directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Emmanuel Lecharny <>
Subject Re: [IMPORTANT] Status of Apache Directory
Date Thu, 02 Nov 2006 22:45:35 GMT
Alex Karasulu a écrit :

> Hi all,

Hi Alex,

> Lately some things have been slowing down and I just want to see were 
> we are and who is working on what.  Who is still around :)?

Well, I guess this is quite normal after a release, as everybody was 
quite under pressure and need to recover a little bit :) (at least, this 
was true for me)

> With the recent emergence of MINA from the Directory TLP to be its own 
> TLP I think it's a good time to recap on what's going on and the 
> current state of affairs.
> Summary:
> 1). New Code Imports
> 2). Documentation and Site
> 3). 1.5.x Branch Roadmap
> 4). OSGi Plan
> 5). Community Growth
> ================
> New Code Imports
> ================
> Mitosis
> -------
> We just imported mitosis into the ADS 1.5.x branch.  Mitosis is a 
> multimaster replication module that Trustin Lee wrote for ApacheDS a 
> while back while working with my company Solarsis Group.  Solarsis 
> contributed the source to the ASF and all IP has been cleared.
> We imported this code but we're working on better integrating it.  It 
> now compiles and we're trying to test it to see if it still works 
> after the 1.0 release.
> We're working with Trustin to transfer as much of the knowledge he has 
> to others on the project currently so others can work in it besides 
> just Trustin, Emmanuel, and I.
> Mitosis needs lots of work. Refactoring and testing under heavy load.
> So volunteers are welcome!

I volunteer, but I won't lead the effort ;)

> Triplesec
> ---------
> An identity management solution built on top of ApacheDS.  Right now 
> tsec has had all documentation and IP cleared for import and it's ready.
> I just have not had the time to move it over and change the package 
> names.
> Not only this but I have to revamp the schema which is a mess and make 
> sure we use Apache Enterprise numbers instead of safehaus ones.  So 
> there is a lot of work that needs to be done here.

I volunteer to, but I must admit that it won't be my first priority. The 
main problem with Triplesec is that it's a plain new piece of software, 
and we need to understand how it works. However, I already changed all 
the headers, and I think that a full code review (I mean, check for 
javadoc, coding rules respect, etc is the best solution to jump into the 

> ==========================
> Documentation and the Site
> ==========================
> We're in really sad shape here.  Almost everything is well out of date 
> and we need to revamp the site to include all the documentation that 
> we have in the ASF confluence which now has all the stuff which was at 
> Safehaus.
> We need someone to take the lead here.  Ersin? Stefan?

Yeah, we badly need someone to take the lead on this ! Men, the site and 
doco sucks ... (ok, may be not that much, but we have bits into xml 
files, bits in moin-moin, bit on safehaus, bits on cwiki, we need to 
reorganize the whole stuff). At this point, a book may be a good idea, 
because not only it will be a great thing to have, but also because it 
will bring some structuration to our doco...

> ====================
> 1.5.x Branch Roadmap
> ====================
> a) refactor
>     i) cleanup jndi code
>     ii) cleanup intercepted methods
>     iii) introduce new request context object
>     iv) introduce a session object
>     v) cleanup misuse of next interceptor and nexus and use 
> the         proper bypass instructions
>     vi) hierarchical partitioning

I will add "inclusion of dynamic schema", using the code provided by 
Norval. This is really painfull to use a plugin to add a new schema ...
The codec need to be reviewed too.

> b) features
>     i) Triggers
>     ii) Better sandbox (security) for stored procedures
>     iii) SASL
>     iv) DSML Service
>     v) Web Interface
>     vi) Transactions
>     vii) Quartz integration

And LdapStudio (pam is working on it), JMX -if we don't have OSGI on the 
first place -,better logging etc...

> Refactor
> --------
> We need a clear roadmap to be defined for the 1.5.x developmental 
> branch.  I agree with Trustin that we need to seriously refactor some 
> nasty JNDI code out of the core to clean it up.

Ohhhh yeah !

> First I think we should refactor the code to better organize it.  Then 
> we need to think about whether or not we're going to cleanup 
> intercepted methods and introduce a session and request (Context) 
> objects instead of each interceptor method having a bunch of arguments.
> In some places we use the next interceptor or the nexus directly 
> instead of using the proper bypass instructions.  With mitosis this 
> will be a problem so we need to cleanup and use this code properly.
> We need to make it so several nexus' can exist and that the nexus 
> itself can store entries.

The first step would be to have a betyter doc. I have started a cwiki 
page about interceptors, but it's far from being finished ...

> Features
> --------
> Next we need to think about features that are to be added and 
> completed in this branch.  Some things have already started but are 
> half complete like the trigger implementation.  We need a time table 
> for that.  We also need a good sandbox for code executed as stored 
> procedures.
> I talked to Trustin about implementing SASL and we think we can shoot 
> for implementing this feature soon after some refactoring of the core.
> Looks like PAM Is working on some DSML code and he would like to 
> integrate it into the core.  PAM please start working in the sandbox 
> with your code right before attempting this merge.
> For DSML and a web based configuration console both at start up for 
> the initial configs and on going configuration we can embed Jetty and 
> have a web console like the one that is in Tsec.
> Transactions anyone? These are not transactions across multiple LDAP 
> operations.  These are transactions that make it so we can recover 
> easily on abrupt shutdowns and to make all operations atomic.  This is 
> a serious issue with triggers.
> We need quartz integration and should pull it out of the mitosis code. 
> We need to do certain time triggered tasks and it makes sense to have it.

We also need a memory partition, and a reorganization of the current 
backend structure. We have a lot of gain to expect from this part. We 
may need to review the way we serialize data into the backend, it's far 
from being optimal.

> =========
> OSGi Plan
> =========
> I have no idea what is going on with the OSGi work.  It seems like 
> John Conlon is the only one left working on it now.
> There are no updates to the community and very little knowledge 
> sharing about this.  We need more status on this if anything.

Who is leading this OSGI stuff ?

> ================
> Community Growth
> ================
> We're still a small community even though we just added 7-8 new 
> committers.  Not many stuck around it seems but it may be too early to 
> tell.
> I'd like to see an effort to increase the number of active committers 
> on this project.  Everyone should help think about how we can do this.
> More people are welcome as usual.

yeah ! We will do our best to help newcommer. Any mail will be quickly 
answered (yes, it will !) and as said somebody in USA (don't remember 
who, gosh ... ;)  : "No one will be left behind".


View raw message