Mailing-List: contact modules-dev-help@httpd.apache.org; run by ezmlm
Precedence: bulk
Reply-To: modules-dev@httpd.apache.org
Received-SPF: pass (athena.apache.org: domain of jmarantz@google.com
 designates 209.85.214.175 as permitted sender)
MIME-Version: 1.0
In-Reply-To: 
 <CANOjuGEq2Fb9+McDnar8eCN=fkNfiXN16jThUb56O7wrJ1xEXw@mail.gmail.com>
References: 
 <CANOjuGEBEJr+uTH=1WoA+7+n+xsMEs8S9uEauHisE8tyi30mVQ@mail.gmail.com>
 <CAGKR+ECATQiw21G-01Vs1Jx4nPykGdm4No7PL4dqXtpH+R=HOg@mail.gmail.com>
 <CANOjuGEq2Fb9+McDnar8eCN=fkNfiXN16jThUb56O7wrJ1xEXw@mail.gmail.com>
From: Joshua Marantz <jmarantz@google.com>
Date: Wed, 1 May 2013 10:15:20 -0400
Message-ID: 
 <CAGKR+EAdE6Q6YCD35MBQdKsvk0tR4v5hhujVGBbAWUCsucymuw@mail.gmail.com>
Subject: Re: Apache Buckets and Brigade
To: "modules-dev@httpd.apache.org" <modules-dev@httpd.apache.org>
Content-Type: multipart/alternative; boundary=089e01228a44b123f404dba8c05e

--089e01228a44b123f404dba8c05e
Content-Type: text/plain; charset=ISO-8859-1

On Wed, May 1, 2013 at 9:48 AM, Sindhi Sindhi <sindhi.for@gmail.com> wrote:

> I doubt if I can use ModPagespeedSubstitute, because
> our string replacement actually uses some business logic. For ex. in
> "oldString", if i find a "old" string at offset 0 i'll replace it with
>

I agree: that configuration-only solution I proposed wouldn't meet your
needs.  Before I give up on it completely, is your business logic going to
always be in C++?  I am just trying to avoid overconstraining the solution.
 For example, would Lua be a suitable vehicle for your business logic? See
https://httpd.apache.org/docs/trunk/mod/mod_lua.html.

I wonder if there's some way mod_lua & an enhanced mod_pagespeed could work
together to provide the substitution with rich business logic.  The reason
I ask is that building & hacking mod_pagespeed's source for your purposes
is a great way to get something working, but I don't have a great answer
for how you'd maintain that over time as mod_pagespeed updates roll out.


> The HTML-centric fetch of data as you mentioned suits the best for me. But
> I dont want mod_pagespeed to actually modify anything in my HTML file, if
> it can give me either the entire HTML file OR HTML-centric fetch of data
> that will solve my problem :)
>

I'm not sure I understand your concern.  mod_pagespeed will not change the
HTML file on disk.  It acts as an Apache output filter changing the bytes
of HTML as they stream out.  As the writer of a mod_pagespeed filter, you
get to interpose C++ handlers for HTML lexical tokens (including
Characters()) and mutate those tokens before they are serialized out to the
next filter.  Is that what you want?

-Josh

>

--089e01228a44b123f404dba8c05e--