This is a continuation of the realloc() problem.
I have an example that can be compiled and run.
I would email it as an attachment to someone
who would look at it. Its too big to post.

Using gdb you can p *ptr and see the corruption of the first structure after a realloc() is called.
You can call realloc () multiple times after and all following structures and
data are correct. It only happens on the first realloc (). The value returned
by realloc () is a pointer to the first structure. The pointer value does
not change.

The work around was to calloc () two structures initally. Set a temporay pointer to stucture, to the second structure tptr = (ptr+1). This effectively bypasses the corrupted first structure. It loads compiles and does what it should do but the work around is a patch to cover up a real problem.
The code comiles and runs correctly without the patch on other
Unix machines?

What we are running is SUSE 11.with the compiler and libs that come with
the distribution.

Then why did you start a new thread?

For those who may have missed it, the starting thread is:
http://www.linuxquestions.org/questi...oblem.-713782/

I would look at it if you like, but only after you have answered my question from your other thread (I don't need to get a big email if my guess at the problem was already correct and you just ignored it).

that is consistent with the error I guessed from your first post.