Re: Problem with STk 3.99.4 on Solaris 2.6

From: Erick Gallesio <Erick.Gallesio_at_unice.fr>
Date: Thu, 9 Dec 1999 15:48:05 +0100 (CET)

Paul Anderson writes:
>
> Erick et al:
>
> I have a user with our STk application (based on 3.99.4) running
> on Solaris 2.6. He sees runtime errors in the application
> that I am finding impossible to reproduce with my machine, even
> if I try to make it as similar as possible to his.
>
> The symptoms differ from run to run. But they seem to fall
> into the following classes:
>
> 1. Messed-up stack or environment. The stack trace shows
> a call to a function, but the arguments to that function
> are wrong. That is, the stack trace shows that the function
> was called with several arguments - a text widget and a few frame
> widgets. However, in the text of the program there is only
> one ocurrence of a call to that function, and its arguments
> are a text widget, an instance of a class, and an integer.
>
> 2. Functions called mysteriously. A run of the program with
> a set of arguments that ought to preclude it from ever
> calling a certain function ends up calling that function
> anyway.
>
> The problem does not appear to be related to either the
> window manager or the X resources, as he can run it
> using an X server on his laptop and it still exhibits
> the same set of problems.
>
> Has anyone seen anything like this before?
>
Hopefully NO : :-)

> Does anyone have any clue as to why this might be happening?
>

It seems to be something which is GC releated. even if I have never
seen the behaviour you describe, my experience is that "when it weird,
it's the GC". I really don't like what you have describe
above. Perhaps to be sure that this is the GC, you could try to
change the number of nitials cells to a very big number and to very
little one and see if you have a difference. Tell me what it gives.


                -- Erick
Received on Thu Dec 09 1999 - 15:50:31 CET