current cvs segfault on Debian hppa

[Gyepi SAM]

On Thu, Oct 03, 2002 at 01:33:10PM -0700, Mark M. Hoffman wrote:
> I think Gyepi is already heading toward "the right way" (tm).

Indeed. This problem has now been solved.
The alignment issue was resolved by using separate buffers for allocating hash nodes and strings.
This was actually in the first iteration of the wordhash code, but I merged them in an attempt to optimize.

Knuth was right: "premature optimization is the root of all evil."

The segment violation was due to lack of 'acceptable' initialization of the user buffer.
I have now changed the interface of wordhash slightly to allow a pointer to an initializer function
to be passed in.

I am in the process of acquiring an account on an hppa machine to prevent a recurrence of such
problems.

-Gyepi 

For summay digest subscription: bogofilter-digest-subscribe at aotto.com