httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Stein <gst...@lyra.org>
Subject Re: Theory of DAV
Date Thu, 21 Sep 2000 03:37:44 GMT
On Wed, Sep 20, 2000 at 08:59:56AM -0500, William A. Rowe, Jr. wrote:
>...
> > Note that two modules are needed: mod_dav and mod_dav_fs. 
> 
> Is there any reason not to build a single loadable binary of the both 
> modules?

I don't think so. It would certainly improve the linkage issue :-)

> > Before 2.0 ships, we will have (at least) a mod_dav_svn to provide a 
> > versioning, DAV-enabled back-end to Apache 2.0. (mod_dav_svn will 
> > ship as part of Subversion)
> 
> Cool (make that a single binary of all three modules?)

Like I said, it would ship separately as part of Subversion (SVN). You'd
have your mod_dav(_fs) DLL, and another mod_dav_svn DLL for the SVN stuff.

> Now the rough stuff...
> 
> The hooks module always declares all hooks are exported from the core.
> This is the current linkage.  That patch I got reamed over 'too many 
> extra arguments' to the hook functions?  We -need- it now, or something 
> like it.  mod_dav exports brand new hooks that are now a pain.

True, but you required a linkage spec in ALL of the declarations. That was a
bit silly :-) ... IMO, have a new extended macro which allows the linkage to
be specified; the current macro would continue to use the default/standard
linkage.

We could also compile DAV directly into the core (and I'd be happy with
that) but I'd doubt that is a concensual point of view :-)

> I'll throw up some proposal, if I can get it to build without revamping
> the declare hooks args, and post up an ApacheModuleDAV (+fs?) tonight.

Cool!

Cheers,
-g

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

Mime
View raw message