Subjects of emails become corrupted if the Subject: MIME header is base64-encoded and folded over multiple lines in RFC violating way

Versions / Builds Affected

20150218

Status

Open

TT / JIRAID

2966

How to Identify

Emails which have the following characteristics: - the MIME Subject: header is base64 encoded - the MIME Subject: header (base64 encoded) is "folded" over multiple lines in an RFC violating way (RFCs mentioned at the bottom) For example: Subject: =?utf-8?B?YWJjZGVmZ2hpamtsbW5vcHFyc3?= =?utf-8?B?R1dnd4eXo=?= (decoded above reads: abcdefghijklmnopqrstuvwxyz) (online decoder: http://coderstoolbox.net/string/#!encoding=base64&action=decode&charset=utf_8) The following symptoms occur: - the subject of such an email is displayed with corrupted characters within GFI Archiver (e.g. in the web page > Archive tab) - the subject of such an email is written with corrupted characters into the database (not the email itself, only the subject reference in the dbo.arc table) - such emails will not successfully folder sync UMPolling wise Based on the sample subject from above, it would appears like this: abcdefghijklmnopqrsp@GWgw��� @ The corrupted subject would be visible in ..\GFI\Store\DebugLogs\UMPoll.ComputeHash.log: 2015-03-17,15:18:27,415,1,"#00000564","#00000004","info ","UMPoll.ComputeHash","abcdefghijklmnopqrsp@gwgw��� @635610460350000000" References: - folding: RFC2822 > 2.2.3. Long Header Fields - encoded MIME headers: RFC2047 > 2. Syntax of encoded-words

Workaround / Fix Details

None available

Required Actions

Escalate with: 1. A set of TS files 2. Sample emails in EML format downloaded from GFI Archiver