List-directed I/O error on Windows 2000

[Intel_C_Intel]

We have a program that ran correctly on NT4.0SP6 that crashes on Windows 2000 when executing the following I/O statement:

READ(cos_line(cind+1:cend-1),*) project_id

Where project_id has been defined as INTEGER(4).

This is not an urgent problem as we have solved it by using:

READ(cos_line(cind+1:cend-1),'(I)') project_id

This avoids the disassembly

call _for_read_int_lis

that was present in the original. Project settings and options are identical between the project as built on W2K and NT, so there may be some kind of disconnect with MSVCRT or something else between the operating systems.

Just thought COMPAQ (or INTEL?) might want to know.

Thanks

John

[Steven_L_Intel1]

John,

We are not aware of any OS-related differences here. What was the exact text of the error message? If you can still reproduce the problem on W2K, please send us a zip archive of your project to vf-support@compaq.com

Steve

[Intel_C_Intel]

Steve

Sorry to have been away so long. When I get some time, I'll try to duplicate the error on a more constrained setting - the original one uses Oracle SQL data and is too big to zip comfortably. the variable "cos_line" is retrieved from the server and decomposed into its constituent variables with internal file reads. The only one that seemed to have problems was the one I mentioned.

Since we have a workaround and enough alligators to go around right now, it may take me a while to get back to this; promise I won't forget, though.

Thanks

[Intel_C_Intel]

Hi John

Thanks for the useful posting that mirrors a similar error on Windows XP (without a straightforward workaround so far). In my case I'd wondered if it was due to using an older version of VF 6.0

I wasn't sure from this whether you'd found any resolution. Confusingly, some of my earlier list directed internal reads worked okay, especially when using static libraries, so it took a while (and a severe deterioration in sanity) to isolate the problem.

Neil.

[Intel_C_Intel]

Neil:

Sorry this took a bit to answer. I haven't done any more work on the List-Directed I/O problem since my posting last August; we had a workaround and I became involved with another project.

The programmer I was helping was using CVF 6.0 at that time. The PC I used then had been upgraded to 6.1B; but the project was large and I never tried the code over on my system. Have you tried to isolate the problem code and feed it some "canned" data as input? If so, you can send in the sample and we can all play with it to see if an upgraded CVF helps (I'm currently on 6.5).

Let me know if I can be of any assistance.

John