Re: IDL 5.5 call_external passing strings [message #27831] |
Tue, 06 November 2001 11:58 |
Paul Probert
Messages: 7 Registered: May 1995
|
Junior Member |
|
|
Mike,
Thanks much. Indeed the only place this is mentioned is in the
export.h
file. The actual documentation is wrong. The "what's new" file is all
marketing tripe, and the relnotes.txt says nothing about this. So the
"10 times faster than fortran" thing didn't work for me today.
Paul Probert
University of Wisconsin
Mike Fitzgibbon wrote:
>
> The problem is that RSI has once again redefined the IDL_STRING structure,
> this time to increase the maximum length. Now it's using a 32 bit integer.
> You will need to recompile your dll to use the external.h from IDL 5.5
> --
> --
> Mike Fitzgibbon MRFitz@ns.arizona.edu
> UofAz, LPL phone:(520)626-4791
> Systems Programmer, Pr. fax: (520)621-6783
|
|
|
Re: IDL 5.5 call_external passing strings [message #27866 is a reply to message #27831] |
Mon, 05 November 2001 15:03  |
mikef
Messages: 9 Registered: May 1998
|
Junior Member |
|
|
In article <3BE7178F.851EE28@facstaff.wisc.edu>,
Paul Probert <pprobert@facstaff.wisc.edu> wrote:
> Folks,
> Under IDL 5.5, if I call into my dll with call_external, passing a
> string
> into the C program as an IDL_STRING structure (as per the documentation,
> and which
> worked fine in 5.3) the pointer to the string, the "s" member of the
> structure,
> now points to invalid memory (address of 0x00000001 in my case), even
> though the slen
> member is right (21 bytes in my case). Is this a bug in 5.5, or what?
> Anybody else
> have this problem?
>
> Paul Probert
> University of Wisconsin
The problem is that RSI has once again redefined the IDL_STRING structure,
this time to increase the maximum length. Now it's using a 32 bit integer.
You will need to recompile your dll to use the external.h from IDL 5.5
--
--
Mike Fitzgibbon MRFitz@ns.arizona.edu
UofAz, LPL phone:(520)626-4791
Systems Programmer, Pr. fax: (520)621-6783
|
|
|