| Re: dlm question [message #32851] |
Wed, 13 November 2002 05:22  |
ronn
Messages: 123
Registered: April 1999
|
Senior Member |
|
|
Hi Eric,
The short answer is that RSI hasn't released the information on how to do
this.
The long answer is that with a little hacking you can figure out how to make
a method call a DLM, but not creating an IDL object. But the "self"
structure that gets passed to the C dlm method is not of much use. That is
you can't use it to call other methods and so on.
It has been while since I looked at this and you have made me curious again.
I am going to look at some old code that I have and if I come up with
anything I will let you know.
-Ronn
--
Ronn Kling
KRS, inc.
email: ronn@rlkling.com
"Application Development with IDL"� programming book updated for IDL5.5!
"Calling C from IDL, Using DLM's to extend your IDL code"!
"Power Graphics with IDL, A Beginner's Guide to Object Graphics", NEW BOOK!
http://www.rlkling.com/
in article PsaA9.5365$Vr1.772164@news20.bellglobal.com, Eric Fournier at
eric.fournier@videotron.ca wrote on 11/12/02 12:00 PM:
> I have a DLM (Dynamically Loadable Modules) question.
>
|
|
|
|
| Re: dlm question [message #32858 is a reply to message #32851] |
Tue, 12 November 2002 16:47  |
Rick Towler
Messages: 821
Registered: August 1998
|
Senior Member |
|
|
I asked this question (in a more obtuse form) a couple of years ago and
received no response. I understand the desire to do this although I would
think that for most applications you could just wrap your dlm routines
within a standard IDL object. This approach certainly isn't as cool as
writing your object as a .dlm but I would guess that your time could be
better spent elsewhere.
But since you piqued my interest again, and your .dlm file gave me an idea,
this is what I have come up with.
First I think we should start with a simple c++ class. Here is the
specification file "test.h":
#include "export.h"
class Test
{
public:
IDL_VPTR IDL_CDECL Init(int argc, IDL_VPTR *argv);
private:
long aprop;
long bprop;
};
Just one public member, Init, and 2 private data members.
Next, we need an implementation file.
#define STRICT
#include <stdio.h>
#include <windows.h>
#include "test.h"
#include "export.h"
#define ARRLEN(arr) (sizeof(arr)/sizeof(arr[0]))
/* Test object Init function*/
IDL_VPTR IDL_CDECL Test::Init(int argc, IDL_VPTR *argv)
{
IDL_ENSURE_SCALAR(argv[0]);
IDL_ENSURE_SCALAR(argv[1]);
aprop = IDL_LongScalar(argv[0]);
bprop = IDL_LongScalar(argv[1]);
return IDL_GettmpLong(1);
}
int IDL_Load(void)
{
static IDL_SYSFUN_DEF2 function_addr[] = {
{(IDL_SYSRTN_GENERIC) Test::Init, "TEST::INIT", 2, 2,
IDL_SYSFUN_DEF_F_METHOD, 0},
};
return IDL_SysRtnAdd(function_addr, TRUE, ARRLEN(function_addr));
}
You will note I have used IDL_SYSFUN_DEF_F_METHOD in the IDL_SYSFUN_DEF2
struct definition. This pretty much has all been a giant guess based on my
limited experience with dlms/c/c++.
But.
I can't compile it. There is a problem with the funct_addr aspect of the
IDL_SYSFUN_DEF2 struct. Well, I don't know exactly where the problem is but
the compiler chokes if the function address is in a class::member form. The
error is:
objectTest.cpp(99) : error C2440: 'type cast' : cannot convert from '' to
'IDL_VARIABLE *(__cdecl *)(void)'
None of the functions with this name in scope match the target type
Any ideas?
-Rick
"Eric Fournier" <eric.fournier@videotron.ca> wrote
> I have a DLM (Dynamically Loadable Modules) question.
>
> I have create a dlm with system routine in the past, but does anyone know
> how to create
> objects and associated methods. I have looked into the idl_dataminer.dlm
and
> it
> declares:
>
> MODULE DATAMINER
> DESCRIPTION IDL DATAMINER support
> VERSION 1.0.1
> SOURCE Research Systems, Inc.
