[Aide] I (ignore changed filename) woes

[Virolainen Pablo]

On Fri, 28 Oct 2005, Virolainen Pablo wrote:

> On Fri, 28 Oct 2005, Marc Haber wrote:
>
>> Hi,
>>
>> on my test host, I am running the CVS snapshot from October 26. I have
>> a rule
>>
>> RotatedLogs = I+n+p+i+u+g+s+b+m+c+md5+sha1+rmd160+haval+gost+crc32+tiger
>>
>> and one
>>
>> /var/log/syslog/syslog\.[0-9][0-9]\.gz$ RotatedLogs
>>
>> However, when my syslog is rotated, and /var/log/syslog/syslog.10.gz
>> moves to /var/log/syslog/syslog.11.gz, this change is shown in the
>> aide output.
>>
>> File: /var/log/syslog/syslog.10.gz
>>  Size     : 863916                            , 1031126
>>  Bcount   : 1696                              , 2024
>>  Mtime    : 2005-10-11 07:45:01               , 2005-10-13 07:44:59
>>  Ctime    : 2005-10-27 07:44:15               , 2005-10-28 07:44:24
>>  Inode    : 180562                            , 180563
>>  MD5      : ZT6Jon9RC19clQ3LueW+fQ==          , VLsNC6qHGe/qRNGde8U77g==
>>  SHA1     : /D31JbLs8uwzXz8dMHqd8sGE2rw=      , kojzzybYEiujhUnVKUBUCwurWDA=
>>  RMD160   : FmadTUIjKhGGZvdrfD1R06SA5Wc=      , CkCt3FS8bELVUipGR0vde01Lv68=
>>  TIGER    : YgHZwaLkEKveImqDc+6EO6QscFtUXbuV  , 7hbMrciBkv6tx/wa7rSJLAxPRCTAeukj
>>  CRC32    : yYMP6g==                          , 4qx5fQ==
>>  HAVAL    : My8zazLZiPHW13j6APssi5ei5LVpGnOhLI/kQqc, i06jS90awcgkXL3W2Wr+ZYZRk80gvBeLc4zl3YF
>>  GOST     : KDCJM2X0BY+jgy7IJG2UJ+39qR7m88epQaC0hsR, P9uCPajBGuED6KEkEs+N0ASDJKTEMdb3uT4tX/r
>>
>> File: /var/log/syslog/syslog.11.gz
>>  Size     : 1024800                           , 863916
>>  Bcount   : 2016                              , 1696
>>  Mtime    : 2005-10-09 07:51:32               , 2005-10-11 07:45:01
>>  Ctime    : 2005-10-27 07:44:15               , 2005-10-28 07:44:24
>>  Inode    : 180561                            , 180562
>>  MD5      : 0oXQIvu+T3X/5AvQ65Mtrg==          , ZT6Jon9RC19clQ3LueW+fQ==
>>  SHA1     : H7qdZT5/xEyqgZxOLLRM4oQB998=      , /D31JbLs8uwzXz8dMHqd8sGE2rw=
>>  RMD160   : bIhGs2R3jUpatgcKKacx2BJtODs=      , FmadTUIjKhGGZvdrfD1R06SA5Wc=
>>  TIGER    : pJuy68VCApxOoLNri09X1tAZKfzY7zkF  , YgHZwaLkEKveImqDc+6EO6QscFtUXbuV
>>  CRC32    : RFE2QQ==                          , yYMP6g==
>>  HAVAL    : I9936r4JkKLw09av5U1BCxCri4awg2Eu+YVq6oq, My8zazLZiPHW13j6APssi5ei5LVpGnOhLI/kQqc
>>  GOST     : CarCEIUtsEVC2DlCSwwBkA2kJ+Dqkd32jBq6B47, KDCJM2X0BY+jgy7IJG2UJ+39qR7m88epQaC0hsR
>>
>> Any idea what I might be doing wrong? Is there any elegant way to
>> debug this?
>>
>> Maybe it'll help if somebody familiar with the aide code could explain
>> - in prose - how the new I setting works internally.
>
> I cannot remember that there is this kind of feature in AIDE. It would be
> nice to have.

Hm.. Someone has implemented this feature (which is good). One propably 
should give warning about if rule has c and I enabled at the same time (or 
just ignore c, if we are checking for changed filenames? For 
implementatio check patch...)

--- gen_list.c  Fri Oct 28 16:33:39 2005
+++ gen_list.c.orig  Mon Oct  3 08:58:31 2005
@@ -1111,7 +1110,7 @@
             oldData->filename,oldData->attr,newData->attr);
      }

-    localignorelist|=ignorelist;
+    localignorelist|=ignorelist|DB_CTIME;

      /* Free the data if same else leave as is for report_tree */
      if(compare_dbline(oldData, newData, localignorelist)==RETOK){


Duke NEMO / C.O.M.A
alias pablo the pallo virolainen