You are viewing a plain text version of this content. The canonical link for it is here.
Posted to ruleqa@spamassassin.apache.org by Tom Hendrikx <to...@whyscream.net> on 2017/02/09 12:12:57 UTC
Re: New masscheck corpus
Hi,
I've finally found some time to setup the masschecks too. This morning I
submitted the first results, which seem to be there according to ruleqa.
Yay! ;)
My initial take was to use the last 2000 ham and spam messages that I
have available handpicked. This unfortunately resulted in the oldest
spam message being almost a year old. I checked my slumbering spamtrap
and discovered some more recent data that could be used. I now manually
verified the stuff in the trap so that is good to go too.
Based on the ruleqa messages from 25th of Jan, I understand that when we
don't have enough samples for a new update, that's mostly due to ham
starvation?
Finally, I found out that the JOBS= setting in the .auto-mass-check.cf
file does not work as advertised. Patch attached.
Kind regards,
Tom
On 25-01-17 10:09, Kevin Golding wrote:
> On Tue, 24 Jan 2017 19:57:35 -0000, David Jones <dj...@ena.com> wrote:
>
>>> From: John McNally <jo...@psfc.coop>
>>
>>> I just ran automasscheck-minimal.sh for the first time this morning
>>> around 2:00 EST and
>>> uploaded the logs to rsync.spamassassin.org. Could one of you please
>>> inspect and see if
>>> it looks reasonable? My username is "jmcnally". Then, I can automate
>>> with cron.
>>
>> I too (user = ena) submitted my first masscheck yesterday and would
>> like confirmation
>> that it looks correct.
>
> The magic place to check is http://ruleqa.spamassassin.org/
>
> Both of you submitted on Jan 23rd, ena only on the 24th
>
> That should suggest you've both got everything running correctly.
>
>>> - What is the best time to schedule my automasscheck cron job? The
>>> NightlyMassCheck
>>> wiki says "at or after 9AM UTC", but is there an optimal time window,
>>> as far as the ruleqa
>>> workflow is concerned?
>>
>> Looking forward to this answer too.
>
> As soon after 9AM UTC as feasible. The day times out at some point and
> anything received after that cutoff won't be included so earlier is
> better in that regard. I believe it's quite a large window to submit in,
> but optimal is early because that collects as much data as soon as
> possible.
>
>>> - My automasscheck run took about 20 minutes on a CentOS 7 VM with 4
>>> CPU's and 12 GB
>>> of RAM. The ham and spam log files were 884 KB and 33 MB, with 502
>>> and 22972 entries,
>>> respectively. Will successive runs continue to process all messages
>>> in the ham and spam
>>> dirs, or just new messages?
>
> All messages. This is because rules evolve, so every day the results
> start with a fresh slate.
>
>> My wrapper script moves the Ham and Spam contents to "Masschecked/Ham"
>> and
>> "Masschecked/Spam" folders after the automasscheck-minimal.sh completes.
>>
>> I saw somewhere on the wiki that mail needs to be moved out of the way
>> after
>> running the automasscheck-minimal.sh script. Can't find it at the
>> moment.
>
> You should expire messages due to age, but not just because they've
> already been checked. If you want to expire them so that you always
> process the most recent X number of messages that's your choice and a
> way to make sure you have a good idea how long a test will take.
> Although once your corpora get to a certain age you find they don't
> really grow since you expire at a similar pace to adding.
>
> As a note also keep in mind that tests run on Saturday take longer
> because they run the network tests too.
Re: Fwd: Re: New masscheck corpus
Posted by John Hardin <jh...@impsec.org>.
On Thu, 9 Feb 2017, Kevin A. McGrail wrote:
> Thanks David.
>
> Anyone got a moment to handle this?
If it can wait until Satudray, sure.
>> From: Tom Hendrikx <to...@whyscream.net>
>> Sent: Thursday, February 9, 2017 6:12 AM
>
>> Hi,
>
>> I've finally found some time to setup the masschecks too. This morning I
>> submitted the first results, which seem to be there according to ruleqa.
>> Yay! ;)
>
>> Finally, I found out that the JOBS= setting in the .auto-mass-check.cf
>> file does not work as advertised. Patch attached.
>
> Attached is Tom's patch plus my minor symlink additions that help when
> running any of the tools in the masses directory.
