Re: Re: Empty message apparently not tokenized

Jem wrote:
>>This is unquestionably a bug in the SP parser.  I just don't have the
>>time to fix it for a few days.  Sorry.  In the meantime if you are using
>>procmail you can add this recipe to eliminate these without even
>>requiring SP.  It's a workaround until I get the real parser fix in
>>place:
> 
> Interesting, I just came back to this list because after a while of leaving 
> my spamprobe setup alone I did try the latest version (still using hashes - 
> very fast). I encountered the exact same problem. It is just a trivial bug, 
> like you said inserting the blank line makes these 'turdlets' properly 
> scored.

Yes I had an if statement in the wrong place.  It immediately stopped 
processing a message if it had an empty body but it should have done so 
AFTER processing the headers.  1.4b fixes that.


> No other bugs that I can see in v1.4a using Hashed-array-cached database.
> 
> On a different note, I have been using the hash approach and have found 
> over many months that even with pretty small hashes (default size, 16 MB) I 
> am getting very good results.

Cool!  What sort of cleanup do you perform on these small hash files? 
I'd like to add a bit more about the hash format in the readme and 
cleanup is an important factor with that format.

All the best,
++Brian


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642