Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: SpamAssassin: devel

[Bug 6962] Implement the missing dump_db_toks() to Redis bayes backend (used by sa-learn --dump)

 

 

SpamAssassin devel RSS feed   Index | Next | Previous | View Threaded


bugzilla-daemon at bugzilla

Jul 30, 2013, 11:02 AM

Post #1 of 5 (36 views)
Permalink
[Bug 6962] Implement the missing dump_db_toks() to Redis bayes backend (used by sa-learn --dump)

https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6962

--- Comment #1 from Mark Martinec <Mark.Martinec [at] ijs> ---
Created attachment 5162
--> https://issues.apache.org/SpamAssassin/attachment.cgi?id=5162&action=edit
implements BayesStore::Redis::dump_db_toks()

--
You are receiving this mail because:
You are the assignee for the bug.


bugzilla-daemon at bugzilla

Jul 30, 2013, 11:08 AM

Post #2 of 5 (34 views)
Permalink
[Bug 6962] Implement the missing dump_db_toks() to Redis bayes backend (used by sa-learn --dump) [In reply to]

https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6962

Mark Martinec <Mark.Martinec [at] ijs> changed:

What |Removed |Added
----------------------------------------------------------------------------
Target Milestone|Undefined |3.4.0

--- Comment #2 from Mark Martinec <Mark.Martinec [at] ijs> ---
trunk:
Bug 6962: Implement the missing dump_db_toks() to Redis
bayes backend (used by sa-learn --dump)Redis.pm
Sending lib/Mail/SpamAssassin/BayesStore/Redis.pm
Committed revision 1508562.

--
You are receiving this mail because:
You are the assignee for the bug.


bugzilla-daemon at bugzilla

Aug 7, 2013, 4:59 AM

Post #3 of 5 (18 views)
Permalink
[Bug 6962] Implement the missing dump_db_toks() to Redis bayes backend (used by sa-learn --dump) [In reply to]

https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6962

--- Comment #3 from Mark Martinec <Mark.Martinec [at] ijs> ---
Btw, the sa-learn --dump and the --backup have a field 'atime' in
their output (the time of the last token access or something like that).
With Redis we rely on its own expiration mechanism, so the atime
is not stored in a database explicitly. It is however possible to
retrieve a remaining time-to-live for each key using a TTL redis call.

Would it be worthwhile to retrieve such TTL time and reconstruct
the atime of tokens (based on the current ttl setting) for the
purpose of a more informative --dump and --backup ? It would
require one more redis call for each token, so it would slow down
the --dump and --backup somewhat.

--
You are receiving this mail because:
You are the assignee for the bug.


bugzilla-daemon at bugzilla

Aug 7, 2013, 6:02 AM

Post #4 of 5 (18 views)
Permalink
[Bug 6962] Implement the missing dump_db_toks() to Redis bayes backend (used by sa-learn --dump) [In reply to]

https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6962

Henrik Krohns <hege [at] hege> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |hege [at] hege

--- Comment #4 from Henrik Krohns <hege [at] hege> ---
(In reply to Mark Martinec from comment #3)
> Would it be worthwhile to retrieve such TTL time and reconstruct
> the atime of tokens

I guess it might have some informational value to someone, so +0.5 from me if
you want. Anyone with a seriously large database would use a direct dumpfile
parser anyway (to be implemented). Marginally slowing down an already slow
native --dump seems fine. ;-)

--
You are receiving this mail because:
You are the assignee for the bug.


bugzilla-daemon at bugzilla

Aug 7, 2013, 7:26 AM

Post #5 of 5 (18 views)
Permalink
[Bug 6962] Implement the missing dump_db_toks() to Redis bayes backend (used by sa-learn --dump) [In reply to]

https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6962

Kevin A. McGrail <kmcgrail [at] pccc> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |kmcgrail [at] pccc

--- Comment #5 from Kevin A. McGrail <kmcgrail [at] pccc> ---
I view Redis as one of the implementations where people are more prone to just
start over a Bayes store from scratch so I don't know that I see the value.
Has anyone requested it? Perhaps just document it as a known issue?

--
You are receiving this mail because:
You are the assignee for the bug.

SpamAssassin devel RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.