Return-Path: Delivered-To: apmail-new-httpd-archive@apache.org Received: (qmail 31490 invoked by uid 500); 2 Aug 2001 14:13:05 -0000 Mailing-List: contact new-httpd-help@apache.org; run by ezmlm Precedence: bulk Reply-To: new-httpd@apache.org list-help: list-unsubscribe: list-post: Delivered-To: mailing list new-httpd@apache.org Received: (qmail 31477 invoked from network); 2 Aug 2001 14:13:05 -0000 Errors-To: Message-ID: <17e901c11b5c$de4b9100$93c0b0d0@roweclan.net> From: "William A. Rowe, Jr." To: References: <3B33B2B2.40800@pacbell.net> <114401c11ab2$348fe170$93c0b0d0@roweclan.net> <3B691549.7040306@pacbell.net> Subject: Re: [PATCH] Re: directory_walk performance Date: Thu, 2 Aug 2001 09:08:23 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 X-Spam-Rating: h31.sny.collab.net 1.6.2 0/1000/N Status: O X-Status: X-Keywords: X-UID: 50 From: "Brian Pane" Sent: Thursday, August 02, 2001 3:54 AM > Here's a patch that implements my pre-merge optimization to > reduce (and in some cases completely eliminate) the calls to > ap_merge_per_dir_configs. Its impacts on directory_walk are > minor; all the real work happens in a new post-config-phase > function. Another glitch. If the source is from the cache, we can't trust that elements of base or add allocated will still in memory when the cache is called :( Not unless we also discard all decendants when we refresh the cache. This implies that we need a protected refcount for each of the cached elements, not only by the request that is relying on the data, but the straight down the chain. Bill