RE: SA-Learn and Windows long file names

["Mark A. DeMichele" <de...@intellipro.com>]

Ok, I stepped trough the sa-learn code with a debugger and discovered
that my e-mail files are getting "grepped" out.  Unfortunately, I'm not
very good at regular expressions yet.  Can someone explain please.  The
problem is in the scan_directory method of  ArchiveIterator.pm.  The
grep is apparently trying to filter out some files, but it's actually
filtering out all my mail.

 

It seems to having something to do with a NON "." or ","  preceding a
non-whitespace.  What the heck is that (sorry, I'm a windows person).

 

# ignore ,234 (deleted or refiled messages) and MH metadata dotfiles

   @files = grep { /^[^,.]\S*$/ } readdir(DIR);

 

 

These are my file names.  (Please don't make fun of these crazy
characters in the file name.  I have no control.  This is apparently how
Active Directory/Exchange Server name the files).  What's odd is that
they encode the asterisk but not the $ or ! or period.

 

%2A%2A%2A%2ASPAM%2A%2A%2A%2A! $1,500 a Day -  Starting Today !%22 DARE
ME%22 scraqwcxnafpsffqm.EML

%2A%2A%2A%2ASPAM%2A%2A%2A%2A! $1.10 A DOSE VIAGRA.EML

 

Anyway,  any insight on why this grep is there would be appreciated.

 

Thanks,

Mark DeMichele

 

 

 

  _____  

From: Mark A. DeMichele 
Sent: Sunday, February 29, 2004 9:42 AM
To: spamassassin-users@incubator.apache.org
Subject: SA-Learn and Windows long file names

 

I've been trying to get sa-learn to learn spam messages stored in an
active directory on an Exchange Server.  I discovered that it's having
problems with long file names.  Exchange names the files using the
subject which is long.  If I copy the messages and rename them to
something shorter, it works ok.  Do you think there's any way this can
be fixed.

Thanks,

Mark DeMichele