Mailing-List: contact apreq-dev-help@httpd.apache.org; run by ezmlm
Precedence: bulk
Received-SPF: pass (asf.osuosl.org: local policy)
Message-ID: <42D67E2A.1080605@vanderbilt.edu>
Date: Thu, 14 Jul 2005 10:00:58 -0500
From: Jeffrey Horner <jeff.horner@vanderbilt.edu>
User-Agent: Debian Thunderbird 1.0.2 (X11/20050602)
MIME-Version: 1.0
To: apreq-dev@httpd.apache.org
Subject: Re: On distributing libapreq2
References: <42D532F3.7030501@vanderbilt.edu>
 <87fyuiw0ra.fsf@gemini.sunstarsys.com>
In-Reply-To: <87fyuiw0ra.fsf@gemini.sunstarsys.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Joe Schaefer wrote:
> Jeffrey Horner <jeff.horner@vanderbilt.edu> writes:
[...]
>>I could distribute the whole source package with mod_R, build it and
>>install just libapreq2.os.*. This seems the simplest, but folks who
>>already have libapreq2 installed AND in use will run into problems.
> 
> I think this is the best route, even after httpd includes libapreq2.
> You could also include an --with-apreq2=/path/to/apreq2-config option 
> in mod_R's configure script, which would use the previously installed 
> version of libapreq2 and ignore the version you're actually shipping
> with mod_R.  Doing that would also help us improve our apreq2-config.in 
> script.
> 

Okay, I'll do it, but my knowledge of autoconf is low. Do you happen to 
have some pointers to docs or such?

Essentially this is what I have to configure my module:

configure --with-apxs=/path/to/apxs --with-R=/path/to/R

Now, if users want to just use the version of libapreq2 that I 
distribute, then configure will have to catch that and make appropriate 
decisions. But if I add --with-apreq2=/path/to/apreq2-config and the 
user chooses it, then configure will have to make appropriate decisions 
based on apreq2-config.

autoconf is a nightmare.

Also, what happens when an executable calls a shared library call? Does 
the entire shared library get loaded into memory, shared memory? If you 
have some pointers to docs on this, I'd be much obliged.

-- 
Jeffrey Horner       Computer Systems Analyst         School of Medicine
615-322-8606         Department of Biostatistics   Vanderbilt University