Re: STORAGE_ERROR : EXCEPTION_STACK_OVERFLOW

["andrew.carroll@okstate.edu" <andrew.carroll@okstate.edu>]

Ahhhhh, it's like a cool breeze on a hot day.  Finally, someone else
gets the results I do.
I've tried several different configurations of from_disc.  I always
seem to get the results you did.  Also, I moved table declarations
around in the tables.txt file to see if I could narrow it down to one
attribute type.  Doing so gives me the same error on different
attribute types.

At first I thought it might be the to_disc function was writing "bad
bytes" to the table file but I have not thought of a way to tell.
This would cause from_disc to fail.  I also tried moving table
declarations around in the tables.txt file and it gives mixed
results.  The date type you mentioned, if you move the declaration to
a different spot in the file, it will read it in fine.  One of the
other attribute types will then be broken.

Hence, I think there are several problems:
1.  somewhere I have a a pointer pointing to something it shouldn't.
It's possibly sharing an attribute in the schema.attributes array.
2.  Maybe?  It could be that the file stream is not an in out
parameter.
3.  It is not dispatching properly.

My goal is to get comp.lang.ada's thoughts.


Thanks.



On Apr 3, 7:49 pm, Georg Bauhaus <bauh...@futureapps.de> wrote:
> On Tue, 2007-04-03 at 20:43 +0100, Simon Wright wrote:
> >               ---------------------------------------------------
> >               Type one of the following at the prompt:
>
> >               ~  QUIT
> >               1  INSERT DATA
> >               2  UPDATE DATA
> >               3  DELETE DATA
> >               4  SHOW RECORDS
> >               For help type 'help'
> >               ---------------------------------------------------
> >               >>1
>
> Erh, got there, finally.
>
> Here is one observation, and a total lack of explanation: In function
> from_disc(...dateattribute), variable "temp", to be used as a
> pointer to the result value, is assigned a pointer to a
> newly allocated dateattribute. On my computer, this pointer isn't
> quite as I would have expected. In GDB, I get STORAGE_ERROR each time
> I try to print a field of the object (temp has an address), even
> though the contents at the address seem plausible by what (gdb) x/60c
> is showing me... It takes a few seconds until gdb dumps core after
> a seg fault in this case.
>
> (A temperature drop of some 15°C rel to yesterday is probably not be
> the only thing that is affecting me, so I'll stop being confused for
> today.)