Re: Hang during learning

   From: "Angus McIntyre" <[email protected]>

   I have a very simple learning script, which consists of:

      {
      learn <osb unique microgroom> (:*:_arg2:)  
      }

   When I invoke it with:

      cat samplefile | crm learn.crm cssfile

   (file and directory names have been changed to protect the innocent), it
   periodically hangs.

   The files being learned are plaintext (they're actually spam emails), and
   the stall occurs with a css file that has been trained on several tens of
   thousands of example files.

   The file being read when the stall occurs varies, so I don't believe the
   problem is caused by something specific to the file contents, path or
   permissions.

   The css file should have ample room on the filesystem to expand if needed.
   The host on which the script is running has a total of 3GB of physical
   memory.

   Deleting the css file and starting over allows processing of the directory
   to complete without problems. However, as I train the css file on more and
   more examples, I will eventually reach a point where the learning script
   hangs again. As far as I have seen the learning process never terminates.

   This is with 20100106-BlameMichelson (TRE 0.8.0 (BSD), but also occurred
   with earlier versions.

   Has anyone else encountered similar problems, and are there any possible
   workarounds?

That's VERY odd.  But very interesting nonetheless.

First, try running cssutil on the cssfile to see if it's 
gone utterly full (it shouldn't have, being microgroomed, but
there's always the chance you've found a bug).

Second- a workaround.  Try increasing the size of the cssfile
itself, with "-S 10000000" to give yourself 10 million slots 
instead of the default 1 million.

See if that helps... it should.

    - Bill Yerazunis

------------------------------------------------------------------------------
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
_______________________________________________
Crm114-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/crm114-general