Bug 4530 - Make SDBM default bayes store
Summary: Make SDBM default bayes store
Status: RESOLVED WONTFIX
Alias: None
Product: Spamassassin
Classification: Unclassified
Component: Plugins (show other bugs)
Version: SVN Trunk (Latest Devel Version)
Hardware: All All
: P5 enhancement
Target Milestone: Future
Assignee: SpamAssassin Developer Mailing List
URL: http://wiki.apache.org/spamassassin/B...
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2005-08-12 20:13 UTC by Duncan Findlay
Modified: 2019-07-30 18:15 UTC (History)
3 users (show)


Attachment Type Modified Status Actions Submitter/CLA Status
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Duncan Findlay 2005-08-12 20:13:05 UTC 
Given that SDBM is *way* faster than DBM, and is even faster than SQL, we should
use SDBM by default.

The only reason not to at the moment is that it screws up existing installs. It
would be nice to have some way of automatically converting or something, but I'm
not sure how this can be done.
Comment 1 Michael Parker 2005-08-12 20:51:25 UTC 
Subject: Re:   New: Make SDBM default bayes store

Suggest we hold on this for the moment.  SDBM has not undergone any
serious testing in a production environment.

It might be the fastest because it's doing everything wrong.

Michael
Comment 2 Justin Mason 2006-09-05 14:48:46 UTC 
I think we should probably do this for 3.2.0.  I've been using SDBM_File on my
scanners for several months now without noticing any bad side effects...
Comment 3 Justin Mason 2006-11-02 15:29:48 UTC 
actually, I heard someone mentioning problems with reading data from an SDBM db
while another process is writing; this works in DB_File and is pretty critical
for SA, so perhaps we should hold off on SDBM use until that's investigated.
Comment 4 Theo Van Dinter 2006-11-02 15:44:48 UTC 
fwiw, DB_File has problems with reads during a write, though it could be
specific to reading the data that's being changed.  See bug 3563.
Comment 5 Justin Mason 2006-12-05 05:53:51 UTC 
moving off to 3.3.0
Comment 6 Justin Mason 2009-06-29 04:28:42 UTC 
is this viable yet?
Comment 7 Justin Mason 2010-01-27 02:20:23 UTC 
moving most remaining 3.3.0 bugs to 3.3.1 milestone
Comment 8 Justin Mason 2010-01-27 03:16:15 UTC 
reassigning, too
Comment 9 Henrik Krohns 2019-07-30 17:32:42 UTC 
Closing old stale bug.

We should really make Redis the default, but why bother, let user configure it..
Comment 10 Benny Pedersen 2019-07-30 18:15:47 UTC 
we should not really depend on specific software, if anything should be defult it should be sqlite, just my one euro

extend sa-awl to support --backup --restore --clear just like sa-learn

that would not be a must but really nice if thats not a problem to migrade to redis :)