> BUILD_DATE AUG 28 2001
> CHECKSUM 378MWGKG
> FUNCTION DB_EXISTS 0 0
> FUNCTION DIALOG_DBCONNECT 1 1 KEYWORDS
> FUNCTION IDLDBDATABASE::GETDATASOURCES 1 1
> FUNCTION IDLDBDATABASE::GETTABLES 1 1
> FUNCTION IDLDBDATABASE::INIT 1 1
>
> I can see that DB_EXISTS a a standard function, but IDLDBDATABASE is an
> object on wich you can
> call obj_new('IDLDBDATABASE') in IDL. Does anyone know how to code such
> thing in the dll associated in the .dlm file.
>
> I know some mecanism exist, IDL_SYSFUN_DEF_F_METHOD is mentionned in the
> export.h.
>
> Any help would be greatly appreciated
>
>
> Eric Fournier
>
>
>
|
|
|
|
|
|
|
|
| Re: dlm question [message #32922 is a reply to message #32851] |
Wed, 20 November 2002 19:52 |
Randy Frank
Messages: 2
Registered: April 1999
|
Junior Member |
|
|
Eric,
Consider the IDL naming scheme for objects and consider the
mechanism for calling a C++ method from C in something like VC++
(take DirectX as an example). This should give you some idea.
Ronn is correct in that RSI has not released this information and
there is no direct way of doing this. That having been said,
with some creativity, one can write some accessor functions that
will get you 90% of the way there, along the lines of what Ronn
suggests (playing with the 'self' reference creatively) until RSI
exposes the necessary functions. I'm not sure what this buys you
over a object written as a .PRO file that happens to call out to
some 'C' functions in a DLM however (an approach that will not
break over time)...
Thanks.
ronn kling wrote:
> Hi Eric,
>
> The short answer is that RSI hasn't released the information on how to do
> this.
>
> The long answer is that with a little hacking you can figure out how to make
> a method call a DLM, but not creating an IDL object. But the "self"
> structure that gets passed to the C dlm method is not of much use. That is
> you can't use it to call other methods and so on.
>
> It has been while since I looked at this and you have made me curious again.
> I am going to look at some old code that I have and if I come up with
> anything I will let you know.
>
> -Ronn
>
>
> --
> Ronn Kling
> KRS, inc.
> email: ronn@rlkling.com
> "Application Development with IDL" programming book updated for IDL5.5!
> "Calling C from IDL, Using DLM's to extend your IDL code"!
> "Power Graphics with IDL, A Beginner's Guide to Object Graphics", NEW BOOK!
> http://www.rlkling.com/
>
>
>
> in article PsaA9.5365$Vr1.772164@news20.bellglobal.com, Eric Fournier at
> eric.fournier@videotron.ca wrote on 11/12/02 12:00 PM:
>
>
>> I have a DLM (Dynamically Loadable Modules) question.
>>
>
>
>
|
|
|
|
| Re: dlm question [message #33015 is a reply to message #32922] |
Thu, 21 November 2002 05:34 |
Randall Skelton
Messages: 169
Registered: October 2000
|
Senior Member |
|
|
On Thu, 21 Nov 2002, Randall Frank wrote:
> Consider the IDL naming scheme for objects and consider the
> mechanism for calling a C++ method from C in something like VC++
> (take DirectX as an example). This should give you some idea.
> Ronn is correct in that RSI has not released this information and
> there is no direct way of doing this. That having been said,
> with some creativity, one can write some accessor functions that
> will get you 90% of the way there, along the lines of what Ronn
> suggests (playing with the 'self' reference creatively) until RSI
> exposes the necessary functions. I'm not sure what this buys you
> over a object written as a .PRO file that happens to call out to
> some 'C' functions in a DLM however (an approach that will not
> break over time)...
I tend to agree-- I've been writing IDL object code that calls my C dlms
for the past year now. Last year I begged and pleaded with RSI to release
this interface (even with a non-disclosure agreement) and had no luck
getting it. I even did my bit to reverse engineer the interface which was
quite successful. I would post what I learned but fear Craig-like
repercussions. My main comment is that while gaining access to the object
API would be nice, the holly grail would involve stable, direct access
(via a C API) to the IDL heap variable. This, along with my request for
proper operator over-loading in IDL objects are my top two requests. Are
any of the lurkers at RSI listening?
Note to DF: Are you ready to compile a new top 10 list?
Cheers,
Randall
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
