directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alex Karasulu <>
Subject Re: help building/extending ApacheDS
Date Mon, 26 Sep 2005 18:07:32 GMT
Thom Park wrote:

>Thanks for this!
>Can you tell me how the schema->java files is being done - is it a java
>routine that does the magic or is it a maven/jelly procedure. I ask as
>it would be nice to be able to do this outside of the maven environment
>and create a standalone utility that could read-in a schema file and
>generate a set of java source files.
It is a plugin that does this: the maven-directory-pluing written for 
just this purpose.  This jelly however is a thin wrapper around the code 
in the apacheds/trunk/plugin project so it could be used for such a 
standalone utility.  However in the future we may remove this means of 
extending the scheme since it is so cumbersome.

Let me give you some background ...

The present schema mechanism was greated for bootstrapping the system 
partition with a minimal set of schema elements.  However we never got 
around to building the schema subsystem.  Part of this was because we 
were waiting for a subentry implementation so we could implement a 
robust schema subsystem properly.  Now that we have subentries 
implemented we can do this properly.  Note that the majority of startup 
time is due to loading and running these schema element producing 
classes generated by the plugin.  We hope to eliminate this in release 
0.9.5 of the server with the proper schema subsystem design.

>That way , I won't need to build the apacheds sub-system all the time,
>and be able to generate and compile these files from the product binary
Well you don't need to rebuild all the time.  You can invoke the plugin 
on a schema in a separate project.  Take a look at how the plugin works 
by inspecting the apacheds/trunk/core/ file.  It shows 
how to incorporate a dot.schema file for generation.  Then what you can 
do is add the jar of your schema to the server.xml under the schema 
section.  This way you do not need to recompile ApacheDS core to run an 
altered schema.

>Perhaps not a big deal to a developer, but for my needs, working from
>the binary distribution rather than the full-blown source distribution
>is a signification streamlining of the whole operation
Yep I understand ... see if you can follow these hints to do that.  If 
not let us know.


View raw message