Leonardo Rodrigues Magalhães wrote:
>
> MrC escreveu:
>> Leonardo Rodrigues Magalhães wrote:
>>
>>> Hello,
>>>
>>> I have banned some file/MIME types in amavisd-2.6.0 using the 'old'
>>> way of doing this, the $banned_filename_re.
>>>
>>> The banned file admin and user notifications, which i enabled,
>>> brings me something like:
>>>
>>> Banned name: multipart/mixed |
>>> application/vnd.ms-powerpoint,.doc,Chaplin.pps
>>> Content type: Banned
>>> Internal reference code for the message is 24067-22/NWynDBsexbaI
>>>
>>>
>>> It's clearly a powerpoint file, because of it's extension (.pps) as
>>> well as it's mime type 'application/vnd.ms-powerpoint'.
>>>
>>> What i'm trying to understand is where that '.doc' cames from !!!
>>>
>>>
>> What type of document does the file(1) utility indicate?
>> What version of file is on your system?
>>
>
> I'm running on a Fedora 8 system, with:
>
> [EMAIL PROTECTED] ~]# rpm -qi file
> Name : file Relocations: (not relocatable)
> Version : 4.21 Vendor: Fedora Project
> Release : 5.fc8 Build Date: Tue 29 Jan 2008
> 06:58:26 AM BRST
>
> which is the latest file package from F8 repositories.
>
> on this system, file returns for .doc and .ppt documents:
>
> [EMAIL PROTECTED] user]# file Defesa.doc
> Defesa.doc: Microsoft Office Document
> [EMAIL PROTECTED] user]# file Projeto\ Final\ II\ VPN\ -\ última.ppt
> Projeto Final II VPN - última.ppt: Microsoft Office Document
> [EMAIL PROTECTED] user]#
>
> hmmmmm ...... seems file returns only 'Microsoft Office Document' .....
>
In the past several versions, the file(1) utility has changed its
opinion many times regarding how to treat PPT documents:
$ file -v
file-4.21
magic file from /usr/share/file/magic
$ file test.ppt
test.ppt: Microsoft Installer
$ file -i test.ppt
test.ppt: \012- application/msword
---
$ file -v
file-4.24
magic file from /usr/local/share/file/magic
$ file ~/test.ppt
/home/cappella/test.ppt: Microsoft Office Document
$ file -i ~/test.ppt
/home/cappella/test.ppt: application/octet-stream
Amavis uses a mapping of full type names to short names, which are then
later referenced in the $banned_filename_re maps. You can see in the
latest amavisd that Microsoft Office Document is mapped to short name
type "doc". This is where ".doc" comes from.
$map_full_type_to_short_type_re = [
...
[qr/^Rich Text Format data\b/ => 'rtf'],
[qr/^Microsoft Office Document\b/i => 'doc'], # OLE2: doc, ppt,
# xls, ...
It is likely that the file formats for PPT, XLS, DOC, .etc have not been
reverse engineered to uniquely distinguish them from each other, and
instead they are all mapped to Microsoft Office Document (in more recent
file versions). The file formats became much more complex with later
versions of Office.
Previous versions of file(1), like 4.21, were very broken in terms of
identification, and there were many false identifications:
$ file -v
file-4.21
magic file from /usr/share/file/magic
$ file -i test.xls
test.xls: \012- application/msword
$ file test.xls
test.xls: Microsoft Installer
Clearly, this Excel spreadsheet is not an Word document. Fedora adds
its own patches to the file utility, of which yours is based on 4.21.
Since your PPT is identified as Microsoft Office Document, it is clear
that Fedora has updated the magic database used for file identification
bringing in line with the more recent 4.24/4.25 releases of file. Yet
the problem still remains - file types from more recent Office pacakges
are identified generically as Microsoft Office Document, and not Excel,
PowerPoint, etc. Given that, I'm not sure what more you can do to
distinguish the types.
Here's a thread which discusses some of the file(1) issues a while ago:
http://groups.google.com/group/mailing.unix.amavis-user/browse_thread/thread/7147b0a90573690c/04ca5171867925c1?lnk=gst&q=powerpoint+file#04ca5171867925c1
MrC
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
AMaViS-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/amavis-user
AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3
AMaViS-HowTos:http://www.amavis.org/howto/
