Re: [Ifile-discuss] Short writeup on ifile

[Clemens Fischer]

"Karl Vogel" <address@hidden>:

> http://www.dnaco.net/~vogelke/.../Filtering_Spam/

a few remarks:

      8  tmp=$tag.$RANDOM.tmp                                                   
       
      9  good=$tag.$RANDOM.good                                                 
       

i think mktemp(1) would be better here, because it assures unique file
names.

  #!/bin/ksh                                                                    
 
  # $Id: mkifiledb,v 1.2 2002/10/18 21:27:13 vogelke Exp $                      
 
  ...
  # local spam.                                                                 
 
  (                                                                             
 
      cd 00-SPAM-MESSAGES/misc                                                  
 
      list=`/bin/ls -d [0-9]* | sort -n`                                        
 
      readmail 3 spam $list                                                     
 
  )                                                                             
 

you don't need the sorting here, do you?  your files even have leading
zeros in their names, so you might as well say:  "readmail 3 spam *"
and leave out the "list=..." part entirely.  if i got your setup
right, you might even be more precise saying "readmail 3 spam ???".

 70  :0 HBD                                                                     
   
 71  * -10^1 Subject:                                                           
   
 72  *   1^1 =[0-9A-F][0-9A-F]=[0-9A-F][0-9A-F]                                 
   
 73  *   1^1 [ ¡¢£$,1tL¥! §!!©ª«¬­®¯°±²³!=µ¶·!>¹º» r s!8¿(B]
 74  *   1^1 [ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏĞÑÒÓÔÕÖ×ØÙÚÛÜİŞß]                                 
   
 75  *   1^1 [àáâãäåæçèéêëìíîïğñòóôõö÷øùúûüışÿ]                                 
   
 76  *   1^1 =[A-F][0-9A-F]=[A-F][0-9A-F]                                       
   
 77  spam-8bit                                                                  
   

you don't have any european friends, right?  :) these rules would
catch non-ASCII from roman, european countries as well.  better than
this is a rule like:

  :0HB
  * 1^0 ^Subject:( *Adv:)|.*(±¤°í)|.*(±¤ °í)
  * 1^0 ^Content-Type:.*charset.*(euc-kr|big5|ks_c_5601-1987|ISO-2022-KR)
  $DEVNULL


clemens