[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

OT: Introspector Future (was Re: [DotGNU]Re: My Favorite soapbox : XML

From: James Michael DuPont
Subject: OT: Introspector Future (was Re: [DotGNU]Re: My Favorite soapbox : XML linkage)
Date: Fri, 7 Jun 2002 11:44:37 -0700 (PDT)

This mail drifts a bit off topic, and I dont want to
continue using up the DotGnu bandwith about the
introspector project. The DotGnu list is supposed to
be for the discussion of the concrete implementation
and technical details of the project. 

Maybe at some far future point the introspector or
some form of it will be included, but we are very far
away from that right now.

I think that that the issue about the Compiler
interface is on topic, but we are moving towards the
issues of the introspector more and more.

This will be my last post about it for the next weeks.

I am still working on beefing up the Test Suite for
CSCC and want to get something published in the next
couple of weeks. That is at least something concrete
that is a direct contribution to the project.

If you are interested in what the plans are read the
notes below.

Thanks for you patience, 

--- Gopal V <address@hidden> wrote:
> If memory serves me right, James Michael DuPont
> wrote:
> > --- Gopal V <address@hidden> wrote:
> > > A company that spends so much time into gcc
> would
> > > have
> > > to at end release their derievative interfaces
> as
> > > GPL.
> > > This means an amount of wasted effort (in
> economics)
> > > 
> > > for the company.
> > No. Not with the CODEDOM interface from Microsoft
> or
> > the GXL language They
> are
> > interfaces that you can put the data into.
> GCC ? ... 

My point is that once you have a even GPLed interface
that extracts the information, the users of such an
interface are shielded. At least that is the argument
of the GXL users, they say that the software that uses
the GXL is older than the GPLed Gcc extractor.

After reading the statements from Rhys and Norbert, I
would say that I will avoid creating any Ast extractor
based on the cscc. 

I have decided that the next step for the introspector
project is to create a proper specification for all
the files and modules. Only that way will a proper
discussion be possible. 

I will remove the gcc as a source of AST data, and
make the input system much more flexible. 

After reviewing the GCC Asts in detail I have realised
that many other systems also supply such
meta-information. This is the interesting realisation.

The sources of information for the introspector could 
also be XML files, from which structure information
can be extracted. 

DTDs and XMLSchemas provide more than enough
interesting sources of information that can be used
instead of C Structure declarations.

The csdoc output would be a safe place to start for
working with pnet.

C# with Csdoc can be used for doing interface
specifications and from that we can generate all types
of wrapper code, for perl and web services.

Perl has a official B::Deparse module that allows for
access to the entire programs structure at run time.
That will provide enough information to get the
introspector running fine.

Postgres also supports the querying of meta-data and
supplies all types of data structures. This could also
be used as an input.

Finally, the gcc_xml ( project has
gotten an official blessing from RMS, because they do
not dump function bodies. This could also be used as a
source of information, also without problems.

The inclusion of VCG for graph layouts and Dia for UML
diagrams, as well as generation of SQL, XML, perl and
later also C# for the data model will provide enough
functionality with endangering any gnu projects.

A SQL database repository for all program meta-data
and annotations will give a flexible and powerful back
end to the system.

The system could also support a webservice/REST like 
architecture for the querying of metadata.

If you are interested in discussing this more, please
contact me or join the maililing list on


James Michael DuPont

Do You Yahoo!?
Yahoo! - Official partner of 2002 FIFA World Cup

reply via email to

[Prev in Thread] Current Thread [Next in Thread]