Kevin A. McGrail wrote:
> Bob Proulx wrote:
> >It looks to me that -D now disables Bayes as a regression in 3.4.0.
>
> Can't recreate and can't see how that's possible. -D turns on debug
> output and I don't see any references to the config option in the
> Bayes library. Not saying you are wrong but it sounds a bit
> unlikely.
I agree that the connection between those two things seems odd. But I
am definitely seeing it.
> And here's a dummy devel system with bayes turned on with no tokens:
Hmm... Thanks for trying to reproduce this. It is useful because if
it isn't reproducible to you then it must be something particular
about my system that is causing this. Knowing that is very useful.
> What does spamassassin -d -t -D < mail.file 2>&1 | grep -i bayes
> show for you?
I picked a spam message and piped it into:
spamassassin -d -t -D 2>&1 | grep -i bayes | tee /tmp/sa.bayes-debug.out
It produced this output:
Feb 21 11:09:22.480 [1714] dbg: plugin: loading
Mail::SpamAssassin::Plugin::Bayes from @INC
Feb 21 11:09:22.670 [1714] dbg: config: fixed relative path:
/var/lib/spamassassin/3.004000/updates_spamassassin_org/23_bayes.cf
Feb 21 11:09:22.670 [1714] dbg: config: using
"/var/lib/spamassassin/3.004000/updates_spamassassin_org/23_bayes.cf" for
included file
Feb 21 11:09:22.670 [1714] dbg: config: read file
/var/lib/spamassassin/3.004000/updates_spamassassin_org/23_bayes.cf
Feb 21 11:09:23.350 [1714] dbg: plugin:
Mail::SpamAssassin::Plugin::Bayes=HASH(0x351ce78) implements 'learner_new',
priority 0
Feb 21 11:09:23.350 [1714] dbg: bayes: learner_new
self=Mail::SpamAssassin::Plugin::Bayes=HASH(0x351ce78),
bayes_store_module=Mail::SpamAssassin::BayesStore::DBM
Feb 21 11:09:23.363 [1714] dbg: bayes: learner_new: got
store=Mail::SpamAssassin::BayesStore::DBM=HASH(0x39bc410)
Feb 21 11:09:23.363 [1714] dbg: plugin:
Mail::SpamAssassin::Plugin::Bayes=HASH(0x351ce78) implements
'learner_is_scan_available', priority 0
Feb 21 11:09:23.364 [1714] dbg: bayes: tie-ing to DB file R/O
/home/bob/.spamassassin/bayes_toks
Feb 21 11:09:23.364 [1714] dbg: bayes: tie-ing to DB file R/O
/home/bob/.spamassassin/bayes_seen
Feb 21 11:09:23.364 [1714] dbg: bayes: found bayes db version 3
Feb 21 11:09:26.701 [1714] dbg: bayes: untie-ing
The content analysis part shows:
Content analysis details: (11.8 points, 5.0 required)
pts rule name description
---- ---------------------- --------------------------------------------------
1.9 URIBL_JP_SURBL Contains an URL listed in the JP SURBL blocklist
[URIs: towns55.eu]
1.7 URIBL_WS_SURBL Contains an URL listed in the WS SURBL blocklist
[URIs: towns55.eu]
1.2 RCVD_IN_BL_SPAMCOP_NET RBL: Received via a relay in bl.spamcop.net
[Blocked - see <http://www.spamcop.net/bl.shtml?67.222.157.193>]
2.7 RCVD_IN_PSBL RBL: Received via a relay in PSBL
[67.222.157.193 listed in psbl.surriel.com]
0.1 UNKNOWN_RCVD Unknown host in received header
0.0 T_HEADER_FROM_DIFFERENT_DOMAINS From and EnvelopeFrom 2nd level mail
domains are different
1.1 MIME_HTML_ONLY BODY: Message only has text/html MIME parts
1.7 URIBL_BLACK Contains an URL listed in the URIBL blacklist
[URIs: towns55.eu]
1.3 RDNS_NONE Delivered to internal network by a host with no rDNS
0.0 T_REMOTE_IMAGE Message contains an external image
Note that there are no BAYES_* rules shown. And the points are
reduced by that amount. If I run the same command removing only the
-D option from it then I get this content analysis.
spamassassin -d -t 2>&1 | tee /tmp/sa.bayes-debug.out
Content analysis details: (13.9 points, 5.0 required)
pts rule name description
---- ---------------------- --------------------------------------------------
1.3 RCVD_IN_BL_SPAMCOP_NET RBL: Received via a relay in bl.spamcop.net
[Blocked - see <http://www.spamcop.net/bl.shtml?67.222.157.193>]
2.7 RCVD_IN_PSBL RBL: Received via a relay in PSBL
[67.222.157.193 listed in psbl.surriel.com]
3.5 BAYES_99 BODY: Bayes spam probability is 99 to 100%
[score: 1.0000]
0.1 UNKNOWN_RCVD Unknown host in received header
0.0 T_HEADER_FROM_DIFFERENT_DOMAINS From and EnvelopeFrom 2nd level mail
domains are different
0.2 BAYES_999 BODY: Bayes spam probability is 99.9 to 100%
[score: 1.0000]
0.7 MIME_HTML_ONLY BODY: Message only has text/html MIME parts
1.7 URIBL_BLACK Contains an URL listed in the URIBL blacklist
[URIs: towns55.eu]
1.2 URIBL_JP_SURBL Contains an URL listed in the JP SURBL blocklist
[URIs: towns55.eu]
1.6 URIBL_WS_SURBL Contains an URL listed in the WS SURBL blocklist
[URIs: towns55.eu]
0.8 RDNS_NONE Delivered to internal network by a host with no rDNS
0.0 T_REMOTE_IMAGE Message contains an external image
Note that without -D it shows both BAYES_99 and BAYES_999 and the
points are appropriate. These were run back-to-back only a minute
apart from each other.
However if you can't recreate then it then must be something specific
to my system. That is useful to me to know. It tells me that I must
chase it down. I will try a pristine system and clean installation
and see if that is different. Hopefully so and then I can A-B compare
until I find the local problem.
Thanks,
Bob
