Mailing-List: contact dev-help@httpd.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@httpd.apache.org
Received-SPF: pass (nike.apache.org: domain of michael@metaparadigm.com
 designates 203.117.131.43 as permitted sender)
Message-ID: <476CDB71.4010500@metaparadigm.com>
Date: Sat, 22 Dec 2007 17:40:01 +0800
From: Michael Clark <michael@metaparadigm.com>
Organization: Metaparadigm Pte Ltd
User-Agent: Mozilla-Thunderbird 2.0.0.6 (X11/20071009)
MIME-Version: 1.0
To: dev@httpd.apache.org
Subject: Re: [RFC] Apache Privilege Separation for WebDAV (now against 2.2.6)
References: <475E8FE6.3080802@metaparadigm.com>
 <4768EC1A.2020400@force-elite.com>
In-Reply-To: <4768EC1A.2020400@force-elite.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Paul Querna wrote:
> How you stubbed out the file io seems fine for the lifetime of 2.2.x 
> and maybe 2.4.x, but in the long run, I believe we need to support 
> some kind of "VFS" layer, to make all IO pluggable. (open file, 
> directory listing, etc etc).

Just thinking aloud on a few other benefits from a pluggable vfs layer.

  * A stackable performance statistics vfs module implementation that 
does timings on filesystem io calls.
  * A statcache stackable vfs modules that just stubs out the stat 
method and caches stat calls for a configurable amount of time (say 10 
seconds).
  * An audit logging stackable vfs module that logs all io calls (mainly 
for debugging).

In a addition to having a way to mount a vfs hook implementation for a 
given path, perhaps also an interface to replace a vfs hook with a 
delegating entry so that it can be stacked.

Michael.