PROPOSAL: Tcl equivalent to .NET interop
Nick Hounsome
in .NET just to get the .NET GUI libraries.
One thing that I do find useful when programming in C# is the .NET
interop capabilities which allow you to call dll functions directly
without having to get into the unpleasantness of C++ on Windows or
writing your own wrapper stuff.
In .NET the interop is based on a comprehensive byte code generation
capability but nothing that elaborate is actually required for this
purpose - just a specification of the input and output C types.
Actually making the calls requires either some assembler :-( or
perhaps pretending that we are calling something with signature
f(int*,int*,int*,int*....? how many? ) :-(
structure handling would be quite easily dealt with by mapping a
structure description to a format for the binary command.
If the specifications could be made so that there was a "simpleish"
transform from the C# spec then we could leverage the existing
knowledge in the pinvoke wiki at http://www.pinvoke.net to call
windows stuff
Hopefully somebody with Windows expertise will be inspired by this and
make it happen - Personally I've always used C# or Tcl on Windows
precisely because it is so difficult to program with anything else.
Aric Bills
discussion, but does FFIDL provide (some of) the functionality you're
requesting?
Nick Hounsome
Yes - that seems to be going in the right direction although it seems
that it is a work in progress especially for Windows which is where it
is most needed e.g. library lookup is no good. (On Linux critcl is
probably the easy way to go).
Darren New
> although it seems that it is a work in progress
It's funny how everything is either a work in progress or abandoned and
hasn't been updated in years. Anyone else notice that? :-)
--
Darren New / San Diego, CA, USA (PST)
His kernel fu is strong.
He studied at the Shao Linux Temple.
Robert Heller
>
> Nick Hounsome wrote:
> > although it seems that it is a work in progress
>
> It's funny how everything is either a work in progress or abandoned and
> hasn't been updated in years. Anyone else notice that? :-)
Most of the universe can be considered either 'work in progress' or
'abandoned and hasn't been updated in [billions of] years', depending on
which of the various comosological theorys that are out there...
>
--
Robert Heller -- 978-544-6933
Deepwoods Software -- Linux Installation and Administration
http://www.deepsoft.com/ -- Web Hosting, with CGI and Database
hel...@deepsoft.com -- Contract Programming: C/C++, Tcl/Tk
Jeff Hobbs
> Nick Hounsome wrote:
>> although it seems that it is a work in progress
>
> It's funny how everything is either a work in progress or abandoned and
> hasn't been updated in years. Anyone else notice that? :-)
Isn't any work either by definition a work in progress or abandoned? To
release patches for something, work is in progress. If you refuse to
work on it anymore ... it is abandoned.
Jeff
Larry W. Virden
I often hear people refer to abandoned work as "stable" work. To me,
it seems that if a software package needs no improvements, fixes,
documentation enhancements or corrections, then the software probably
isn't very complex, or isn't being used very much.
Donal K. Fellows
> I often hear people refer to abandoned work as "stable" work. To me,
> it seems that if a software package needs no improvements, fixes,
> documentation enhancements or corrections, then the software probably
> isn't very complex, or isn't being used very much.
There is virtue in not being very complex; it means that the software
can actually be fixed to the point where it is correct.
Donal.