RE: [Gcl-devel] Various questions about GCL

[Goffioul Michael]

 

1) Would it be possible to make GCL embedded into a C library (preferably shared) such that

a 3rd party application could make use of GCL as an embedded computation engine? This

would mean to have an API to initialize/terminate the GCL engine, to convert between C types

and LISP objects, to "eval()" or "funcall()" arbitrary code. Our final target is actually to embed

a LISP engine into a MEX file such that our framework written in LISP can be accessed from

the MATLAB user interface. 

 

As you may have already discovered, GCL is not set up to do this.  My guess is that if compiler::link were to be altered appropriately and the necessary initialise/terminate functions written, it would be possible to produce a "GCL in a box" in the form of a DLL or Unix shared library.

Would anyone be interested to do it? (I guess it's mainly a matter of cut-n-paste from the

"main" function). 

 

I don't know how MATLAB and MEX work, but if the DLL had to be a COM server rather than a simple C based DLL on Windows, then things might (or might not) be much tougher; I have a faint memory that COM servers have to be reentrant.   If COM is needed, you should seriously consider Corman Common Lisp which is supplied as an in-process COM server.

No, it simply has to be a normal DLL linked against MATLAB libraries. 

 

2) Is there any FFI framework?

 

Yes, but compared to other functional programming language implementations it is very primitive.  Examples may be seen in the xlib, opengl, JAPI and MPI extensions to GCL.

Is there any documentation about this? (I saw in gcl_japi.lsp that it uses "defentry", but I couldn't

find any doc about it; I couldn't find the opengl or MPI extensions in the code, and "make" in

xgcl-2 fails...). Can it be used to define C-functions located in a shared library (to be loaded at

run-time)? 

 

3) (make-hash-table :test #'equalp) is not supported

I'm not an ANSI expert but I assume we will add that if it is part of the spec, or if you really need it.

I don't know if it's ANSI, but it's in the doc. See (describe 'make-hash-table). 

 

4) Code like the following does not compile:

(unless (find-package "MYPKG")

  (defpackage "MYPKG" (:use "COMMON-LISP")))

(in-package "MYPKG")

(export 'f1)

(defun f1 (x) (* x x))

 

It complains at "in-package" statement. If I enclose the first 2 top-level forms (unless, in-package)

into an "(eval-when (compile load eval) ...)" form, it compiles but doesn't load (after restarting GCL),

still complaining at "in-package" statement.

(Note: I'm using anonymous CVS version under Linux)

You're probably talking about the warning ";; Warning: The package operation (IN-PACKAGE "MYPKG") was in a bad place." shown below.  This is  a long standing problem and is, I believe something we mean to address during the move towards releasing version 2.7.0:

Yes, that's the error I get. 

5) Would GCL compile/work under Win32/MSVC6?

I would be pleased to help you if you went ahead and tried this.  I would suggest simply building under MSYS for the convenience of the build system, using the VC6 compiler and linker instead of gcc/ld.  The alternative would be to write nmake compatible makefiles and run under a Windows command prompt, but you would then always be one step behind the main build system.

I can have a look when I boot Windows again.

 

Thanks for your answers.

Michael.