Attachment filename gets corrupted

[xming]

Hi,

We have upgraded to 11 GA, and all those small issues are solved, except this one: the attachemnet filename gets corrupted, actually the file extension is chopped off so a lot of people cannot open the files automatically and they have to guess the file type

original MIME header

Code: Select all

Content-Type: image/jpeg;
   name="=?BIG5?B?s+zFUSAxMDggUENTIFNPQ0tFVCBTRVQgUElDLmpwZw==?="
Content-Transfer-Encoding: base64
X-Attachment-Id: 0.1
Content-Disposition: attachment; filename="=?BIG5?B?s+zFUSAxMDggUENTIFNPQ0tFVCBTRVQgUElDLmpwZw==?="

MIME header after passing thru scalix

Code: Select all

--scalix-part-0022a22bb1=_01
Content-Type: image/jpeg
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
   filename="=?UTF-8?Q?=E5=96=AC=E9=9C=B8_108_PCS_SOCKET_SET_PIC.?="

Any work arounds?

[kanderson]

I know there's an issue in Bugzilla right now where filenames are truncated if they are longer than ~50 characters. This seems to only affect the webclient.

Is that consistent with what you are seeing?

[xming]

This is consistent with almost all attachments encoded in Big5, and it's not only the webmail but also with the outlook connector.

[florian]

based on quick analysis, this looks like a bug....

I've opened http://bugzilla.scalix.com/show_bug.cgi?id=14605 to track.

Florian.

[xming]

Thank you florian,

Why does scalix even touch the mime headers? Isn't it possible to leave the filename encoded in Big5? Or you you want everything in UTF?

[florian]

I think we need the filename in UTF-8 for Outlook to work properly; also, not sure about web browsers - do they typically understand Big5? Whatever, UTF-8 should work on more platforms than Big5 and the conversion should be lossless, so..... I think this is probably a relatively minor conversion logic problem and hope we have time to look into it soon.

Cheers,
Florian.

[xming]

Thank you for the explanation, I would expect if someone is interested reading BIG5 encoded files then his PC has that :)

Anyway converting to UTF8 is better but not needed as the original sender sent it this way. Just my 2 cents.

I will follow the bugzilla and thanks again

xming