Bug 3188 - local characters counted as "illegal" in subject
Summary: local characters counted as "illegal" in subject
Status: RESOLVED WORKSFORME
Alias: None
Product: Spamassassin
Classification: Unclassified
Component: Libraries (show other bugs)
Version: 2.63
Hardware: All All
: P5 major
Target Milestone: Undefined
Assignee: SpamAssassin Developer Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2004-03-17 22:05 UTC by Garry Glendown
Modified: 2005-01-21 02:19 UTC (History)
0 users


Attachment Type Modified Status Actions Submitter/CLA Status
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Garry Glendown 2004-03-17 22:05:34 UTC 
The function check_illegal_chars excludes a single instance of three special
characters on subject (#, pound, cent) - anyway, for users located in countries
or speaking languages that use special non-7-bit-characters like "äöüß" (or the
dozens of accented chars) this might prove to be a problem. I know that the RFCs
call for other encoding than 7bit for those characters, but not all mailers seem
to follow that RFC - there seem to be instances where special characters are not
escaped.
Question is whether this should be taken into account or not - while I do think
that on from and to headers there should not be any special characters, I have
yet to see any spam where any of the special chars could have identified it as
spam (contrary, spam that uses funny chars for an eye catcher 99 out of 100
cases has the correct encoding, getting it through that spam check ...)
Comment 1 Daniel Quinlan 2004-07-27 00:28:00 UTC 
From: and To: are no different than Subject: in this case.

I agree that maybe the rule should go.  It's better to have per-user
checks for this sort of thing.
Comment 2 Daniel Quinlan 2005-01-21 11:19:27 UTC 
Well, we've taken this under advisement.

I tuned the callers of check_illegal_chars to have higher limits, which
helps, and getting more European email in our nightly/release corpora would
help, but there's not much else to do with this bug.