>
> Please take a few minutes to commit these patches to show the community
> that our input/help is valued. The current automasscheck-minimal.sh is 4
> years old and has some bugs that Tom fixed. An update to this file will show
> the rest of the potential masscheck helpers that this tool is important and
> still maintained.
>
> Thanks,
> Dave
--
John Hardin KA7OHZ http://www.impsec.org/~jhardin/
jhardin@impsec.org FALaholic #11174 pgpk -a jhardin@impsec.org
key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
Usually Microsoft doesn't develop products, we buy products.
-- Arno Edelmann, Microsoft product manager
-----------------------------------------------------------------------
3 days until Abraham Lincoln's and Charles Darwin's 208th Birthdays
Re: Fwd: Re: New masscheck corpus
Posted by "Kevin A. McGrail" <KM...@PCCC.com>.
On 2/12/2017 8:16 PM, John Hardin wrote:
> On Thu, 9 Feb 2017, Kevin A. McGrail wrote:
>
>> Thanks David.
>>
>> Anyone got a moment to handle this?
>
> Checked in on trunk, merged to 3.4 branch.
Rockin!
Re: Fwd: Re: New masscheck corpus
Posted by "Kevin A. McGrail" <KM...@PCCC.com>.
On 2/12/2017 8:16 PM, John Hardin wrote:
> On Thu, 9 Feb 2017, Kevin A. McGrail wrote:
>
>> Thanks David.
>>
>> Anyone got a moment to handle this?
>
> Checked in on trunk, merged to 3.4 branch.
Rockin!
Re: Fwd: Re: New masscheck corpus
Posted by John Hardin <jh...@impsec.org>.
On Thu, 9 Feb 2017, Kevin A. McGrail wrote:
> Thanks David.
>
> Anyone got a moment to handle this?
Checked in on trunk, merged to 3.4 branch.
>> From: Tom Hendrikx <to...@whyscream.net>
>> Sent: Thursday, February 9, 2017 6:12 AM
>
>> Hi,
>
>> I've finally found some time to setup the masschecks too. This morning I
>> submitted the first results, which seem to be there according to ruleqa.
>> Yay! ;)
>
>> Finally, I found out that the JOBS= setting in the .auto-mass-check.cf
>> file does not work as advertised. Patch attached.
>
> Attached is Tom's patch plus my minor symlink additions that help when
> running any of the tools in the masses directory.
>
> Please take a few minutes to commit these patches to show the community
> that our input/help is valued. The current automasscheck-minimal.sh is 4
> years old and has some bugs that Tom fixed. An update to this file will show
> the rest of the potential masscheck helpers that this tool is important and
> still maintained.
>
> Thanks,
> Dave
>
>
>
--
John Hardin KA7OHZ http://www.impsec.org/~jhardin/
jhardin@impsec.org FALaholic #11174 pgpk -a jhardin@impsec.org
key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
North Korea: the only country in the world where people would risk
execution to flee to communist China. -- Ride Fast
-----------------------------------------------------------------------
Today: Abraham Lincoln's and Charles Darwin's 208th Birthdays
Re: Fwd: Re: New masscheck corpus
Posted by John Hardin <jh...@impsec.org>.
On Thu, 9 Feb 2017, Kevin A. McGrail wrote:
> Thanks David.
>
> Anyone got a moment to handle this?
If it can wait until Satudray, sure.
>> From: Tom Hendrikx <to...@whyscream.net>
>> Sent: Thursday, February 9, 2017 6:12 AM
>
>> Hi,
>
>> I've finally found some time to setup the masschecks too. This morning I
>> submitted the first results, which seem to be there according to ruleqa.
>> Yay! ;)
>
>> Finally, I found out that the JOBS= setting in the .auto-mass-check.cf
>> file does not work as advertised. Patch attached.
>
> Attached is Tom's patch plus my minor symlink additions that help when
> running any of the tools in the masses directory.
>
> Please take a few minutes to commit these patches to show the community
> that our input/help is valued. The current automasscheck-minimal.sh is 4
> years old and has some bugs that Tom fixed. An update to this file will show
> the rest of the potential masscheck helpers that this tool is important and
> still maintained.
>
> Thanks,
> Dave
--
John Hardin KA7OHZ http://www.impsec.org/~jhardin/
jhardin@impsec.org FALaholic #11174 pgpk -a jhardin@impsec.org
key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
Usually Microsoft doesn't develop products, we buy products.
