[Kolab-devel] slapd.conf.template

[Thomas Spuhler]

On Sunday 17 January 2010 01:00:00 pm Richard Bos wrote:
> Op zondag 17 januari 2010 17:22:07 schreef Thomas Spuhler:
> > On Sunday 17 January 2010 09:33:14 am Richard Bos wrote:
> > > Op zondag 17 januari 2010 15:20:09 schreef Thomas Spuhler:
> > > > > Perhaps you should use back_monitor.la to load the monitor.  What
> > > > > happens when the reverse the order of the back_ modules, like this:
> > > > > moduleload     back_monitor
> > > > > moduleload     back_bdb
> > > > > moduleload     refint
> > > > > moduleload     unique
> > > >
> > > > It works when I drop the modulepath line
> > >
> > > What about the .la suffix?  Did you try it with that?  Like:
> > > moduleload     back_monitor.la
> >
> > yes, I did
But it did not make a difference
> > But when deleting the modulepath line it's not needed
>
> I'm confused now.  If you add the suffix '.la' it works, is that what you
> say? Than why not use it like that?  Hmm, I see; you still need to patch
> the file. So you either need to patch it by removing the modulepath line,
> or you need to patch the modulenames and add the suffix '.la'.
>
> What happens if you remove the modulepath and use the suffix '.la'? 
> Perhaps that is a generic solution for all?
>
> Otherwise the code should be changed to something like:
> @@@if ldapserver_moduleload@@@
> @@@if ldapserver_modulepath@@@
> modulepath     @ldapserver_modulepath@
> @@@endif@@@
> moduleload     back_bdb
> moduleload     back_monitor
> moduleload     refint
> moduleload     unique
> @@@endif@@@
OK will work on this
>
> I'm not sure, if nested @@@if's are possible.  If you investigate the
> slapd.conf further, you'll find more nested @@@if's so I think it works. 
> It might be necessary to replace more @@@if ldapserver_modulepath@@@ into
> @@@if ldapserver_moduleload@@@ if you want to activate this change.
>
> I actually find it better to use a separate variable that is used to enable
> module loading.  For example: if the module loading is disabled, the
> modulepath will now be preserved.  With the current code it will be altered
> to 'false'.

-- 
Thomas