memory access woes in bogofilter or db-4.0.14?

[David Relson]

At 09:25 PM 1/8/03, Graham Wilson wrote:

>On Wed, Jan 08, 2003 at 05:34:06PM +0100, Matthias Andree wrote:
> > IMHO, the db_setvalue can be killed and every code be switched to change
> > db_setvalue_and_* -- it's the db-storing that figures if the timestamp
> > needs to be saved.
>
>why dont we just call it db_setvalue (instead of db_setvalue_and_date)?

Graham,

At the moment, master designer that I am, I created two paths to the output 
capability.

One path is for database maintenance use, i.e. bogoutil.  This path is used 
by load_file() and sets token (string), count (int), and timestamp (int).

The other path is for bogofilter's register_words() function, or more 
precisely the db_increment() and db_decrement() functions.  These functions 
neither know nor care that the wordlists have timestamps.  So they use a 
path with token and count, but no timestamp.

This second set of function calls could be changed to use the 3 argument 
function and provide the default timestamp (already defined in bogofilter) 
as the additional argument.

When I wrote the code I thought it better to have the two 
functions.  Others seem to prefer a single function approach.  At the 
moment, it matters very little to me which way we go.

David