.texas.net (mikedoug@207.207.0.39) by taz.hyperreal.org with SMTP; 4 Feb 1998 08:00:14 -0000 Received: (from mikedoug@localhost) by staff2.texas.net (8.8.5/8.7.5) id CAA19233; Wed, 4 Feb 1998 02:00:12 -0600 (CST) Message-ID: <19980204020012.47146@texas.net> Date: Wed, 4 Feb 1998 02:00:12 -0600 From: Michael Douglass To: new-httpd@apache.org Subject: Re: apache/linux modules References:

Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.88 In-Reply-To: ; from Dean Gaudet on Mon, Feb 02, 1998 at 04:13:05PM -0800 Sender: new-httpd-owner@apache.org Precedence: bulk Reply-To: new-httpd@apache.org On Mon, Feb 02, 1998 at 04:13:05PM -0800, Dean Gaudet said: > You'll waste some memory this route, but for the 90% case that isn't an > issue. Why would it be wasting memory with this route? I know that, at least under Solaris, .so code segments are placed in shared memory (hence 'shared libraries')... Wouldn't this actually save memory since having 100 servers running will have only a fraction of the amount of code space used for the modules as normal? To me this would be a good think. Is that what you mean by the 90% case??? And if that is the case, then this concept and idea should be touted as a thing to do. Although I _hardly_ would want a static-only or dynamic-only concept... One in which you have static modules and have the option of loading static modules seems a nice compromise. Or am I lost? -- Michael Douglass Texas Networking, Inc. anyway, I'm off, perl code is making me [a] crosseyed toady