[
https://bro-tracker.atlassian.net/browse/BIT-1238?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18304#comment-18304
]
Brian O'Berry edited comment on BIT-1238 at 10/11/14 10:28 AM:
---------------------------------------------------------------
After looking at
[http://www.gnu.org/software/tar/manual/html_node/Standard.html], experimenting
some more with regex101.com, and running {code}perl -ne
'/[[:print:]\x00]{100}(?:[0-7]{1,7} ?\x00){3}/ and print "yup\n";'{code} on
various tar files, the pcre shown in the perl command seems like a better
pattern for the {{file-tar}} signature. The changes from the original pcre
include:
* don't capture the name field group -- remove the parenthesis from
{code}([[:print:]\x00]{100}){code}
* don't capture the mode, uid, gid group -- remove the outer parenthesis from
{code}(([[:digit:]\x00\x20]){8}){code}
* change the group pattern for mode, uid, gid to be non-capturing -- prefix the
enclosed pattern with {{?:}}
* change the possible digits for mode, uid, gid to octal -- change
{{\[:digit:\]}} to {{0-7}}
* move the space and null ({{\x00}}) bytes for mode, uid, gid outside the group
pattern so they match only at the end
* change the count value on the mode, uid, gid group from 8 to range 1-7 to
account for moving the optional space and required null byte outside the group
* make the trailing space optional for the mode, uid, gid group , so tar files
created by both gtar and bsdtar match correctly
was (Author: boberry):
After looking at
[http://www.gnu.org/software/tar/manual/html_node/Standard.html], experimenting
some more with regex101.com, and running {code}perl -ne
'/[[:print:]\x00]{100}(?:[0-7]{1,7} ?\x00){3}/ and print "yup\n";'{code} on
various tar files, the pcre shown in the perl command seems like a better
pattern for the {{file-tar}} signature. The changes from the original pcre
include:
* don't capture the name field group -- remove the parenthesis from
{code}([[:print:]\x00]{100}){code}
* don't capture the mode, uid, gid group -- remove the outer parenthesis from
{code}(([[:digit:]\x00\x20]){8}){code}
* change the group pattern for mode, uid, gid to be non-capturing -- prefix the
enclosed pattern with {{?:}}
* change the possible digits for mode, uid, gid to octal -- change
{{\[:digits:\]}} to {{0-7}}
* move the space and null ({{\x00}}) bytes for mode, uid, gid outside the group
pattern so they match only at the end
* change the count value on the mode, uid, gid group from 8 to range 1-7 to
account for moving the optional space and required null byte outside the group
* make the trailing space optional for the mode, uid, gid group , so tar files
created by both gtar and bsdtar match correctly
> High false-positive for application/x-tar signature
> ---------------------------------------------------
>
> Key: BIT-1238
> URL: https://bro-tracker.atlassian.net/browse/BIT-1238
> Project: Bro Issue Tracker
> Issue Type: Problem
> Components: Bro
> Affects Versions: 2.3
> Reporter: Brian O'Berry
> Labels: file, mime, signature
>
> The following signature in base/frameworks/files/magic/general.sig frequently
> triggers on text files in our environment, and includes a strength value
> higher than GNU and POSIX tar signatures in libmagic.sig.
> {code}
> signature file-tar {
> file-magic /([[:print:]\x00]){100}(([[:digit:]\x00\x20]){8}){3}/
> file-mime "application/x-tar", 150
> }
> {code}
--
This message was sent by Atlassian JIRA
(v6.4-OD-05-009#64003)
_______________________________________________
bro-dev mailing list
[email protected]
http://mailman.icsi.berkeley.edu/mailman/listinfo/bro-dev
