bogoutil -m STILL not removing "singletons"?
David Relson
relson at osagesoftware.com
Tue Sep 9 20:09:18 CEST 2003
On Tue, 9 Sep 2003 10:14:25 -0700 (PDT)
Benji Tittle <benji at tittle.net> wrote:
>
> Armed with a better understanding of bogoutil (thanks, Chris Wilkes),
> I've tried this again... but I'm STILL don't seem to be getting
> results.
>
> Here's the new sequence of commands & output. I started with a
> database freshly rebuilt from my corpora. Single wordlist.db file,
> 8224768 bytes.
>
> $ bogoutil -d ./wordlist.db | wc -l
> 224841
> $ bogoutil -m ./wordlist.db -c1
> (c_get): Successful return: 0
> $ bogoutil -d ./wordlist.db | wc -l
> 224840
>
> I double-count my ham, so there should be no ham singletons, but I'm
> having trouble believing that I had only ONE singleton in my entire
> spam corpus!
>
> Size was unchanged at first. I then compacted the database with:
> $ bogoutil -d ./wordlist.db | bogoutil -l wordlist.db.new
>
> After compaction the db was 8220672... a reduction of only 4k. What
> does the database compaction actually do, anyway? Because I *did*
> compact the database before doing any of this. Is "singleton removal"
> an
> undocumented feature of a "-d | -l" compaction?
>
> I should mention that the bogoutil -m command returns the "(c_get)"
> line instantly... i.e. it doesn't seem like it's actually doing
> anthing.
Benji,
This time you're right :-)
It appears that when the database API was revised a month or so back,
some of the maintenance code got broken. There appears to be a
conflicting use of return codes in some of the routines.
I'll see what I can do.
David
--
David Relson Osage Software Systems, Inc.
relson at osagesoftware.com Ann Arbor, MI 48103
www.osagesoftware.com tel: 734.821.8800
More information about the Bogofilter
mailing list