Re: MIME headers in Procmail

Quoting John Levine <[email protected]>:

> As I expect everyone is aware, you can use MIME hackery in mail
> headers to include non-ASCII text in just about everything other than
> the actual e-mail address.  Writing patterns to match MIME stuff is
> painful, since there's about a dozen character sets with somewhat
> different encoding, and if they do base64 encoding, the coded values
> change depending on where in the string the characters are.
>
> So it would be handy if there were an option to interpret all the MIME
> and turn all the text into UTF-8 which (assuming they are 8 bit clean)
> existing regex code should handle.
>
> Anyone done that?  Thought about it?

Well, after having my email address used by a spammer to sign up on  
thousands of forums in multiple countries, I found this worked pretty  
well to standardize the Subject line:

SUBJECT=`formail -xSubject: |/usr/local/bin/perl -MEncode -ne 'print  
encode ("utf8",decode ("MIME-Header",$_ )) '`
so I could delete anything that said 'Welcome To' in the subject.  I  
assume this is what you're shooting for.

The problem is, a user friendly interface generally isn't set up to do  
variable matching.  It's just going to do a header match.  So there's  
a little re-write there.

I'm also not sure, as an English Speaker, what would happen with other  
languages when that conversion is done.  Also, if that user friendly  
interface were modified to do variable matches instead of header  
matches....  Would it be more or less reliable?  What happens to  
umlauts and such?

That's about as far as I've thought about it - but I've had user  
complaints about filter accuracy due to the UTF8 issue, I'm just not  
sure what to do about it...  :)

Rick