Deb package

David Relson relson at osagesoftware.com
Wed Jan 29 01:45:39 CET 2003


At 07:36 PM 1/28/03, Matthias Andree wrote:

>David Relson <relson at osagesoftware.com> writes:
>
> > I'm committing the patch to cvs because I'm sure it's needed.  Looks
> > like there'll be a 10.1.3-bug-fix ...
>
>How about using something along the lines of:
>
>#include <inttypes.h>
>/* ^ that's POSIX */
>
>use uint32_t for the database?

We _could_ do that, for some future release.

>32 bit integers will be a problem once token counts cross the billion
>(10^9) and approach 2^32, but to fix this for good, we'd have to store
>information about the "value" format in a special token in the data base
>itself (to allow queries if the DB is accessed from machines of
>different architecture, NFS for instance), and add an abstraction
>layer. Probably 0.12 stuff though.

I think we're orders of magnitude away from _needing_ 32bit counts.  Given 
32bit counts, I imagine the wordlists will be very, very large.  At the 
moment, Greg is having difficulties as his wordlists cross into tens of 
megabytes.





Return-Path: <>
Delivered-To: relson at osagesoftware.com
Received: from bncextmail.bancredito.com (unknown [200.42.204.6])
	by osagesoftware.com (Postfix) with ESMTP id D517C2838C
	for <relson at osagesoftware.com>; Tue, 28 Jan 2003 19:48:49 -0500 (EST)
Received: by bncextmail.bancredito.com with Internet Mail Service (5.5.2653.19)
	id <DQ6QJ4GY>; Tue, 28 Jan 2003 20:58:47 -0400
Message-ID: <1FB51AE25956D61180990004ACB8CE11DE8E04 at EXCHANGE3.tricom.net>
From: System Administrator <postmaster at bancredito.com>
To: relson at osagesoftware.com
Subject: Undeliverable: Re: t.lock2 test failures
Date: Tue, 28 Jan 2003 20:58:46 -0400
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
X-MS-Embedded-Report: 
Content-Type: multipart/mixed;
	boundary="----_=_NextPart_000_01C2C731.93EB93C0"
X-Bogosity: No, tests=bogofilter, spamicity=0.000000, version=0.10.2.tst.20030210
	  int  cnt    prob   spamicity  histogram
	 0.00   38  0.017396  0.008273  ######################################
	 0.10   11  0.156209  0.029654  ###########
	 0.20   12  0.250460  0.065428  ############
	 0.30    7  0.357741  0.094141  #######
	 0.40    0  0.000000  0.094141  
	 0.50    0  0.000000  0.094141  
	 0.60   11  0.649503  0.180496  ###########
	 0.70    4  0.729500  0.211310  ####
	 0.80    3  0.834211  0.239114  ###
	 0.90    0  0.000000  0.000000  

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_000_01C2C731.93EB93C0
Content-Type: text/plain

Your message

  To:      MSR Development Corp. Sales Dept.
  Subject: Re: t.lock2 test failures
  Sent:    Tue, 28 Jan 2003 20:48:33 -0400

did not reach the following recipient(s):

c=DO;a= ;p=BANCO NACIONAL D;o=TRICOM;dda:SMTP=CABREU at TRICOM.COM.DO; on Tue,
28 Jan 2003 20:49:09 -0400
    The recipient name is not recognized
	The MTS-ID of the original message is: c=do;a= ;p=banco nacional
d;l=EXCHANGE30301290049D2S44ALL
    MSEXCH:IMS:BANCO NACIONAL DE CREDITO:TRICOM:EXCHANGE3 0 (000C05A6)
Unknown Recipient



------_=_NextPart_000_01C2C731.93EB93C0
Content-Type: message/rfc822

Message-ID: <4.3.2.7.2.20030128194806.00c1ef00 at mail.osagesoftware.com>
From: Mail Delivery System <Mailer-Daemon at gluck.debian.org>
To: "MSR Development Corp. Sales Dept." <cabreu at tricom.com.do>
Subject: Re: t.lock2 test failures
Date: Tue, 28 Jan 2003 20:48:33 -0400
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
X-MS-Embedded-Report: 
Content-Type: text/plain


At 06:56 PM 1/28/03, Clint Adams wrote:

 > > The arm architecture was one where test t.lock2 failed.  Go into the
 > > bogofilter-0.10.1.2/test directories (whatever it's named) and run the
 > > following commands:
 >
 >Once again, the problem on arm magically disappears.  But I've got more
 >problems for you.

t.lock2 is a non-deterministic test.  It runs multiple (10) copies of
bogofilter - at one time - to see if locking works or fails.  Unfortunately
system load, the scheduler, and other factors prevent consecutive runs
being identical to one another.  However, it _is_ an indicator of how well
database locking works. 

------_=_NextPart_000_01C2C731.93EB93C0--



More information about the bogofilter-dev mailing list