Re: 6.1 issue

On 04/10/2009, at 3:01 AM, Erik Olsson wrote:

After being unable to get a successful compile on a 6.1 machine due to ccsid
issues I used trigraphs for curly braces etc. The compilation then went
fine, but...

I have the following code.

fb = _Rreadf(fp, &buf, sizeof(buf), __NO_LOCK);
if (fb->num_bytes != sizeof(buf))
??<
cout << "That did not work" << endl;
_Rclose(fp);
return -1;
??>

The bizarre thing is that the if statement always evaluates to true (even if
the read finds a record, and even if i change the comparison to "==". That
truly defies the rules of logic).
Other if statements in the same program work as expected. It's created with
CRTBNDCPP.
I have tried the program successfully on a V5R4 machine.

Guess it's time to check this with IBM?

I think the exclamation mark is a variant character. Depending on your source file CCSID you may need to use an alternative token (e.g., not_eq). See page 12 of the VRM610 ILE C/C++ Language Reference.

Just some other thoughts:

What is the value of fb=>num_bytes after a successful read?

What does errno contain when the line of code fails?

Is the code above a direct copy/paste of the real code or do you have a dangling semicolon after the if test?

Did you run the identical 610 code on 540 or did you re-enter it?

Is the source file CCSID and run time CCSID the same on 540 and 610?

Is the run-time language ID the same on 540 and 610?

Did you try compiling for 540 on the 610 system? Did it make a difference?

Regards,
Simon Coulter.
--------------------------------------------------------------------
FlyByNight Software OS/400, i5/OS Technical Specialists

http://www.flybynight.com.au/
Phone: +61 2 6657 8251 Mobile: +61 0411 091 400 /"\
Fax: +61 2 6657 8251 \ /
X
ASCII Ribbon campaign against HTML E-Mail / \
--------------------------------------------------------------------