Mailing-List: contact dev-help@directory.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Apache Directory Developers List" <dev@directory.apache.org>
Received-SPF: neutral (hermes.apache.org: local policy)
Subject: Re: Help with Apache ASN.1 grammar compilation
From: Emmanuel Lecharny <elecharny@apache.org>
To: Apache Directory Developers List <dev@directory.apache.org>
In-Reply-To: <42B33F95.3010609@ericsson.com>
References: <20050617065138.69769.qmail@web32604.mail.mud.yahoo.com>
	 <1118991806.755.25.camel@wkslx01.iktek.com>
	 <42B27EFF.1030309@bellsouth.net> <1119001113.8015.2.camel@portable>
	 <42B2E4DA.40601@bellsouth.net> <1119025069.8015.11.camel@portable>
	 <42B32359.5040107@bellsouth.net>
	 <1119037207.4995.5.camel@wkslx01.iktek.com> <42B33F95.3010609@ericsson.com>
Content-Type: text/plain; charset=utf-8
Organization: iktek
Date: Fri, 17 Jun 2005 23:49:51 +0200
Message-Id: <1119044991.7400.28.camel@wkslx01.iktek.com>
Mime-Version: 1.0
Content-Transfer-Encoding: 8bit

Hi Paul, 

> As someone who started investigating the project primarially because
> of the ASN.1 compiler, it is a bit misleading to have virtually
> nothing but the shell.  What would be of much more value (in my
> opinion) is a bit more description of the design goals, methods,
> implementation ideas, etc.  

Yep. I agree. The fact is that we needed an ASN.1 codec for LDAP grammar
first, and it derived to an ASN.1 compiler, which is a good idea.

Currently, The LDAP ASN.1 codec is being rewrited from scratch, which
will bring more ideas for a futur ASN.1 compiler. If you feel the
courage to browse the sandbox, you'll see the seed of the new codec,
which decode LDAP grammar (almost finished) and also SPNEGO grammar.
It's much more Compiler oriented, as it's based on a state-automaton
approach, with actions being called at each state. The next step will
probably to générate the state-automation automatically from the ASN.1
grammar, using the current engine to run the automaton. This is
obviously one of our target.

Another objective is to be able to generate different kinds of codec :
BER, DER, CER or PER (don't bet that iot could be done before the end if
this year !)

> Having some sort of information in the format of a wiki (similar to
> the ASN.1 wiki, but with more compiler-centric details) might also go
> a ways towards enticing additional contributions.

You are damn right ! We will try to dedicate some time soon on this
subject. The point is that writing a clear documentation is not a piece
of cake, especially for me - you may have noticed that I'm not an
english native ;( - and takes a LOT of time. Any contribution is
welcomed, even criticism, and we will do our best to deal with it. JIRA
is really a cool tool if you want to contribute.

> I've thought about looking into ways of contributing on this front,
> but I'm limited to when I'm not at work.

Any contribution is GOOD. Even limited in time. Reading doco, and found
erros in it (bad syntax, errors, improvment), reading the code, testing
the code, writing some benchmarks, or whatever limited tasks are really
really usefull. We don't have time, nobody does, so sharing those tasks
is a way to progress faster. Feel free to contribute ! You are welcome !
(btw, I'm also working, so my contribution is limited to nights and
week-end, as far as it does not kill my social life ... so it's pretty
much like between 11PM and 2AM !)

Emmanuel Lécharny