On Wed, 29 Apr 2009, Mindaugas Kavaliauskas wrote:

Hi,

>   FOR nI := nI + 1 TO snCount

The problem is cause by
   nI := nI + 1 => nI += 1
optimization. After converting the expressions we still try to access
original nI expression which does not longer exists.
I'll fix it.

> BTW, this sample GPF hbrun of 1.0.1 stable version.

It's access to unused (freed) memory so the bad final results are random.
GPF, infinite loop, etc.
If you are using Linux then recompile Harbour with -g GCC option and
then execute harbour inside valgrind redirecting stderr to file.
You will have quite good bug report for such code.

> I have more GPF's on 1.0.1 but they are not stable, but no GPF on
> current SVN.

If you have any other code example which exploits some problems with
current Harbour compiler code then please send them. It will greatly
help to fix the problem.
You can verify them using valgrind. It's not necessary to reproduce
GPF which can be repeatable only randomly.

best regards,
Przemek
_______________________________________________
Harbour mailing list
[email protected]
http://lists.harbour-project.org/mailman/listinfo/harbour

Reply via email to