-- Arno Edelmann, Microsoft product manager
-----------------------------------------------------------------------
3 days until Abraham Lincoln's and Charles Darwin's 208th Birthdays
Re: Fwd: Re: New masscheck corpus
Posted by John Hardin <jh...@impsec.org>.
On Thu, 9 Feb 2017, Kevin A. McGrail wrote:
> Thanks David.
>
> Anyone got a moment to handle this?
Checked in on trunk, merged to 3.4 branch.
>> From: Tom Hendrikx <to...@whyscream.net>
>> Sent: Thursday, February 9, 2017 6:12 AM
>
>> Hi,
>
>> I've finally found some time to setup the masschecks too. This morning I
>> submitted the first results, which seem to be there according to ruleqa.
>> Yay! ;)
>
>> Finally, I found out that the JOBS= setting in the .auto-mass-check.cf
>> file does not work as advertised. Patch attached.
>
> Attached is Tom's patch plus my minor symlink additions that help when
> running any of the tools in the masses directory.
>
> Please take a few minutes to commit these patches to show the community
> that our input/help is valued. The current automasscheck-minimal.sh is 4
> years old and has some bugs that Tom fixed. An update to this file will show
> the rest of the potential masscheck helpers that this tool is important and
> still maintained.
>
> Thanks,
> Dave
>
>
>
--
John Hardin KA7OHZ http://www.impsec.org/~jhardin/
jhardin@impsec.org FALaholic #11174 pgpk -a jhardin@impsec.org
key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
North Korea: the only country in the world where people would risk
execution to flee to communist China. -- Ride Fast
-----------------------------------------------------------------------
Today: Abraham Lincoln's and Charles Darwin's 208th Birthdays
Fwd: Re: New masscheck corpus
Posted by "Kevin A. McGrail" <KM...@PCCC.com>.
Thanks David.
Anyone got a moment to handle this?
>From: Tom Hendrikx <to...@whyscream.net>
>Sent: Thursday, February 9, 2017 6:12 AM
>Hi,
>I've finally found some time to setup the masschecks too. This morning I
>submitted the first results, which seem to be there according to ruleqa.
>Yay! ;)
>Finally, I found out that the JOBS= setting in the .auto-mass-check.cf
>file does not work as advertised. Patch attached.
Attached is Tom's patch plus my minor symlink additions that help when
running any of the tools in the masses directory.
Please take a few minutes to commit these patches to show the community
that our input/help is valued. The current automasscheck-minimal.sh is 4
years old and has some bugs that Tom fixed. An update to this file will show
the rest of the potential masscheck helpers that this tool is important and
still maintained.
Thanks,
Dave
Fwd: Re: New masscheck corpus
Posted by "Kevin A. McGrail" <KM...@PCCC.com>.
Thanks David.
Anyone got a moment to handle this?
>From: Tom Hendrikx <to...@whyscream.net>
>Sent: Thursday, February 9, 2017 6:12 AM
>Hi,
>I've finally found some time to setup the masschecks too. This morning I
>submitted the first results, which seem to be there according to ruleqa.
>Yay! ;)
>Finally, I found out that the JOBS= setting in the .auto-mass-check.cf
>file does not work as advertised. Patch attached.
Attached is Tom's patch plus my minor symlink additions that help when
running any of the tools in the masses directory.
Please take a few minutes to commit these patches to show the community
that our input/help is valued. The current automasscheck-minimal.sh is 4
years old and has some bugs that Tom fixed. An update to this file will show
the rest of the potential masscheck helpers that this tool is important and
still maintained.
Thanks,
Dave
Re: New masscheck corpus
Posted by David Jones <dj...@ena.com>.
>From: Tom Hendrikx <to...@whyscream.net>
>Sent: Thursday, February 9, 2017 6:12 AM
>Hi,
>I've finally found some time to setup the masschecks too. This morning I
>submitted the first results, which seem to be there according to ruleqa.
>Yay! ;)
>Finally, I found out that the JOBS= setting in the .auto-mass-check.cf
>file does not work as advertised. Patch attached.
Attached is Tom's patch plus my minor symlink additions that help when
running any of the tools in the masses directory.
Please take a few minutes to commit these patches to show the community
that our input/help is valued. The current automasscheck-minimal.sh is 4
years old and has some bugs that Tom fixed. An update to this file will show
the rest of the potential masscheck helpers that this tool is important and
still maintained.
Thanks,
Dave