DCOM and Google's Summer of Code

[Ivan Pechorin]

Hello!

I'm in trouble trying to estimate effort needed to implement
marshalling interoperable with MS.
Is it a right task for 2-2.5 months of Google's "Summer of Code"?

Best regards,
Ivan Pechorin.

> Hello!
>
> I'm in trouble trying to estimate effort needed to implement
> marshalling interoperable with MS.
> Is it a right task for 2-2.5 months of Google's "Summer of Code"?

I think DCOM is way above student level.

There might be some other tasks.

Ciao, Marcus

[Ivan Pechorin]

> > I'm in trouble trying to estimate effort needed to implement
> > marshalling interoperable with MS.
> > Is it a right task for 2-2.5 months of Google's "Summer of Code"?
>
> I think DCOM is way above student level.
> There might be some other tasks.

Marcus, thank you for realistic estimation. I'm interested to work on
Wine's DCOM in a long term. Please, could you suggest some
intermediate tasks?

[Mike Hearn]

On Wed, 01 Jun 2005 23:40:29 +0400, Ivan Pechorin wrote:
> Marcus, thank you for realistic estimation. I'm interested to work on
> Wine's DCOM in a long term. Please, could you suggest some
> intermediate tasks?

There's a page we're writing here:

http://wiki.winehq.com/SummerOfCode

It might prove interesting.

DCOM is stupidly hard, although that said I'm a student myself and Rob
only recently stopped being a student. Raw RPC is a bit easier but still ...

I think the ASP or theming tasks might be fun. Alternatively, maybe some
Direct3D work?

[Ivan Pechorin]

> DCOM is stupidly hard, although that said I'm a student myself and Rob
> only recently stopped being a student. Raw RPC is a bit easier but still ...
>
> I think the ASP or theming tasks might be fun. Alternatively, maybe some
> Direct3D work?

Mike, I need Wine's DCOM implemented to port some complex app to
Linux. Google's offer is just a reason to start working on it a bit
earlier than I planned.

[Mike Hearn]

On Thu, 02 Jun 2005 00:26:14 +0400, Ivan Pechorin wrote:
> Mike, I need Wine's DCOM implemented to port some complex app to
> Linux. Google's offer is just a reason to start working on it a bit
> earlier than I planned.

Ah, I see. If you are definitely going to work on it, then maybe Rob and
myself should create a wiki page with all the work that needs to be done.

At the moment much work remains to be done. For instance:

* Make the RPC runtime compatible with Microsofts marshalling format/DCE
 * Possibly implement the necessary parts of widl
 * Implement the OXID resolver, a proper endpoint mapper, maybe a proper
   ROT
 * Implement complex/simple pinging
 * Maybe implement IRemoteActivation/ISystemActivator

You may wish to work with or contract CodeWeavers to help you with this
project.

thanks -mike

[Robert Shearman]

Ivan Pechorin wrote:
> Hello!
>
> I'm in trouble trying to estimate effort needed to implement
> marshalling interoperable with MS.
> Is it a right task for 2-2.5 months of Google's "Summer of Code"?

DCOM is a much larger project requiring some degree of infrastructure 
work that would be difficult for a person new to the Wine codebase and 
new to COM.
So DCOM is probably out of reach, but the underlying protocol, DCE/RPC 
is probably right for 2-2.5 months of work. The work will involve 
parsing MIDL format strings and marshaling/unmarshaling data correctly 
to/from a buffer. The project should be easier for a new, but bright 
coder who hasn't yet built up a knowledge of Win32 / POSIX APIs than 
many other projects. I could provide a test program that exercises and 
certain number of Ndr types and say "make this work". We could vary the 
number of types to vary the amount of work.

[Ivan Pechorin]

> > I'm in trouble trying to estimate effort needed to implement
> > marshalling interoperable with MS.
> > Is it a right task for 2-2.5 months of Google's "Summer of Code"?
> DCOM is a much larger project requiring some degree of infrastructure
> work that would be difficult for a person new to the Wine codebase and
> new to COM.
> So DCOM is probably out of reach, but the underlying protocol, DCE/RPC
> is probably right for 2-2.5 months of work.

Robert, thank you for clarification.

I'm interested in Wine's DCOM in a long term because I have a big
task: I need to port to Linux some complex commercial middleware and a
couple of apps based on it. These apps employ DCOM heavily.
> The work will involve
> parsing MIDL format strings and marshaling/unmarshaling data correctly
> to/from a buffer. The project should be easier for a new, but bright
> coder who hasn't yet built up a knowledge of Win32 / POSIX APIs than
> many other projects.
> I could provide a test program that exercises and
> certain number of Ndr types and say "make this work". We could vary the
> number of types to vary the amount of work.

It will be very helpful if you provide this test program.

Robert, thank you for accurate and clear definition of an intermediate goal.

[Jakob Eriksson]

Ivan Pechorin wrote:
> Hello!
>
> I'm in trouble trying to estimate effort needed to implement
> marshalling interoperable with MS.
> Is it a right task for 2-2.5 months of Google's "Summer of Code"?

I think it's much more work than that.

//Jakob