Details
-
Bug
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
Description
The handling of problems when processing messages through JSieve is not robust.
The case I like to describe is caused when a ParseException is produced while processing a message (for example, being hit by MIME4J-253),
org.apache.james.mime4j.field.address.ParseException: Encountered "\r" at line 1, column 42.
Was expecting one of:
":" ...
"<" ...
<DOTATOM> ...
"\"" ...
<DOTATOM> ...
"\"" ...
The current logic produces a new message being preceded in the subject with [SIEVE ERROR]. This message enters the JSieve processing chain again and again is processed, and another message now with the subject starting with [SIEVE ERROR][SIEVE ERROR] is produced. And so goes on and on several hundred of times. That fills the mailbox and it is not desired.
An example of a message that relates to MIME4J-253 is
https://www.mail-archive.com/server-dev@james.apache.org/msg59652.html
And the header that is causing the exception:
From: =?utf-8?Q?R=C3=A9mi_Kowalski_=28JIRA=29?=
<server-dev@james.apache.org>