V4RB 351 Var Char Error

[Robert Brenstein]

on 12/2/03 8:52 PM, Cindy Brown at cindy at kowhaiprogramming.com wrote:

>  Hi Ruslan,
>
>>  1) at least during last 2 years was not found any major bugs in VarChar as I
>>  remember.
>
>  I need to put this in perspective. We haven't had any real crashes with
>  Varchar or anything else over the last year or so. The databases have been
>  running really well. It's just been these last two weeks. Of course, the
>  database has been more intensively used as it is school report time. The
>  crashes have happened at six different sites. I have been able to sort them
>  out (by putting a replacement database in) but it just seems unusual that it
>  is happening all of a sudden. I am now using REALbasic 5.2 which is much
>  faster than 4.5 with the server sockets and I wonder if that has something
>  to do with it.

Cindy, Ruslan,

The description of 351 varchar troubles reads like deja vu for me. If 
you check the list archives for threads "351 again" and "351 again 
p.s." from a year ago, you will see that I had a database that worked 
fine for several months and then when users really started pounding 
at it (in preparation for the next semester), I got hit with 351's 
popping up randomly in different varchar fields. No crashes were 
involved whatsoever.

My problems went away when I switched to Valentina 1.9.6. However, 
there was nothing ever clearly identified as the cause or a fix. I am 
not truly sure whether the problem is fixed (kernel had some code 
changes for varchars) or I am just plainly lucky while the problem 
still lingers in the kernel. As I recall, the errors were occuring 
ONLY under the heavier usage on the live server. I ran a number of 
simulations of user activity (script-generated requests sent to a 
private server) but they never produced any errors.

Of course, the similarity of symptoms could be purely coincidental.

Robert Brenstein