Mailing-List: contact new-httpd-help@apache.org; run by ezmlm
Precedence: bulk
Reply-To: new-httpd@apache.org
Date: Sat, 27 May 2000 15:14:58 -0700 (PDT)
From: Greg Stein <gstein@lyra.org>
To: new-httpd@apache.org
Subject: moving globals (was: RE: cvs commit: ...)
In-Reply-To: <000701bfc825$d3a42b80$345985d0@corecomm.net>
Message-ID: <Pine.LNX.4.10.10005271512140.32359-100000@nebula.lyra.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Sat, 27 May 2000, William A. Rowe, Jr. wrote:
>...
> > >     This patch also moves the following data from http_main 
> > to http_config:
> > >   
> > >       const char *ap_server_argv0;
> > >       const char *ap_server_root;
> > >       ap_array_header_t *ap_server_pre_read_config;
> > >       ap_array_header_t *ap_server_post_read_config;
> > >       ap_array_header_t *ap_server_config_defines;
> > 
> > Why?
> 
> Unless someone has STRONG feelings, I am looking to isolate the code
> module http_main.c to allow it to link into a seperate binary.  These
> args are the only thing in my way.  In this, the Win32 implementation
> becomes Apache.exe (of http_main.c alone) dynamically linked to all
> the remaining core server code.

This is premised on http_main.c being able to link into a separate binary.
I think that you might run into problems over time because that is not
apparent in the source organization. People may end up introducing new
things to http_main.c that break this.

I don't have a particular solution to that, but it is a concern.

In any case, the move of the variables seems okay. Globals are bunk in the
first place, so whereever they go, they're still bogus :-)

Cheers,
-g

-- 
Greg Stein, http://www.lyra.org/