If you don't use mass-check or don't know what it is, you can ignore
this message.
Okay, we have a new mass-check. Many things are the same, but there are
lots of new features and how you use it has significantly changed.
Before, the model of mass check was to run separate mass-checks, one for
spam and one for ham. To make use of additional processors or get
initial spam and ham results at the same time, you typically ran two or
more mass-checks at the same time. Messages were basically checked in
the order given to mass-check (although directories could be sorted by
filename).
Now, you run one mass-check for everything and tell it about all of your
ham and spam at once. It scans all of the messages, sorts them by the
received date (well, a reasonable guess), and then interleaves ham and
spam so that a 50/50 mix is used (at least, until you run out of the
smaller set of the two, ham or spam).
In addition, if you want to run more than one check process at a time,
you can specify a -j option (like "-j 2") to use additional processors
or do network mass-checks. I found that on my fast uniprocessor box,
a "-j 8" achieved about a 6x speed-up when doing network checks.
The command line usage is a bit tricker, but I'll give a few short
examples.
Before, your mass-check might have looked like this:
mass-check --mh /my/spamdir > spam.log
mass-check /my/spam.mbox >> spam.log
mass-check --mh /my/gooddir > ham.log
mass-check --single /my/good.email >> ham.log
Now, you would do this:
mass-check spam:dir:/my/spamdir spam:mbox:/my/spam.mbox \
good:dir:/my/gooddir good:file:/my/good.email
And the results get written to spam.log and ham.log.
Or, you can just stuff the targets into a file and use the -f option.
To check a few messages before, you might have done:
mass-check --single /tmp/1 /tmp/2
(output goes to stdout)
And now, you'd just do:
mass-check --file /tmp/1 /tmp/2
(--file makes the output go to stdout, it's assumed that since you're
not telling it whether this is spam or ham that you're doing some sort
of test)
But, you can mix it up and use different types of targets unlike before:
mass-check --file /tmp/1 --mbox /tmp/3.mbox --file /tmp/2
At this point, I'd suggest you update CVS and try it out. The rest of
the usage is located at the top of the mass-check script. Some other
things to note:
- there's a --net option so no more editing of mass-check to turn on
network tests
- "nonspam" is dead as a term in mass-check. everything is "ham" now.
- tar file support is gone for the moment
- "new" and "cur" of maildirs have to be specified separately
Please check the usage of every option before you assume it's broken.
Also, if you want to read the entire saga of these changes, read bug
1203.
All that being said, if you have any problems or questions, post a
message to the SAdev mailing list or submit a bug and I will get to it
ASAP.
Daniel
--
Daniel Quinlan Linux, open source, and
http://www.pathname.com/~quinlan/ anti-spam consulting
-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
